Protein S-glutathionylation confers cellular resistance to ferroptosis induced by glutathione depletion

Abstract

Ferroptosis is one of the most critical biological consequences of glutathione depletion. Excessive oxidative stress, indicated by an elevated oxidized glutathione (GSSG)/reduced glutathione (GSH) ratio, is recognized as a key driver of ferroptosis. However, in glutathione depletion-induced ferroptosis, a marked decrease in total glutathione levels (including both GSH and GSSG) is frequently observed, yet its significance remains understudied. Protein S-glutathionylation (protein-SSG) levels are closely linked to the redox state and cellular glutathione pools including GSH and GSSG. To date, the role of protein-SSG during cell ferroptosis induced by glutathione depletion remains poorly understood. Here, we demonstrated that upregulation of CHAC1, a glutathione-degrading enzyme, acted as a key regulator of protein-SSG formation and exacerbated glutathione depletion-induced ferroptosis. This effect was observed in both in vitro and in vivo models, including erastin-induced ferroptosis across multiple cell lines and acetaminophen overdose-triggered ferroptosis in hepatocytes. Deficiency of CHAC1 resulted in increased glutathione pools, enhanced protein-SSG, improved liver function, and attenuation of hepatocyte ferroptosis upon acetaminophen challenge. These protective effects were reversed by CHAC1 overexpression. Using quantitative redox proteomics, we identified glutathione pool-sensitive S-glutathionylated proteins. As an important example, we discovered that ADP-ribosylation factor 6 (ARF6) was regulated by S-glutathionylation during glutathione depletion-induced ferroptosis. Our findings revealed that CHAC1 upregulation reduced the S-glutathionylation of ARF6, resulting in decreased ARF6 levels in lysosomes. This, in turn, enhanced the localization of the transferrin receptor (TFRC) on the cell membrane and increased transferrin uptake, ultimately compromising the protective role of ARF6 in ferroptosis induced by glutathione depletion. Targeting TFRC using GalNAc-siTfrc mitigated acetaminophen-induced liver injury in vivo. In conclusion, our study provide evidence that availability of glutathione pools affects protein S-glutathionylation and regulates protein functions to influence the process of ferroptosis, which opens an avenue to understanding the cell ferroptosis induced by glutathione depletion.

Keywords: ADP-Ribosylation factor 6; CHAC1; Ferroptosis; Protein S-Glutathionylation; Transferrin.

