Survival analyses using TCGA patient cohorts
The bulk RNA-seq expression data from The Cancer Genome Atlas (TCGA) was deconvoluted with PRISM13 in order to separate signals from EOCs, immune cells, and fibroblasts. Clinical data for TCGA cohort12 was downloaded from cBioPortal database (http://www.cbioportal.org/study/summary?id=ov_tcga_pub). Analysis was carried out on R version 4.0.3. Using the Molecular SIGnature Data Base (https://www.gsea-msigdb.org/) iron related signatures were evaluated and those with significant p-value were selected:
GO:0034755: GOBP_REGULATION_OF_IRON_ION_TRANSMEMBRANE_TRANSPORT; gene members: iron-sulfur cluster assembly enzyme (ISCU), homeostatic iron regulator (HFE), interferon gamma (IFNG), microRNA 210 (MIR210), ATPase copper transporting alpha (ATP7A), beta-2-microglobulin (B2M), transferrin (TF)
GO:0098711: GOBP_IRON_ION_IMPORT_ACROSS_PLASMA_MEMBRANE; gene members: iron-sulfur cluster assembly enzyme (ISCU), STEAP2 metalloreductase (STEAP2), interferon gamma (IFNG), microRNA 210 (MIR210), solute carrier family 39 member 8 (SLC39A8).
GO:0008198: GOMF_FERROUS_IRON_BINDING; gene members: cysteine dioxygenase type 1 (CDO1), egl-9 family hypoxia inducible factor 2 (EGLN2), egl-9 family hypoxia inducible factor 3 (EGLN3), DnaJ heat shock protein family (Hsp40) member C24 (DNAJC24), alkB homolog 2, alpha-ketoglutarate dependent dioxygenase (ALKBH2), alkB homolog 3, alpha-ketoglutarate dependent dioxygenase (ALKBH3), ferrochelatase (FECH), iron-sulfur cluster assembly enzyme (ISCU), 3-hydroxyanthranilate 3,4-dioxygenase (HAAO), frataxin (FXN), ferritin heavy chain 1(FTH1), ferritin light chain (FTL), diphthamide biosynthesis 3 (DPH3), ferritin heavy chain 1 pseudogene 19 (FTH1P19), phytanoyl-CoA 2-hydroxylase (PHYH), ferritin heavy chain like 17 (FTHL17), acid phosphatase 5, tartrate resistant (ACP5), egl-9 family hypoxia inducible factor 1 (EGLN1), tet methylcytosine dioxygenase 2 (TET2), hypoxia inducible factor 1 subunit alpha inhibitor (HIF1AN), synuclein alpha (SNCA), transferrin (TF), tyrosine hydroxylase (TH), FTO alpha-ketoglutarate dependent dioxygenase (FTO), alkB homolog 1, histone H2A dioxygenase (ALKBH1), ferritin mitochondrial (FTMT), hephaestin (HEPH).
Signature expression scores for each deconvoluted cell type in each sample were calculated with AUCell version 1.12.064, which calculates the activity of a gene set using a rank-based method. Six genes (FTH1P19, MIR210, BOLA2B, ERFE, UQCRFS1P1 and CYP2G1P) present in the studied iron related signatures were not found in the bulk RNA-sequencing expression data. Gene set objects in R were created with the R package GSEABase version 1.52.165. Samples were generally divided into high and low signature score groups with the cutoffs set according to score tertiles: lowest tertile were annotated as “Low” and highest tertile were annotated as “High”. For the signatures "GOBP_REGULATION_OF_IRON_ION_TRANSMEMBRANE_TRANSPORT" and "GOBP_IRON_ION_IMPORT_ACROSS_PLASMA_MEMBRANE", the scores in EOC’s and fibroblasts were very low. Therefore, “Positive” samples were classified as samples with a non-zero score whereas “Negative” samples were classified as samples with a score of 0.
Survival analysis was done using Cox proportional hazards model from the R package survival version 3.2-1066. Proportional hazards assumptions of the Cox regression were tested with cox.zph-function. Kaplan-Meier curves were plotted for visualization purposes using R package survminer version 0.4.967. p-values for survival analyses were adjusted using p.adjust-function from base R stats-package with method set to “fdr”.
Patient-derived specimens
Plasma samples from cancer-free women were obtained from the New York Blood Center, while malignant ascites fluid samples were obtained from patients with Stage III-IV HGSOC were procured through Surgical Pathology at Weill Cornell Medicine and Memorial Sloan-Kettering Cancer Center. All specimens were acquired with informed consent, classified as surgical discard, and kept totally de-identified for subsequent experimental analyses. The ascites fluid underwent initial processing by centrifugation at 4°C for 10 minutes at 400 rcf, with subsequent separation of supernatants from cell pellets and filtration through 0.22-μm filters to eliminate cellular debris. Processed samples were cryopreserved at -80°C in small aliquots to minimize freeze-thaw cycles. Sample information is described in Table S1.
