Arsenic trioxide suppresses lung adenocarcinoma stem cell stemness by inhibiting m6A modification to promote ferroptosis

Abstract

Arsenic trioxide (ATO) is well known for its inhibitory effects on cancer progression, including lung adenocarcinoma (LUAD), but the molecular mechanism remains elusive. This study aimed to investigate the roles of ATO in regulating LUAD stem cells (LASCs) and the underlying mechanisms. To induce LASCs, cells cultured in an F12 medium, containing B27, epidermal growth factor, and basic fibroblast growth factor, induced LASCs. LASCs stemness was assessed through tumor sphere formation assay, and percentages of CD133+ cells were detected by flow cytometry. The Cell Counting Kit-8 method was used to assess LASCs viability, while reactive oxygen species (ROS) and iron ion levels were quantitated by fluorescence microscopy and spectrophotometry, respectively, and total m6A levels were measured by dot blot. Additionally, LASCs mitochondrial alterations were analyzed via transmission electron microscopy. Finally, the tumorigenicity of LASCs was assessed using a cancer cell line-based xenograft model. Tumor sphere formation and CD133 expression were used to validate the successful induction of LASCs from A549 and NCI-H1975 cells. ATO significantly inhibited proliferation, reduced ZC3H13 expression and total m6A modification levels, and increased ROS and iron ion content, but repressed sphere formation and CD133 expression in LASCs. ZC3H13 overexpression or ferrostatin-1 treatment abrogated LASCs stemness inhibition caused by ATO treatment, and interference with ZC3H13 inhibited LASCs stemness. Furthermore, the promotion of LASCs ferroptosis by ATO was effectively mitigated by ZC3H13 overexpression, while interference with ZC3H13 further promoted ferroptosis. Moreover, si-ZC3H13 promoted ferroptosis and impaired stemness in LASCs, which ferrostatin-1 abrogated. Finally, ZC3H13 overexpression alleviated the inhibitory effects of ATO on LASCs tumorigenicity. Taken together, ATO treatment substantially impaired the stemness of LUAD stem cells by promoting the ferroptosis program, which was mediated by its ZC3H13 gene expression inhibition to suppress m6A medication.

Introduction

Lung adenocarcinoma (LUAD) is the most prevalent histological subtype of lung cancer, which is a non-small-cell lung cancer (NSCLC) subtype and accounts for nearly 40% of all lung cancer cases [1]. LUAD is usually derived from the abnormal transformation of alveolar epithelial type II cells that secrete mucus [1] and is well known for its 5-year mortality of 60%-99% [2,3]. The high global incidence of LUAD is closely associated with various risk factors such as cigarette smoking, carcinogen exposure in work conditions, and environmental pollution [3]. Recent progress during the past decades has demonstrated that mutations in a set of driver genes, such as genetic epidermal growth factor receptor, Kirsten rat sarcoma viral oncogene, and v-Raf murine sarcoma viral oncogene homolog B (BRAF), greatly facilitated LUAD initiation and progression [3-5]. The elucidation of genetic mutations driving LUAD development guided the recent application of multiple targeted therapies such as erlotinib and gefitinib [1]. However, the incidence and mortality of LUAD remain at high levels, which necessitate the development of new therapeutic reagents and a full understanding of molecular pharmaceutical mechanisms.

Arsenic trioxide (As₂O₃ [ATO]) is a major type of toxic arsenic that has been used in ancient Chinese medicine for treating multiple severe human disorders such as rheumatic diseases, syphilis, and psoriasis. ATO was recently discovered as a potent anticancer drug for acute promyelocytic leukemia and several solid cancer types with low toxicity [6]. Additionally, the effective anticancer roles of ATO have been recently mediated by its regulation of various biological processes such as cell apoptosis, oxidative stress, angiogenesis, cell cycle arrest, and cancer stem cell activities [6,7]. Importantly, a potent therapeutic reagent for lung cancers has characterized ATO, but the responses of different lung cancer subtypes to ATO significantly vary [7]. In particular, early investigation revealed that ATO could repress the growth of both NSCLC and small cell lung cancer by inhibiting angiogenesis via modulating the vascular endothelial growth factor and Dll4-Notch signaling cascades [8]. The inhibitory effects of ATO on LUAD are mediated by multiple mechanisms, including suppression of E2F1 and thymidylate synthase [9,10]. Furthermore, ATO inhibited the viability of cancer stem-like cells (CSLC) in lung cancer through Gli1 suppression, which is a major transcription factor in the Hedgehog pathway [11]. However, the effects of ATO on LUAD stem cells (LASCs) and the underlying mechanisms remain unclear.

Ferroptosis is a newly identified form of programmed cell death, which is generally induced by iron accumulation and phospholipid peroxidation and widely exists in various species, including mammals, higher plants, fungi, and protozoa [12,13]. Recent research has revealed that ferroptosis integrated extracellular and intracellular signals, such as heat and radiation exposure, redox hemostasis, selenium metabolism, and multiple cellular signaling pathways [12]. Ferroptosis initiation and progression are mediated and regulated by interactions between many molecular machinery and signaling components [14]. Among them, glutathione peroxidase 4 (GPX4) is a phospholipid hydroperoxidase responsible for reducing phospholipid and cholesterol hydroperoxide, which serves as the key inhibitor of phospholipid peroxidation, and GPX4 inhibition is a key downstream event during ferroptosis [12,14]. Moreover, ferroptosis plays an essential role in various physiological and pathological processes, including tissue damage, inflammation, immune surveillance, and tumor development suppression [12,15]. Therefore, ferroptosis is prevalently evaded during lung cancer progression [16,17], and ferroptosis induction has been shown to mediate NSCLC inhibition by artemisinin derivatives [18]. However, little is known about the roles of ferroptosis in LUAD stem cells and its suppression by ATO. Moreover, N6-methyladenosine (m6A) modification catalyzed by methyltransferase-like protein 3 (METTL3) could initiate ferroptosis during LUAD development [19], but whether or not ATO regulates m6A modification to inhibit LUAD remains unclear.

This study aimed to explore the possible roles of ATO in modulating the ferroptosis program to regulate LASCs activity, as well as the implication of m6A modification regulation in such processes. These investigations provide a new perspective on the pharmaceutic mechanisms underlying LUAD inhibition by ATO, which may provide new clues for developing novel therapeutic reagents for lung cancer treatment.

