The NSC34 cells were purchased by Tebu-BIO [Magenta, (MI) Italy]. Cells were used between 5–20 passages. Cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM) in the presence of 10% inactivated fetal bovine serum and penicillin/streptomycin at 37°C, 5% CO2 and 95% air, in a cell culture incubator. SH-SY5Y cells were purchased and grown as previously published (Guida et al., 2015a). Methylmercury (II) chloride (MeHg) (cod: 442534 stock solution 100 mM) and Necrostatin-1 (Nec) (cod. N9037; stock solution 20 mM) were purchased from Sigma both obtained from Sigma–Aldrich (St. Louis, MO, United States) and were dissolved as previously reported (Guida et al., 2018). Culture media and sera were purchased from Invitrogen (Milan, Italy). All chemicals were diluted in cell culture medium. Synthetic oligonucleotides were from Eurofins Genomics. For Nec experiments NSC34 cells were pre-incubated for 2 h in full medium with the drug at 1, 5, 10, and 20 μM, followed with MeHg 100 nM for 24 h. For MTT and LDH assays, cells were plated in 24 well plates at a density of 1 × 10⁴ cells/well; for qRT-PCR, they were plated in 60 mm plates at a density of 5 × 10⁴ cells/plate; for Western blot, Immunoprecipitation and ChIP, they were plated in 100 mm plates at 6 × 10⁶ cells/plate.

siRNA for REST (siREST) (EMU029201), siRNA for Sp1 (siSp1) (EMU061231) and siRNA for KMT2A (siKMT2A) (EMU180041) were purchased from Sigma-Aldrich, whereas the negative control (siCTL) (001910-10-05) was purchased from Dharmacon. NSC34 at 60% confluence were transfected with siRNAs (25 nM) by using Lipofectamine 2000 in Optimem medium, according to the protocol by the manufacturer (Invitrogen Srl, S. Giuliano Milanese Italy). For constructs transfection in NSC34 cells have been used the following plasmids: (1) pcDNA3.1, (2) pF151 pcDNA3.1(+)SOD1WT and (3) pF155 pcDNA3.1(+)SOD1-G93A, that were a gift from Elizabeth Fisher (Addgene plasmid # 26397 and # 26401) (Stevens et al., 2010). The amount of DNA constructs used for transient transfection (60% confluence) was: 500 ng DNA in 24-multiwell plates for MTT and LDH assays, 10 μg of DNA in 100 mm dishes for quantitative real time PCR (Q-PCR), Western blot and ChIP analysis. 24 h after transfection low serum medium containing vehicle or 100 nM MeHg was added. Transfection efficiency was evaluated by western blot analysis to demonstrate the presence of the human SOD1 protein and by Q-PCR to assess the levels of REST, Sp1 and KMT2A mRNAs after siRNAs treatment (Chong et al., 2021).

Cellular vitality is related to mitochondrial function which was assessed by measuring the activity of mitochondrial dehydrogenases. The soluble compound 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) was used as substrate. Mitochondrial dehydrogenases reduce MTT by converting it into formazan, a salt that is insoluble in water and optically active at 540 nm. The amount of formazan produced is directly proportional to the metabolic activity of mitochondrial dehydrogenases and therefore, indirectly, to cell viability. At the end of the exposure to MeHg, the culture medium was removed and the cells were incubated in 2 ml of a PBS solution containing MTT at a concentration of 0.5 mg/ml for 1 h at 37°C in an atmosphere consisting of 5% CO2 and 95% air. At the end of the hour, the incubation was interrupted by removing the solution and adding 1 ml of dimethyl sulfoxide (DMSO) to solubilize the formed formazan salts. The results were obtained by spectrophotometric analysis and were expressed as the percentage of survival of cells treated with MeHg compared to the values obtained from cultures exposed to the vehicle (Veh) to which the 100% survival value was attributed.

