The efficacy of MC3343 was tested in 11 human patient-derived EWS cell lines, including 8 conventional, and 3 novel cell lines from patient-derived xenografts (PDX). The IC₅₀ values, calculated after 72h of cell exposure to the compound, were similar in all the cell lines, ranging from 2 to 5 µM ( Table 1 ). The agent induced a dose-dependent reduction of the expression of DNMT1, but the effect on DNMT3a was much milder ( Figure 1A ), indicating a more specific mechanisms of action when compared to the conventional inhibitor of DNA methylation, 5AzadC, that was used as a control. Consistently, MC3343 was similar to or even more potent than 5AzadC in inhibiting DNMT enzyme activity ( Figure 1B ). However, when we used a genome-wide methyl-CpG array to compare the DNA methylation status of three EWS cell lines (TC-71, A-673 and PDX-EW#4-C) after 72h treatment with either MC3343 or 5AzadC, a profound discrepancy between the two drugs was observed. In fact, while cell exposure to 5AzadC resulted in extensive reduction of CpG methylation in all three cell lines, cells treated with MC3343 revealed a methylation pattern indistinct from that observed in untreated cells ( Figure 1C ). This indicates a different mechanism of action between the two compounds that deserves to be further analyzed, because although 5AzadC is approved by the US FDA for the treatment of myelodysplasia and other malignancies, it has demonstrated unsatisfactory results in the treatment of EWS patients (39).

The depletion of DNMT1 leads to the removal of the enzyme from replication forks, which was demonstrated to result in DNA strand breaks and the activation of genotoxic stress pathway signaling, leading to further DNA damage, γH2AX foci formation and cell death (40, 41). Thus, we analyzed the impact of MC3343 treatment on EWS cells on these biological processes. Using double labeling with propidium iodide to mark S-phase cells and bromodeoxyuridine (BrdU) to label cells actively synthesizing DNA, we discovered that MC3343 treatment triggers a significant accumulation of cells in the G1 ( Figure 2 ). As expected, the percentage of cells in G1 phase was higher in the faster proliferating TC-71 and A-673 cells (doubling time: 18-20h) compared to LAP-35 and PDX-EW#4-C (doubling time: 40-50h) (p < 0.05 Student’s t test). Besides effects on the cell cycle, MC3343 induced the formation of histone variant γH2AX foci, a characteristic marker of cells undergoing replication stress and DNA strand breaks (42) with much faster kinetics as compared to 5AzadC. In fact, treatment of EWS cells with MC3343 led to an induction of γH2AX foci starting after 3h with a peak at 6h; while in cells treated with 5AzadC, the appearance of γH2AX foci formation occurred only after 24h and not in all cell lines ( Figure 3 and Supplementary Figure S1 ). We also evaluated the impact of MC3343 on genomic instability in four EWS cell lines. Control and treated cells were cultured for 6 to 24 hours, and the impact on genome stability monitored by evaluating micronuclei, a well-recognized marker of genomic instability (43) ( Figures 4A, B ). This effect was better represented in p53 mutated TC-71 and A-673 cells than in p53wt LA-P35 and PDX-EW#4-C cells (p< 0.0001 Student’s t test). In addition, the percentage of cells with micronuclei following exposure to MC3343 was maintained in p53 mutated cells even after removing the compound for at least two cell doublings (rescue for 48h-96h), indicating that in p53 mutated cells (TC-71 and A-673) micronuclei are inherited and persist into the daughter cells. In contrast, when the p53 response is functional (LAP-35 and PDX-EW#4-C), exposure to MC3343 led to activation of p53-dependent signaling ( Figure 5A ) and cells that were more prone to apoptosis, as demonstrated by annexin labeling ( Figure 5B ). Representative western blottings indicated that inactivation of DNMT1 by MC3343 induced phosphorylation of p53 on Ser15, a residue that is targeted by ATM and whose phosphorylation has been correlated with accumulation of p53 and activation of its downstream targets in response to DNA damage (44). Moreover, upregulation of p21 and cleavage of pro-Caspase 3 and PARP1 were consistently observed in p53 wt MC3343-treated PDX-EW#4-C but not in p53 mut TC-71 cells.