Material and methods

Cell culture and treatment

The LUAD cell lines A549 (#CC0202) and NCI-H1975 (#CC0206) were obtained from the CellCook Company (Guangzhou, China) and separately cultured in F-12K medium (#BL311A; Biosharp, Hefei, China) and RPMI 1640 (#CM2017, CellCook), supplemented with 10% fetal bovine serum (FBS, #A3160802; Gibico), at 37°C in a humidified atmosphere with 5% CO₂. Short tandem repeat profiling was performed to validate LUAD cell line identity. The induction of A549 and NCI-H1975 cell development toward LASCs was performed as described below, which were cultured in a medium containing 0.625, 1.25, 2.5, 5, 10, 20, or 40 mM ATO (#H20080664; SL Pharm, Beijing, China) followed by normal culture for another 24 or 48 h.

Stem cell induction and sphere formation

LASCs were induced from A549 and NCI-H1975 cell lines by culturing with a medium containing B27, EGF, and basic fibroblast growth factor (bFGF). Briefly, cultured A549 and NCI-H1975 cells in the logarithmic growth phase were washed twice with PBS solution and treated with 1.0 ml of 0.25% trypsin solution for an appropriate time, followed by the addition of 10% FBS to terminate trypsin digestion. Cells were then collected by centrifugation at 800 rpm for 5 min, seeded in 6-well plates (30-50 cells/well), and cultured under 37°C and 5% CO₂ for approximately 6-8 days until cell clone formation in the 6-well plates. Single-cell clones from A549 and NCI-H1975 cells were then collected and digested with 0.25% trypsin solution again, which were collected by centrifugation at 800 rpm for 5 min, resuspended in F12 medium containing 2% B27, 20 ng/ml of EGF, and 20 ng/ml of bFGF, and planted in new 6-well plates (10³ cells/well), followed by culture at 37°C for another week until suspended stem cell spheres in regular sizes and shapes were clear under microscopy.

Cell surface marker detection

The expression of the cell surface biomarker protein CD133 in induced LASCs was assessed quantitatively by the percentages of CD133-positive cells using flow cytometry. In brief, approximately 10⁶ LASCs were first blocked in 100 μl of blocking buffer (PBS solution containing 2% BSA and 1% FBS) for 15 min in darkness, which were then incubated in darkness with diluted antibodies, including CD133 APC Mab (#FAB11331A-100; R&D Systems) and APC Mouse IgG2a Isotype Control (#E-AB-F09802E; Elabscience Biotechnology) for 30 min at 4°C. Subsequently, these LASCs cells were washed twice with 3 ml of PBS solution, resuspended in 500 μl of PBS solution, and finally analyzed by ACEA NovoCyte flow cytometry.

Cell viability detection

The viability of cultured LASCs was evaluated in this study using the Cell Counting Kit-8 (CCK-8) (#G4103; ServiceBio, Wuhan, China) following the manufacturer’s instructions. Briefly, cultured LASCs in the algorithmic growth phase were first washed once with PBS solution, digested with appropriate 0.25% trypsin solution, resuspended (3 × 10⁴ cells/ml), seeded in 96-well plates (100 μl/well), and cultured at 37°C for 24 and 48 h, which were then mixed with 10 μl of CCK-8 solution followed by culturing at 37°C for another 2 h. Finally, the viability of cultured LASCs was determined by measuring the OD450 values on a microplate reader.

Quantitative reverse transcription PCR (RT-PCR)

This study detected the relative mRNA levels of target genes by quantitative RT-PCR. Briefly, total RNA samples from cultured cancer stem cells or tumor tissues were prepared using Triquick Reagent (Trizol Substitute) (#R1100; Solarbio, Beijing, China) following the manufacturer’s instructions. The cDNA samples were then synthesized from 2 μg of total RNA using the HiScript III RT SuperMix for qPCR (#R323-01; Vazyme, Nanjing, China) following the manufacturer’s instructions. Subsequently, target gene expression was detected via real-time quantitative PCR using the ChamQ Universal SYBR qPCR Master Mix (#Q711-02; Vazyme, Nanjing, China) following the manufacturer’s instructions. The relative expression levels were finally calculated using the standard 2^-∆∆Ct method. Table 1 lists the primers used for quantitative PCR assay in this study.

Table 1.

Information of primers used in quantitative PCR assay

Gene ID	Primer sequences (5’ to 3’)	Product length (bp)
METTL14-F	GGGGTTGGACCTTGGAAGAG	156
METTL14-R	CCCATGAGGCAGTGTTCCTT
METTL16-F	ACTTTTGCATGTGCAACCCTC	232
METTL16-R	GCCAGGCTGCATTTCTTTCC
WTAP-F	ACCAACGAAGAACCTCTTCCC	204
WTAP-R	AGACTCCTGCTGTTGTTGCT
ZC3H13-F	CCTGGGGCTGCACTCTTAAA	150
ZC3H13-R	AAGTAAAATGAAACTACCTTTGGGT
FTO-F	TTGCCCGAACATTACCTGCT	92
FTO-R	TGTGAGGTCAAACGGCAGAG
ALKBH5-F	GCCGTCATCAACGACTACCA	208
ALKBH5-R	ATCCACTGAGCACAGTCACG
GPX4-F	TGAAGATCCAACCCAAGGGC	84
GPX4-R	AGCCGTTCTTGTCGATGAGG
GAPDH-F	GAGTCAACGGATTTGGTCGT	185
GAPDH-R	GACAAGCTTCCCGTTCTCAG

Western blot

Total proteins were extracted from cultured LASCs or tumor tissues using RIPA lysis buffer (#P0013B; Shanghai, Beijing, China) following the manufacturer’s instructions. The concentration of protein samples was determined using the BCA method using a protein quantitation kit (#BL521A; Biosharp, Hefei, China). Subsequently, an appropriate volume of protein samples was boiled at 98°C for 5 min, separated through SDS-PAGE, and blotted onto polyvinylidene fluoride membranes, which were then blocked with 5% lipid-free milk solution and incubated with the primary and secondary antibodies following standard procedures. Finally, protein expression was developed using enhanced chemiluminescent substrates (#BL520A; Biosharp, Hefei, China). The primary antibodies used in this study are as follows: anti-METTL14 (#A8530; ABCLONAL), anti-METTL16 (#A15894; ABCLONAL), anti-Wilms’ tumor 1-associating protein (WTAP) (#A14695; ABCLONAL), anti-ZC3H13 (#TD4623; ABMART), anti-FTO (#A1438; ABCLONAL), anti-ALKBH5 (#A11684; ABCLONAL), and anti-GAPDH (#60004-1-Ig; proteintech).

