Plasmids and reagents

The plasmid pLenti-CMV-METTL16-FLAG-GFP-Puro, p3*FLAG-CMV-SENP3 and pWPXLd-FLAG-SENP3 were sourced from the Public Protein/Plasmid Library (Nanjing, China). The pENTER-LTF was purchased from Vigene Biosciences (Shandong, China). LTF-GFP construction involved the insertion of the open reading frame (ORF) of LTF from the pENTER-LTF plasmid into the linearized pEX-1 vector (GenePharma, Shanghai, China). Mutated plasmids (METTL16-PPF185/186/187AAG, SENP3-C532A, LTF K57R, LTF K405R, LTF K118R and LTF 2KR) were generated via high-fidelity PCR-based site-directed mutagenesis using pLenti-CMV-METTL16-FLAG-GFP-Puro, p3*FLAG-CMV-SENP3 and LTF-GFP as templates (Vazyme, Nanjing, China), and mutations were confirmed via Sanger sequencing. pT3-EF1a-MYC (Addgene plasmid # 92046;http://n2t.net/addgene:92046;RRID:Addgene_92046) [17] and px330-sgp53 (Addgene plasmid # 59910;http://n2t.net/addgene:59910;RRID:Addgene_59910) [18] were acquired from Addgene. pLentiCRISPR-sgMettl16, px459-sgSenp3 and px459-sgLtf harboring two targets linked by gRNA scaffold were purchased from Tsingke Biotech Co., Ltd (Beijing, China). The MYC-SUMO3 and UBC9 were constructed by inserting the ORF of SUMO3 and UBC9 into the linearized pEX-3 vector (GenePharma). All of the cloning PCR primers and sgRNA sequences are provided in Supplementary Table S1. The plasmids were extracted using a NucleoBond Xtra Midi kit (MACHEREY-NAGEL, Germany).

RSL3 ((1 S,3R)-RSL3, HY-100218A, MedChemExpress, USA); Fer-1 (Ferrostatin-1, S7243, Selleck, TX, USA); DFO (Deferoxamine, D9533,Sigma-Aldrich, Merck KGaA, Germany), FAC (Ammonium iron(III) citrate, F5879, Sigma); MG132 (HY-13259, MedChemExpress); ActD (Actinomycin D, HY17559, MedChemExpress); CHX (Cycloheximide, S7418, Selleck); Leupeptin (HY-18234, MedChemExpress); NEM (N-Ethylmaleimide, HY-D0843, MedChemExpress); Sorafenib (S7397, Selleck); Geneticin (G418 Sulfate, 108321-42-2, Selleck).

Antibodies

Primary antibodies against the following proteins were obtained from Cell Signaling Technology (MA, USA): METTL16 (87538 S); SENP3 (5591 S); anti-rabbit IgG, HRP-linked Antibody (7074); anti-mouse IgG, HRP-linked Antibody (7076); Mouse (G3A1) mAb IgG1 Isotype Control (5415); Rabbit (DA1E) mAb IgG XP® Isotype Control (3900); from Sigma: MYC-Tag(SAB1305535); Anti-FLAG®M2 (F1804); from Abcam (Cambridge, UK): METTL16 (ab252420); IGF2BP2 (ab128175); LTF (ab109216); 4-HNE (ab46545); MDA (ab43066); GPX4 (ab125066); NRF2 (ab62352); from Proteintech (Wuhan, China): β-actin (66009-1-Ig); GFP-Tag (66002-1-Ig); HA-Tag (66006-2-Ig); METTL3 (15073-1-AP); METTL14 (26158-1-AP); ACSL4 (22401-1-AP); from ABclonal (Wuhan, China): LTF (A12902); FTH1 (A19544).

