Oxidative Stress and Nuclear Reprogramming: A Pilot Study of the Effects of Reactive Oxygen Species on Architectural and Epigenetic Landscapes

Cell genome integrity is continuously threatened by various sources, both endogenous and exogenous. Oxidative stress causes a multitude of damages, severely affecting cell viability, fidelity of genetic information inheritance, and determining profound alterations in gene expression. Epigenetics represents a major form of gene expression modulation, influencing DNA accessibility to transcription factors and the overall nuclear architecture. When assessing the stress-induced epigenome reprogramming, widely diffused biochemical and molecular approaches commonly fail to incorporate analyses such as architectural chromatin alterations and target molecules precise spatial localization. Unveiling the significance of the nuclear response to the oxidative stress, as well as the functional effects over the chromatin organization, may reveal targets and strategies for approaches aiming at limiting the impact on cellular stability. For these reasons, we utilized potassium bromate treatment, a stressor able to induce DNA damages without altering the cellular microenvironment, hence purely modeling nuclear oxidative stress. By means of high-resolution techniques, we described profound alterations in DNA and histone epigenetic modifications and in chromatin organization in response to the reactive oxygen species.


Introduction
Free radicals effect on organisms is a central topic in biology and it is intensively studied because of its importance in several fields, such as cellular signaling, ageing, and human pathologies [1]. In case of imbalance in the redox homeostasis, an oxidative stress condition is established with deleterious effects on biomolecules and detrimental consequences on cells and organisms [2]. High ROS levels, capable of overwhelming cellular antioxidant defenses, can be determined by exogenous sources, such as radiations, anticancer drugs and transition or heavy metals, or by endogenous causes [3].
ROS are important players in the initiation and progression of multiple forms of cancer and inflammatory diseases [4]. Furthermore, high levels of ROS have been found in metabolic diseases such as obesity and type 2 diabetes (T2D), whose complexity is reflected by the many organs and tissues contemporaneously affected; in endocrine diseases, such as Cushing's syndrome; in autoimmune thyroid diseases: for instance, Hashimoto's thyroiditis and Graves' disease have been associated to oxidative stress, but also alteration in the pattern of DNA methylation. Overall, ROS may contribute to the pathogenesis through both genetic and epigenetic mechanisms, for example by altered gene expression modulation due to hypermethylation or hypomethylation. For these reasons, oxidative stress may represent a relevant therapeutic target, as antioxidants may effectively be beneficial in associated with transcriptional silencing, such as trimethylation of lysine 9 on histone H3 (H3K9me3); on the contrary, several epigenetic markers are instructive for gene expression, for instance acetylation of lysine 9 on histone H3 (H3K9Ac) [28][29][30][31][32][33].
Methylation and demethylation are frequently observed to be deregulated in various diseases and cancer types, due to their importance in gene expression regulation. DNA methylation can change metabolism toward the development of multiple chronic syndromes [34]. Peroxisome proliferator-activated receptor alpha (PPAR-α) is a nuclear receptor expressed in active metabolic tissues such as liver, heart, brown fat, and skeletal muscle, where it regulates the expression of genes involved in lipid metabolism. A positive correlation has been found between the PPARA (gene encoding PPAR-α) DNA methylation level and the metabolic syndrome index, triglyceride concentrations and HOMA-IR (HOmeostatic Model Assessment for Insulin Resistance) [35]. Furthermore, adipose tissue inflammation, which is an important factor in the development of obesity and metabolic syndrome, has been associated with decreased DNA methylation of the gene encoding tumor necrosis factor (TNF) in subjects with metabolic syndrome compared to healthy subjects [36]. Indeed, TNF and many other inflammatory mediators are commonly found to be upregulated in insulin resistance and obesity [37]. Given the crucial role played by chromatin organization in modulating gene expression regulation, elucidating its architectural characteristics is of critical importance to understand genome function in health and disease. Therefore, this research aimed at investigating DNA methylation and demethylation, histones post-translational modifications, and the influence of epigenetics on chromatin organization under the effects of oxidative stress, which is notably implicated in several diseases, aging, and carcinogenesis.