Dot blotting

The total levels of m6A in cultured LASCs and tumor tissues were quantitated by the dot blotting assay as follows. Briefly, total RNA samples were extracted from cancer stem cells or tumor tissues from naked mice as described above, which were then diluted into a concentration gradient, denatured at 95°C followed by immediate chilling, and transferred onto a nylon membrane. Subsequently, these membranes were subjected to automatic cross-linking under ultraviolet radiation, removal of unbound RNA samples, incubation with diluted antibodies recognizing N6-methyladenosine (m6A) (#ab208577; ABCAM) at 4°C under gentle shaking, three washes with washing buffer for 5 min each, followed by incubation with mouse anti-rabbit IgG-HRP antibodies (#BL003A; Biosharp, Hefei, China), and wash with washing buffer again. Finally, nylon membranes were developed using Western Blotting Substrate (#BL520A; Biosharp, Hefei, China).

Reactive oxygen species (ROS) quantitation

The ROS levels in cancer stem cells and tumor tissue slides were measured using the ROS Detection Kit (#BL714A; Biosharp, Hefei, China) following the manufacturer’s instructions. Briefly, LASCs or tumor tissue slides were incubated in darkness with appropriate volumes of H2DCFDA working solution for 30 min at 37°C, were then washed twice with serum-free culture medium, and finally observed under a fluorescence microscope with a fluorescein isothiocyanate optical filter. At least three biological replicates were used for statistical analysis.

Iron ion detection

The total iron ion (Fe) contents in LASCs or tumor tissues were quantitatively detected using the Tissue Iron Assay Kit (#A039-2-1; Nanjing Jiancheng Bioengineering Institute, Nanjing, China) following the manufacturer’s protocol. Briefly, cancer stem cells or tumor tissues were collected and homogenized on ice, which were centrifuged at 2000 rpm for 15 min, and the supernatant was collected for quantitating protein concentration as described above. Subsequently, the supernatants were mixed with 1.5 ml of dipyridine solution, boiled at 98°C for 5 min, cooled to room temperature, and centrifuged at 3500 rpm for 10 min. The OD520 values were detected on a spectrophotometer. Three biological repeats were performed for statistical analysis of iron ion levels.

Transmission electron microscopy

The mitochondrial morphology in cultured LASCs was evaluated using transmission electron microscopy. Briefly, the cultured LASCs after specified treatments were collected and fixed in 2.5% glutaraldehyde for 3 h at 4°C, washed 6 times with 0.1 M of PBS solution at 4°C for 30 min, incubated with 1% osmic acid (#18456; TED PELLA, USA) for 1 h at 4°C, washed three times with 0.1 M of PBS solution for 5 min each, dehydrated using a graded series of ethanol and acetone solutions, and immersed in resin for 4 h. Subsequently, ultrathin sections were made from fixed cell masses, which were finally observed using transmission electron microscopy (Japan Electron Optics Laboratory Co., Ltd., JEM-1400 PLUS). At least three repeats of transmission electron microscopy were performed to evaluate mitochondrial morphological alterations.

Cell transfections

The coding sequences of the ZCH13 gene were purified by PCR from cultured LASCs, which were then ligated with the pCDH-CMV-EF1A-EGFP-T2A-PURO plasmid (#P0268; MiaoLingBio, Wuhan, China) for the overexpression of the ZC3H13 gene. GeneCFPS Company (Wuxi, China) synthesized siRNA fragment 1 sequences targeting ZC3H13 (5’-CAGCAAUUAUAGAAGGUCACCAGAA-3’), siRNA fragment 2 sequences targeting ZC3H13 (5’-GGGAAUUAAUGAAGCUGGAACAAGA-3’), and the corresponding negative control (NC) sequences (5’-CAGATTAAGATTGGACCACAACGAA-3’) to silence ZC3H13 expression in cultured LASCs. The above recombinant plasmid or siRNA sequences were transfected into LASCs cells using the Lip2000 Transfection Reagent (#BL623B; Biosharp, Hefei, China) following the manufacturer’s instructions.

In vivo tumorigenicity

The tumorigenicity of LASCs was evaluated using the cancer cell line-based xenograft (CDX) model. In brief, 18 BALB/c nude mice were randomly divided into three groups (6 mice per group), including the NC group, the ATO group, and the ZC3H13 + ATO group. Mice in the ATO group and the ZC3H13 + ATO group were intraperitoneally injected with 1 μmol of ATO (#H20080664; SL Pharm, Beijing, China) every 3 days, which was done three times. Subsequently, nude mice were subcutaneously injected with 3 × 10⁶ LASCs or ZC3H13-overexpressing LASCs at the rear flank, followed by intraperitoneal injection of 1 μmol of ATO every 3 days, which lasted for 4 weeks, until visible tumor tissues were formed in nude mice. Finally, the tumor tissues formed in nude mice were measured and collected for the following assays. The same treatments were administered to mice in the NC group, except that mice were injected with the same volume of normal saline before the injection of LASCs other than ATO.

Immunofluorescence

The in-situ expression of CD133 antigen in tumor tissues was analyzed using the immunofluorescence method. Briefly, tumor tissues derived from LASCs in nude mice were first immersed in sucrose solution and then in a 1:1 mixture of sucrose and tissue embedding medium (#G6059-110ML; ServiceBio, Wuhan, China) for 2 h at room temperature, followed by two further rounds of immersion at room temperature in tissue embedding medium for 4 h and 6 h, respectively. Tissue slides (10 μm in thickness) were then made from the above tumor tissues, fixed with 4% PFA (in PBS) for 10 min at room temperature, washed twice with 0.1% Triton X-100 in PBS for 5 min, blocked with 3% blocking serum diluted in PBS for 1 h at room temperature, washed three times with PBS for 2 min, incubated with diluted primary antibodies (in 3% BSA solution) overnight at 4°C, washed again three times with PBS, and incubated with fluorescence-conjugated secondary antibody solution in darkness for 1 h at room temperature. Finally, tissue slides were counterstained with DAPI solution in darkness and were then mounted and analyzed under fluorescence microscopy.