Cell death was evaluated with the lactate dehydrogenase (LDH) cytotoxicity kit from Cayman, as previously reported (Guida et al., 2018). NSC34 cells have been pre-treated for 2 h with Nec at 1, 5, 10, and 20 μM or with vehicle, then they have been exposed to MeHg 100 nM for 24 h. Cells treated with 1% Triton X-100 (Sigma) were used as positive control of cell death.

Total RNA was extracted using Trizol according to the manufacturer’s instructions and quantified spectrophotometrically. The retro-transcription was performed as previously reported (Formisano et al., 2014). The reaction mixture was incubated for 60 min at 42°C and 2 min at 94°C, followed by 35 cycles of 30 s at 95°C, 30 s at 60°C and 30 s at 72°C. Normalization of the results was performed by β-Actin (Actin). The oligonucleotide sequences were: REST (m): (FW 5′-GTTCAGCACGTGCGAACTCA-3′ and RV 5′-TCCGCATGTGTCGCGTTA-3′); for Sp1 (m): (FW 5′-GCCTGTTAGGAGGTCCCTGAA-3′ and RV 5′-AGGCTGCCCATTTGTACTCATTTA-3′); for Sp3 (m): (FW 5′-CCTGACACAGACCCCTGTTGA-3′ and RV 5′-CAGTG CTCGCATCTGTGGAA-3′); for KMT2A (m): (FW 5′-GACCGAATCGGACTAAACACTTTC-3′ and RV 5′-GTGACCTTGCCTAGTAATCGAACTT-3′); for CREB (m): (FW 5′-TACTTCTCCATTCTCTCCTTCTCA-3′ and RV 5′-TGTAC GCGAGCGAGAGG-3′), for JunD (m): (FW 5′-GCTCTCT CTCCTCCCGACA-3′ and RV 5′-GGCGAACCAAGGATTAC GGA-3′) for β-Actin (m): (FW 5′-GTCGTACCACTG GCATTGTG-3′ and RV 5′-AGAGGGTAATCGTGCGTGAC-3′). Changes in gene expression between groups were represented as the mean of the relative quantification (RQ) values, that was calculated as the difference in threshold cycle (ΔCt) between the target gene and the reference gene (2^–ΔΔCT = RQ).

Protein expression was evaluated using standard Western Blotting methods. After the treatments, the cells were collected and centrifuged for 30 min at 14,000 RPM. The cells were washed only once with cold PBS, then a lysis buffer was added: Tris–HCl pH 7.5 (1 M), NaF (0.5 M), NaCl (4 M), PMSF (100 mM), Sodium Vanadate (100 mM), NP-40 (20%) plus protease inhibitor cocktail. The determination of the protein content of the samples was carried out according to the Bradford method. Aliquots containing 80 μg of proteins were fractionated by polyacrylamide gel electrophoresis (BIORAD Laboratories, Milan, Italy) in SDS (0.1%) (SDS-PAGE) and the separated proteins were transferred to a nitrocellulose filter. The filters were blocked for 2 h at room temperature in a blocking solution prepared with 0.1% Tween-20 (FlukaChemie GmbH, Italy), in 5% casein (Bio-Rad Laboratories, Milan, Italy), dissolved in TBS tween (Bio-Rad Laboratories, Milan, Italy).

Immunoprecipitation of KMT2A in MeHg-treated SOD1-G93A cells was performed as previously described (Guida et al., 2015b). Briefly, cell lysates (1 mg) were immunoprecipitated overnight at 4°C using KMT2A antibody (3 μg) and IgG as negative control. Protein A/G-agarose beads (25 μl) (sc-2003 Santa Cruz Biotechnology) were used to precipitate the bound protein. The antibodies used were: anti-REST (07-579, Millipore, 1:1,000); anti-Sp1 (sc-14027, Santa Cruz Biotechnology, 1:500); anti-KMT2A (sc-374392, Santa Cruz Biotechnology, 1:500); H3K4-3me (PA5-17420, Invitrogen, 1:1,000); anti-caspase-8 (sc-6136, Santa Cruz Biotechnology, 1;1,000); RIPK1 (sc-41169, Santa Cruz Biotechnology, 1:500); anti-SOD1 (cod: PA5-27240, Thermo Fischer Scientific, 1:1,000) and anti-Tubulin (T5168, Sigma-Aldrich, 1:10,000). Total lysates of differentiated SH-SY5Y cells, that basally express REST, Sp1; KMT2A, H3K4-me3, RIPK1 and caspase-8 (Das et al., 2009; Guida et al., 2015a, 2017b; Wang et al., 2019) were used as positive control for all antibodies used for Western Blot experiments (Supplementary Figure 3). Total lysate of NSC34 cell transfected with SOD1-G93A construct is the positive control for SOD1 antibody (Figure 1A).