We then evaluated the impact of the DNMTi MC3343 on cytotoxicity induced by DNA damaging agents, such as doxorubicin, a widely-used genotoxic drug that intercalates between DNA base pairs and inhibits the enzyme topoisomerase II (45), and talazoparib, a PARPi that functions by impairing DNA repair mechanisms (46, 47). The pretreatment of cells with the MC3343 increased the cytotoxicity of both doxorubicin and talazoparib ( Figure 6A ). Combination indexes ranged from moderate to strong synergism ( Figure 6B ) and the formation of micronuclei that are prone to rupture (γH2AX positive) was significantly higher in combined treatments ( Figures 6C, D ).

MC3343 was prepared as previously reported (36, 37). Stock solutions of 5-Aza-2′-deoxycytidine (5azadC) (A3656, Sigma-Aldrich, St. Louis, MA, U.S.), Doxorubicin (D1515, Sigma-Aldrich) and Talazoparib (BMN673) (S7048, Selleckchem, Houston, TX, U.S.) were prepared and stored according to the manufacturer’s instructions.

DNMT activity was quantified with an EpiQuick DNA Methyltransferase Activity/Inhibition Assay Kit (P-3001-1, Epigentek Inc., Farmingdale, NY, U.S.). Cells were treated with MC3343 or 5azadC (3-10 µM) for 48 h, and 5µg of nuclear extracts, isolated with EpiQuick Nuclear Extraction Kit (OP-0002-1, Epigentek Inc), were added to each reaction well, according to the manufacturer’s protocol. Absorbance was determined using a microplate spectrophotometer at 450 nm (GloMax luminometer, Promega, Madison, WI, U.S.). The data are expressed as the mean ± SE of three independent experiments.

A panel of 11 EWS human cell lines were analyzed. A-673 (CRL-1598), SK-ES-1 (HTB-86), SK-N-MC (HTB-10), and RD-ES (HTB-166) were provided by the American Type Culture Collection (ATCC, Manassas, VA, U.S.); TC-71 and 6647 were kindly provided by T.J. Triche (Children’s Hospital, Los Angeles, CA, U.S.). LAP-35 and IOR/CAR, were obtained in the Experimental Oncology Lab of the Rizzoli Institute as previously described (58, 59). The cell lines PDX-EW#2-C, PDX-EW#4-C and PDX-EW#5-C were obtained from patient-derived xenografts (PDXs) as previously described (60). All cell lines were tested for mycoplasma contamination every 3 months (LT07-318, MycoAlert Mycoplasma Detection Kit, Lonza, Basel, Switzerland). Cell lines were authenticated (STR profiling) by analysis of the following loci: AMEL, D3S1358, TH01, D21S11, D18S51, D10S1248, D1S1656, D2S1338, D16S539, D22S1045, VWA, D8S1179, FGA, D2S441, D12S391, D19S433 and SE33 (last control December 2017 and July 2018; POWERPLEX ESX 17 Fast System, Promega, Madison, WI, U.S.). All cell lines were immediately amplified to constitute liquid nitrogen stocks and were never passaged for more than 1 month upon thawing. Cells were maintained in Iscove’s Modified Dulbecco’s Medium (ECB2072L, IMDM, Euroclone, Milan, Italy) supplemented with 10% heat-inactivated FBS (ECS0180L, Euroclone), penicillin (20 U/mL) and streptomycin (100 µg/mL) (ECB3001D, Euroclone) in a 37°C humidified at 5% CO₂.

To perform cell viability experiments, 2-5x10⁵ cells were plated in 6-well dishes and treated after 24h with the indicated drugs for 72h, roughly equivalent to at least two doubling times of each cell line, before Trypan blue (T8154, Sigma-Aldrich) vital cell counting or DNA methylation studies. In parallel, cells were treated with medium containing dimethyl sulfoxide (DMSO) (D2438, Sigma-Aldrich) as a control. The highest final concentration of DMSO in the medium was <0.005%, which had no effect on cell growth.