Statistical analysis

Quantitative data obtained from at least biological replicates are presented as mean ± standard deviation and analyzed using Statistical Package for the Social Sciences version 20.0 software. Differences between groups were evaluated by Student’s t-test or analysis of variance as appropriate, and a P-value of < 0.05 indicated significant differences.

Results

ATO inhibited proliferation and m6A modification in LASCs

We first established a cellular LASCs model by inducing A549 cells and NCI-H1975 cells using a sphere formation medium (F12 containing B27, EGF, and bFGF) to investigate the cellular regulatory effects of ATO on LASCs, as introduced in the Materials and Methods section. We observed the significant tumor sphere formation from cultured A549 cells after treatment with sphere formation medium through the sphere formation assay for 1, 3, 5, and 7 days in a treatment time-dependent manner (Figure 1A), as well as that from cultured NCI-H1975 cells after treatment with sphere formation medium for 1, 3, and 7 days in a treatment time-dependent manner (Supplementary Figure 1A). Our flow cytometry assay revealed that the percentages of stem cell marker gene CD133-positive cells in induced LASCs were significantly upregulated after inductive treatment for 7 days (Figure 1B and Supplementary Figure 1B). These results revealed that we successfully induced LASCs formation from A549 and NCI-H1975 cells, which were used for the following assays.

Figure 1.

Inhibition of proliferation and m6A formation by ATO in LASCs from A549 cells. (A) Tumorsphere formation in A549 cells induced by treatment with the sphere formation medium. Tumor sphere formation after treatment for 1, 3, 5, and 7 days was evaluated using the sphere formation assay. (B) Increase in CD133-positive cells in A549 cells treated with sphere formation medium. The percentages of CD133+ cells were measured by flow cytometry. (C) Suppression of LASCs cell viability by ATO treatment for 24 or 48 h. LASCs from A549 cells were treated with 0, 0.625, 1.25, 2.5, 5, 10, 20, or 40 mM of ATO, followed by detection of cell viability by the CCK-8 method. (D and E) Effects of ATO treatment on the expression of m6A regulator genes in LASCs from A549 cells. The mRNA (D) and protein (E) levels of major m6A writers and erasers in LASCs from A549 cells were analyzed by quantitative RT-PCR and western blotting, respectively. (F) Decrease in total m6A content in LASCs from A549 cells induced by ATO treatment. The total m6A levels in LASCs from A549 cells were determined using the dot blot method. ATO: arsenic trioxide; LASCs: lung adenocarcinoma stem cells; NC: negative control; METTL14/16: methyltransferase-like protein 14/16; WTAP: Wilms’ tumor 1-associating protein; ZC3H13: zinc finger CCCH domain-containing protein 13; FTO: fat mass and obesity-associated gene; ALKBH5: alkylated DNA repair protein alkB homolog 5; *P < 0.05.

Subsequently, these induced LASCs were treated with 0, 0.625, 1.25, 2.5, 5, 10, 20, or 40 mM of ATO for 24 or 48 h, followed by the measurement of LASCs proliferation using the CCK-8 assay (Figure 1C and Supplementary Figure 1C). We revealed that ATO treatment induced significant inhibition of LASCs proliferation rates in a concentration-dependent manner (Figure 1C and Supplementary Figure 1C). The half-maximal inhibitory concentration (IC50) of ATO in LASCs from A549 cells was nearly 40 μM, and the IC50 of ATO in LASCs from NCI-H1975 cells was almost 24 μM (Figure 1C and Supplementary Figure 1C). We detected the expression of major m6A-catalyzing genes in LASCs from A549 cells treated with ATO to analyze the roles of ATO in regulating m6A modification in LASCs. We found that the expression of m6A writers methyltransferase-like protein 14 (METTL14), METTL16, WTAP, and ZC3H13 (zinc finger CCCH domain-containing protein 13), as well as m6A erasers fat mass and obesity-associated gene (FTO) and alkylated DNA repair protein alkB homolog 5 (ALKBH5) genes in LASCs from A549 cells, were greatly reduced by ATO treatment using quantitative PCR assay (Figure 1D). Among them, ZC3H13 expression showed the greatest ATO treatment induced decrease (Figure 1D). Additionally, ATO treatment in LASCs from A549 cells remarkably downregulated the protein level of ZC3H13 (Figure 1E). Congruently, we observed via dot blot that the total m6A content in LASCs from A549 cells was significantly reduced by ATO treatment compared with the NC group (Figure 1F). Moreover, we revealed that ATO treatment significantly decreased the total m6A content and the expression of ZC3H13 in LASCs from NCI-H1975 cells compared with the negative control group (Supplementary Figure 1D-F). These results revealed that ATO treatment effectively suppressed cell proliferation and m6A modification in LASCs.

ATO promotes ferroptosis and suppresses stemness in LASCs

We evaluated the influences of ATO treatment on the ferroptosis and stemness of LASCs to gain more insight into the mechanisms of LASCs regulation by ATO. We found by the tumorsphere formation assay that ATO treatment effectively repressed the sphere-forming capacities of LASCs compared with the NC group (Figure 2A and Supplementary Figure 2A). Additionally, flow cytometry revealed that ATO treatment greatly reduced the percentage of CD133+ LASCs compared with the NC group (Figure 2B and Supplementary Figure 2B). In contrast, we observed that ATO treatment significantly elevated the contents of reactive oxygen species (ROS) in the LASCs (Figure 2C). Similarly, the iron ion levels in the LASCs demonstrated substantial elevation after ATO treatment (Figure 2D). Correspondingly, ATO treatment greatly reduced the transcription level of the GPX4 gene in LASCs (Figure 2E and Supplementary Figure 2C). Consistently, our subsequent western blot revealed that ATO treatment significantly downregulated the protein abundances of GPX4 in LASCs (Figure 2F and Supplementary Figure 2D). Additionally, through transmission electron microscopy, we observed that ATO treatment caused significantly abnormal morphological alterations of mitochondria in LASCs, including mitochondrial membrane condensation, crystal membrane density elevation, shape irregularities, and mitochondrial crista reduction (Figure 2G). These results revealed that ATO can induce ferroptosis and repress stemness in LASCs.

Figure 2.

