Dynamic and Temporal Transcriptomic Analysis Reveals Ferroptosis-Mediated Antileukemia Activity of S-Dimethylarsino-Glutathione: Insights into Novel Therapeutic Strategy

Abstract

Open AccessCCS ChemistryRESEARCH ARTICLE30 Apr 2021Dynamic and Temporal Transcriptomic Analysis Reveals Ferroptosis-Mediated Antileukemia Activity of S-Dimethylarsino-Glutathione: Insights into Novel Therapeutic Strategy Xiaohan Xu†, Haibo Wang†, Hongyan Li and Hongzhe Sun Xiaohan Xu† Department of Chemistry, CAS-HKU Joint Laboratory of Metallomics on Health and Environment, The University of Hong Kong, Hong Kong 999077 †X. Xu and H. Wang contributed equally to this work.Google Scholar More articles by this author , Haibo Wang† Department of Chemistry, CAS-HKU Joint Laboratory of Metallomics on Health and Environment, The University of Hong Kong, Hong Kong 999077 †X. Xu and H. Wang contributed equally to this work.Google Scholar More articles by this author , Hongyan Li Department of Chemistry, CAS-HKU Joint Laboratory of Metallomics on Health and Environment, The University of Hong Kong, Hong Kong 999077 Google Scholar More articles by this author and Hongzhe Sun *Corresponding author: E-mail Address: [email protected] Department of Chemistry, CAS-HKU Joint Laboratory of Metallomics on Health and Environment, The University of Hong Kong, Hong Kong 999077 Google Scholar More articles by this author https://doi.org/10.31635/ccschem.021.202000685 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesTrack Citations ShareFacebookTwitterLinked InEmail S-dimethylarsino-glutathione (ZIO-101, darinaparsin®) exhibits broader antitumor activity and less toxicity than arsenic trioxide (ATO), a clinically used drug for acute promyelocytic leukemia (APL) treatment. However, the mechanisms of action underlying antileukemia activities of ZIO-101 remain largely unknown. Herein, by integrating dynamic transcriptomic analysis and biochemical characterizations, we for the first time delineated that ZIO-101 exerted antiproliferative effects against leukemia cells via activating ferroptosis pathway, a newly discovered iron-dependent programmed cell death, at the early stage, as evidenced by abnormally elevated intracellular iron contents and lipid peroxidation. We further demonstrated that silencing heme oxygenase 1 (HMOX1), an important iron homeostasis-related gene, effectively attenuated ferroptosis induced by ZIO-101, with iron accumulation and lipid peroxidation being significantly alleviated. Significantly, we discovered that ZIO-101 and kinase inhibitors (Dasatinib/Dactolisib) could synergistically kill leukemia cells, with a combination index of <1.0 under all the tested drug concentrations. Our findings regarding the ferroptosis-mediated antileukemia activity of ZIO-101, based on the dynamic and temporal transcriptomic analysis, provide promising approaches to combat drug-resistant leukemia by combining ZIO-101 with kinase inhibitors. The methodology may be further exploited for uncovering the modes of action of other drugs. Download figure Download PowerPoint Introduction Ferroptosis is a newly discovered programmed cell death characterized by intracellular iron accumulation and lipid peroxidation. It was first proposed by Dixon et al.1 in 2012, and since then, this nonapoptotic cell death pathway has attracted considerable attention. Accumulated evidence has shown that ferroptosis plays a critical role in many human diseases, including cancer, neurodegenerative disorders, cardiopathy, and so forth.2–4 The disturbance of intracellular lipid peroxidation metabolism and iron homeostasis is a primary underlying mechanism of ferroptosis.5,6 Inhibition of glutathione peroxidase 4 (GPX4), a cytotoxic lipid peroxides buffering enzyme, by system Xc- (cystine/glutamate antitransporter) dysfunction can induce lipid reactive oxygen species (ROS) accumulation and lead to the generation of ferroptosis.7 In addition, disruption of intracellular iron homeostasis can severely influence