In sequential drugs combination experiments, cells were plated and pre-treated with MC3343 (3-5 µM) for 24h then the medium was replaced with the drugs at indicated concentrations (doxorubicin 3-30 ng/ml, talazoparib 1-10 nM) or fresh IMDM medium without drugs.

For cell cycle analysis after 24h of treatment, 5-bromo-2’-deoxyuridine (BrdU) (B5002, Sigma-Aldrich) labeling was used to assess cell proliferation. Cells were processed for indirect immunofluorescence using anti‐BrdU mAb (1:8, 347580, BD-Biosciences, Franklin Lakes, NJ, U.S.) primary antibody and anti‐mouse FITC secondary antibody (1:100, 31569, Thermo Fisher Scientific, Waltham, MA, U.S.) and 20 μg/mL propidium iodide (P4170, Sigma-Aldrich) for the analysis of DNA content prior to flow cytometry analysis (Sysmex, Partec GmbH Münster, Germany). Detection and quantification of apoptosis was obtained by cell staining with Annexin V and 7-amino-actinomycin D (7-AAD) (A1310, Thermo Fisher Scientific) using the PE Annexin V detection kit I (Life Technologies Corporation, Monza, Italy) according to the manufacturer’s instructions. Samples were analyzed on an Attune NxT Acoustic Focusing Cytometer (Life Technologies Corporation), and data were analyzed using Attune Cytometric 2.6 software (Life Technologies Corporation). Cells were classified as early apoptotic, late apoptotic or dead if they were Annexin V+/7-AAD−, Annexin V+/7-AAD+ or Annexin V−/7-AAD+, respectively.

DNA was extracted from cell pellets by PureLink genomic DNA kit (K182001, Thermo Fisher Scientific). The DNA methylation profiles of the samples were obtained using the Infinium Methylation EPIC Array (850 K). DNA samples were assessed for their quality using a NanoDrop^® ND-2000 UV-Vis Spectrophotometer (Thermo Fisher Scientific). The samples were separated in agarose gels. Those with intact genomic DNA, showing no smear in the gel, were selected for subsequent experiments. Intact genomic DNA was diluted to 50 ng/μL based on Quant-iT PicoGreen (Invitrogen) quantitation. Concentrations were adjusted according to these results. For bisulfite conversion, 600 ng of input gDNA was required. Bisulfite-modified gDNA was prepared using the EZ DNA Methylation kit (Zymo Research) according to the manufacturer’s instructions. Conversion reagent was added, followed by subsequent incubation in a thermocycler to denature the samples. CT-converted DNA was washed and de-sulfonated with de-sulfonation buffer, after which the DNA was washed again and eluted with 12 μL elution buffer. The whole-genome amplification process required 250 ng of input bisulfite-converted DNA (MA1) and created a sufficient quantity of DNA (1,000X amplification) for use on a single BeadChip in the Infinium methylation assay (Illumina RPM and MSM). After amplification, the product was fragmented using a proprietary reagent (FMS), precipitated with 2-propanol (plus precipitating reagent; PM1), and re-suspended in formamide-containing hybridization buffer (RA1). The DNA samples were denatured for 20 min at 95°C and placed in a humidified container for a minimum of 16 h at 48°C, allowing CpG loci to hybridize with the 50-mer capture probes. Following hybridization, the BeadChip/Te-Flow chamber assembly was placed on a temperature-controlled Tecan flow-through chamber rack, and subsequent washing, extension, and staining were performed by adding reagents to the Te-Flow chamber. For the allele-specific single-base extension assay, primers were extended by polymerase and labeled nucleotide mix (TEM), and then stained by repeated application of STM (staining reagent) and ATM (anti-staining reagent). After staining, the slides were washed with low-salt wash buffer (PB1), immediately coated with XC4, and imaged using the iScan System (Illumina). The iScan System has a two-color (532 nm/658 nm) confocal fluorescent scanner with 0.54 μm pixel resolution. The scanner excited the fluorophores generated during signal amplification/staining of the allele-specific (one color) extension products on the BeadChips. Image intensities were extracted using Illumina’s GenomeStudio Software.