Modulation of LASCs from A549 cells ferroptosis and stemness by ATO. (A) Effects of ATO treatment on the sphere-forming capacities of LASCs. LASCs were treated with 20 mM of ATO for 48 h, and the capacity of tumorsphere formation was assessed by the tumorsphere formation assay. (B) Decreased percentage of CD133-positive LASCs caused by ATO treatment. The percentages of CD133+ cells were analyzed using the flow cytometry method. (C) Promotion of ROS production in LASCs induced by ATO treatment. LASCs were treated with 20 mM of ATO for 24 h, and the ROS levels in LASCs were quantitated by fluorescence microscopy. (D) Increases in iron ion levels in LASCs after ATO treatment. The iron ion contents in the LASCs were detected by the colorimetric assay. (E and F) Inhibition of GPX4 gene expression by ATO treatment in LASCs. The mRNA (E) and protein (F) levels of GPX4 in LASCs were detected by quantitative RT-PCR and western blotting, respectively, after ATO treatment at 20 mM for 24 or 48 h. (G) Morphological aberrations in LASCs mitochondria induced by ATO treatment. The morphological features of mitochondria in LASCs were observed via transmission electron microscopy. LASCs: lung adenocarcinoma stem cells; NC: negative control; ATO: arsenic trioxide; GPX4: glutathione peroxidase 4; GAPDH: glyceraldehyde-3-phosphate dehydrogenase; *P < 0.05.

ZC3H13 overexpression or ferrostatin-1 (Fer-1) treatment mitigates ATO-induced inhibition of LASCs stemness

We overexpressed the m6A writer gene ZC3H13 in LASCs or inhibited ferroptosis progression in LASCs by Fer-1 to assess the mediating roles of m6A modification and ferroptosis in treatment induced LASCs stemness suppression. Through quantitative RT-PCR, we demonstrated that the mRNA levels of the ZC3H13 gene in LASCs under ATO treatment were greatly elevated by ZC3H13 gene overexpression, but not by Fer-1 treatment (Figure 3A and Supplementary Figure 3A). The following western blot assay revealed similar alterations of ZC3H13 proteins in LASCs treated with the combination of ATO with Fer-1 or ZC3H13 overexpression (Figure 3B and Supplementary Figure 3B). Additionally, we revealed that the total m6A contents in the LASCs under ATO treatment were greatly increased by ZC3H13 gene overexpression, but not influenced by Fer-1 treatment (Figure 3C). Importantly, we showed that tumor sphere formation in LASCs treated with ATO was greatly promoted by Fer-1 treatment or ZC3H13 overexpression compared with cells treated with ATO alone (Figure 3D and Supplementary Figure 3C). Consistently, we found significantly increased percentages of CD133-positive LASCs under ATO treatment by Fer-1 treatment or ZC3H13 overexpression, in contrast to LASCs treated with ATO alone (Figure 3E and Supplementary Figure 3D). These results indicated that LASCs stemness inhibition by ATO treatment was mediated by the repression of m6A modification and the promotion of ferroptotic cell death.

Figure 3.

Effects of ZC3H13 overexpression and Fer-1 on the stemness of LASCs from A549 cells. A. ZC3H13 mRNA levels in LASCs that overexpress the ZC3H13 gene or treated with Fer-1. ZC3H13 mRNA levels in LASCs were analyzed by quantitative PCR. B. ZC3H13 protein abundances in LASCs with ZC3H13 gene overexpression or under the treatment. Western blotting was performed to assess ZC3H13 protein levels. C. Total m6A level changes in LASCs under ATO treatment combined with ZC3H13 overexpression or Fer-1 treatment. m6A levels in LASCs were detected via the dot blot method. D. Promotion of the sphere formation capacity of LASCs under ATO treatment by ZC3H13 overexpression or Fer-1 treatment. The tumorsphere-forming capacity of LASCs was assessed by the sphere formation assay. E. Recovery of the percentages of CD133-positive LASCs under ATO treatment by ZC3H13 gene overexpression or Fer-1 treatment. Percentages of CD133+ LASCs cells were quantitated by flow cytometry. Fer-1: ferrostatin-1; ZC3H13: zinc finger CCCH domain-containing protein 13; LASCs: lung adenocarcinoma stem cells; NC: negative control; ATO: arsenic trioxide; GAPDH: glyceraldehyde-3-phosphate dehydrogenase; *P < 0.05.

ZC3H13 overexpression abrogates LASCs ferroptosis enhancement caused by ATO treatment

We subsequently assessed the alteration of ferroptosis status in LASCs with ZC3H13 overexpression under ATO treatment to validate the implication of m6A modification in regulating LASCs ferroptosis by ATO treatment. We observed that ROS production in LASCs treated with ATO was significantly lowered by Fer-1 treatment or ZC3H13 overexpression compared with LASCs treated with ATO alone (Figure 4A). Similarly, the iron ion levels in the LASCs under ATO treatment were greatly reduced by Fer-1 treatment or ZC3H13 overexpression compared with the ATO treatment (Figure 4B). Moreover, we observed by both quantitative RT-PCR and western blot assays that Fer-1 treatment or ZC3H13 overexpression remarkably promoted GPX4 gene expression in LASCs treated with ATO, in contrast to the ATO treatment group (Figure 4C, 4D and Supplementary Figure 4A, 4B). Additionally, our transmission electron microscopy method confirmed that ATO treatment induced abnormal morphological alterations of mitochondria in LASCs which were substantially alleviated by ZC3H13 overexpression or Fer-1 treatment (Figure 4E).

Figure 4.

Alleviation of ATO-induced LASCs ferroptosis by ZC3H13 overexpression. (A) ROS content downregulation in LASCs from A549 cells treated with ATO by ZC3H13 overexpression or Fer-1 treatment. ROS levels in LASCs from A549 cells were determined using fluorescence microscopy. (B) Decrease in iron ion content in LASCs from A549 cells under ATO treatment by Fer-1 treatment or ZC3H13 overexpression. Iron ion contents in LASCs from A549 cells were measured using the colorimetric method. (C and D) Elevated GPX4 gene expression in LASCs from A549 cells under ATO treatment by Fer-1 treatment or ZC3H13 overexpression. GPX4 mRNA (C) or protein (D) levels in LASCs from A549 cells were detected by quantitative PCR and western blotting, respectively. (E) The alleviation of ATO-induced LASCs mitochondrial aberrations by ZC3H13 overexpression or Fer-1 treatment. LASCs were obtained from A549 cells. LASCs: lung adenocarcinoma stem cells; NC: negative control; ATO: arsenic trioxide; Fer-1: ferrostatin-1; ZC3H13: zinc finger CCCH domain-containing protein 13; GPX4: glutathione peroxidase 4; GAPDH: glyceraldehyde-3-phosphate dehydrogenase; *P < 0.05.