normal biological processes.8 Proteins involved in cellular iron regulation and distribution, such as transferrin receptor protein 1 (TFR1),9 heme oxygenase 1 (HMOX1),10 and heat shock protein beta-1 (HSPB1),11 have been validated to play important roles in the induction of ferroptosis. Consistently, ferroptosis can be effectively inhibited by iron chelators and substances that suppress lipid peroxidation, such as ferrostatin-1 (Fer-1), vitamin E, and liproxstatin-1.7 To maintain abnormal proliferation, cancer cells exhibit a significantly increased iron requirement compared with normal cells, which renders them more vulnerable to ferroptosis.12 Given that the effectiveness of apoptosis inducers in cancer treatment is relatively limited due to both intrinsic and acquired drug resistance, numerous studies have focused on discovering novel ferroptosis inducers to combat cancers recently.12 It has been reported that cisplatin can induce ferroptosis in different cancer cells and the combination of cisplatin and erastin (the classic ferroptosis-inducing agent) exhibits significant synergism in killing cancer cells.13 Very recently, auranofin, an antirheumatoid arthritis drug, has been demonstrated to mitigate systemic iron overload and induce ferroptosis through inhibiting thioredoxin reductase activity.5 Therefore, targeting ferroptosis may serve as a promising alternative strategy for the development of novel cancer therapies. The therapeutic utilizations of arsenic-based drugs can be traced back to ancient societies.14,15 Arsenic trioxide (ATO) has been successfully used for the treatment of leukemia.16,17 However, its severe toxicity to human organs significantly limits its wider therapeutic applications,18 which has aroused great interests in exploring low-toxic alternatives of ATO.19 In recent years, organic arsenic compounds, with some already in clinical trials, have been extensively explored due to their more potent antiproliferative efficacy but lower toxicity compared with inorganic arsenic agents.20 S-dimethylarsino-glutathione (ZIO-101, darinaparsin®), a promising organic arsenic agent with improved antitumor behaviors toward a variety of malignancies in vitro and in vivo, is now under phase I/II clinical trial with both oral and intravenous administration.21 In spite of great efforts, the modes of action of ZIO-101 still remain largely elusive. Given that metalloanticancer drugs often act on various targets through multifaceted mechanisms,22–24 understanding their modes of action may in turn extend their pharmaceutical potential.25–27 For example, the combination of two or more therapeutic interventions that have distinctive mechanisms of action can minimize the development of drug resistance and reduce severe side effects.28 It has been reported that ZIO-101 exerts its therapeutic effects by inducing apoptosis in cancer cells.21,29 However, it is unknown whether ZIO-101 can also activate other types of programmed cell death pathways, for example, ferroptosis and autophagy. Herein, we carried out a genome-wide transcriptional analysis to explore the mechanisms of action underlying the antileukemia activities of ZIO-101. Surprisingly, ferroptosis was identified for the first time as the most prominently activated pathway in NB4 cells [acute promyelocytic leukemia (APL)] treated with ZIO-101 at the early stage. Furthermore, knockdown of HMOX1, an iron homeostasis-related gene that was dramatically upregulated by ZIO-101, significantly ameliorated the abnormal iron accumulation and lipid peroxidation, which unveiled the mechanisms of ferroptosis. The ferroptosis-mediated antileukemia activity of ZIO-101 offers more potential for its utilizations in combination therapies to fight cancer. Interestingly, we found that ZIO-101 and kinase inhibitors showed synergistic effects against NB4 cells, providing a novel therapeutic strategy to combat drug-resistant APL. Experimental Methods NB4 cells were harvested after being treated with 5 μM ZIO-101 (Toronto Research Chemicals, Toronto, ON, Canada) for 0, 6, 12, and 24 h. The whole-cell RNA in each sample was then extracted for genome-wide transcriptional analysis. Afterward, bioinformatics analysis, including principal component analysis (PCA), pathway enrichment analysis, and so forth, were conducted to explore dynamic transcriptomic alterations in leukemia cells mediated by ZIO-101. Subsequently, biochemical characterizations, such as real-time polymerase chain reaction (PCR), western blotting, flow cytometry, gene knockdown, cell viability assay, and so forth, were employed to illustrate the vital roles that ferroptosis and HMOX1 played in the antileukemia activities of ZIO-101. More experimental details are available in the Supporting Information. Results and Discussion ZIO-101 tunes transcriptome in leukemia cells We first carried out a genome-wide transcriptional analysis to investigate the underlying mechanisms of antileukemia activities of ZIO-101. NB4 cells were treated with 5 μM ZIO-101 based on cell viability data ( Supporting Information Figure S1) for 0, 6, 12, and 24 h to explore dynamic transcriptomic alterations mediated by ZIO-101 in leukemia cells. Six biological replicates were prepared for each ZIO-101 treatment group, and three of them were combined into one sample, of which the whole-cell RNA was extracted and quantified. Supporting Information Table S1 summarizes the concentration, purity, and integrity [RNA integrity number (RIN)] of extracted total RNA in each sample. All RNA samples showed good quality and were delivered to Gene Denovo Biotechnology Co. (Guangzhou, China) for RNA library preparation (Illumina, San Diego, CA) and sequencing on an Illumina HiSeq2500 machine. Bioinformatics analysis was conducted with DESeq230 and edgeR31 software. PCA was performed to obtain an overview of the transcriptomic variations.32 As shown in Figure 1a, based on PC1 (60.5%) and PC2 (30.3%), the samples were separated into four distinct groups, suggesting significant transcriptomic differences among four ZIO-101 treatment groups. The high repeatability of RNA samples was verified by pairwise Pearson’s correlation coefficients of two replicates among four ZIO-101 treatment groups (Figure 1b). The dynamic effects of ZIO-101 were investigated by studying differential gene expressions across a 24 h-RNAseq time course. Compared with the control group, we observed significant gene expressions after ZIO-101 exposure for 6 h. The largest numbers of differentially expressed genes (DEG) were detected at 24 h, with 2763 and 741 genes being up- and downregulated, respectively (Figure 1c). The total numbers of DEG in three comparisons are displayed in the Venn diagram (Figure 1d) with 829 dysregulated genes being identified in all three ZIO-101 treated groups. Unsupervised hierarchical clustering analysis (HCA) (Figure 1e) of four treatment groups showed that they clustered into four distinct groups, which coincided with PCA results. Taken together, we demonstrate that, in general, a dynamic and temporal response of NB4 cells to ZIO-101 treatment can be distinguished from each other using large-scale gene expression analyses.33,34 Figure 1 | Time-resolved transcriptomic analysis of ZIO-101-stimulated alterations in NB4 cells. (a) PCA of the four transcriptomes. (b) Pairwise Pearson’s correlation coefficients of the sequencing data from four ZIO-101-treated groups in two replicates. (c) The numbers of the up- and downregulated genes in NB4 cells treated with ZIO-101 for 6, 12, and 24 h. (d) A Venn diagram showing the numbers of DEG in three comparisons: 6 h vs ctrl, 12 h vs ctrl, and 24 h vs ctrl. (e) A heatmap showing expression levels of genes in the four transcriptomes. (f) Volcano plots for expressed genes in three comparisons, that is, 6 h vs ctrl, 12 h vs ctrl, and 24 h vs ctrl. x- and y-axes represent the log2(FC) (FC = fold change) for the two samples and −lg(FDR) (FDR = false discovery rate), respectively. Red (upregulated) and