DNA Methylation and Demethylation Pattern Alteration following KBrO 3 Treatment
Ultrastructural imaging followed by semi-quantitative analyses revealed significant alterations in the presence of 5-methylcytosine and in its oxidized form 5-hydroxymethylcytosine. Our results highlight a significant reduction in the presence of 5mC. On the contrary, we report an increment in 5hmC abundancy ( Figure 1).
Coherently with 5mC reduction, our data show that KBrO 3 treatment significantly reduces DNA methyltransferase 3a (DNMT3a) presence and lower values are also maintained in the recovery phase. Consistently with the reported increment in 5hmC abundancy, our analysis indicates a significant increase in TET2 levels ( Figure 2). Coherently with 5mC reduction, our data show that KBrO3 treatment significantly reduces DNA methyltransferase 3a (DNMT3a) presence and lower values are also maintained in the recovery phase. Consistently with the reported increment in 5hmC abundancy, our analysis indicates a significant increase in TET2 levels ( Figure 2). Coherently with 5mC reduction, our data show that KBrO3 treatment si reduces DNA methyltransferase 3a (DNMT3a) presence and lower value maintained in the recovery phase. Consistently with the reported incremen abundancy, our analysis indicates a significant increase in TET2 levels ( Figure   ( Figure S1). (d,e) Histograms representing the abun of DNMT3a and TET2 per nucleus. Asterisks identify colloidal gold counting-revealed sign differences. Mean with SEM are reported (*: p Value < 0.05; ***: p Value < 0.001).

The Histone Code Reflects Cell Transcriptional Activity
Then, we examined the distribution and performed statistical analyses asse transcriptionally permissive or repressive histone markers: acetylated H3K9 methylated H3K9, and tri-methylated H3K27, respectively.
Potassium bromate induces a statistically significant reduction in the amount repressive histone markers H3K9me3 and H3K27me3. On the contrary, our data that the treatment determines an increase in the transcription permissive hi modification H3K9Ac (Figure 3).  Figure S1). (d,e) Histograms representing the abundancy of DNMT3a and TET2 per nucleus. Asterisks identify colloidal gold counting-revealed significant differences. Mean with SEM are reported (*: p Value < 0.05; ***: p Value < 0.001).

The Histone Code Reflects Cell Transcriptional Activity
Then, we examined the distribution and performed statistical analyses assessing transcriptionally permissive or repressive histone markers: acetylated H3K9, tri-methylated H3K9, and tri-methylated H3K27, respectively.
Potassium bromate induces a statistically significant reduction in the amount of the repressive histone markers H3K9me3 and H3K27me3. On the contrary, our data show that  Figure S1). (e-g) Histograms representing the abundancy of the histone PTMs per nucleus. Asterisks identify colloidal gold counting-revealed significant differences. Mean with SEM are reported (**: p Value < 0.01; ***: p Value < 0.001).

EZH2 Localization Is Altered following Oxidative Insult
Next, we focused our attention on EZH2, analyzing its expression and cellular localization by combining qPCR and super-resolution microscopy.
Since qPCR indicates that EZH2 mRNA abundancy is not directly affected by the experimental conditions, we investigated whether the protein localization may be altered. Dual stimulated emission depletion (STED) microscopy observations allowed precise EZH2 protein localization as well as structural analysis of the chromatin organization. Analysis of untreated cells shows that EZH2 preferentially localizes at the borders of heterochromatin and in nuclear regions mostly devoid of chromatin. When assessing the oxidative stress condition, a slight decrease in EZH2 nuclear abundancy is revealed; the signal also appears mostly diffused in the nucleoplasm compared to the spot-like signal condition typical of untreated cells. However, differently from the control, EZH2 is also clearly visible in the cytoplasm, that is verified also during the recovery ( Figure 4).  Figure S1). (e-g) Histograms representing the abundancy of the histone PTMs per nucleus. Asterisks identify colloidal gold counting-revealed significant differences. Mean with SEM are reported (**: p Value < 0.01; ***: p Value < 0.001).