Cell lines

Huh7, PLC/PRF/5, Hepa1-6, Hep3B, HepG2, HCCLM3, MHCC97-H, SK-HEP-1 (used exclusively for METTL16 protein analysis due to its endothelial origin), and HEK293T cell lines were cultured in Dulbecco’s Modified Eagle Medium (DMEM, Gibco, Grand Island, NY, USA) and SNU387 cell lines were maintained in RPMI-1640 (Corning, USA). Both media were supplemented with 10% fetal bovine serum (FBS, Sigma) and 1% penicillin/streptomycin (Gibco). Cells were incubated at 37 °C with 5% CO2. All cell lines were sourced from the Cell Bank, Chinese Academy of Sciences (Shanghai, https://www.cellbank.org.cn), and tested negative for mycoplasma contamination.

Cell transfection

Plasmids or siRNAs were transfected using Lipofectamine 3000 Reagents (Invitrogen, USA) following the manufacturer’s protocols. The small interference RNAs (siRNAs) and non-specific control (NC), provided by GenePharma, are listed in Supplementary Table S2. For stable HCC cell lines, lentiviruses designed for overexpressing METTL16 and SENP3, and for RNA interference against METTL16 and SENP3 (based on siMETTL16 and siSENP3 sequences) were prepared by GenePharma. HCC cells were infected with these lentiviruses in the presence of 4 µg/mL polybrene and selected with 2.5 µg/mL puromycin. To establish stable cell lines expressing LTF/LTF 2KR, cells were transfected with GFP-LTF/LTF 2KR plasmids and selected with 700 µg/mL G418 (Geneticin) to isolate single-cell-derived clones with successful LTF/LTF 2KR integration and overexpression.

Cell proliferation assays

Cell viability was measured using the Cell Counting Kit-8 reagent (CCK-8, Dojindo, Japan) according to the manufacturer’s instructions. Triplicate wells containing 3 × 103 cells each were plated in 96-well plates, followed by the addition of 10µL CCK-8 reagent per well and a 75-minute incubation at 37 °C. Absorbance was measured at 450 nm using a spectrophotometer. For colony formation assays, 1 × 103 cells per well were seeded in 6-well plates, with media changed every 4 days. Colonies, typically formed in approximately 10–17 days, were fixed with 4% paraformaldehyde, stained with 0.5% crystal violet, and subsequently photographed and counted.

Iron, MDA, and GSH level measurement

Following treatment, 2 × 106 HCC cells, mouse liver tissues, or subcutaneous xenografts were harvested and resuspended in PBS or saline for further analysis. Iron levels were determined using the Iron Assay Kit (A039-2-1, Jiancheng Bioengineering Institute, China), while MDA levels were assessed with the MDA Assay Kit (A003-4-1, Jiancheng Bioengineering Institute, China), and GSH levels were measured using the GSH Assay Kit (A006-2-1, Jiancheng Bioengineering Institute, China). Assays were conducted according to the manufacturer’s protocols.

ROS and lipid peroxidation measurement

Lipid peroxidation was assessed using BODIPY™ 581/591 C11 (D3861, Invitrogen). The intracellular ROS levels were measured using the ROS Assay Kit with the DCFH-DA probe (S0033M, Beyotime, China). BODIPY™ 581/591 C11 (10 mM) and DCFH-DA (10 mM) probes were diluted at ratios of 1:2000 and 1:1000, respectively, in a serum-free medium. Cells were incubated with the probes at 37 °C in the dark for 20–30 min. Following incubation, cells were washed with PBS, and fluorescent images were captured with an inverted microscope. HCC cells were subsequently resuspended and quantified with a CytoFLEX flow cytometer and reanalyzed using FlowJo software.

Labile iron pool and total iron assays

The labile iron pool within HCC cells was assessed utilizing Calcein-AM (HY-D0041, MedChemExpress). After collection, cells were washed twice with PBS and then incubated with 0.125µM Calcein-AM in either a serum-free medium or PBS for 15 min at 37 °C. Fluorescence measurements were performed using a fluorescence microscope and a CytoFLEX flow cytometer. The fluorescence intensity of Calcein-AM increases inversely with free iron content. Moreover, the presence of free Fe2+ in HCC cells was determined by employing FerroOrange (F374, Dojindo). Following three washes with Hank’s Balanced Salt Solution (HBSS), cells were exposed to 1 µM FerroOrange for 30 min at 37 °C. Fluorescence intensity was subsequently measured using both a fluorescence microscope and a CytoFLEX flow cytometer.