blue (downregulated) dots indicate the genes with significant differences, while the grey dots correspond to genes with no significant difference. (g) Top 20 enriched KEGG pathways in three comparisons, that is, 6 h vs ctrl, 12 h vs ctrl, and 24 h vs ctrl. (h) The FDs of gene expression levels in NB4 cells under ZIO-101 treatment determined by RT-PCR. Eight representative genes were selected to validate the transcriptional analysis results. Download figure Download PowerPoint Significantly regulated genes (Figure 1f) in ZIO-101 treated NB4 cells were classified by gene ontology (GO) analysis to clearly illustrate their biological significance. Biological regulation, metabolic process, signaling, and cell killing are vital biological processes that have been tuned by ZIO-101 ( Supporting Information Figure S2). The Database for Annotation, Visualization, and Integrated Discovery (DAVID) was utilized to analyze the enriched Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways, and the results indicated that ZIO-101 activated certain crucial cell death pathways in NB4 cells after 6 h-exposure, including ferroptosis, transcriptional misregulation in cancers, mitogen-activated protein kinases (MAPK) signaling pathway, and so forth (Figure 1g). Slight changes in the top 20 enriched KEGG pathways were observed with the increase of ZIO-101 incubation time. Some pathways, for example, the NF-κB signaling pathway, the MAPK signaling pathway, and the nucleotide-binding, oligomerization domain (NOD)-like receptor signaling pathway, were enriched more dramatically in NB4 cells exposed to ZIO-101 for 12 and 24 h. q-Value was employed as the indicator of the enrichment of pathways (Figure 1g). Among the top 20 enriched pathways, the q-value of “ferroptosis” was minimal in NB4 cells treated with ZIO-101 for 6 and 12 h, showing that ferroptosis was the most prominently activated pathway in NB4 cells treated with ZIO-101 for 6 and 12 h, but not for 24 h. In other words, ZIO-101 initiated cell death through activating ferroptosis at the early stage. In NB4 cells treated with ZIO-101 for 24 h, apoptosis was one of the top 20 KEGG pathways that were prominently activated, indicating that apoptosis was induced at 24 h, which coincided with the previous report.21 To verify the accuracy and repeatability of the genome-wide transcriptional analysis, we chose eight representative genes involved in important pathways (i.e., ferroptosis, the MAPK signaling pathway, and the NF-κB signaling pathway) and examined their expression levels by real-time PCR (RT-PCR). We found that the changes in expression levels of these genes obtained by RT-PCR were coincident with the transcriptome analysis data (Figure 1h and Supporting Information Table S2). ZIO-101 activates ferroptosis in NB4 cells at the early stage Ferroptosis, a new type of programmed cell death that depends on intracellular iron homeostasis, plays a vital role in regulating the development of cancers. Activation or inhibition of ferroptosis has become a promising strategy in cancer therapy.7 Transcriptional analysis indicates that ferroptosis is the most significantly activated pathway in NB4 cells upon exposure to ZIO-101 for 6 and 12 h. In total, 21 genes involved in ferroptosis were significantly regulated compared with the untreated group (as control) (Figure 2a and Supporting Information Table S3). Protein–protein interaction (PPI) analysis of these genes by Search Tool for the Retrieval of Interacting Genes/Proteins (STRING)35 are presented by Cytoscape36 and displayed in Supporting Information Figure S3. Figure 2 | ZIO-101 induces ferroptosis in NB4 cells at the early stage. (a) Heatmap of 21 significantly regulated genes involved in ZIO-101-induced ferroptosis. (b) Dynamic fold changes of expression levels of four iron homeostasis-related ferroptotic genes in NB4 cells induced by ZIO-101 (n = 3). (c) Protein expression level and relative band intensity of