EZH2 Localization Is Altered following Oxidative Insult
Next, we focused our attention on EZH2, analyzing its expression and cellular localization by combining qPCR and super-resolution microscopy.
Since qPCR indicates that EZH2 mRNA abundancy is not directly affected by the experimental conditions, we investigated whether the protein localization may be altered. Dual stimulated emission depletion (STED) microscopy observations allowed precise EZH2 protein localization as well as structural analysis of the chromatin organization. Analysis of untreated cells shows that EZH2 preferentially localizes at the borders of heterochromatin and in nuclear regions mostly devoid of chromatin. When assessing the oxidative stress condition, a slight decrease in EZH2 nuclear abundancy is revealed; the signal also appears mostly diffused in the nucleoplasm compared to the spot-like signal condition typical of untreated cells. However, differently from the control, EZH2 is also clearly visible in the cytoplasm, that is verified also during the recovery (Figure 4).    Figure S1). (d) Histograms representing the normalized fluorescence intensity of EZH2 signal in the nucleus and in the cytoplasm. Note the nuclear labelling reduction and the simultaneous cytoplasm signal increment during KBrO3 stress. Asterisks identify statistically significant differences (*: p Value < 0.05; **: p Value < 0.01; ***: p Value < 0.001).

Oxidative Stress and Chromatin Reorganization
Then, we investigated the effects of KBrO3-induced oxidative stress on chromatin organization. First, transmission electron microscopy analysis highlighted the formation of several, small, condensed chromatin clusters. It is noteworthy to report that these data confirm the previously reported results obtained with STED microscopy. At the same time, observation of the cytoplasm did not highlight severely compromised organelles, as their ultrastructure is left mainly unaltered, hence supporting data according to which KBrO3 does not alter the cellular microenvironment.
Based on the chromatin rearrangement, live cell imaging has been performed to monitor the discussed chromatin reorganization step-by-step. The timelapse displays the gradual chromatin condensation in clusters during the experimental condition. Notably, after removal of the oxidizing agent, the nuclear compartment reverts to the initial organization ( Figure 5).  Figure S1). (d) Histograms representing the normalized fluorescence intensity of EZH2 signal in the nucleus and in the cytoplasm. Note the nuclear labelling reduction and the simultaneous cytoplasm signal increment during KBrO 3 stress. Asterisks identify statistically significant differences (*: p Value < 0.05; **: p Value < 0.01; ***: p Value < 0.001).

Oxidative Stress and Chromatin Reorganization
Then, we investigated the effects of KBrO 3 -induced oxidative stress on chromatin organization. First, transmission electron microscopy analysis highlighted the formation of several, small, condensed chromatin clusters. It is noteworthy to report that these data confirm the previously reported results obtained with STED microscopy. At the same time, observation of the cytoplasm did not highlight severely compromised organelles, as their ultrastructure is left mainly unaltered, hence supporting data according to which KBrO 3 does not alter the cellular microenvironment.
Based on the chromatin rearrangement, live cell imaging has been performed to monitor the discussed chromatin reorganization step-by-step. The timelapse displays the gradual chromatin condensation in clusters during the experimental condition. Notably, after removal of the oxidizing agent, the nuclear compartment reverts to the initial organization ( Figure 5).

Oxidative Stress Induction Assessment
As further confirmation about the efficacy of the approach, we tested the efficie of the KBrO3 treatment by measuring cellular nitrotyrosine level as biomarker of oxida stress. Widefield fluorescence quantitative analysis highlighted a significant incremen the nitrotyrosine labelling during the potassium bromate treatment compared to control and recovery conditions ( Figure 6). This relative measurement confirms effectiveness of the KBrO3 as oxidative stress inducer.