Mice and experimental murine ovarian cancer models
Female mice were housed in pathogen-free microisolator cages at the animal facilities of Weill Cornell Medicine and used at 8-12 weeks of age. Mice were handled in compliance with the Institutional Animal Care and Use Committee procedures and guidelines under protocol 2011-0098. Wild type C57BL/6J and B6.Cg-Rag2tm1.1Cgn/J were purchased from the Jackson laboratory. IL-15 translational reporter mice (IL-152A-eGFP) were kindly provided by Dr. Ross Kedl68. Parental ID8 cells and the aggressive ID8-Defb29/Vegfa derivate were cultured and used as previously described19,69. Both cell lines were obtained under MTA from Drs. K. Roby and J. Conejo-Garcia, respectively. The PPNM cell line (Trp53−/−R172HPten−/−Nf1−/−MycOE) was generously provided by Dr. R. Weinberg under MTA24. The MP cell line was generated as described70. For tumor implantation, 1.5 × 106 ID8-based ovarian cancer cells suspended in 200 ml of sterile PBS was intraperitoneally (i.p.) injected into mice. Alternatively, PPNM cells were suspended in PBS containing Matrigel (Corning Matrigel matrix, Cat# 47743-716) at 1:1 ratio, and 200 ml of the mix containing 1.0 × 106 cells were administered i.p. into WT mice, as reported24. Metastatic progression, ascites accumulation, and host survival were monitored over time. Tumor burden in the peritoneal cavity was assessed by live bioluminescent imaging. Briefly, PPNM-bearing mice were given a single i.p. injection of VivoGlo luciferin (2 mg/mouse. Promega, Cat# P1042) and then imaged on a Xenogen IVIS Spectrum In Vivo imaging system at the Weill Cornell Research Animal Resource Center. All cell lines were verified for mycoplasma contamination and maintained under prophylactic plasmocin supplementation (Invivogen, Cat# ant-mpt).
Nucleic acid extraction and quantitative PCR analyses
Mouse or human total RNA was isolated using the RNeasy Mini kit (Qiagen, Cat# 74106) or QIAzol lysis reagent (Qiagen, Cat# 79306) according to the manufacturer’s instructions. 0.1-1 μg of RNA was reverse transcribed to generate cDNA using the qScript cDNA synthesis kit (Quantabio, Cat# 95047). Quantitative RT-PCR was performed using PerfeCTa SYBR green fastmix (Quantabio, Cat# 95071) on a QuantStudio 6 Flex real-time PCR system (Applied Biosystems). Normalized gene expression was calculated by the comparative threshold cycle method using ACTB for human or Actb for mouse as endogenous controls. For cytosolic and mitochondrial DNA extraction, mitochondria were fisrt obtained using the mitochondria isolation kit for cultured cells (Thermo scientific, Cat# 89874) according to the manufacturer’s instructions. Then, DNA from cytosolic and mitochondrial fractions were extracted using the Dneasy blood & tissue kit (Qiagen. Cat# 69504) according to manufacturer’s instructions. Levels of mt-Cox1 was compared to that of genomic 18S.
Mouse RNA-sequencing
ID8-Defb29/Vegfa cells were treated with vehicle or deferiprone (Sigma-Aldrich, Cat# 379409) for 12 hours, and total RNA was subsequently isolated using the RNeasy MinElute kit (Qiagen, Cat# 74204) according to the manufacturer’s instructions. Quality control checks were conducted on all samples using an Agilent Bioanalyzer 2100 to ensure RNA integrity.
mRNA libraries were generated and sequenced at the Weill Cornell Genomics Resources core facility. Raw sequencing data underwent pre-processing and analysis using the Partek software, following a standard pipeline. RNA-seq data were aligned to the mm10 genome using GSNAP method, and Partek E/M was used to estimate read counts at gene level, leveraging ensemble transcriptome information. To assess differential gene expression between experimental groups, a GSA (Gene Set Analysis) method was applied. Gene expression changes were considered statistically significant if they met the False Discovery Rate (FDR) threshold of less than 5%. Additionally, gene set enrichment analysis was performed using Ingenuity Pathway Analysis (IPA, Qiagen) with a focus on "Upstream regulators." Regulators with a statistical significance level (P-value) lower than 10-6 and with predicted activation or inhibition states (Z-score > 2) were reported. This analysis identified key molecular regulators associated with the observed gene expression changes induced by the experimental treatments. RNA-seq data were deposited under NCBI GEO Accession number (TBD).
Single cell RNA-sequencing
Human: For the analysis of single-cell RNA sequencing (scRNA-seq) data, preprocessed counts from 22 HGSOC tumor specimens11 were downloaded from Gene Expression Omnibus (GEO) with accession code GSE165897. Individual cell scores for each signature were obtained using Ucell (v1.3.1)71 and a pairwise Wilcoxon test was performed to compare the average scores between fibroblasts, immune, and epithelial ovarian cancer cells (EOCs) in each sample.