HMOX1, FTL, MAP1LC3B, NCOA4, and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) determined by western blot in NB4 cells treated with ZIO-101 for 0, 6, 12, and 24 h (n = 3). Bar graph shows the quantification of protein levels normalized to untreated control group. (d) ROS production measured by CM-H2DCFDA staining via flow cytometry (n = 3). Left: Flow cytometric histogram of ZIO-101-treated NB4 cells. Right: Fluorescent intensity normalized to untreated control group. (e) Lipid ROS production measured by lipid peroxidation sensor (C11-BODIPY-581/591 dye) staining in ZIO-101-treated NB4 cells via flow cytometry (n = 3). (f) Western blot showing elevated 4-HNE protein adducts in ZIO-101-treated NB4 cells with quantification (n = 3). (g) Measurements of intracellular iron contents in NB4 cells treated with ZIO-101 for 0, 6, 12, and 24 h (n = 3). (h) Cell viability of NB4 cells treated with 5 μM ZIO-101, 5 μM ZIO-101 + 10 μM Fer-1, and 5 μM ZIO-101 + 10 μM DFO for 12 h (n = 3). The results are displayed as mean ± SEM (b–h). Two-tailed t-test was utilized for all comparisons between two groups. *p < 0.05, **p < 0.01, ***p < 0.001. Download figure Download PowerPoint Among the 21 regulated genes, 11 and 10 genes were identified being up- and downregulated, respectively, upon exposure of NB4 cells to ZIO-101 for 6 and 12 h. Importantly, we found that genes encoding proteins closely related to cellular iron homeostasis were significantly activated, while the expression levels of genes involved in the glutathione (GSH)/glutathione peroxidase 4 (GPX4) pathway were barely changed ( Supporting Information Figure S4). The dynamic fold changes of expression levels of iron homeostasis-related genes are shown in Figure 2b. HMOX1 was one of the most dramatically upregulated genes, with the expression level being elevated approximately 300-, 100-, and 5-fold in NB4 cells after being treated with ZIO-101 for 6, 12, and 24 h, respectively. As the mediator of the first step of heme catabolism, HMOX1 cleaves heme to generate biliverdin, CO, and Fe2+.37 Overexpression of HMOX1 has been reported to display antiproliferative, anti-inflammatory, and anticancer properties.38 HMOX1 upregulation can also change intracellular iron homeostasis and distribution by enhancing heme degradation and ferritin production in different cells.39,40 The gene encoding ferritin light chain (FTL) protein that played an indispensable role in the storage and delivery of iron as well as the maintenance of iron homeostasis41 was upregulated approximately fivefold. As an autophagic marker, microtubule-associated proteins 1A/1B light chain 3B (MAP1LC3B) is an important iron homeostasis-relevant protein and closely related to ferroptosis since the generation of ferroptosis is accompanied by increased autophagy.42,43 The gene-encoding MAP1LC3B was upregulated eightfold. For nuclear receptor coactivator 4 (NCOA4), the expression level was slightly decreased to 0.5- and 0.6-fold in NB4 cells treated with ZIO-101 for 6 and 12 h, while at 24 h, the expression level was elevated 1.6-fold. Protein expression levels of these genes were further examined by western blot (Figure 2c). The expression of HMOX1, FTL, LC3B, and NCOA4 was increased 1.8-, 1.2-, 1.6-, and 1.1-fold, respectively, in NB4 cells exposed to ZIO-101 for 6 h and the changes became more significant, that is, 3.5-, 1.7-, 2.5-, and 1.6-fold, respectively, after 12 h. Such an increase remained until 24 h for HMOX1 (4.6-fold) and NCOA4 (2.1-fold). For FTL and LC3B, the expression levels were slightly decreased from 12 to 24 h. In general, ZIO-101 treatment significantly increased the expression levels of iron homeostasis genes coded proteins, consistent with the changes at transcriptomic levels. It has been reported that upregulating HMOX1 can elevate the expression of FTL, which leads to the release of a large amount of iron, causing the abnormal increase of intracellular iron contents and lipid ROS.22,44 Based on our transcriptional analysis, we reasoned that ferroptosis