Oxidative Stress Induction Assessment
As further confirmation about the efficacy of the approach, we tested the efficiency of the KBrO 3 treatment by measuring cellular nitrotyrosine level as biomarker of oxidative stress. Widefield fluorescence quantitative analysis highlighted a significant increment in the nitrotyrosine labelling during the potassium bromate treatment compared to the control and recovery conditions ( Figure 6). This relative measurement confirms the effectiveness of the KBrO 3 as oxidative stress inducer.

Discussion
In this work we assessed the effect of KBrO3-induced oxidative stress on nuclear epigenetic phenomena and chromatin modifications. We selected HeLa cells for their high metabolic rate, in order to maximize the response at the level of gene expression modulation; moreover, owing to its high expression profile, this model could partially mimic accessible chromatin dynamics already identified during endocrine pancreas development and several other endocrine cells differentiation [38,39].
Taken together, our data demonstrate that KBrO3 treatment-induced oxidative stress profoundly alters chromatin organization and several epigenetic patterns in cell nuclei. After evaluating treatment efficiency measuring ROS level by nitrotyrosine assessment, a promising biomarker for oxidative stress [40,41], we have shown significant reductions in 5mC and in its corresponding DNA-targeting enzyme DNMT3a. Moreover, to better understand epigenetic regulation of chromatin structure, we investigated histone PTMs during the experimental conditions, considering both PTMs instructive for gene silencing (H3K9me3 and H3K27me3) and transcription (H3K9Ac). Coherently with data obtained from DNA methylation analysis, statistically significant drops in the histone markers for silenced chromatin regions H3K9me3 and H3K27me3 have been described. On the contrary, a raise in H3K9Ac levels has been identified, further confirming the idea that KBrO3 induces a positive gene expression epigenetic profile. Finally, live cell imaging allowed the progressive observation of small heterochromatin clumps formation during the treatment, as well as their dissolving in the recovery, suggesting chromatin rearrangement for specific silencing and activation of genes according to their function during this source of stress. Paired with dual STED approach, which associates analyses conducted at the level of DNA with findings about Polycomb Repressive Complex 2 catalytic subunit EZH2, we revealed a characteristic nuclear distribution of EZH2, which mainly localizes in areas devoid of heterochromatin clumps and at the borders of these