Mouse: The cellular fraction of peritoneal lavage samples was isolated from mice bearing ID8-Defb29/Vegfa ovarian cancer for 35 days and subjected to scRNA-seq. Library preparation, sequencing, and raw data preprocessing were performed at the Weill Cornell Medicine Epigenomics Core Facility using the Illumina HiSeq2500 platform. Subsequently, the reads underwent processing and analysis utilizing Seurat 4.1.272. After subsetting low quality data (min.cells=3, min.features=200, nFeature_RNA min=200 max=2500, percent.mt<5), 6,502 cells were obtained and log-normalized. Non-linear dimensional reduction using Uniform manifold approximation and projection (UMAP) with default parameters (dims=1:10, n_neighbors = 30, metric = cosine, n_trees = 50), was used to obtain cell clusters. The top 10 significant cluster’s markers were found (FindAllMarkers min.pct = 0 logfc.threshold = 0) to corroborate cell types. Lineage putative genes were used to identify the immune cell and the tumor cell clusters were designed according to the differential expressed genes using ingenuity pathway analysis (IPA, Qiagen) for the top cell function based on the following pre-determined categories: Invasive: cellular movement category [p-value= 1.05 x 10-23, z-score 2.81 (92 members upregulated)], proliferative: cell cycle category [p-value= 1.24 x10-14, z-score 2.23 (39 members upregulated)], Stem-like: embryonic development category [p-value= 5.76 x10-12, z-score 1.87 (63 members upregulated)], also the expression of putative cancer stem cell genes was evaluated: glutathione S-transferase mu 2 (Gstm2, p-value 2.8 x10-297, avg_Log(FC)= 1.22, aldehyde dehydrogenase 1 family member A1 (Aldh1a, p-value: 6.5 x10-179, avg_Log(FC)= 0.68), SRY-box transcription factor 9 (Sox9, p-value: 3.63 x10-137, avg_Log(FC)= 0.62), SRY-box transcription factor 4 (Sox4, p-value: 4.94 x10-70, avg_Log(FC)= 0.38), Chemo-resistant: embryonic development category p-value= 2.70 x10-22, z-score 1.61 (68 members upregulated) this cluster also exhibited high expression of the multidrug resistance ATP binding cassette subfamily B member 1 (Abcb1b, p-value: 1.20 x10-257, avg_Log(FC)= 0.68).
Gene sets related to iron metabolism were obtained from the Molecular Signatures Database v7.5.1 (MSigDB) for Mus musculus in the ontology gene sets (C5), and signatures related to disease were excluded. Single Sample Gene Set Enrichment Analysis (ssGSEA) was used to calculate enrichment scores for iron-related gene sets. These scores were then obtained using Escape v.1.20 (enrichIt) and plotted for each cell type. Mouse scRNA-seq data were deposited under NCBI GEO Accession number (TBD).
Cytokine and chemokine quantification
Human undiluted ascites samples were submitted to Eve Technologies™ Assay Services for analysis using the Human Cytokine/Chemokine 71-Plex Discovery Assay® Array. To obtain peritoneal lavage samples from tumor-bearing mice, we injected 5 ml of sterile PBS, and the liquid was aspirated using a 10-ml syringe with a 20G1 ½ needle. During collection the needle was detached to avoid mechanical red blood cells lysis. Subsequently, the samples were concentrated utilizing 3K Amicon tubes (Millipore, Cat# UFC800324), and all samples were normalized to a final protein concentration of 5 mg/ml. Mouse samples were submitted to Eve Technologies™ Assay Services for analysis using the Mouse Cytokine/Chemokine 44-Plex Discovery Assay® Array.
Ovarian cancer organoid development, characterization, and maintenance
Patient-derived fresh tissue samples were collected with written informed patient consent with the approval of the Institutional Review Board (IRB #1305013903) at Weill Cornell Medicine. Patient-derived tumor organoids (PDTO) lines were developed as described73 with some modifications. Briefly, fresh tissue samples were washed three times with transport media [DMEM (Gibco, Cat# 11971025) with 1X Glutamax (Invitrogen, Cat# 35050079), 100 U/mL penicillin, 100 μg/mL streptomycin (Gibco, Cat# 15140148), Primocin 100 μg/mL (InvivoGen, Cat# ant-pm), and 10 μmol/L Rock inhibitor Y-27632 (Selleck Chemical Inc., Cat# S1049)], and placed in a sterile 3-cm petri dish (Falcon) for mechanical dissection into smaller pieces (~2 mm diameter) prior to enzymatic digestion. Enzymatic digestion was done with collagenase media [DMEM (Gibco, Cat# 11971025, 100 U/mL penicillin, 100 μg/mL streptomycin (Gibco, Cat# 15140148), 250 U/mL collagenase IV (Life Technologies, Cat# 17104019), 100 μg/mL Primocin (InvivoGen, Cat# ant-pm), and 10 μmol/L Rock inhibitor Y-27632 (Selleck Chemical Inc., Cat# S1049)] in a volume of at least 20 times the tissue volume and incubated on a shaker at 200 rpm at 37°C until the digestion solution turned cloudy, typically 30-45 minutes. The suspension was centrifuged at 300 rcf for 3 