Quantitative real-time polymerase chain reaction (qRT-PCR)

Cells and tissues were lysed for RNA extraction using Trizol (Invitrogen). 1 µg RNA was reverse transcribed into cDNA with the PrimeScript™ RT reagent kit (Takara Bio, Japan). qRT-PCR was carried out with the TB Green® Premix Ex Taq™ (Tli RNaseH Plus) kit (Takara Bio, Japan) on an ABI Q6 PCR system. Gene expression was quantified using the 2−ΔΔCT method with the human β-actin gene as the endogenous control. Primer sequences are provided in the Supplementary Table S3.

Dual-luciferase reporter assay

Dual-luciferase reporters, including SENP3-CDS WT-Luc, SENP3-CDS MUT-Luc, SENP3-3’UTR WT-Luc and SENP3-3’UTR MUT-Luc, were generated using the pGL3 vector by Sangon Biotech Co., Ltd (Shanghai, China). HCC cells, at 70–80% confluence in 24-well plates, were co-transfected with 500 ng of each plasmid and 10 ng of pRL-SV40 Renilla luciferase reporter plasmid. After 24–48 h, cells were harvested, and luciferase activity was measured using the dual-luciferase assay kit (Promega, Wisconsin, USA) and the Glomax-multi Luminometer.

MeRIP/RIP assay

The methylated RNA immunoprecipitation (MeRIP) assay was conducted using the Magna MeRIP m6A Kit (Millipore, USA) to quantify m6A modification in RNA. Initially, genomic DNA was eliminated using DNase, followed by RNA fragmentation. Magnetic beads were incubated with either 10 µg of anti-m6A antibody or IgG, then immunoprecipitated with RNA samples. Subsequently, m6A-enriched RNA fragments were eluted, reverse transcribed, and quantified via RT-qPCR using primers designed for m6A sites (see Supplementary Table S4). For the RNA immunoprecipitation (RIP) assay, the Magna RIP™ Kit (Millipore, USA) was utilized following the manufacturer’s instructions. Cell lysates from 4 × 106 cells were sonicated and subjected to immunoprecipitation at 4 °C overnight with primary antibodies against METTL16 or IGF2BP2, with a homologous IgG serving as the isotype control. The relative enrichment of m6A or SENP3 in each sample was determined by normalizing the Ct values obtained from samples immunoprecipitated with m6A/METTL16/IGF2BP2 antibodies against the Ct values from the corresponding input fractions. As described in the prior studies [19], the formulas used for analysis are presented as follows: ΔCTRIP = CTRIP – CTinput; ΔCTIgG = CTIgG – CTinput; ΔΔCT = ΔCTRIP – ΔCTIgG; fold enrichment = 2–ΔΔCT. The resultant relative fold enrichment values were normalized to those of the corresponding control group.

Western blot assay

Cells and tissues were cracked by RIPA lysis buffer (20–188, Merck Millipore, USA) supplemented with protease inhibitor cocktail (4693159001, Roche, Swiss), followed by centrifugation at 12,000 × rpm at 4 °C for 15 min. The resulting mixtures were then separated using SDS-PAGE and electro-transferred onto a nitrocellulose filter (NC) membrane. To block nonspecific binding, the membrane was incubated in a solution of 5% non-fat milk with 0.1% Tween-20 in PBS (PBST) for 1 h. After incubating with specific primary and corresponding secondary antibodies, the protein bands were visualized with the Odyssey Infrared imaging system (LI-COR Biosciences, Lincoln, NE, USA) or Tanon 5200 chemiluminescent imaging system (Tanon, Shanghai, China).

mRNA and protein stability assays

Cells were treated by ActD and CHX for 0, 4, and 8 h. The total RNA and protein were extracted from the treated cells. The obtained results were normalized to the values measured at 0 h, serving as the baseline for comparison across various time points.