Discussion
In this work we assessed the effect of KBrO 3 -induced oxidative stress on nuclear epigenetic phenomena and chromatin modifications. We selected HeLa cells for their high metabolic rate, in order to maximize the response at the level of gene expression modulation; moreover, owing to its high expression profile, this model could partially mimic accessible chromatin dynamics already identified during endocrine pancreas development and several other endocrine cells differentiation [38,39].
Taken together, our data demonstrate that KBrO 3 treatment-induced oxidative stress profoundly alters chromatin organization and several epigenetic patterns in cell nuclei. After evaluating treatment efficiency measuring ROS level by nitrotyrosine assessment, a promising biomarker for oxidative stress [40,41], we have shown significant reductions in 5mC and in its corresponding DNA-targeting enzyme DNMT3a. Moreover, to better understand epigenetic regulation of chromatin structure, we investigated histone PTMs during the experimental conditions, considering both PTMs instructive for gene silencing (H3K9me3 and H3K27me3) and transcription (H3K9Ac). Coherently with data obtained from DNA methylation analysis, statistically significant drops in the histone markers for silenced chromatin regions H3K9me3 and H3K27me3 have been described. On the contrary, a raise in H3K9Ac levels has been identified, further confirming the idea that KBrO 3 induces a positive gene expression epigenetic profile. Finally, live cell imaging allowed the progressive observation of small heterochromatin clumps formation during the treatment, as well as their dissolving in the recovery, suggesting chromatin rearrangement for specific silencing and activation of genes according to their function during this source of stress. Paired with dual STED approach, which associates analyses conducted at the level of DNA with findings about Polycomb Repressive Complex 2 catalytic subunit EZH2, we revealed a characteristic nuclear distribution of EZH2, which mainly localizes in areas devoid of heterochromatin clumps and at the borders of these clusters, with a significant reduction of its level and a contemporary increment in the cytoplasmic signal during the stress condition. While our data clearly identify a form of EZH2 migration from the cell nucleus to the cytoplasm, qPCR does not highlight statistical differences in its gene expression level. We hypothesize this discrepancy to be due to the experimental parameters and the specificity of KBrO 3 treatment as not excessively altering the cellular microenvironment. Indeed, Polycomb Repressive Complexes have already been identified as involved in epigenetic response to certain type of stress [42,43], hence we could speculate that by increasing the exposure time or concentration of potassium bromate a rapid increment in EZH2 would be detected also by qPCR. Remarkably, our experiment shows a trend consisting in a constant gene expression increment, almost reaching significant values the longer the time point advanced from the stress source exposure. Since other groups reported similar expression parameters and, notably, a cytoplasmic EZH2 relocalization [44,45], it is tempting to speculate a constant increment in EZH2 mRNA abundancy (as already identified by microscopy and western blot quantifications at the protein level), but further studies are required to clarify this question. Together with the findings from these groups, our data suggest a main role of EZH2 and of the Polycomb-group family in reshaping the nuclear architecture and in gene expression regulation during the experimental conditions.
Our data agree with numerous other findings. First, our results confirmed extensive chromatin reorganization under oxidative stress conditions, as described by Madugundu and colleagues, validating the observation of a specific increment in 5mC oxidation products [46]. Moreover, our data are also in accordance with Kietzman and colleagues, who described the importance of ROS in altering the epigenetic landscape, affecting DNA methylation, post-translational histone modifications, and non-coding RNA transcripts and their effect on cardiovascular diseases [47]. Furthermore, the chromatin rearrangement following oxidative stress shown here is in agreement with various works [48,49], that reported profound nuclear reorganization caused by oxidative stress in neurodegenerative models. ROS effect on histones represents another important topic: since histones are the most abundant chromatin proteins, any change in their PTMs have a critical outcome on chromatin structure, hence influencing genome stability and gene expression [50]. We linked KBrO 3 effect with a general reduction in repressive PTMs, while transcription-instructive ones are statistically increased. Coherently, a similar gene expression upregulation scenario was described by García-Giménez and colleagues, although they focused their attention on histone H3 glutathionylation only [51]. Furthermore, multiple studies described a putative role for ROS as enhancers of histone deacetylases activity [52,53]; nonetheless, this aspect is still largely debated [50]. Thus, this work integrated in a series of research activities highlights the relevance-gaining topic of chromatin reorganization after stress stimuli: preservation of genome stability and gene expression regulation are essential features concerning significant issues such as the way cells respond to damage, pathophysiological conditions, or aging [54,55].
The process of mRNA transcription constitutes a critical step of the gene flow information and employs multiple essential regulators to produce a faultless cellular proteome to maintain healthy physiology of the whole organism. Therefore, it is not surprising that any abnormalities, caused by external or internal insults, may contribute to a perturbed homeostasis and consequently human diseases. Acute oxidative stress is one of the major causes of such imbalances and it is known to induce epigenetic changes and dysregulation of gene expression in pathological cases and in ageing [56,57]. For instance, obesity has gained huge importance in the past decades as the number of people with condition is constantly rising. Obesity holds its roots in a complex lifestyle-gene interaction, deeply interconnected to epigenetic alterations. Since preventive measures are often reported to fail, interest in new approaches is rising, for example in the field of nutraceuticals using antioxidants to overcome acute oxidative stress [58,59]. Similarly, ROS are key players in development of low-grade inflammation, which in turn is involved in the physiopathology of polycystic ovary syndrome (related to insulin-resistance, immune dyscrasia and much more), as well as in metabolic syndrome-related manifestation, as atherosclerosis, hypertension, and T2D: hence, a profound understanding of the link between onset of such conditions and their genetic bases (for instance, at the level of the histone code and overall chromatin organization) could be reflected in discovery of new targets and development of new therapies aiming at ameliorating patients' life conditions [60,61]. Inhibiting oxidative damage may have therapeutic benefits, thus administration of vitamins, polyphenols, and antioxidant molecules is reasonably reported to have ameliorating effect on patients [62][63][64][65]. Moreover, several epigenetic drugs are already being tested and in the near future their combination with antioxidants treatments may pave the way for novel possibilities [66][67][68]. For instance, in the recent years interest in chronic oxidative stress is rising, motivated by the modern focus on space technology and transportation. Of particular interest will be the impact on the health of astronauts: combining basic data from research in genetic and medical fields with validated antioxidant-integrated dietary regime could be a critical [69][70][71]. Integration of endocrine tissue-, gene-, and oxidative stress-specific studies of epigenetic modifications will deepen the knowledge and strengthen the therapeutical approaches, positively impacting drug design and therapies to ameliorate conditions and increase the lifespan of patients and people with high risk of endocrine diseases [72,73].
48 h before experiments, cells were grown in 25 cm 2 flasks. At approximately 70% confluence, cells were subjected to potassium bromate treatment by replacing the culture medium with fresh medium containing 40 mM KBrO 3 for 1 h at 37 • C; for the recovery condition, after the treatment cells were allowed to recover in fresh medium for 1 h at 37 • C [16].