Fig. 1

Reduced protein-SSG is associated with decreased glutathione pools by CHAC1 induction in multiple cell types undergoing glutathione deprivation-induced ferroptosis. (A) Western blot analysis of S-glutathionylated proteins and CHAC1 in eight human cell lines treated with DMSO or erastin (10 μM, 20 μM). β-actin was used as an internal reference. (B) Quantification of GSH level (Data are mean ± SEM of n = 3 biologically independent samples, ∗∗∗P < 0.001, and ∗∗∗∗P < 0.0001, by one-way ANOVA). (C) Quantification of GSSG levels (Data are mean ± SEM of n = 3 biologically independent samples, ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, and ∗∗∗∗P < 0.0001 by one-way ANOVA). (D) RT-qPCR analysis of CHAC1 mRNA and Western blot analysis of CHAC1 protein levels. mRNA values were compared to those of DMSO-treated T84 cells (Data are mean ± SEM of n = 3 biologically independent samples, ∗∗P < 0.01, ∗∗∗P < 0.001, and ∗∗∗∗P < 0.0001 by one-way ANOVA). (E) RT-qPCR analysis of CHAC1 mRNA levels in H1299 cells transfected with CHAC1 siRNA and then treated with 10 μM erastin (Data are mean ± SEM of n = 3 biologically independent samples, one-way ANOVA). (F) Quantification of GSH and GSSG levels in H1299 cells transfected with CHAC1 siRNA and then treated with 10 μM erastin (Data are mean ± SEM of n = 3 biologically independent samples, one-way ANOVA). (G) Western blot analysis of CHAC1 and S-glutathionylated proteins in H1299 cells transfected with CHAC1 siRNA and treated with 20 μM erastin. GAPDH was used as an internal reference. The relative expression levels of these proteins were normalized by their respective DMSO controls (Data are mean ± SEM of n = 3 biologically independent experiments, one-way ANOVA). (H) Relative cell viability of H1299 cells treated with erastin (10 μM, 20 μM) for 24 h, normalized by DMSO control, as measured by the CellTiter-Glo® Luminescent Cell Viability Assay (Data are mean ± SEM of n = 4 biologically independent samples, t-test). (I) Cell viability of PMHs and Hepa1-6 cells treated with erastin (10 μM, 20 μM) for 24 h, normalized by DMSO control (Data are mean ± SEM of n = 5 biologically independent samples, one-way ANOVA). (J) Western blot analysis of S-glutathionylated proteins in PMHs and Hepa1-6 cells treated with DMSO, or 10 μM or 20 μM erastin. GAPDH was used as an internal reference. (K) RT-qPCR analysis of Chac1 mRNA levels in PMHs and Hepa1-6 cells treated with DMSO, or 10 μM erastin for 12 h. (Data are mean ± SEM of n = 3 biologically independent samples, t-test). (L) Western blot analysis of CHAC1-FLAG protein levels in PMHs and Hepa1-6 cells infected with Ad-GFP or Ad-CHAC1 adenoviruses. β-ACTIN was used as an internal reference. Cell viability of PMHs and Hepa1-6 cells infected with Ad-GFP or Ad-CHAC1 adenoviruses and treated with DMSO or erastin (10 μM, 20 μM) for 24 h (Data are mean ± SEM of n = 5 biologically independent samples, $ for one-way ANOVA and # for t-test). (M) Western blot analysis of CHAC1 protein levels in Hepa1-6 cells were transfected with NC siRNA and Chac1 siRNA and treated with DMSO or 10 μM erastin. β-ACTIN was used as an internal reference. Cell viability of Hepa1-6 cells infected with NC siRNA and Chac1 siRNA and treated with DMSO or 10 μM erastin for 24 h (Data are mean ± SEM of n = 6 biologically independent samples, one-way ANOVA). PMH, primary mouse hepatocyte; GSH, glutathione; GSSG, oxidized glutathione.

Fig. 2

Decreased glutathione pools by CHAC1 reduces protein-SSG and aggravates APAP-induced hepatotoxicity and ferroptosis in primary mouse hepatocytes. (A) The heatmap shows the relative levels of upregulated and downregulated genes in the liver tissues of mice treated with 300 mg/kg APAP for 3 h, 750 mg/kg APAP for 3 h, 300 mg/kg APAP for 6 h, and 750 mg/kg APAP for 6 h compared to the saline group. The heatmap was ranked by fold change of 750 mg/kg APAP for 3 h group versus the saline group for 3 h. (Fold change ≥1.5, P < 0.05) (n = 3 mice/group). (B) Volcano plots shows upregulated and downregulated genes from RNA transcriptome data of the group treated with 750 mg/kg APAP for 6 h compared to the saline group (Fold change ≥1.5, P < 0.05; n = 3 mice/group). (C) RT-qPCR analysis of Chac1 mRNA levels in mouse liver tissues treated with saline or 300 and 750 mg/kg APAP for 3, 6, or 12 h. (Data are mean ± SEM of n = 7 mice/group, one-way ANOVA). (D) Immunohistochemical staining of CHAC1 in liver sections from healthy controls (n = 9) and patients with AILI (n = 9), with quantification of immunohistochemical scores (Data are mean ± SEM, t-test). The black arrow indicates positive staining. Scale bar = 100 μm. (E) RT-qPCR analysis of Chac1 mRNA levels in PMHs with or without 20 mM APAP challenge for 1, 3, 6, and 12 h (Data are mean ± SEM of n = 3 biologically independent samples, t-test). (F) Western blot analysis of S-glutathionylated proteins in PMHs from Chac1^+/+ and Chac1^−/− mice treated with 20 mM APAP for 3 h. GAPDH was used as an internal reference. The statistical chart shows the relative expression levels of S-glutathionylated proteins (Data are mean ± SEM of n = 3 mice, t-test). (G) Western blot analysis of S-glutathionylated proteins in PMHs infected with Ad-GFP or Ad-CHAC1 at multiplicities of infection (MOI) of 0.1, 1, 2.5, 5, and 10 for 12 h and, followed by treatment with 20 mM APAP for 6 h. GAPDH was used as an internal reference. (H) Quantification of GSH and GSSG levels and ratio of GSSG/GSH in PMHs infected with Ad-GFP or Ad-CHAC1 adenovirus at MOIs of 1 and 10 for 12 h, followed by treatment with 20 mM APAP for 6 h (Data are mean ± SEM of n = 4 biologically independent samples, t-test). (I) PMHs isolated from Chac1^+/+ and Chac1^−/− mice were treated with 20 mM APAP for 12 h. Relative cell viability was measured using the CellTiter-Glo Luminescent Cell Viability Assay and normalized by DMEM control (Data are mean ± SEM of n = 5 biologically independent samples, t-test). (J) MDA levels in PMHs infected with Ad-GFP or Ad-CHAC1 adenovirus at MOI = 10 for 36 h and then treated with 20 mM APAP for 12 h (Data are mean ± SEM of n = 3 biologically independent samples, t-test). (K) PMHs from Chac1^+/+ and Chac1^−/− mice were treated with 20 mM APAP for 12 h. Representative images of C11 Bodipy 581/591 fluorescence. Red fluorescence represents non-lipid oxidation, and green fluorescence represents lipid oxidation. The statistical chart shows the ratio of green to red fluorescence (Scale bars = 200 μm, data are mean ± SEM of n = 3 biologically independent samples, t-test). APAP, acetaminophen; AILI, APAP-induced liver injury; PMH, primary mouse hepatocyte; GSH, glutathione; GSSG, oxidized glutathione; MDA, determination of malondialdehyde; Chac1^−/−, Chac1-deficient; Chac1^+/+, wild-type controls.