Co-immunoprecipitation (Co-IP)

The total proteins were lysed using IP lysis buffer (20 mM Tris, pH 8.0; 150 mM NaCl; 1.0% NP-40; 1 mM ethylenediaminetetraacetic acid (EDTA)) supplemented with protease inhibitor (4693159001, Roche) and phosphatase inhibitor (4906837001, Roche). The lysates were then centrifuged at 12,000 rpm at 4 °C for 15 min. After centrifugation, the supernatants were mixed with GFP-Nanoab-Agarose (LABLEAD Biotech, Beijing, China) overnight at 4 °C or incubated with IP antibody at 4 °C for 2 h, followed by mixed with Dynabeads™ Protein G beads (10003D, Thermo Fisher Scientific, Waltham, MA, USA) overnight at 4 °C. The beads were washed three times using pre-cooled IP buffer and then resuspended in 2 x SDS loading buffer (Beyotime). The analysis of protein-protein interactions was performed using western blot assay.

IP coupled with mass spectrometry

Total proteins were extracted from the cells, followed by immunoprecipitation (IP) using SENP3 antibodies and Dynabeads™ Protein G beads as previously described. The SENP3-associated proteins were separated by gel electrophoresis, and the gels were stained using a Fast Silver Stain kit (P0017S, Beyotime Biotech, Beijing, China) according to the manufacturer’s instructions. Subsequently, the silver-stained bands of interest were excised and processed for mass spectrometric analysis using an Ultimate 3000 nano ultra-performance liquid chromatography-tandem Q Exactive plus mass spectrometry system (LuMing Biotech, Shanghai, China). The information on SENP3-associated proteins was listed in Supplementary Table S5.

SUMOylation and ubiquitination assays

To assess SUMOylation, HCC cells were cultured in 10 cm plates, transfected with the indicated plasmids, and lysed using SUMO lysis buffer (20 mM Tris, pH 8.0; 150 mM NaCl; 1.0% NP-40; 1 mM EDTA; 1.0% SDS; 20 µM NEM) with protease inhibitors and PMSF. Following a 30-minute incubation on a 4 °C shaker and subsequent sonication until the lysate became fluid, the protein samples were centrifuged. The supernatant was then incubated with GFP-Nanoab-Agarose beads overnight at 4 °C. Afterward, the beads were washed five times with SUMO lysis buffer without SDS and boiled for 10 min in SDS loading buffer. The eluted proteins were subjected to Western blot analysis. For ubiquitination, transfected HCC cells were treated with MG132 (40 µM for 4 h) before lysis to inhibit proteasomal degradation. Whole-cell lysates were extracted using UB lysis buffer (20 mM Tris, pH 8.0; 150 mM NaCl; 1.0% NP-40; 1 mM EDTA) and processed as described for the SUMOylation assay.

Mouse models

The mice were housed in a controlled environment with air filtration, temperature maintained at 22–24 °C, lighting controlled, and humidity between 40% and 70%. They were provided with unlimited access to a standard diet.

Male athymic nude mice (5–6 weeks old) were obtained from GemPharmatech Laboratories (Shanghai, China) for establishing a xenograft model of HCC. The mice were randomly divided into different groups before injection. 1.0 × 106 stable HCCLM3 shNC, shMETTL16-1 or shMETTL16-2 cells per mouse were subcutaneously injected into nude mice (12 mice per group). All mice in each group received intraperitoneal injections of either DMSO or RSL3 (5 mg/kg) every 2 days. Additionally, 2.0 × 106 Hep3B stable LV-VEC + shNC, LV-METTL16 + shNC, LV-VEC + shSENP3 or LV-METTL16 + shSENP3 cells per mouse were subcutaneously injected into nude mice (5 mice per group). Male C57BL/6 mice (5–6 weeks old) obtained from GemPharmatech Laboratories received injections of 6.0 × 105 stable Hepa1-6 cells sgNC or sgMettl16 (10 mice per group). Similarly, all mice in each group were administered intraperitoneal injections of either DMSO or RSL3 (10 mg/kg) every 2 days. Furthermore, 1.0 × 106 stable HCCLM3 shNC + LV-VEC, shMETTL16 + LV-VEC, shNC + LV-SENP3 or shMETTL16 + LV-SENP3 per mouse were subcutaneously injected into nude mice (6 mice per group). Tumor volume was monitored every two days using the formula: Tumor volume = (Length × Width2)/2. The tumors were resected for weighting and photographing after the mice were euthanized.

