Identification of a set of genes potentially responsible for resistance to ferroptosis in lung adenocarcinoma cancer stem cells

Abstract

Scientific literature supports the evidence that cancer stem cells (CSCs) retain inside low reactive oxygen species (ROS) levels and are, therefore, less susceptible to cell death, including ferroptosis, a type of cell death dependent on iron-driven lipid peroxidation. A collection of lung adenocarcinoma (LUAD) primary cell lines derived from malignant pleural effusions (MPEs) of patients was used to obtain 3D spheroids enriched for stem-like properties. We observed that the ferroptosis inducer RSL3 triggered lipid peroxidation and cell death in LUAD cells when grown in 2D conditions; however, when grown in 3D conditions, all cell lines underwent a phenotypic switch, exhibiting substantial resistance to RSL3 and, therefore, protection against ferroptotic cell death. Interestingly, this phenomenon was reversed by disrupting 3D cells and growing them back in adherence, supporting the idea of CSCs plasticity, which holds that cancer cells have the dynamic ability to transition between a CSC state and a non-CSC state. Molecular analyses showed that ferroptosis resistance in 3D spheroids correlated with an increased expression of antioxidant genes and high levels of proteins involved in iron storage and export, indicating protection against oxidative stress and low availability of iron for the initiation of ferroptosis. Moreover, transcriptomic analyses highlighted a novel subset of genes commonly modulated in 3D spheroids and potentially capable of driving ferroptosis protection in LUAD-CSCs, thus allowing to better understand the mechanisms of CSC-mediated drug resistance in tumors.

Fig. 1. Lung adenocarcinoma CSCs are resistant to ferroptosis activator RSL3.

A Representative images at 10x magnification of BBIRE-T248, PUC30, PUC36, and PUC37 primary cultures grown in adherent (2D) and in non-adherent (3D) conditions. The scale bar represents 50 μm. B Real-time PCR analyses of stemness-associated genes in primary cells cultured in 2D and 3D conditions. Data were representative of three independent experiments; values are expressed as mean ± SEM and are statistically significant if *p < 0.05 (paired Student’s t-test). C, D Dose-response curves of sensitivity to RSL3 at 72 h in BBIRE-T248, PUC30, PUC36, and PUC37 primary cultures (C) and in stable cell line NCI-H460 (D) grown in 2D and 3D conditions. E Comparison of dose-response curves of sensitivity to RSL3 in all cell lines grown in adherence (2D). F Workflow of the sphere disruption experiment and subsequent evaluation of sensitivity to RSL3 (Created with BioRender.com). G Dose-response curves of sensitivity to RSL3 at 72 h in BBIRE-T248 2D, 3D, and disrupted 3D cells. Data were representative of three independent experiments; values are expressed as mean ± SEM and are statistically significant if *p < 0.05 (Welch’s t-test) for the comparison between 2D and 3D (C, D) or for the comparison between 3D and disrupted 3D (G). The IC₅₀ (half maximal inhibitory concentration) values are reported in the graphs.

Fig. 2. LUAD-CSCs resist to RSL3-induced ferroptosis by counteracting lipid and mitochondrial ROS production.

A Workflow of measurement of lipid peroxidation, mitochondrial peroxidation, and cell death in primary and stable cell lines grown in 2D and 3D conditions following RSL3 and Ferrostatin-1 treatment (Created with BioRender.com). B Density curves of lipid peroxidation of a representative experiment of BBIRE-T248, PUC30, PUC36, PUC37, and NCI-H460 cell lines treated with DMSO (UNT), 1 μM RSL3 (RSL3), and 1 μM RSL3 + 2 μM Ferrostatin-1 (RSL3 + FER-1) for 2 h. C Histograms of three independent experiments of C11-BODIPY (581/591) staining to detect lipid peroxides. D Density curves of mitochondrial peroxidation of a representative experiment of BBIRE-T248, PUC30, PUC36, PUC37, and NCI-H460 cell lines treated with DMSO (UNT), 1 μM RSL3 (RSL3), and 1 μM RSL3 + 2 μM Ferrostatin-1 (RSL3 + FER-1) for 2 h. E Histograms of three independent experiments of MitoSOX staining to detect mitochondrial peroxides. F Histograms of three independent experiments of SYTOX Blue staining to quantify cell death (R: RSL3; F: Ferrostatin-1). All data were expressed as mean ± SEM and are statistically significant if *p < 0.05 and very significant if **p < 0.01 (paired Student’s t-test).