Transmission Electron Microscopy Sample Preparation
Control and treated cells were harvested by mild trypsinization with 0.25% trypsin in PBS containing 0.05% ethylene diamine tetraacetic acid (EDTA), centrifuged at 800 rpm for 5 and then fixed for morphological analysis or for immunocytochemistry.
In both cases, ultrathin sections (60-80 nm) were cut on a Reichert OM-U3 ultramicrotome and collected on nickel grids (200 Mesh). The specimens were observed with a JEM 1200 EX II (JEOL, Peabody, MA, USA) electron microscope operating at 100 kV and equipped with a MegaView G2 CCD camera (Olympus OSIS, Tokyo, Japan).

Transmission Electron Microscopy Immunocytochemistry and Sample Staining
Grids for immunocytochemical analyses were first incubated in a drop of normal goat serum (NGS) diluted in PBS and then incubated over-night (ON) in a drop of primary antibody (Table 1). Then, grids were rinsed twice in PBS/TWEEN20 for 5 and twice in PBS for 5 , incubated in NGS and finally in the proper colloidal gold-particle conjugatedsecondary antibodies for 30 at RT. Grids were rinsed twice in PBS for 5 and twice in dH 2 O for 5 .

Antibody (Host) Target Dilution
Anti-Mouse 12 nm gold particle conjugated (goat) Mouse IgG 1:20 in PBS 115-205-068-AB_2338730 (Jackson ImmunoResearch) Staining for immunocytochemistry was performed according to the EDTA regressive technique [74]. Grids were incubated in 4% uranyl acetate solution at RT for 2 min, followed by a 30 s treatment in 0.2 M EDTA at RT. Then, grids are incubated in the lead citrate solution for 2 min at RT. Every step is followed by rapid rinsing in dH 2 O.
Staining for morphological analysis was achieved via uranyl acetate and lead citrate staining [75]. Grids were incubated in a 4% uranyl acetate solution at room temperature (RT) for 10 min. Then, several rinses were rapidly performed in dH 2 O. Subsequently, grids were incubated in lead citrate at RT for 2 min, followed by rapid rinses in dH 2 O.