Fig. 3

Decreased glutathione pools by CHAC1 reduces protein-SSG and aggravates APAP-induced hepatotoxicity and ferroptosis in APAP-injured mice liver. (A) Quantification of GSH in the liver tissues of Chac1^+/+ Ad-GFP, Chac1^−/− Ad-GFP, and Chac1^−/− Ad-CHAC1 mice treated with saline or 300 mg/kg APAP for 2 and 6 h (Data are mean ± SEM of n = 3 and 5 mice/group, t-test). (B) Quantification of GSSG in the liver tissues of Chac1^+/+ Ad-GFP, Chac1^−/− Ad-GFP, and Chac1^−/− Ad-CHAC1 mice treated with saline or 300 mg/kg APAP for 2 and 6 h (Data are mean ± SEM of n = 3 and 5 mice/group, t-test). (C) Western blot analysis of S-glutathionylated proteins in the liver tissues of Chac1^+/+ Ad-GFP and Chac1^−/− Ad-GFP mice treated with 300 mg/kg APAP for 2 and 6 h. GAPDH was used as an internal reference, followed by quantification of relative levels of S-glutathionylated proteins after 6 h of APAP treatment (Data are mean ± SEM of n = 3 mice/group, t-test). (D) Western blot analysis of S-glutathionylated proteins and CHAC1-FLAG protein in the liver tissues of Chac1^−/− Ad-GFP and Chac1^−/− Ad-CHAC1 mice treated with 300 mg/kg APAP for 6 h. GAPDH was used as an internal reference. Statistical chart shows the relative expression levels of S-glutathionylated proteins 6 h after APAP treatment (Data are mean ± SEM of n = 4 mice/group, t-test). (E) Serum levels of ALT and AST in Chac1^+/+ Ad-GFP, Chac1^−/− Ad-GFP, and Chac1^−/− Ad-CHAC1 mice treated with 300 mg/kg APAP for 6 h (Data are mean ± SEM of n = 5 mice/group, t-test). (F) H&E staining, 4-HNE protein adduct staining in liver tissues of Chac1^+/+ Ad-GFP, Chac1^−/− Ad-GFP and Chac1^−/− Ad-CHAC1 mice treated with 300 mg/kg APAP for 6 h. Scale bars = 200 μm. The area of liver injury and immunohistochemical score for 4-HNE protein adduct staining were quantified (Data are mean ± SEM of n = 5 mice/group, t-test). APAP, acetaminophen; GSH, glutathione; GSSG, oxidized glutathione; ALT, alanine aminotransferase; AST, aspartate aminotransferase; H&E, haematoxylin and eosin; Chac1^−/−, Chac1-deficient; Chac1^+/+, wild-type controls.

