2.1. Cell culture
The MV411 (ATCC, USA) cell line was cultured in Iscove's Modified Dulbecco’s Medium (IMDM, Hyclone) containing 10% fetal bovine serum (Gibco#10099141), and molm13 (DSMZ, Germany) cell line was cultured in RPMI 1640 Medium (Gibco# C11875500BT) containing 10% fetal bovine serum. The NK92-MI cell line (Procell#CL-0533) was cultured in a Special medium for NK cells (Procell, #CM-0533). All the cell lines were maintained at 37℃ in a 5% CO2 atmosphere and had been authenticated by the Cell Line Authentication Services, using Short Tandem Repeat (STR) analysis and a regular screen for Mycoplasma contamination by RT-PCR.
2.2. Treatment paradigm
AML cell lines (MV411 and MOLM13) were divided into different groups according to the dose of ATO. The control group was exposed to 1% DMSO and the ATO group was treated with 3 µM or 6 µM ATO; According to the application of ATO or Z-DEVD-FMK, AML cells were divided into ATO or ATO + DEVD groups. The ATO group was treated with 5 µM ATO, and the ATO + DEVD group was treated with both 5 µM ATO and 25 µM Z-DEVD-FMK (caspase-3 specific inhibitor) for the in vitro experiment.
2.3. Animal experiments and ethical statements
The animals in our study were approved by the Institutional Animal Care and Use Committee of The First Affiliated Hospital of Sun Yat-Sen University (No.  040). To establish nude mouse xenograft models, female nude mice (four weeks old; average weight, 15g) were obtained from Experimental Animal Laboratories (Zhuhai, China). MV411 cells (AML cell line) were resuspended in PBS (1×105 cells/mice) and then injected into the right foreleg of each mouse subcutaneously at an inoculation depth of 1 cm. After the tumor grew to about 100 mm3, mice were randomly separated into 2 groups of 5 mice and then treated with either the vehicle (PBS, Gibco) or ATO (3 mg/kg) via intraperitoneal injection and their body weight was measured once per day. About 15 days after administration, all mice were sacrificed by cervical dislocation. Tumors, spleens, and livers were dissected. Tumor sizes in mice were measured every day using a digital caliper, and all tumor volumes were calculated using the following formula: volume = length × (width) 2/2.
2.4. Annexin V and PI staining
A total of 1 × 106 AML cells were treated with DMSO or ATO (3µM, 6µM) in the presence or absence of 25 µM Z-DEVD-FMK (caspase-3 specific inhibitor) for 24 h. The cells were then analyzed using the annexin V-FITC Apoptosis Kit (KeyGEN#KGA108) according to the manufacturer’s protocol.
2.5. Cell preparation and flow cytometric analysis
Single-cell suspensions were prepared from the tumors, spleens and peripheral blood of nude mouse xenograft models washed with PBS and filtered through a cell strainer (200 µm). The mononuclear cell–containing supernatants were obtained using 55% Percoll and OptiPrep gradient centrifugation, transferred, and washed twice before antibody staining. Single-cell suspensions were then stained with fluorochrome-conjugated monoclonal antibodies (eBioscience, USA) against the following cell surface markers: APC-conjugated CD49b and AF700-conjugated CD16. The images of all samples were acquired on a Beckman Coulter (Beckman, USA). Data were analyzed using CytExpert software.
2.6. Quantitative real-time reverse transcription PCR (qRT-PCR)
Total RNA was extracted from cells using TRIzol Reagent (Gibco, Carlsbad, CA, USA), and 1 µg total RNA (mRNA) was reverse-transcribed into cDNA using Evo M-MLV RT Premix (AG11706, Accurate Biology). The real-time qPCR was performed with SYBR Green Premix Pro Taq HS qPCR Kit (#AG11701, Accurate Biology), with three replicates set for each well. The mRNA expression relative to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was determined using the 2−ΔΔCT method. The primers were synthesized by RuiBiotech (Beijing, China) as the sequences described in Table S1.
2.7. LDH release assay and ELISA
LDH was measured using a Cytotoxicity LDH Assay Kit-WST (Dojindo#CK12) according to the manufacturer's instructions. The absorbance was calculated at 490 nm. IL-18, IL-1β, HMGB1, and GZMB were measured using a Quantitative IL-18/IL-1β/HMGB1/GZMB ELISA kit (MEIMIAN, China) according to the manufacturer's instructions. The absorbance was calculated at 450 nm. At least three replicated experiments were performed.