Fig. 3. LUAD-CSCs show an increased expression of a set of antioxidant genes.

A Gene expression of antioxidant genes by RT-PCR in BBIRE-T248, PUC30, PUC36, and PUC37 cell lines. Data from three independent experiments are expressed as fold change of the mean of gene expression ± SEM and are statistically significant if *p < 0.05 and very significant if **p < 0.01 (paired Student’s t-test). β-Actin was used for normalization. SLC7A11 solute carrier family 7 member 11, GCLC glutamate-cysteine ligase catalytic subunit, GPX4 glutathione peroxidase 4, NFE2L22 nuclear factor erythroid-derived 2-like 2, AKR1C2 aldo-keto reductase family 1 member C2, HMOX-1 heme oxygenase-1, NQO1 NAD(P)H quinone dehydrogenase 1. B Western blotting analysis of GPX4 and β-Actin in the four LUAD primary cell lines analysed. The panels show the bands of three independent experiments and the relative fold change of densitometric analyses of GPX4 and β-actin proteins. C Intracellular content of GSH and GSSG in BBIRE-T248, PUC30, PUC36, and PUC37 cell lines. Data from three independent experiments are expressed as mean ± SEM and are statistically significant if *p < 0.05 and very significant if **p < 0.01 (paired Student’s t-test). D Western blotting analysis of GPX4 and β-ACTIN in the LUAD primary cell line BBIRE-T248, following treatment with 1 μM RSL3 alone or in combination with Ferrostatin-1 (Fer-1) for 2 h.

Fig. 4. LUAD-CSCs modulate other molecular pathways potentially involved in ferroptosis.

A, B Gene expression of genes involved in iron metabolism (A) and lipid metabolism (B) by RT-PCR in BBIRE-T248, PUC30, PUC36, and PUC37 cell lines. Data from three independent experiments are expressed as fold change of the mean of gene expression ± SEM and are statistically significant if *p < 0.05 and very significant if **p < 0.01 (paired Student’s t-test). β-actin was used for normalization. FTH1 Ferritin heavy chain 1, TRFC Transferrin receptor 1, SLC40A1 Solute Carrier Family 40 Member 1, SCD1 Stearoyl-CoA desaturase 1, FADS2 Fatty acid desaturase 2, ACSL4 Acyl-CoA synthetase long-chain family member 4. C, D Representative images of morphological aspect of BBIRE-T248 cells after treatment with RSL3 (1 μM) alone or in combination with Ferrostatin-1 (2 μM) for 2 h by transmission electron microscopy (C) and by inverted microscope (D). UNT untreated, Fer-1 Ferrostatin-1. (10x microscopic enlargement).

Fig. 5. Differential expression analysis (DEA) of genes transcriptionally modulated in 3D spheroids versus 2D cells in BBIRE-T248, PUC36, PUC30, and PUC37 primary cell lines.

A Venn diagram of commonly up and downregulated genes. B Heatmap of commonly up and downregulated genes. Only genes with log2Fold Change (FC) >1 or <−1 and adjusted p_val <0.05 were considered significantly upregulated (red) or downregulated (blue). C Gene expression of selected genes commonly upregulated in 3D spheroids of the four primary cell cultures (BBIRE-T248, PUC30, PUC36, and PUC37) by RT-PCR. Data from three independent experiments are expressed as fold change of the mean of gene expression ± SEM and are statistically significant if *p < 0.05 and very significant if **p < 0.01 (paired Student’s t-test). β-actin was used for normalization. GGT1 Gamma-glutamyltransferase 1, SEPW1 Selenoprotein W1, MUC1 Mucin 1. D Schematic model showing the putative role that each commonly upregulated gene in 3D cultures of the four tested primary lines may play in the protection from ferroptosis (Created with BioRender.com).