Fig. 4

Redox proteomic analysis reveals that the S-glutathionylation of Cys90 on ARF6 is regulated by CHAC1 in ferroptotic PMHs induced by APAP. (A) Flowchart outlining the key experimental procedures for proteomic analysis of S-glutathionylation. (B) Scatter plot illustrating the distribution of differential modification sites, sorted by the ratio of Ad-GFP + APAP/Ad-GFP. Red dots indicating up-regulation of significant differences, blue dots indicating down-regulation of significant differences, and grey dots indicating no significant differences (CV < 0.1, fold change ≥1.2). (C) Scatter plot showing the distribution of differential modification sites, sorted by the ratio of Ad-CHAC1 + APAP/Ad-GFP + APAP. Red dots indicating up-regulation of significant differences, blue dots indicating down-regulation of significant differences and grey dots indicating no significant differences (CV < 0.1, fold change ≥1.2). (D) Venn diagram showing differentially modified sites under both APAP stimulation and CHAC1 overexpression (Fold change ≥1.2). (E) The heat map illustrating the union of differential modification sites in Ad-GFP, Ad-GFP + APAP, Ad-CHAC1, and Ad-CHAC1 + APAP comparison groups (CV < 0.1, fold change ≥1.2). (F) Scatter plot showing differentially modified sites under both APAP stimulation and CHAC1 overexpression; the order was sorted by the ratio of Ad-GFP + APAP/Ad-GFP (CV < 0.1, fold change ≥1.2). (G) Two-stage mass spectrometry of the glutathionylated peptide from ARF6 in PMHs. The secondary mass spectrum shows fragment ion information of the ARF6 C90 peptide segment. (H) Histogram showing the relative modification abundance of ARF6 C90 in different treatment groups, with glutathionylated peptides identified and quantified by LC-MS/MS (All values were normalized by the mean of the AdGFP-CON group, data are mean ± SEM of n = 2 biologically independent samples). (I) IP assay showing the expression of S-glutathionylated ARF6 in 293T cells overexpressing Myc-tagged ARF6. Whole cell lysates were used to confirm the expression of ARF6. (J) Two-stage mass spectrometry of the glutathionylated peptide from ARF6 in 293T cells overexpressing Myc-tagged ARF6. The secondary mass spectrum shows fragment ion information of the ARF6 C90 peptide segment. PMH, primary mouse hepatocyte; IP, immunoprecipitation.

Fig. 5

ARF6 serves the protective role in ferroptosis. (A) PMHs were transfected with Arf6 siRNA and treated with 20 mM APAP. Relative cell viability was measured using the CellTiter-Glo® Luminescent Cell Viability Assay and normalized by DMEM control (Data are mean ± SEM of n = 6 biologically independent samples, one-way ANOVA). (B) PMHs were transfected with Arf6 siRNA and treated with 20 mM APAP. Representative images of calcein-AM/PI live/dead cell staining (Green: live cells; Red: dead cells, Scale bars = 200 μm). Followed by ratio of dead to living cells (Data are mean ± SEM of n = 3 biologically independent samples, one-way ANOVA). (C) Representative images of the C11 Bodipy 581/591 fluorescent probe used to detect the formation of lipid peroxides (Red: non-oxidized lipids; Green: oxidized lipids; Scale bars = 200 μm). The statistical chart shows the ratio of green to red fluorescence (Data are mean ± SEM of n = 3 biologically independent samples, one-way ANOVA). (D) Representative images of the FerroOrange fluorescent probe used to detect labile ferrous ions (Red: FerroOrange; Blue: DAPI; Scale bars = 200 μm). The statistical chart shows the relative Fe²⁺ fluorescence intensity (Data are mean ± SEM of n = 3 biologically independent samples, one-way ANOVA). APAP, acetaminophen.