2.8. Nucleoplasmic separation of cellular proteins
Collect cells (500rpm, 3min), discard the supernatant and wash twice with pre-cooled PBS. Add 200 µl of CER I mixture with protease inhibitor for every 4*10^6 suspension cells, vortex for 15 s to fully lyse the cells, and left on ice for 10 min. Add CER II, vortex for 10 s, and left on ice for 1 min. Then vortex vigorously for 10 s and centrifuge at 13,200 rpm at 4°C for 5 min. The supernatant fraction was the cytoplasmic fraction of protein. The precipitated fraction was added to NER (with protease inhibitor), vortexed at high speed for 15s, and left on ice for 10min. Repeat the vortex-rest operation until the precipitate was dissolved. Add 5× loading buffer, denature the protein at 100°C for 5 min, and store at -20°C.
2.9. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE)
AML cells were treated with or without ATO and serum-free cultured for 24h. The supernatant of the culture medium was collected and concentrated 40 times. The samples with preheating (100°C for 10 min) were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS‐PAGE) on 10% agarose gels, which were peeled off and stained by Coomassie Brilliant Blue (P0017F, Beyotime).
2.10. Western blot analysis
All the proteins were extracted from cultured cells and mouse tissues by RIPA (Beyotime, Shanghai, China), using a protease inhibitor and phosphorylase inhibitor cocktail (Bimake, Shanghai, China). The protein concentration was measured via BCA Protein Assay Kit (Pierce, Rockford, IL, USA). Total cell lysate per sample was loaded on Omni-PAGE Hepes-Tris Gels (Yamei, Shanghai, China) and separated by electrophoresis. The western blot was blotted onto a piece of 0.2 um PVDF membrane using a wet transfer device (Bio-Rad). Block the blot in a 5% milk powder solution diluted in TBST (Tris-buffered saline/0.05% Tween 20), incubated with indicated antibody, and signal detected with horseradish peroxidase (HRP) -conjugated secondary antibody and enhanced chemiluminescence (ECL) detection (Bio-Rad). The following antibodies were used: anti-DFNA5/GSDME (Abcam, #ab215191), anti-cleaved-caspase-3 (Cell Signaling Technology, #9664), anti-HMGB1 (Proteintech, #10829-1-AP), anti-Granzyme B (Proteintech, #13588), anti-p38 (Novus, #96907), anti-CTSB (Cell Signaling Technology, #31718T), anti-MAPKAPK2 (Proteintech, #13949), anti-p53 (Proteintech, #60283). Equal protein loading was assessed using mouse anti-β-actin and anti-GAPDH antibodies (BOSTER, TA-09, China).
2.11. Co-culture system
AML cells were cultured in a 12-well plate in the bottom transwell (0.4µm) chamber while NK92-MIs were co-cultured in the top chamber. According to the presence or absence of treatment with ATO and the presence or absence of co-culture with MV411, they were divided into control group, MV411 + NK group, or ATO + MV411 + NK group. NK92-MI cells in the control group were treated with 0.3% as control, NK92-MI cells in the MV411 + NK group were co-cultured MV411 with NK92-MI, and NK92-MI cells in the ATO + MV411 + NK group were co-cultured MV411 with NK92-MI with 3 µM ATO added to the MV411 culture in the bottom chamber.
2.12. Cell viability assay
After the specified treatment, the ability of cell proliferation of NK-92MI cells in different groups was determined using the Cell Counting Kit-8 (CCK-8) assay reagent (Dojindo#CK04, Japan) according to the manufacturer’s protocol. The absorbance was calculated at 450 nm. Three replicated experiments were performed for each well.