minutes and the cell pellet was washed once with washing media [Advanced DMEM (Gibco, Cat# 12491023), 100 U/mL penicillin, 100 ug/mL streptomycin (Gibco, Cat# 15140148), 1x Glutamax (Cat# 35050079), and 1x HEPES (Invitrogen, Cat# 5630130)]. The cells were resuspended in a small volume of tissue-type specific primary culture media: Advanced DMEM (Gibco, Cat# 12491023) with 1x glutamax (Cat# 35050079), HEPES (Invitrogen, Cat# 5630130), B27 (Gibco, Cat# A1486701), 100 U/mL penicillin, 100 μg/mL streptomycin (Gibco, Cat# 15140148), 100 μg/mL Primocin (InvivoGen, Cat# ant-pm), 10% noggin conditioned media, 10% R-spondin conditioned media, 10 mM Nicotinamide (Sigma-Aldrich, Cat# Nicotinamide), 1.25 mM N-acetylcysteine (Sigma-Aldrich, Cat# A0737), 1ng/mL Recombinant Human FGF-b (Peprotech, Cat# 100-18B), 20ng/mL Recombinant Human FGF-10 (Peprotech, Cat# 100-26), 1 μM PGE2 (R&D Systems, Cat# 2296), 10 μM SB202190 (Sigma-Aldrich, Cat# S7067), 50ng/mL Mouse Recombinant EGF (Invitrogen, Cat# 315-09), 10 μM Y-27632 (Selleck Chemical Inc., Cat# S1049), 10 ng/mL Heregulin Beta-1 (Peprotech, Cat# 100-03), 500 nM A-83-01 (Tocris, Cat# 2939), and 100 uM Beta Estradiol (Sigma, Cat# E8875). Up to ten 100 μL drops of Matrigel/cell suspension were distributed into a 6-well cell suspension culture plate. The drops were allowed to polymerize for 30 min inside the incubator at 37°C and 5% CO2 and afterwards, 3-mL tumor type–specific primary culture media were added per well. Fresh culture media was replaced every 3 to 4 days. PDTOs at approximately 300 to 500 μm were passaged using TrypLE Express (Gibco, Cat# 12604013) for 10 minutes in the water bath at 37°C. Single cells and small cell clusters were replated according to the procedure described above. Monthly mycoplasma screening was performed using the PCR Mycoplasma Detection Kit (abm, Cat# G238).
Treatment of ovarian cancer organoids with deferiprone
50,000 cells obtained from disaggregated organoids were plated in a 100-μL cell culture media and Matrigel mix (v/v 2:1) in 6-well plates. Three domes were plated in each well of a 6-well plate. Plates were incubated for 30 minutes at 37°C to allow polymerization of the drops, and then 3 mL of tissue specific primary culture media described above was added to each well. After 72 hours of plating, the media was aspirated, replaced with 3 mL PBS, and incubated for 10 minutes. This PBS wash was repeated one more time and then 0, 100, 150, or 200 mM deferiprone (Sigma-Aldrich, Cat# 379409) was added in 3 mL of fresh media made without B27. After an incubation of 96 hours, cells were dissociated as described above and collected for downstream analysis.
In vitro drug treatments
Ovarian cancer cell lines were seeded and allowed to attach overnight. Subsequently, deferiprone (Sigma-Aldrich, Cat# 379409) or aphidicolin (Sigma-Aldrich, Cat# A0781) were added at a 2X concentration in fresh medium and incubated for the specified time intervals. After the incubation periods, cells were rinsed with PBS and then detached for downstream analysis. For treatment with inhibitors H-151 (1 mM, Invivogen, Cat# inh-h151), ATR inhibitor (1 mM, AZD6738, Selleckchem, Cat# S7693), ATM inhibitor (100 nM, AZD0156, Selleckchem, Cat# S8375), CHK1/2 inhibitor (300 nM, AZD7762, Selleckchem, Cat# S1532), cells were pre-incubated for 2 hours before treatment with deferiprone or aphidicolin. For all the experiments working solutions of deferiprone were freshly prepared in culture medium at a concentration of 10 mM, followed by the preparation of dilutions at the specified concentrations in culture medium.
MTT assay for cytotoxicity and synergy
Ovarian cancer cells were seeded at a density of 3x103 cells per well in 96-well plates containing 100 µL of culture medium. These plates were then incubated overnight at 37°C with 5% CO2. On the following day, 100 µL of deferiprone (2x concentration) was added and incubated for the specified time points. After the treatment, the cells were rinsed with RPMI medium without phenol red. Then, 100 µL of RPMI without phenol red and 50 µL of 1 mg/mL Thiazolyl Blue Tetrazolium Bromide (MTT) in sterile water were added to each well. The cells were incubated for 2 hours, followed by centrifugation, and washed with cold PBS. The plates were dried, and 100 µL of DMSO was added. After incubating for 30 minutes with shaking, the absorbance was measured at 540 nm using a Spectramax iD3 instrument (Molecular Devices). For in vitro synergy analysis, serial dilutions of cisplatin (Sigma, Cat#479306) and deferiprone (Sigma-Aldrich, Cat# 379409) were prepared and mixed to the desired concentrations and incubated for 48 hours, then MTT was performed to calculate the percentage of viability and the Bliss synergy scores were calculated using synergy finder 3.074.