Fer-1 treatment mitigates alterations in LASCs ferroptosis and stemness induced by si-ZC3H13

We assessed the changes in ferroptosis and stemness of LASCs under si-ZC3H13-1 (or si-ZC3H13-2) and Fer-1 treatment to further analyze the roles of ferroptosis in improving m6A-regulated LASCs stemness. The quantitative PCR method revealed that si-ZC3H13-1 (or si-ZC3H13-2) greatly downregulated the mRNA levels of the ZC3H13 gene in LASCs, which the combination of Fer-1 treatment could not change (Figure 5A and Supplementary Figure 5A). Corresponding alterations of the ZC3H13 proteins in LASCs were observed after treatment with si-ZC3H13-1 (or si-ZC3H13-2) and Fer-1 (Figure 5B and Supplementary Figure 5B). Consistently, we found that si-ZC3H13-1 (or si-ZC3H13-2) transfection substantially repressed the total m6A content in the LASCs, which could not be recovered by Fer-1 treatment (Figure 5C). Moreover, si-ZC3H13-1 (or si-ZC3H13-2) transfection greatly increased the total ROS content in LASCs, which the combination of Fer-1 treatment significantly abrogated (Figure 5D). Furthermore, si-ZC3H13-1 (or si-ZC3H13-2) transfection remarkably elevated the total iron ion levels in the LASCs, which Fer-1 treatment had substantially inhibited (Figure 5E). In contrast, si-ZC3H13-1 (or si-ZC3H13-2) transfection significantly downregulated the GPX4 gene expression in LASCs, which the combination of Fer-1 treatment has effectively elevated (Figure 5F, 5G and Supplementary Figure 5A, 5B). Additionally, si-ZC3H13-1 (or si-ZC3H13-2) transfection induced abnormal mitochondrial changes in LASCs, which the combination of Fer-1 treatment has significantly alleviated (Figure 5H). Conversely, we observed that si-ZC3H13-1 (or si-ZC3H13-2) transfection greatly reduced both the sphere-forming capacity and CD133+ cell percentages in the LASCs, which the combination of Fer-1 treatment has greatly recovered (Figure 5I, 5J and Supplementary Figure 5C, 5D). These results confirmed that ZC3H13-mediated m6A modification enhances LASCs stemness by inhibiting ferroptosis.

Figure 5.

Abrogation of si-ZC3H13-induced ferroptosis enhancement and stemness inhibition in LASCs from A549 cells by Fer-1 treatment. (A and B) Expression levels of ZC3H13 gene in LASCs treated with si-ZC3H13-1 (or si-ZC3H13-2) and Fer-1. ZC3H13 mRNA (A) and protein (B) levels in LASCs were quantitated by quantitative RT-PCR and western blotting, respectively. (C) Changes in total m6A modification levels in LASCs treated with a combination of si-ZC3H13-1 (or si-ZC3H13-2) and Fer-1. Total m6A levels in LASCs were analyzed using the dot blot assay. (D) Effects of Fer-1 treatment on ROS content in LASCs transfected with si-ZC3H13-1 (or si-ZC3H13-2). Fluorescence microscopy was used for the quantitation of ROS content in LASCs. (E) Iron ion level downregulation in LASCs treated with si-ZC3H13-1 (or si-ZC3H13-2) by Fer-1 treatment. The iron ion contents in LASCs were compared using the colorimetric method. (F and G) Promotion of GPX4 gene expression by Fer-1 treatment in LASCs with ZC3H13 knockdown by si-ZC3H13-1 (or si-ZC3H13-2). The mRNA (F) and protein (G) levels of the GPX4 gene in LASCs were detected by quantitative PCR and western blotting, respectively. (H) Modulation of mitochondrial functions by Fer-1 treatment in LASCs with silenced ZC3H13 expression. Morphological alterations of mitochondria in LASCs were assessed by transmission electron microscopy. (I) Recovery of the sphere-forming capacity of ZX3H13-silenced LASCs by Fer-1 treatment. Tumor sphere formation by LASCs was evaluated using the sphere formation assay. (J) Elevation of CD133-positive cell percentages in ZX3H13-silenced LASCs by Fer-1 treatment. Flow cytometry was performed to quantify CD133-positive LASCs percentages. ZC3H13: zinc finger CCCH domain-containing protein 13; LASCs: lung adenocarcinoma stem cells; NC: negative control; si-ZC3H13-1: siRNA fragment 1 of ZC3H13; si-ZC3H13-2: siRNA fragment 2 of ZC3H13; Fer-1: Ferrostatin-1; GAPDH: glyceraldehyde-3-phosphate dehydrogenase; GPX4: glutathione peroxidase 4; *P < 0.05.

Interference with ZC3H13 further inhibited the ATO-induced LASCs stemness inhibition

ATO downregulates ZC3H13 expression, and interference with ZC3H13 combined with ATO treatment can further decrease the regulation of ZC3H13 mRNA and protein expression (Figure 6A, 6B and Supplementary Figure 6A, 6B). ATO downregulated the modification level of total RNA m6A, and interference with ZC3H13 combined with ATO treatment further downregulated the modification level of total RNA m6A (Figure 6C). ATO inhibited stem cell tumor cluster formation, and interference with ZC3H13 combined with ATO treatment further decreased stem cell tumor cluster formation (Figure 6D and Supplementary Figure 6C). ATO inhibited the proportion of CD133-positive stem cells, and interference with ZC3H13 combined with ATO treatment further decreased the proportion of CD133-positive stem cells (Figure 6E and Supplementary Figure 6D). Moreover, ATO decreased GPX4 expression at the mRNA and protein levels, and interference with ZC3H13 combined with ATO treatment further inhibited the mRNA and protein expression of GPX4 (Supplementary Figure 6A, 6B). The results revealed that ATO inhibits LASCs stemness and ferroptosis, and interference with ZC3H13 combined with ATO treatment further decreases LASCs stemness and ferroptosis.

Figure 6.