2.13. Hematoxylin and eosin (H&E) staining and Immunohistochemistry (IHC)
A portion of mice’s xenograft tumor tissue was fixed with 4% paraformaldehyde (PH 7.4). We gradually Dewaxed as followed: Xylene I for 20 min; Xylene II for 20 min; 100% ethanol I for 5 min; 100% ethanol II for 5 min; 75% ethanol for 5 min; Stained sections with Hematoxylin solution for 3–5 min, rinsed with tap water. Then we treated the section with Hematoxylin Differentiation solution, Hematoxylin Scott Tap Bluing, and then Eosin dye for 5 min. We dehydrated as followed: 100% ethanol I for 5 min; 100% ethanol II for 5 min; 100% ethanol III for 5 min; Xylene I for 5 min; Xylene II for 5 min; Finally, we sealed with neutral gum, observed with a pathological section scanner, acquired images and analyzed them; IHC was performed on tumor tissue sections from Mouse xenograft formalin-fixed paraffin-embedded tumor tissue sections of 4-µm thickness to analyze the expression of KI67. The sections were de-paraffinized and rehydrated. Endogenous peroxidase activity was blocked with 3% hydrogen peroxide and incubated at room temperature in darkness for 25 minutes. The sections are placed in PBS (PH7.4) and shaken on a decolorizing shaper 3 times for 5 minutes each. Then, the sections were incubated overnight at 4℃ with an anti-KI67 antibody (1:300 dilution, Servicebio). After washing off the antibodies, the sections were incubated with a secondary antibody (HRP labeled) from the corresponding species of primary antibody and incubated at room temperature for 50 minutes. Finally, diaminobenzidine substrate (DAB) was applied to sections and counterstained with hematoxylin to counterstain the nucleus for microscopic examination after resin sealing. Immunohistochemical staining intensity was semi-quantified using Image J software that compared the intensity of staining with the proportion of the stained cells. The level of KI67 expression was determined by the intensity.
2.14. Immunofluorescence (IF)
IF was performed on tumor tissue sections from Mouse xenograft formalin-fixed paraffin-embedded tumor tissue sections of 4-µm thickness to analyze the expression of Granzyme B and CD16. The following primary antibodies were used: rabbit anti-Granzyme B (Proteintech#13588, 1:200), rabbit anti-CD16 (1:200, eBioscience, USA). The sections were made and dewaxed with xylene and ethanol of different concentrations. Antigen retrieval was performed in EDTA antigen retrieval buffer (pH 8.0) and maintained at a sub-boiling temperature for 8 min and then followed by another sub-boiling temperature for 7 min. 3% hydrogen peroxide was added to the sections to block the activity of endogenous peroxidase. 3% BSA to cover the marked tissue to block non-specific binding for 30 min and then Incubating slides with primary antibody (diluted with PBS appropriately) overnight at 4 ℃. After washing, the secondary antibody (FITC) was added at room temperature for 50 min. The visualization of the nuclei was done by adding DAPI, and then the sections were mounted after adding the anti-fluorescence quencher. Images were collected by Fluorescent Microscopy. DAPI glows blue by UV excitation wavelength 330–380 nm and emission wavelength 420 nm; FITC glows green by excitation wavelength 465–495 nm and emission wavelength 515–555 nm; CY3 glows red by excitation wavelength 510–560 nm and emission wavelength 590 nm.
2.15. Label-free quantitative LC/MS proteomics analysis
The cell culture supernatants of AML cells were collected in ATO and NC groups, the AML xenograft tumors of the ATO and PBS groups were isolated from mice (n = 3 per group), and the label-free quantification proteomics analysis was performed (Novogene, China). Briefly, proteins were extracted, and the concentrations were determined by Bradford protein assay (Bio-Rad, USA). After proteins were digested with Trypsin Gold (Promega), shot-gun proteomics analyses were performed using an EASY-nLC 1200 UHPLC system (Thermo Fisher Scientific) coupled with an Orbitrap Q Exactive HF-X mass spectrometer (Thermo Fisher Scientific) operating in the data-dependent acquisition mode. Data were searched against the UniProt database for Mus musculus(85,165 sequences, downloaded 18 January 2019) database by the search engine Proteome Discoverer 2.2 (Thermo Fisher Scientific). To analyze the protein family and pathway, the functional analysis was conducted using the interproscan program against the databases of COG (Clusters of Orthologous Groups) and KEGG (Kyoto Encyclopedia of Genes and Genomes). DEPs were used for Volcanic map analysis, cluster heat map analysis, and enrichment analysis of GO, IPR, and KEGG. The probable protein-protein interactions were predicted using the STRING server.
2.16. Statistical analysis
The statistical analysis of data was performed using GraphPad Prism 8. The comparison between the other two groups of data was analyzed by an unpaired t-test. The comparison among multiple groups was carried out by one-way analysis of variance. Body weight and tumor volume at different time points were compared using repeated-measures analysis of variance. P < 0.05 was considered to be significant. Graphs and error bars represent means ± SD of independent biological experiments unless stated otherwise. For all experiments, the investigators were not blinded.