Seahorse analyses
ID8-Defb29/Vegf-A were plated at a density of 2.5x105 cells per well overnight, next day cells were treated with vehicle or deferiprone (Sigma-Aldrich, Cat# 379409) at 100 or 200 mM for 6 hours, then the cells were washed and non-buffered XF base medium (Agilent, Cat# 102353-100) containing 25 mM glucose (Sigma, Cat# G7021), 2 mM L-glutamine (Gibco, Cat# 25030081), and 1 mM sodium pyruvate (Gibco, Cat# 11360070) pH= 7.4 was added. Cells were plated and OCR measurements were analyzed on an XFe96 Extracellular Flux Analyzer (Agilent). After, basal OCR measurements were obtained, an OCR trace was recorded in response to oligomycin (1 μM), carbonyl cyanide-p-(trifluoromethoxy) phenylhydrazone (FCCP, 1 μM), and rotenone and antimycin (0.5 μM each) following the XF Cell Mito Stress test kit (Agilent, Cat# 103010-100). After analysis, the cell numbers of each well were determined by nuclear DNA staining with Hoechst 33342 (Sigma, Cat# H3570) and OCR values were normalized accordingly. Finally, metabolic parameters were calculated using the seahorse Agilent Wave software (Agilent). At least ten technical replicates per treatment were examined.
Generation of IRF3-deficient ovarian cancer cell lines using CRISPR/Cas9
a 20-nucleotide sgRNAs directed against murine Irf3 (NM_016849.4) were designed to target the genomic sequences CCAGTGGTGCCTACACCCCG (IRF-3 sgRNA#1) and TGAACCGGAAAGAAGTGTTG (IRF-3sgRNA#2), (the 3 additional nucleotides highlighted in bold represent the protospacer adjacent motif, or PAM). These target sequences correspond to exon 3 of the murine Irf3 cDNA and were chosen using the Broad Institute’s CRISPick tool (https://portals.broadinstitute.org/gppx/crispick/public). As a control, a scrambled sgRNA (sg Ctrl) was used harboring a 20-nucleotide sequence that is computationally designed to be non-targeting within the murine genome. The RNA sequence for this non-targeting control was CGUUAAUCGCGUAUAAUACG. To generate Irf3-deficent Ovarian cancer lines, ID8-Defb29/Vegf-A were electroporated with ATTO-550-labeled sgRNA-Cas9 complexes using the Neon transfection system, according to the manufacturer’s protocol (IDT, Cat# 1075931). All materials for sgRNA-Cas9 complex generation were purchased from Integrated DNA Technologies (IDT) and prepared as instructed in the IDT protocols using NEON transfection system. Cells were electroporated with either scrambled sgRNA-Cas9 complexes, or the two Irf3 sgRNA-Cas9 complexes with sequences described above. 24 hours post-electroporation, fluorescently labeled ATTO-550+ single cells for each condition were sorted by FACS, expanded, and screened for Irf3 ablation separately. To screen for IRF-3 ablation, western blot analysis using rabbit anti-mouse IRF-3 (cell signaling technology Cat#D83B9) was performed on total protein isolated from cells electroporated with sgRNA-Cas9 complexes containing the Irf3 sgRNA-Cas9 described above. Following knock-out confirmation in 16 clones from the sgRNA#1 and 16 clones from the sgRNA#2, random clones from the IRF3 sgRNA#2 and the sgCtrl were used for deferiprone experiments.
Iron quantification
For total iron measurements, 50 μL of ascites fluid was digested in 50 μL 50% HNO3 water with 0.1% digitonin. Digested ascites samples were subsequently diluted 1:50 in 0.2% HNO3 water and 20 μL of sample or iron standard measured by graphite furnace atomic absorption spectroscopy. For heme iron measurements, a method based on the conversion of the heme moiety to its fluorescent porphyrin derivative by the removal of heme iron under acidic reducing conditions was used 75,76. 200 μL of ascites fluid was centrifuged at 1000 rfc for 5 mins to remove cellular debris. Ascites supernatants were divided into two 100 μL aliquots and 500 μL 2M oxalic acid added to each. One aliquot was heated to 95°C for 30 minutes to release iron from heme and generate fluorescent protoporphyrin IX. The other aliquot was left at room temperature for 30 minutes. Both aliquots were centrifuged for 10 min at 1,000 rfc at 4°C to remove debris. 200 μL of the heated and unheated aliquots were placed into a black 96-well-clear bottomed cell culture microplate (Greiner bio-one, Cat# 655090, Lot: E19083A9) and the fluorescence read at ex404 nm / em630 nm. The background fluorescence of the unheated aliquot was subtracted from the heated aliquot and the extinction co-efficient of heme at 630nm (1 μM heme = 15,200 fluorescence units) was used to determine heme concentrations as previously described76.