Effects of interference with ZC3H13 on the stemness of LASCs from A549 cells. A. ZC3H13 mRNA levels in LASCs with ZC3H13 combined with ATO treatment. ZC3H13 mRNA levels in LASCs were analyzed by quantitative PCR. B. ZC3H13 protein abundances in LASCs with ZC3H13 combined with ATO treatment. Western blotting was performed to evaluate ZC3H13 protein levels. C. Alterations of total m6A levels in LASCs interference with ZC3H13 combined with ATO treatment. m6A levels in LASCs were detected via the dot blot method. D. Promotion of the sphere formation capacity of LASCs interference with ZC3H13 combined with ATO treatment. The tumorsphere-forming capacity of LASCs was evaluated using the sphere formation assay. E. Percentages of CD133-positive LASCs after interference with ZC3H13 combined with ATO treatment. Percentages of CD133+ LASCs cells were quantitated by flow cytometry. ZC3H13: zinc finger CCCH domain-containing protein 13; LASCs: lung adenocarcinoma stem cells; NC: negative control; si-ZC3H13-1: siRNA fragment 1 of ZC3H13; si-ZC3H13-2: siRNA fragment 2 of ZC3H13; ATO: arsenic trioxide; GAPDH: glyceraldehyde-3-phosphate dehydrogenase; *P < 0.05.

Interference with ZC3H13 promotes LASCs ferroptosis

ATO promoted ROS activity, and interference with ZC3H13 combined with ATO treatment further enhanced ROS activity (Figure 7A). ATO upregulated intracellular iron concentration, and interference with ZC3H13 combined with ATO treatment further increased the regulation of intracellular iron concentration (Figure 7B). ATO inhibited GPX4 expression, and interference with ZC3H13 combined with ATO treatment further repressed GPX4 mRNA and protein expression (Figure 7C, 7D). ATO could damage mitochondria, whereas interference with ZC3H13 combined with ATO treatment could further damage mitochondria (Figure 7E). The results demonstrated that ATO can promote ferroptosis and that interference with ZC3H13 combined with ATO treatment can further enhance ferroptosis.

Figure 7.

Effects of interference with ZC3H13 on ferroptosis in LASCs from A549 cells. (A) ROS content downregulation in LASCs that interfere with ZC3H13 combined with ATO treatment. ROS levels in LASCs were determined using fluorescence microscopy. (B) Decrease in iron ion contents in LASCs that interfere with ZC3H13 combined with ATO treatment. Iron ion contents in LASCs were measured using the colorimetric method. (C and D) GPX4 gene or protein expression in LASCs that interfere with ZC3H13 combined with ATO treatment. GPX4 mRNA (C) or protein (D) levels in LASCs were detected by quantitative PCR and western blotting, respectively. (E) Enhancement of ATO-induced LASCs mitochondrial aberrations by interference with ZC3H13 combined with ATO treatment. LASCs: lung adenocarcinoma stem cells; NC: negative control; si-ZC3H13-1: siRNA fragment 1 of ZC3H13; si-ZC3H13-2: siRNA fragment 2 of ZC3H13; ATO: arsenic trioxide; ZC3H13: zinc finger CCCH domain-containing protein 13; GPX4: glutathione peroxidase 4; GAPDH: glyceraldehyde-3-phosphate dehydrogenase; *P < 0.05.

ZC3H13 overexpression mitigated the LASCs tumorigenicity inhibition by ATO treatment

We finally applied a cancer cell line-based xenograft model to assess the tumorigenicity of LASCs to further validate the functions of ZC2H13-mediated m6A modification in inhibiting LUAD development by ATO. We observed significantly lower tumor volumes that developed from LASCs in naked mice treated with ATO than those in the NC group, whereas ZC3H13 overexpression partially recovered tumor volumes in mice treated with ATO (Figure 8A, 8B). Similarly, ATO treatment significantly reduced the weights of tumors formed in naked mice, which were partially recovered by ZC3H13 overexpression in LASCs (Figure 8C). Additionally, we revealed that the expression of the ZC3H13 gene in the tumor tissues of the ATO group was remarkably lower than that in the NC group, but ZC3H13 overexpression in LASCs significantly elevated the expression of ZC3H13 in tumor tissues formed in the naked mice (Figure 8D). The abundance of ZC3H13 protein in the tumors formed in naked mice was also markedly increased by ZC3H13 gene overexpression in LASCs (Figure 8E).

Figure 8.

ATO impaired LASCs tumorigenicity by inhibiting ZC3H13 to promote ferroptosis. (A-C) Effect of ZC3H13 overexpression on the volumes and weights of tumors formed in ATO-treated naked mice. The volume (A and B) and weights (C) of tumors formed in naked mice were measured weekly after LASCs from A549 cells injection, which lasted for four consecutive weeks. (D and E) Alterations in ZC3H13 gene expression in tumors derived from ZC3H13-overexpressing LASCs in naked mice treated with ATO. ZC3H13 mRNA and protein levels in tumors were detected by quantitative PCR and western blotting, respectively. (F) Elevation of total m6A content in tumors developed from ZC3H13-overexpressing LASCs in naked mice treated with ATO Total m6A levels in tumor tissues were analyzed using the dot bot. (G) Changes in CD133+ cells in tumors derived from ZC3H13-overexpressing LASCs in naked mice treated with ATO. Immunofluorescence was performed to assess CD133+ cells in mice tissues. (H) ROS contents in tumor tissues developed from ZC3H13-overexpressing LASCs in naked mice under ATO treatment. (I) Influence of ZC3H13 overexpression on the iron ion level of tumor tissues in mice treated with ATO. Iron ion levels in tumor tissues were measured using the colorimetric method. (J and K) ZC3H13 overexpression-induced alterations of GPX4 expression in tumors developed from LASCs in naked mice. GPX4 mRNA and protein levels in tumor tissues were measured by quantitative PCR and western blotting, respectively. LASCs: lung adenocarcinoma stem cells; NC: negative control; ATO: arsenic trioxide; ZC3H13: zinc finger CCCH domain-containing protein 13; GPX4: glutathione peroxidase 4; GAPDH: glyceraldehyde-3-phosphate dehydrogenase; *P < 0.05.