could be induced by ZIO-101 in NB4 cells through stimulating the expression of HMOX1 and other iron-regulating proteins since lipid peroxidation and iron accumulation are two major biochemical features of ferroptosis.7 To verify our hypothesis, we first measured intracellular ROS levels using 2,7-dichlorodihydrofluorescein diacetate CM-H2DCFDA as a fluorescent probe via flow cytometry (Figure 2d). For unstained cells and the CM-H2DCFDA stained control group, the calculated median fluorescence intensity (MFI) values were 230 and 6668 a.u., respectively. After treatment with ZIO-101 for 6 and 12 h, the cellular ROS levels increased to around 130% (MFI = 8682 a.u.) and 185% (MFI = 12,435 a.u.), demonstrating ZIO-101-induced ROS production in NB4 cells in a time-dependent manner.45 Subsequently, we measured the lipid ROS levels in NB4 cells treated with 5 μM ZIO-101 for 0, 6, 12, and 24 h using flow cytometry and lipid peroxidation sensor (C11-BODIPY-581/591 dye). The ratio of the signal from the 510 to 590 channel was used to quantify lipid ROS levels in cells. In control cells, the signal was in the red channel, and the ratio of 510/590 was very low since most lipids were nonoxidized. When the cells were treated with ZIO-101 for 6 and 12 h, the ratios were significantly increased 1.9- and 2.1-fold, respectively. However, the ratio decreased 0.7-fold upon 24 h treatment, demonstrating that ZIO-101 treatment evidently increased the lipid ROS level in NB4 cells at the early stage (Figure 2e). We further measured the expression level of 4-hydroxynonenal (4-HNE) protein adducts in NB4 cells to determine the changes of intracellular lipid ROS production since 4-HNE, an α,β-unsaturated hydroxyalkenal, is a byproduct of lipid peroxidation in cells. As shown in Figure 2f, compared with the control group, the expression level of 4-HNE protein adducts was elevated 1.5- and 1.6-fold, respectively, after NB4 cells were treated with ZIO-101 for 6 and 12 h. However, no obvious change of 4-HNE expression level was observed in NB4 cells exposed to ZIO-101 for 24 h, which further illustrated that ZIO-101-induced lipid peroxidation in NB4 cells at the early stage. Besides, changes of iron contents in NB4 cells induced by ZIO-101 were assessed using a standard Iron Assay Kit. After treatment with ZIO-101 for 6, 12, and 24 h, the iron concentration in NB4 cells increased to around 160%, 183%, and 168%, respectively, demonstrating that ZIO-101 significantly induced the excessive accumulation of iron in NB4 cells (Figure 2g). To further validate that ZIO-101 could lead to ferroptotic cell death in NB4 cells through inducing abnormal intracellular iron accumulation and excessive lipid ROS, we measured the viability of NB4 cells exposed to the combination of ZIO-101 and ferroptosis inhibitor or iron chelator. As displayed in Figure 2h, the viability of NB4 cells decreased to approximate 63% after treatment with 5 μM ZIO-101 for 12 h. However, the inhibitory effect of ZIO-101 was significantly reduced by adding 10 μM Fer-1, a promising and selective ferroptosis inhibitor that prevents intracellular lipid ROS accumulation.1 Moreover, addition of 10 μM iron-chelating agent deferoxamine (DFO) to NB4 cells under 5 μM ZIO-101 treatment also increased the viability from 63% to 81%, illustrating that cell death induced by ZIO-101 could be effectively inhibited by decreasing intracellular iron concentrations. Since ferroptosis is a programmed cell death arising from an iron-dependent accumulation of lipid ROS,1 the aforementioned results collectively confirm that ZIO-101 activates ferroptosis in NB4 cells at the early stage. HMOX1 knockdown alleviates ZIO-101-mediated ferroptosis Because HMOX1 is critical for iron homeostasis, a decisive factor in ferroptosis, and it was dramatically upregulated in NB4 cells upon ZIO-101 treatment both at transcriptomic and protein levels, we anticipated that HMOX1 would play a vital role in ZIO-101-induced ferroptosis. To validate this