Widefield and Confocal Microscopy Sample Preparation and Immunocytochemistry
Treated and control HeLa cells were grown on coverslips and fixed with 4% formalin for 20 at RT. Cells were then post-fixed with 70% ethanol at −20 • C for at least 24 h. Samples were rehydrated for 10 in PBS at RT and processed for DNA staining or immunocytochemistry. DNA was stained with either 0.1 µg mL −1 Hoechst 33258 for 5 , followed by two rinses in PBS of 5 each or SPY505-DNA 20X (Spirochrome, Stein am Rhein, Switzerland) followed by a quick rinse in dH 2 O. For immunocytochemistry, cells were incubated in the blocking buffer for 20 and then in the primary antibody (Table 2) for 1 h at RT. Then, after three washes in PBS of 5 each, samples were incubated with the proper secondary antibody for 45 at RT. After 5 washes in PBS of 5 each, DNA was stained with Hoechst 33258 or SPY505-DNA 20X as already described. Coverslips were mounted using Mowiol and observed with a Leica TCS SP8 STED 3X (Wetzlar, Germany).

Live Cell Imaging
HeLa cells were grown on coverslip and stained with SPY505-DNA 1X in complete medium. After incubating 1 h at 37 • C, cells were observed with a Leica TCS SP8 STED 3X equipped with the incubation chamber for in vivo imaging. Cells were treated for potassium bromate stress and for the correspondent recovery condition as previously indicated.

Stimulated Emission Depletion Microscopy Sample Preparation and Immunocytochemistry
Treated and control HeLa cells were grown and processed as already described for confocal microscopy. Samples were observed with a Leica TCS SP8 STED 3X equipped with a 660 nm depletion laser.
Samples were boiled at 96 • C for 3 min, centrifuged at 18,000× g and then separated by SDS page in a discontinuous system composed of a stacking gel (3%) and a running gel (6%, 10%, and 15%) according to the proteins molecular weight. Precision Plus Protein Standards All Blue (Bio-Rad, Hercules, CA, USA) were loaded as ladder. Electrophoresis was performed in an electrophoretic chamber MINI-PROTEAN III (Bio-Rad) applying a constant voltage of 150 V for 60 and 120 min.
Proteins were transferred from the gel on a polyvinylidene difluoride (PVDF) (Bio-Rad) using an electrophoretic chamber MINI-PROTEAN III (Bio-Rad) applying a constant electric current of 200 mA for 2 h. After 60 incubation in a solution of 5% BSA in Trisbuffered saline (TBS), PVDF was incubated in primary antibodies (see Tables 1 and 2) diluted 1:1000 in BSA 5% in TBS, NaN3 0.02%, Tween20 0.1% ON at 4 • C. After five rinses in washing buffer (PBS + Tween20 0.1%) for 5 , PVDF was incubated with the proper horseradish-peroxidase-conjugated secondary antibody diluted 1:2000 (anti-rabbit) or 1:5000 (anti-mouse) in washing buffer for 60 at RT. After rinsing in washing buffer, proteins were finally visualized using an enhanced chemiluminescence (ECL) kit and the light emission was detected using a ChemiDoc XRS (Bio-Rad).

Real Time PCR (qPCR)
Treated and control cells were detached by mild trypsinization and centrifuged at 800 rpm for 5 . RNA was extracted (RNeasy Kit, Qiagen, Hilden, Germany) and quantified using the Nanodrop spectrophotometer (Bio-Rad).
mRNA was retrotranscribed using the Superscript IV protocol (Invitrogen), heating at 65 • C for 5 50 ng/uL random examers, 10 mM dNTPs mix and template RNA in RNase-free water. After incubation on ice, the Superscript IV buffer, 100 mM dithiothreitol, ribonuclease inhibitor and the Superscript IV reverse transcriptase (200 U/Ul) were added and the thermal-cycler was run under the following parameters: 23 • C for 10 , 55 • C for 10 and 80 • C for 10 . The analysis was conducted assessing EZH2 expression using RPS18 and GADPH as reference genes. Primers were selected using following Primer3 setup: 18-20-23 nucleotides length, 58-60-63 • C melting temperature with a maximum difference of 2 • C between forward and reverse primers, 100-300 product size range, resulting in finale 281 product size. EZH2 forward sequence (5 to 3 ): GCGGATAAAGACCCCAC-