Proteomic analyses
Ascites samples were separated using SDS-PAGE and stained with SimplyBlueTMSafeStain (Thermo scientific, Cat# LC6065), eight bands were identified and cutted. Individual bands were submitted to the Weill Cornell proteomics & metabolomics core facility. Proteins were then concentrated by centrifugation and buffer exchange using the Amicon Ultra-0.5 Centrifugal Filter Unit with Ultracel-3 membrane (Millipore, Cat# UFC5003), according to the manufacturer’s protocol, with the exchange buffer consisting of 4 M urea, 1 M thiourea and 50 mM TEAB at pH 8.5. Proteins were reduced with 10 mM dithiothreitol, incubated at 34 °C for 1 h, then alkylated with 58 mM iodoacetamide for 45 min in the dark at room temperature and then quenched by a final addition of 36 mM dithiothreitol. The solutions were then diluted with 50mM ammonium bicarbonate (pH 8.0) to a final buffer concentration of 1M urea before trypsin digestion. Each sample was digested with 0.8 μg of trypsin for 18 h at 37 °C. The digestion was stopped by addition of TFA to a final pH 2.2–2.5. The samples were then desalted with SOLA HRP SPE Cartridge (Thermo Scientific, Cat# 60109-001). First, the cartridges were conditioned with 1 × 0.5 mL 90% methanol, 0.1% TFA and equilibrated with 2 × 0.5 mL 0.1% TFA. The samples were diluted 1:1 with 0.2% TFA and were run slowly through cartridges. After washing with 2 × 0.5 mL of equilibration solution, peptides were eluted by 1 × 0.5 mL of 50% ACN, 0.1% TFA and dried in a speed-vacuum centrifuge. Samples were reconstituted in 60 μL of 50% ACN, 0.1% TFA and loaded onto columns right after the equilibration step allowing slow flow-through. Cartridges were washed three times with 1.0 mL of equilibration solution and peptides were eluted twice with 0.6 mL of 50% ACN, 0.1% TFA, after which they were dried down in a speed-vacuum centrifuge for further use. The nano-LC–MS/MS analysis was carried out using UltiMate3000 RSLCnano (Dionex) coupled to an Orbitrap Fusion (Thermo-Fisher Scientific) mass spectrometer equipped with a nanospray Flex Ion Source. Each sample was reconstituted in 22 μL of 0.5% formic acid and 10 μL was loaded onto an Acclaim PepMap 100 C18 trap column (5 μm, 100 μm × 20 mm, 100 Å, Thermo Fisher Scientific) with nanoViper Fittings at 20 μL/min of 0.5% formic acid for on-line desalting. After 2 min, the valve switched to allow peptides to be separated on an Acclaim PepMap C18 nano column (3 μm, 75 μm × 25 cm, Thermo Fisher Scientific). Mobile phase A consisted of 2% ACN, 0.1% formic acid in water, mobile phase B was 95% ACN, 0.1% formic acid in water and the 120 min gradient was as follows: 5% to 23% to 35% B at 300 nl/min (3 to 83 to 123 min, respectively), followed by a 9-min ramping to 90% B, a 9-min hold at 90% B and quick switch to 7% B in 1 min. The column was re-equilibrated with 5% B for 20 min before the next run. A 10-fmol injection of standard BSA digest mixture with a short 30-min gradient was run for quality-control purposes. The Orbitrap Fusion instrument was operating in positive-ion mode with nano-spray voltage set at 1.7 kV and source temperature at 275 °C. External calibration for Fourier transform, ion trap and quadrupole mass analyzers was performed before the analysis. The Orbitrap full MS survey scan (m/z 400–1,800) was followed by top-3-s, data-dependent higher collision dissociation (HCD) product-dependent electron-transfer dissociation (ETD) MS/MS scans for precursor peptides with 2–8 charges above a threshold-ion count of 50,000 with normalized collision energy of 32%. Mass spectrometry survey scans were acquired at a resolving power of 120,000 (full-width at half maximum at m/z 200), with automatic gain control (AGC) = 2 × 105 and maximum injection time (maximum IT) = 50 ms, and HCD MS/MS scans at resolution 30,000 with AGC = 5 × 104, maximum IT = 60 ms and with Q isolation window (m/z) at 3 for the mass range m/z 105–2,000. Dynamic exclusion parameters were set at 1 within 60-s exclusion duration with ± 10 p.p.m. exclusion-mass width. The product-ion trigger list consisted of peaks at 204.0867 Da (HexNAc oxonium ion), 138.0545 Da (HexNAc fragment), 366.1396 Da (Hex-HexNAc oxonium ions) and 274.0927 Da (dehydrated N-acetylneuraminic acid). If one of the HCD product ions in the list was detected, two charge-dependent ETD MS/MS scans with HCD supplementary activation (SA for electron transfer and higher-energy collision dissociation (EThcD) scan) on the same precursor ion were triggered and collected in a linear ion trap. For doubly charged precursors, the ETD reaction time was set at 150 ms and the SA energy was set at 25%, and the same parameters set at 125 ms and 20%, respectively, were used for higher-charged precursors. For both ion-triggered scans, the fluoranthene ETD reagent target was set at 2 × 105, the AGC target at 1 × 104, maximum IT at 105 ms and isolation window at 3. All data were acquired under Xcalibur 3.0 operation software and Orbitrap Fusion Tune Application v.2.1 (Thermo Fisher Scientific). All mass spectrometry and MS/MS raw spectra from each sample were searched using Byonics v.2.8.2 (Protein Metrics) using Homo sapiens protein database containing 133,840 sequences and downloaded from Uniprot TrEMLB on 4 January 2016. The peptide search parameters were as follows: two missed cleavages for full trypsin digestion with fixed carbamidomethyl modification of cysteine, variable modifications of methionine oxidation and deamidation on asparagine and glutamine residues. The peptide-mass tolerance was 10 p.p.m. and fragment-mass tolerance values for HCD and EThcD spectra were 0.05 Da and 0.6 Da, respectively. The maximum number of common and rare modifications were both set at two. Identified peptides were filtered for maximum 2% FDR or 50 hits to the reverse database. The total abundance calculated as the sum of the intensity for each protein in all bands and the percentage was calculated from the total intensity.