hypothesis, we knocked down HMOX1 in NB4 cells using HMOX1-specific siRNA and the transfection reagent Lipofectamine™ 2000.46 Compared with the negative control group, the expression level of HMOX1 was reduced to 27% in NB4 cells transfected with 10 nM HMOX1-specific siRNA for 24 h (Figure 3a). Figure 3 | HMOX1 knockdown alleviates ZIO-101-activated ferroptosis in NB4 cells. (a) HMOX1 expression levels determined by western blot and relative band intensity in NB4 cells transfected with HMOX1 siRNA or negative control siRNA. (b) ZIO-101-induced fold changes of expression levels of ferroptosis-related genes determined by RT-PCR in NB4 cells with siRNA transfection (n = 3). (c) Comparison of the cell viability upon ZIO-101 treatment in NB4 cells with siRNA transfection (n = 3). (d) Changes of intracellular iron contents induced by ZIO-101 in NB4 cells with siRNA transfection (n = 3). (e) Expression levels of 4-HNE protein adducts determined by western blot in NB4 cells with siRNA transfection upon ZIO-101 treatment. (f) ROS production measured by CM-H2DCFDA staining in ZIO-101-treated NB4 cells with siRNA transfection via flow cytometry (n = 3). (g) Lipid ROS production measured by lipid peroxidation sensor (C11-BODIPY-581/591 dye) staining in ZIO-101 treated NB4 cells with siRNA transfection via flow cytometry (n = 3). (h) A diagram for displaying the proposed mechanisms of action of ferroptosis in NB4 cells caused by ZIO-101 via upregulating HMOX1. The results are displayed as mean ± SEM (a–g). Two-tailed t-test was utilized for all comparisons between two groups. *p < 0.05, **p < 0.01, ***p < 0.001. Download figure Download PowerPoint We then examined the influence of HMOX1 knockdown on the expression levels of ferroptosis-related genes. NB4 cells transfected with negative control siRNA or HMOX1-specific siRNA were incubated with 5 μM ZIO-101 for 12 h and the fold changes of expression levels of ferroptosis relevant genes were measured by RT-PCR (Figure 3b). The expression level of HMOX1 was reduced by 48% after NB4 cells were transfected with specific siRNA. In contrast, HMOX1 knockdown had no significant effect on the expression of the three other ferroptotic genes (FTL, LC3B, and NCOA4). Generally, ZIO-101 significantly stimulated the upregulation of these genes; however, the changes were less obvious in HMOX1 knockdown cells compared with the negative control group. In detail, 155- and 16-fold changes for the expression of HMOX1 and FTL were induced by ZIO-101 in the control cells; whereas dramatic drops in these changes, that is, 53- and 3-fold changes, respectively, were observed in HMOX1 knockdown cells. Consistent with this, changes in the expression of FTL and LC3B at protein levels under ZIO-101 treatment in HMOX1 knockdown cells determined by western blot were less significant than those in negative control cells ( Supporting Information Figure S5), demonstrating that HMOX1 knockdown effectively alleviated the upregulation of ferroptotic genes induced by ZIO-101. To explore whether HMOX1 knockdown could affect ZIO-101-induced ferroptotic cell death, we measured the cell viability upon ZIO-101 treatment in NB4 cells after siRNA transfection. Since ferroptosis is the most prominently activated pathway in NB4 cells under ZIO-101 treatment at the early stage, and the IC50 value of ZIO-101 to NB4 cells after incubation for 7 h is approximately 11.7 μM,29 we investigated the viability of NB4 cells treated with 20 μM ZIO-101 for 8 h. As shown in Figure 3c, the viability was determined to be 44% after NB4 cells transfected with HMOX1 siRNA were treated with 20 μM ZIO-101 for 8 h, which was higher than the control group (35%), suggesting that silencing HMOX1 could effectively inhibit ZIO-101-induced cell death. We further investigated the effects of HMOX1 knockdown on iron contents and ROS levels in NB4 cells to intuitively show the functional role of HMOX1 in ZIO-101-induced ferroptosis. The changes in the iron contents in NB4 cells transfected