Hepatocyte-specific Mettl16 conditional knockout mice (designated Mettl16f/f, Alb-Cre mice) were generated using CRISPR/Cas9 technology by mating Mettl16-floxed (Mettl16f/f) mice (loxP sequences inserted at the ends of exons 2 and 5, GemPharmatech Co., Ltd) with Alb-Cre transgenic mice (Model Organisms Center, Inc, Shanghai, China). Additionally, Mettl16f/f mice were injected with 100 µL AAV suspension in saline (AAV8-TBG-Cre/Ctrl virus titer > 1.5 × 1012 vg/mL, Vigene Biosciences Co., Ltd) in 4 weeks via the tail vein to generate the Mettl16f/f, AAV8-TBG-Cre and Mettl16f/f, AAV8-TBG-Ctrl mice.

For detecting Mettl16 function in de novo mice HCC model, a sterile solution containing 12 µg of pT3-EF1a-MYC (MYC), 10 µg of px330-sgp53 (sgp53) and 3 µg of transposon SB13 transposase-encoding plasmid was prepared in 2 mL 0.9% NaCl solution. For assessing Mettl16/Senp3/Ltf axis function, the solution included 12 µg of pT3-EF1a-MYC-IRES-Mettl16 (or 12 µg of the control vector pT3-EF1a-MYC), 10 µg of sgp53, 3 µg of transposon SB13 transposase-encoding plasmid and 10 µg of px459-sgSenp3(sgSenp3) or 10 µg of px459-sgLtf (sgLtf) (or 10 µg of the control vector px459) in 2 mL 0.9% NaCl. Each mouse was injected with a volume equivalent to 10% of its body weight, administered within 5–7 s.

The serum biochemistry assay

At the end of the experiment, all mice were euthanized. Blood samples were collected and centrifuged at 2500 rpm for 15 min. The serum was then separated and stored at -80 °C until analysis. Serum levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), triglycerides (TG), cholesterol (CHO), cholinesterase (CHE), and γ-glutamyltransferase (γ-GGT) were quantified using commercial assay kits (Jiancheng Bioengineering Institute, China), following the manufacturer’s protocols.

Immunohistochemistry (IHC), Hematoxylin and Eosin (H&E)

Tissues from mouse models were fixed in 4% paraformaldehyde (Beyotime) and embedded in paraffin. For the IHC staining assay, sections were deparaffinized, rehydrated, and boiled for antigen retrieval. Blocking was done with 5% BSA, followed by overnight incubation with primary antibodies at 4 °C, and a 1-hour room temperature incubation with horseradish peroxidase-conjugated secondary antibodies. Antigen detection was enhanced using a diaminobenzidine (DAB)-based chromogenic system. H&E staining was performed according to the standard procedures.

Human HCC tissue specimens and human tissue microarray (TMA)

The study utilized tumors and matched adjacent normal control samples from 37 patients with HCC, sourced from Shanghai East Hospital, Tongji University School of Medicine, China, with approval from the Ethics Committee of Shanghai East Hospital, and all participating patients provided informed consent. Upon collection, the tissues were promptly preserved in liquid nitrogen and stored at -80 °C until required for analysis.