The protein bands under examination were visualized using a chemoluminescence system (ECL Western Detection Kit; Amersham, Arlington Heights, IL, United States). The bands were quantized by densitometry. Normalization of the results was ensured by hybridizing the blots with anti-tubulin antibody.

Experiments were performed as previously reported (Guida et al., 2018). Briefly, NSC34 cells were cross-linked with 1% formaldehyde and then the reaction was stopped by adding glycine 0.125 M. Samples were washed four times in cold PBS containing anti-protease cocktail and lysis buffer. Chromatin was fragmented into 200–500 bp by sonication. Cells chromatin lysates were incubated overnight with 5 μg of antibody for Sp1 (sc-14027) and KMT2A (sc-374392) from Santa Cruz Biotechnology, H3K4-me3 (PA5-17420, Thermo Fisher Scientific), and RNA Pol II (05-623, Merck Millipore). Normal rabbit IgG (sc-2357, Santa Cruz Biotechnology) was used as negative control. After immunoprecipitation, the DNA-histone complex was collected with 50 μl of salmon sperm DNA/protein A or G-agarose beads for 2 h (16-157, 16-201, Merck Millipore) and subsequently processed as previously reported (Guida et al., 2018). The oligonucleotides used for the amplification of immunoprecipitated DNA recognizing A region of REST mouse promoter sequence were already published (Ravache et al., 2010). Binding activity was normalized for the total input of chromatin and graphically represented as percentage of vehicle.

All experiments were performed in triplicate. The data were represented as means ± SD. For the analysis an unpaired Student’s t test for statistic between two experimental groups and the one-way ANOVA with Tukey’s post hoc test for more than two experimental conditions were used.

NSC34 cells were transiently transfected separately with the following plasmids: (1) empty vector (EV), (2) the vector overexpressing the wild type SOD1 gene (SOD1-WT) and (3) the vector overexpressing the G93A mutant SOD1 gene (SOD1-G93A). As shown in Figure 1A, SOD1-WT and SOD1-G93A cells showed 48 and 58% increase in SOD1 protein expression, respectively, compared to EV group, thus confirming the transfection efficiency. As expected, cell survival significantly decreased after 72 h transfection in SOD1-G93A cells, but not in SOD1-WT and EV cells, as revealed by MTT (Figure 1B) and LDH assays (Supplementary Figures 2A–C). To study the possible role of G93A mutation in accelerating MeHg-induced cell death, EV, SOD1-WT and SOD1-G93A cells were exposed for 24, 48, and 72 h to 100 nM of MeHg. This concentration (100 nM) of MeHg was used because it reduced cell viability after 72 h, but not at 24 and 48 h, in NSC34 motor neuron-like cells transfected with EV (Supplementary Figures 1A–C). Notably, SOD1-G93A experimental group showed a significant reduction in cell vitality already after 24 h of 100 nM MeHg exposures. By contrast, in EV and SOD1-WT cells a significant reduction in cell survival was observed only after 72 h of MeHg exposure, as revealed by MTT (Figures 2A–C) and LDH assays (Supplementary Figures 2D–F). Since 100 nM MeHg treatment for 24 h damaged approximately 50% of motor neuron-like NSC34 SOD1-G93A cells, this concentration and this time point were chosen for our experiments. In accordance with MeHg-induced cell death, REST mRNA and protein expression increased at 24, 48 and 72 h in SOD1-G93A cells, whereas in EV and SOD1-WT cells REST mRNA and protein increased only after 72 h of MeHg treatment (Figures 2D–I). These results suggest that SOD1-G93A cells are more sensitive to MeHg-induced neurotoxic effect via REST up-regulation than SOD1-WT cells. In support to this result, REST was knocked-down by a specific siRNA (siREST) transfection that reduced its mRNA level of 70% compared to siCTL (Supplementary Figure 3). Indeed, siREST reverted MeHg-induced REST protein increase (Figures 3A–C) and prevented SOD1-G93A cell death at 24 h, whereas this effect was exerted at 72 h in EV and SOD1-WT cells (Figures 3D–I).