Flow cytometry
Analyses were conducted using fluorochrome-conjugated antibodies purchased from BioLegend, unless stated otherwise. Cells were washed with PBS, Fc-gamma receptor-blocked using TruStain fcXTM (anti-mouse CD16/32, 93) and then stained for surface markers at 4°C in the dark for 30 minutes using the following antibodies: anti-CD45 (30-F11, 1:200), anti-CD3 (17A2, 1:200), anti-CD4 (RM4-5, 1:200), anti-CD8α (53-6.7, 1:200), anti-CD11c (N418, 1:200), anti-I-A/I-E (M5/114.15.2, 1:400; Tonbo biosciences), anti-CD11b (M1/70, 1:200), anti-F4/80 (BM8, 1:200), anti-Ly6g (1A81 1:200), anti-Ly6c (HK1.4, 1:200), anti-NK1.1 (PK136, 1:200), anti-TER-119 (TER-119, 1:200), anti-CD19 (6D5, 1:200), anti-CD27 (LG.3A10, 1:200), anti-CD71 (CY1G4, 1:200), anti-ULBP-1/MULT-1 (237104; R&D systems; 1:50), Rat IgG2A Isotype Control (54447; R&D systems), anti-CD262 [DR5, TRAIL-R2] (MD5-1, 1:200) and Anti-IgG Isotype Ctrl (HTK888). Cells were then washed and stained with DAPI (Biolegend) or LIVE/DEADTM Fixable Near-IR dead cell stain for live/dead discrimination (Invitrogen). Flow cytometry was performed on a LSRII or a Fortessa-X20 instruments (BD Biosciences). Cell populations were sorted from peritoneal lavage or ascites samples from Ovarian cancer -bearing mice using a FACSAria sorter (BD Biosciences) or a Sony MA900 (Sony) at the WCM CLC Flow Cytometry Core Facility, and flow cytometry data were analyzed using FlowJo v.10 (TreeStar).
Immunohistochemistry
Omentum samples were collected from tumor bearing mice and embedded in paraffin, then slides were generated in the Microscopy and Imaging Core Facility of Weill Cornell Medicine and stained for NKp46 in the Center of Comparative Medicine and Pathology-Laboratory of Comparative Pathology (LCP) of Memorial Sloan Kettering Cancer Center/Weill Cornell Medicine. Slides were scanned and processed in a Axioscan 7 instrument (Zeiss), and the images were analyzed in FIJI (ImageJ). NKp46 stained area and total tissue area were calculated by the color deconvolution function and NKp46 staining were expressed as percentage of total tissue area and the average of 3-5 fields per sample was reported.
Western blotting
Cancer cells were washed twice in 1X cold PBS and cell pellets were lysed using RIPA lysis and extraction buffer (Thermo Fisher Scientific, Cat# 89900) supplemented with a protease and phosphatase inhibitor tablet (Millipore, Cat# 11697498001 and Roche, Cat# 04906837001). Homogenates were centrifuged at 14,000 rpm for 30 min at 4°C, and the supernatants were collected. Protein concentrations were determined using a BCA protein assay kit (Thermo Fisher Scientific, Cat# 23225). Equivalent amounts of protein were separated via SDS-PAGE and transferred onto PVDF membranes following standard protocols. The following antibodies were used: anti-beta actin (Cell Signaling Technologies, Cat# 8457), anti-pTBK1 (Cell Signaling Technologies, Cat# 5483), anti-TBK1 (Cell Signaling Technologies, Cat# 3504), anti-pIRE3 (Cell Signaling Technologies Cat# 4947), anti-IRE3 (Cell Signaling Technologies, Cat# 4302), HRP-Conjugated Beta Actin Monoclonal Antibody (Thermo Fisher Scientific, Cat# MA5-15739-HRP), and goat anti-rabbit secondary antibodies conjugated with HRP (Thermo Fisher Scientific, Cat# 32460). SuperSignal West Pico (Thermo Fisher Scientific, Cat# 34580) or Femto chemiluminescent substrates (Thermo Fisher Scientific, Cat# 34095) were used to image blots in an iBright CL1000 instrument (Thermo Fisher Scientific), band intensity was measured using ImageJ Fiji.