More importantly, we found a significantly lower total m6A modification in the tumors of the ATO group than that in the NC group, which was then substantially elevated by ZC3H13 overexpression in the LASCs (Figure 8F). Our subsequent immunofluorescence assay revealed that the number of CD133+ cells in tumor tissues of the ATO group was significantly less than that of the NC group, which was then elevated by ZC3H13 overexpression in LASCs (Figure 8G). In contrast, the ROS contents in the tumors of the ATO group were significantly higher than those of the NC group, which was then greatly reduced by ZC3H13 overexpression in LASCs (Figure 8H). Similar changes in iron ion levels were detected in tumors derived from ZC3H13-overexpressing LASCs in naked mice treated with ATO (Figure 8I). Additionally, we revealed that both the mRNA and protein GPX4 gene levels in tumors of the ATO group were effectively decreased by ATO treatment in naked mice, which were then substantially elevated by ZC3H13 overexpression in LASCs (Figure 8J, 8K). Together, these results indicated that ATO could effectively impair the tumorigenicity of LASCs in naked mice by inhibiting ZC3H13 expression.

Discussion

ATO was applied for the clinical treatment of multiple human disorders in ancient China and other countries for a long time [6,20]. However, the greatest attention has only been paid to this chemical drug since the recent discovery of its potency in suppressing APL and other malignancy diseases [6,20,21]. The possibility of ATO in regulating cancer stem cells of LUAD, as well as the underlying mechanisms, remains elusive despite extensive research revealing the potency of ATO in inhibiting lung cancer development and progression. This study revealed that ATO effectively inhibited the stemness of LASCs, accompanied by ferroptosis improvement and m6A modification inhibition. Our subsequent investigations clarified that LASCs stemness inhibition by ATO mediated the suppression of ZC3H13 expression to reduce m6A modification, thereby enhancing ferroptosis. Furthermore, we revealed that ZC3H13-mediated m6A medication improves the stemness of LASCs by inhibiting ferroptosis. Finally, we confirmed, using a CDX model, that ATO treatment substantially repressed the tumorigenicity of LASCs by inhibiting ZC3H13 expression to modulate m6A formation and ferroptosis progression. These observations indicate that ATO inhibits LUAD progression by regulating m6A modification and ferroptosis to suppress LASCs stemness.

Cancer stem cells (CSCs) refer to the specialized subgroup of cancer cells featuring high tumorigenic properties such as high invasion, migration, high self-renewal, and differentiation capacities [22-24]. Since their first discovery in the 1970s, CSCs have become a research hotspot during the past decades in the cancer biology field due to the elucidation of CSCs as key drivers and main mediators of cancer initiation, high metastasis rates, immune escape, chemotherapy resistance, and disease relapse [22,25]. Therefore, developing effective inhibitors, specifically targeting CSCs, has long been regarded as a promising strategy for obtaining new therapeutic drugs with a high possibility of improving long-term outcomes. Previous research has revealed that ATO inhibits CSC activity in hepatocellular carcinoma by modulating NF-B signaling or targeting the Minichromosome maintenance protein 7 (MCM7) complex [26,27]. Additionally, ATO could repress the activities of CSLC in lung cancer [11], but its CSC regulation in LUAD remains unknown. This study confirmed, for the first time, that ATO significantly inhibited LASCs stemness, as evidenced by the remarkable decreases in sphere-forming capacity and CD133+ cell percentages, indicating the high possibility of ATO being applied as an effective anti-LUAD treatment, especially for those with strong chemotherapy resistance.

As introduced above, ferroptosis is a newly characterized type of programmed cell death characterized by the involvement of phospholipid peroxidation, ROS overproduction, and iron ion accumulation [12,14]. Additionally, ferroptosis program progression has been usually evaded in the initiation and development of various cancer types, including lung cancer [16,17]. In contrast, ferroptosis has recently repressed the invasiveness, metastasis, and chemotherapy tolerance of many types of CSCs, including ovarian, colorectal, and breast cancers [28-30]. However, the roles of ferroptosis in LUAD pathogenesis in cancer stem cell regulation remain largely unknown. This study revealed that ATO treatment induced significant ROS production and iron ion accumulation in LASCs, demonstrating that ATO can induce the ferroptosis program in LASCs. Moreover, GPX4 generally protects cells against ferroptosis because of its capability to repress phospholipid peroxides; thus, GPX4 inhibition has been applied as a common ferroptosis biomarker [31,32]. Additionally, we observed that ATO treatment significantly decreased GPX4 expression in LASCs. Furthermore, the inhibition of ferroptosis development using Fer-1 alleviated the suppression of sphere formation and CD133+ cell percentage in LASCs by ATO treatment, providing strong evidence of the mediating roles of ferroptosis in ATO-induced LASCs inhibition, which provides clues for the potential application of other ferroptosis regulators in LUAD treatment.

N6-methyladenosine (m6A) is a key type of internal transcript modification catalyzed by multiple methyltransferases, such as METTL3/14, WTAP, ZC3H13, and vir like m6A methyltransferase associated, which can be reversed by several demethylases, including FTO and ALKBH5 [33,34]. M6A modification serves as an important layer of the post-transcriptional mechanism regulating eukaryotic gene expression [35], which is closely associated with multiple physiological processes and disease development, including lung cancer, because of its significant impacts on RNA splicing, stability and degradation, translation initiation, and elongation [33,35,36]. In particular, m6A modification promotes cell proliferation and LUAD metastasis by modulating the Wnt/β-catenin signaling cascades [37]. Other extensive research in recent years has also supported the significant implication of m6A modification in LUAD pathogenesis [38-40]. More importantly, m6A catalyzed by METTL3 inhibited ferroptosis to improve LUAD development and progression [19]. In this study, we revealed that ATO treatment significantly repressed m6A writer ZC3H13 expression and m6A formation in LASCs, which essentially mediated the functions of ATO in promoting ferroptosis and repressing stemness in LASCs. Additionally, an in vivo tumorigenicity assay in naked mice further validated the mediating roles of ZC3H13-induced m6A modification and ferroptosis in LUAD inhibition by ATO treatment. These results revealed the feasibility of targeting m6A modification in CSCs as a new strategy for malignant disease prevention.

In summary, we explored the anti-LUAD pharmaceutical mechanism of ATO using a cellular LASCs model in this study and demonstrated that ATO effectively suppressed LASCs proliferation and stemness through inhibition of ZC3H13-mediated m6A modification to promote the ferroptosis program. These observations provided new insights into the molecular mechanism underlying the inhibitory effects of ATO on LUAD development and progression, which also guided the application of ATO and other m6A modifiers in the clinical management of patients with lung cancer.

Disclosure of conflict of interest

None.