Since it has been reported that REST gene is transcriptionally modulated by the two Sp family transcription factors Sp1 and Sp3, CREB, and JunD (Guida et al., 2015a, 2017b), we evaluated the effect of MeHg exposure to modulate the mRNA levels of the above mentioned transcriptional factors in SOD1-G93A cells. We found that 24 h exposure to 100 nM MeHg induced an increase in Sp1, but not in CREB, JunD and Sp3 mRNA expression (Figures 4A–D). Coherently, 24 h of MeHg exposure determined a significant increase also in Sp1 protein expression (Figure 4E). In order to clarify the possible role of Sp1 in the modulation of REST gene in MeHg-induced toxicity in SOD1-G93A cells, Sp1 was silenced by transfection with a specific siRNA (siSp1). The transfection efficiency was demonstrated by 40% reduction of Sp1 mRNA levels (Figure 4F). Notably, siSp1 counteracted the effect of MeHg on REST mRNA increase (Figure 4G).

Recently it has been demonstrated that Sp1 interacts with the epigenetic enzyme KMT2A in the breast cancer cell line MCF7 cells (Xu et al., 2017). The coimmunoprecipitation assay performed in SOD1-G93A cells exposed for 24 h to 100 nM MeHg revealed that Sp1 physically interacted with KMT2A (Figure 5A upper panel). The presence of KMT2A protein in the immunoprecipitated pool has been validated by immunoblot with KMT2A antibody (Figure 5A upper panel). On the basis of the interaction between Sp1 and KMT2A we examined the possibility that KMT2A could modulate REST expression in SOD1-G93A cells. Silencing of KMT2A (siKMT2A) significantly reduced its mRNA expression of 37% and the trimethylation of the lysine-4 on the histone-3 (H3K4-me3) protein levels of 60% (Figures 5B,C), thus confirming the transfection efficiency. It is noteworthy that siKMT2A counteracted MeHg-induced REST mRNA and protein up-regulation (Figures 5D,E). Interestingly, Chromatin Immunoprecipitation (ChIP) assay revealed that in MeHg-treated SOD1-G93A cells siSp1 reduced the Sp1 and KMT2A binding on REST A promoter region, by inducing a reduction of two well-known markers of gene transcription H3K4-me3 and RNA polymerase II (RNA-Pol II) (Figures 6A–D).

It has been previously demonstrated that necroptosis contributes to motor neuron death in amyotrophic lateral sclerosis (Yuan et al., 2019) and that the neurotoxicant PCB-95 activates the necroptotic pathway in cortical neurons via REST up-regulation (Guida et al., 2017b). It is widely documented that for the activation of necroptosis two conditions must be satisfied: (1) the activation of Receptor InterActing Serine/Threonine Kinase 1 (RIPK1) and (2) the blockade of caspase 8 activation. To this aim, we evaluated the effect of REST knocking-down on RIPK1 protein expression and caspase-8 activation in MeHg-treated SOD1-G93A cells. Notably, siREST: (1) counteracted the MeHg-mediated RIPK1 increase and (2) induced the activation of the caspase 8, thus preventing the activation of the necroptotic pathways (Figures 7A,B). Importantly, MTT and LDH assays demonstrated that the pre-treatment with the specific inhibitor of necroptosis Necrostatin 1 (Nec) (1, 5, 10, and 20 μM) determined in MeHg-treated SOD1-G93A cells a significant reduction in cell death (Figures 7C,D).