In vivo treatments
Wild type B6, B6.Cg-Rag2tm1.1Cgn/J (Rag2) or IL-152A-eGFP reporter mice were implanted via i.p. injection with 1.5 x 106 ID8-Defb29/Vegfa, or 1.0 x 106 PPNM ovarian cancer cells. After 7 days, mice were i.p. (or by oral gavage when specified) treated every day with 150 mg/kg of deferiprone (Sigma-Aldrich, Cat# 379409) and/or once per week 5 mg/kg cisplatin (Accord, Cat# 16729-288-11) during a span of 4 weeks, both drugs were prepared in sterile human grade saline and filtered. For the acuteadministration approach, mice bearing ovarian cancer for 21 days were i.p. treated with one dose of cisplatin (5mg/kg; Accord, Cat# 16729-288-11) and/or deferiprone (150 mg/kg; Sigma-Aldrich, Cat# 379409) every day for a span of seven consecutive days. For in vivo blocking of type-I IFN signaling two approaches were used: Survival approach: mice bearing ID8-Defb29/Vegfa tumors were i.p. treated every 3 days with isotype control antibodies (Bioxcell, Cat# BE0083) or anti-IFNAR blocking antibodies (Bioxcell, Cat# #BE0241) at 200 µg/mouse, starting 3 days after tumor implantation and until the mice reach end-point criteria. Simultaneously, the mice received cisplatin and/or deferiprone as mentioned above. Acute administration approach with IFNAR blockade: Mice bearing ID8-Defb29/Vegfa tumors for 18 days received a dose of isotype control antibodies (Bioxcell, Cat# BE0083) or anti-IFNAR blocking antibodies (Bioxcell, Cat# #BE0241) at 200 µg/mouse, at day 21 mice were i.p. treated with one dose of cisplatin (5mg/kg; Accord, Cat# 16729-288-11) and/or deferiprone (150 mg/kg; Sigma-Aldrich, Cat# 379409) every day for a span of seven consecutive days. At day 28 the mice were humanely sacrificed, and peritoneal lavage was performed to collect the infiltrating cells for downstream analysis. For in vivo depletion of NK cells, mice were i.p. treated every 6 days with PK136 (Bioxcell, Cat# BE0036) and mouse IgG2a isotype control (Bioxcell, Cat# BE0085) at 200 µg/mouse, starting 6 days after tumor implantation flowed by treatment with cisplatin and deferiprone as mentioned above. Initial preliminary in vivo experiments were conducted using deferiprone provided by ApoPharma Inc. under MTA. Subsequent therapeutic and mechanistic experiments were performed using deferiprone from Sigma-Aldrich (Cat# 379409)
Dietary iron experiments
Isocaloric modified AIN-93G rodent diet (Research Diets) containing 3, 45, or 300 ppm of iron-citrate were provided ad libitum one week prior to tumor challenge. 3.0 x 106 ID8 parental cells were i.p implanted into wild type C57BL6/6J mice and mouse weight and status were monitored weekly or daily, until end-point when animals were humanely euthanized.
Mitochondrial iron content
To assess mitochondrial iron content, we utilized Mito-Ferro Green (#M489, Dojindo) following the manufacturer's guidelines. In brief, the cells were seeded overnight to allow attachment, underwent two rounds of washing with Hank's Balanced Salt Solution (HBSS) and were subsequently incubated with 5 μM Mito-Ferro Green (200 ml), prepared in HBSS for 30 minutes at 37 °C. The cells were washed twice with HBSS and 100 μM deferiprone (200 ml) was added to the cells and incubated for 30 minutes at 37 oC in 5% CO2 incubator. The cells were washed twice with HBSS, subsequently 100 μM ammonium iron (II) sulfate (200 ml) was added to the cells and incubated for 1 hour at 37 oC in 5% CO2 incubator. To visualize the cell nuclei, DAPI was used as a counterstain. After three HBSS washes, fluorescence was observed using a fluorescence microscope at the Microscopy and Imaging Core Facility of Weill Cornell Medicine. The fluorescence signal was quantified using ImageJ software.
Illustrations
Illustrations and schemes were created with BioRender.com under publication license.
Statistical analyses
All statistical analyses were performed using the GraphPad Prism software (Version 10.0.3). Significance for pairwise correlation analyses was calculated using the Spearman’s or Pearson correlation coefficient (r or r2). Comparisons between two groups were assessed using unpaired two-tailed Student’s t-test. Multiple comparisons were assessed by one-way ANOVA including Tukey’s multiple comparisons test or two-way ANOVA with Šidák’s multiple comparison test with single pooled variance. Host survival rates were compared using the Log-Rank (Mantel-Cox) test. For violin plots all data points, median and quartiles were showed, for bar plots data are presented as mean ± SEM. Unless otherwise stated, exact significant p-values are shown, and non-significant values were omitted.