Preliminar evaluation of epigenetic modifications followingi vitro hydroquinone exposure in models of human acute promyelocytic leukemia cells (HL-60) and of human umbilical cord mesenchimal stem cell (hUCMSC)

Benzene, a volatile aromatic hydrocarbon, is extensively used in industry even though it is recognized as a myelotoxin with leukemogenic activity, representing a significant occupational risk. Metabolism of benzene plays a fundamental role in its toxicity, and among the different metabolites, hydroquinone (HQ) is one of the most important, as it accumulates in the bone marrow where it can induce several genetic and epigenetic changes. Nevertheless, the actual mechanisms behind the carcinogenetic effects of benzene and HQ and the role of epigenetic alterations in the process of tumorigenesis are not fully clarified yet. Aberrant patterns of DNA methylation, including loss of imprinting (LOI), gene-specific hyper- or hypomethylation and global hypo-methylation are all common in several tumor types, including AML, and are important for transcriptional repression or activation of cancer-associated genes. Most of the studies on the epigenetic effects of benzene conducted so far have been centered on DNA methylation (measured on repeated elements LINE-1 and Alu, used as a surrogate for the entire genome), whereas only very recently a few investigations have analyzed the effects of environmental chemicals on histone modifications. Several studies have shown a strong correlation between DNA methylation and histone modifications, indicating a cross-talk between DNMTs and histone modifications. For example, trimethylation of H3K27 (H3K27me3) can be associated with increased DNA methylation while de novo methylation does not occur in regions where bi-or trimethylation of the histone 3 lysine 4 (H3K4me2 or H3K4me3) is observed. Conversely, the simultaneous presence of H3K27me3 and H3K4me3 is recognized as a bivalent mark; it was firstly described during development and it appears to play a role in poisoning non-active genes for aberrant transcription in cancer. Most notably, epidemiological studies on healthy subjects have shown a difference in the pattern of methylation between subjects exposed to benzene (gasoline station attendants) compared to office workers, with a significant reduction in the methylation of LINE-1 and Alu. This profile of global hypomethylation following exposure to low doses of benzene appears to be qualitatively comparable to what observed in AML and other neoplasms. Thus, in the present work we attempted to reproduce in vitro the conditions of chronic exposure to benzene found in the peripheral blood of subjects exposed. We started from “low” doses of hydroquinone in vitro (below 15 μM) and we proceeded by progressively lowering the concentration to explore the cellular response to different types of exposure and concentrations, to evaluate if HQ might be able to alter the epigenetic signature in two different biological systems, thereby describing a poorly explored step in the mechanism of toxicity associated with benzene exposure. We finally set up a chronic treatment 4 weeks-long with HQ 1 μM, corresponding to 110 ng/mL, a concentration within the amount of total HQ (between 20 and 120 ng/mL, corresponding to 2–16 ng/mL of free HQ) found in the blood of subjects exposed to airborne benzene ranging from 1 mg/m3 (around 0,3 ppm) to 80mg/m3 (around 25 ppm). This concentration was used on a stabilized cell line of human acute promyelocytic leukemia (hAPML), HL60, analyzed as a model of hematopoietic cells; exploring the epigenetic events occurring in chromatin, we found the instauration of the distinctive signature combining the repressive H3K27me3 and the activating H3K4me3, with the gradual increase in H3K4me3 levels, on LINE-1 promoter region. We observed the absence of statistically significant variations in DNA methylation and expression levels of LINE-1, despite a decrease in protein levels of UHRF1, DNA methyl-transferases and histone methyl-transferases. Moving onto a model of normal stem cells, we had to further lower the concentration to avoid cytotoxicity. Thus, human umbilical cord mesenchymal stem cells (hUCMSC) were treated for 4 weeks with HQ 0,1 μM to verify if long-term exposure to low doses of HQ could be able to alter the epigenetic signature in staminal cells outside the hematopoietic compartment. Surprisingly, we found a progressive increase of H3K4me3 in our control cells, with the instauration of the bivalent mark at the third week; the fourth week was non-evaluable, since the cytotoxic effect was prevalent despite the reduced concentration. Most interestingly, at the third week we observed the peculiar inversion in the levels of H3K27me3 and H3K4me3 on LINE-1 promoter region, with higher H3K27me3 in HQ treated cells as compared to the control, and opposite to what observed in HL60. On the other hand, as for HL60 cells no statistically significant variations in DNA methylation was appreciable. These differences were seen not only on chromatin but also on the profile of mRNA expression: preliminary Principal Component Analysis showed a significant distance between controls in the three biological replicates under consideration. In conclusion, in vitro treatment with low-dose HQ determined the instauration of a poisoned state of chromatin in LINE-1 sequences in both the models considered, suggesting that prolonged exposure could cause persistent epigenetic alterations.