A human HCC microarray (TMA, #HLivH180Su09) containing 90 paired HCC tissues and their corresponding matched normal control samples, was obtained from Shanghai Outdo Biotech (Shanghai, China). Two independent pathologists, blinded to the clinical data, evaluated the immunostaining of the samples. They assessed each cancer specimen based on staining intensity (0 = no staining, 1 = weak staining, 2 = moderate staining, 3 = strong staining) and the percentage of stained cells (0, positive cells ≤ 10%; 1, 11–25%; 2, 26–50%; 3, 51–75%; and 4, > 75%). Expression scores, calculated by multiplying intensity and percentage scores, were classified into low (negative (expression scores ≤ 2) and weak positive (2 < expression scores ≤ 4)) and high expression levels (moderate positive (4 < expression scores ≤ 6) and strong positive (expression scores > 6)).

Isolation and culture of tumor organoids

The noncancerous portion of hepatic carcinoma tissue was excised, and hepatic carcinoma cells were isolated utilizing a combination of mechanical disruption and enzymatic digestion. Briefly, the HCC tissue of appropriate dimensions underwent mincing, washing, and subsequent incubation at 37 °C with Tumor Tissue Digestion Solution (K601003, BioGenous Technologies, Suzhou, China) on an orbital shaker for 1–2 h. The resultant suspension was filtered through a 100 μm nylon cell strainer to remove tissue fragments. After centrifugation and resuspension, cells were seeded onto Matrigel Organoid Culture ECM (M315066, BioGenous Technologies) within 24-well plates and overlaid with the Hepatocellular Carcinoma Organoid Kit (K2105-HCC, BioGenous Technologies). The culture medium was replenished every 4 days, and organoids were passaged every 1–4 weeks. Following treatment with RSL3 (2.5µM) or Sorafenib (10µM) for 10 days, organoids were examined, photographed and quantified. Organoid formation capacity was assessed based on the number of viable organoids with diameters exceeding 50 μm and quantified utilizing ImageJ software.

Bioinformatics analysis

Expression levels of METTL16, SENP3, and LTF, along with survival analysis in cancer, were conducted utilizing gene expression and clinical information from The Cancer Genome Atlas (TCGA) project (https://portal.gdc.cancer.gov) and the supplementary information provided by Gao, Q., et al [20]. The public datasets of ICGC (https://dcc.icgc.org/projects/LIRI-JP), GSE104462, GSE109211, GSE104580, GSE45436, GSE182607, GSE181515, GSE182593, GSE248769 and GSE262114 were employed. Similar genes to METTL16 were analyzed using GEPIA2 (http://gepia2.cancer-pku.cn/#similar) and are listed in the Supplementary table S6. METTL16 and SENP3 expression in the TCGA-LIHC cohort from different cancer stages was analyzed using UALCAN (https://ualcan.path.uab.edu/). scRNA-seq data from GSE181515 were analyzed through scLiverDB (https://guolab.wchscu.cn/liverdb). Drug activity data and mRNA expression levels of METTL16 were assessed via CTRP (https://portals.broadinstitute.org/ctrp/). Ferroptosis genes were obtained from FerrDb V1(http://www.zhounan.org/ferrdb) and Ferroptosis signature genes were provided by Wu, A., et al. (Supplementary table S6) [21]. To predict potential SUMOylated residues in LTF, bioinformatic platforms including GPS-SUMO (http://sumosp.biocuckoo.org), SUMOplot™ (https://www.abcepta.com/sumoplot), and JASSA (http://www.jassa.fr/index.php?m=jassa) were used. Raw data for determining the genetic dependence of METTL family members in cancers were obtained from https://depmap.org/portal/ and https://score.depmap.sanger.ac.uk/.

Statistical analysis

Data are expressed as mean ± SD. Statistical significance was determined using the Student t-test or Mann-Whitney test. The χ2 test was employed for categorical data comparisons. ANOVA or Kruskal-Wallis test was used for comparisons among multiple groups. Statistical significance for survival was calculated using the Log-rank test. Spearman and Pearson’s correlation analyses were performed for correlation analysis. GraphPad Prism 9 software was used to create the graphs and for the statistical analysis (GraphPad Software, Inc.). Significance values were set at *, P < 0.05; **, P < 0.01; ***, P < 0.001, n.s., not significant.