Methylation array data was processed with R statistical language using methods from the Bioconductor. Data quality was assessed using the standard pipeline from the minfi package: Low quality probes (detection pvalue > 0.01) and low-quality samples (> 0.05% low quality probes) were removed from the analysis and known SNPs, sex chromosomes and cross-hybridizing probes were excluded. The data was normalized using ‘ssNoob’ normalization method and, after normalization, m-values and beta-values were calculated. Differential methylation analysis was performed with the limma package (~ 0 + treatment + cell line; single sites, p<0.01) and DMRcate (regions, fdr < 0.01).

Immunofluorescence was performed on adherent cells grown on coverslips, fixed in cold methanol for 7 minutes at -20°C and permeabilized with 0.15% Triton X-100 (T8787, Sigma-Aldrich) in PBS, blocked with 4% Bovine serum albumin (BSA) (A4503, Sigma-Aldrich)/PBS for 1 hour at room temperature and incubated with anti-phospho-histone H2A.X (Ser139) antibody (1:100, 9718, Cell Signaling Technology, Danvers, MA, U.S). Nuclei were counterstained by bisbenzimide Hoechst 33258 (0.5 µg/ml, B2883, Sigma-Aldrich, St. Louis, MA, U.S.). Cell fluorescence was then evaluated using a Nikon Eclipse 90i microscope (Nikon Instruments, Florence, Italy). For quantification purposes, γH2AX spots and micronuclei were quantified from at least 20 random fields (~500 cells), expressed as γH2AX mean fluorescence intensity (MFI) and percentage of cells with micronuclei respectively. Images acquisition and processing were conducted using the NIS-Elements A.R. 3,10 software.

EWS cells were lysed with RIPA buffer (89900, Thermo Fisher Scientific) containing protease inhibitors (A32955, Thermo Fisher Scientific) and phosphatase inhibitors (A32957, Thermo Fisher Scientific). Equivalent amounts of total cell lysates were separated by 8-12% SDS-PAGE under denaturing conditions and transferred onto nitrocellulose membrane (0.45 μm, Bio-Rad, Hercules, CA, U.S.). Membranes were incubated overnight with the following primary antibodies: anti-DNMT1 (1:2000, A300-041A-M lot# 1, Bethyl Laboratories, Montgomery, TX, U.S), anti-DNMT3a (1:1000, sc-20703 lot # L0412, Santa Cruz Biotechnology, Dallas, TX, U.S.), anti-phospho-ATM (S1981) (1:2000, ab36810 lot# gr3185919-27, Abcam, Cambridge, UK), anti-p21 (1:1000, sc-6246 lot # E0511, Santa Cruz Biotechnology), anti-phospho-p53 (Ser15) (1:2000, 9284 lot#9, Cell Signaling Technology), anti-p53 (1:3000, MCA1701 lot #410, Bio-Rad), anti-PARP1 (1:5000, 9542 lot#15, Cell Signaling Technology), anti-cleaved Caspase-3 (Asp175) (1:500, 9661 lot # 38, Cell Signaling Technology) and anti-GAPDH (1:10000, 2118 lot # 14, Cell Signaling Technology). After washes with 1x Tris-buffered saline/Tween20, the membranes were incubated with secondary anti-rabbit or anti-mouse antibodies conjugated to horseradish peroxidase (GE Healthcare, Piscataway, NJ, U.S.) and revealed by SuperSignal West Pico PLUS Chemiluminescent Substrate (34580, Thermo Fisher Scientific).

All statistical analyses were performed using Prism version 7.0 (GraphPad Software, La Jolla, CA, US). IC₅₀ (concentration required to inhibit cell proliferation by 50%) values were calculated from the linear transformation of dose-response curves. Comparison between two groups were evaluated with two-tailed Student’s t-tests. Experimental data including more than 2 groups were analyzed using one-way or two-way ANOVA. The data were considered statistically significant at p <0.05. Drug interactions were analyzed by the combination index (CI) calculated with the fractional product method based on that described by Chou et al. (61). Subclassification of CI values was used (0.1– 0.3 strong synergism, 0.3–0.7 synergism, 0.7–0.85 moderate synergism, 0.85–0.9 slight synergism, 0.9–1.1 nearly additive).