Fig. 6

S-glutathionylation of ARF6 at Cys90 enhances its trafficking to lysosomes, resulting in reduced plasma membrane localization of TFRC and diminished transferrin uptake. (A) Representative fluorescence images of ARF6-KO H1299 cells expressing ARF6-WT-mCherry, ARF6-C90A-mCherry, or ARF6-C90D-mCherry (Blue: DAPI, Red: mCherry, Green: LysoPrime probe. Scale bars = 50 μm). (B) Subcellular fractionation of HEK293T cells expressing ARF6-WT-Myc, ARF6-C90A-Myc, or ARF6-C90D-Myc. Western blot analysis of TFRC and ARF6 protein levels in total cell lysates, lysosomal, cytoplasmic and plasma membrane fractions. β-ACTIN, LAMP1, Na⁺/K⁺-ATPase were used as internal references. Relative TFRC and ARF6 protein levels in those fractions were quantified (Data are mean ± SEM of n = 3 biologically independent experiments, one-way ANOVA). (C) Representative fluorescence images of 293T cells expressing ARF6-WT-Myc, ARF6-C90A-Myc, or ARF6-C90D-Myc. (Blue: DAPI, Green: Alexa Fluor® 647-labeled transferrin probe, Red, ARF6-mCherry. Scale bars = 20 μm). (D) Relative transferrin uptake (Data are mean ± SEM of n = 10 biologically independent samples, one-way ANOVA). (E) Relative cell viability of ARF6-KO H1299 cells treated with erastin (10 μM) for 24 h, normalized by DMSO control, as measured by the CellTiter-Glo® Luminescent Cell Viability Assay (Data are mean ± SEM of n = 6 biologically independent samples, t-test). TFRC, transferrin receptor.

Fig. 7

TFRC mediates the exacerbation of APAP-induced hepatocyte ferroptosis caused by CHAC1 upregulation. (A) Representative fluorescence images showing transferrin uptake in Chac1^+/+ and Chac1^−/− PMHs treated with 20 mM APAP. (Blue: DAPI, Green: Alexa Fluor® 647-labeled transferrin probe. Scale bars = 20 μm). The statistical chart shows the relative transferrin uptake (Data are mean ± SEM of n = 10 biologically independent samples, one-way ANOVA). (B) Representative fluorescence images showing transferrin uptake in PMHs infected with Ad-GFP or Ad-CHAC1 adenoviruses (MOI = 10), followed by treatment with 20 mM APAP. (Red: Alexa Fluor® 546-labeled transferrin probe, Blue: DAPI, Scale bars = 50 μm).The statistical chart shows the relative transferrin uptake (Data are mean ± SEM of n = 10 biologically independent samples, one-way ANOVA). (C) Representative fluorescence images showing Fe²⁺ in PMHs infected with Ad-GFP or Ad-CHAC1 adenoviruses (MOI = 10), followed by treatment with or without 20 mM APAP (Red: FerroOrange fluorescent probe, Green: GFP, Scale bars = 200 μm). The statistical chart shows the relative Fe²⁺ fluorescence intensity (Data are mean ± SEM of n = 3 biologically independent samples, t-test). (D) PMHs were transfected with NC siRNA or Tfrc siRNA, followed by infection with Ad-GFP or Ad-CHAC1 adenoviruses (MOI = 10) and treatment with 20 mM APAP. Representative images of PI dead cell staining (Green: GFP, Red: dead cells, Scale bars = 100 μm). The statistical chart shows the ratio of dead cells (Data are mean ± SEM of n = 4 biologically independent samples, t-test). (E) RT-qPCR analysis of Tfrc mRNA levels in WT mice transfected with GalNAc-siNC or GalNAc-siTfrc, followed by treatment with 300 mg/kg APAP for 6 h, normalized by Saline control (Data are mean ± SEM of n = 3 and 6 mice/group, one-way ANOVA). (F) Western blot analysis of TFRC proteins in WT mice transfected with GalNAc-siNC or GalNAc-siTfrc and treated with 300 mg/kg APAP for 6 h. GAPDH was used as an internal reference. Followed by quantification of relative levels of TFRC, normalized by saline control (Data are mean ± SEM of n = 3 and 6 mice/group, one-way ANOVA). (G) Serum levels of ALT in WT mice transfected with GalNAc-siNC or GalNAc-siTfrc and treated with 300 mg/kg APAP for 6 h (Data are mean ± SEM of n = 3 and 6 mice/group, one-way ANOVA). (H) Serum levels of AST in WT mice transfected with GalNAc-siNC or GalNAc-siTfrc and treated with 300 mg/kg APAP for 6 h (Data are mean ± SEM of n = 3 and 6 mice/group, one-way ANOVA). (I) H&E staining, in liver tissues of WT mice transfected with GalNAc-siNC or GalNAc-siTfrc and treated with 300 mg/kg APAP for 6 h. The area of liver injury were quantified (Data are mean ± SEM of n = 3 and 6 mice/group, one-way ANOVA). (J) Western blot analysis of 4-HNE proteins adduct levels in WT mice transfected with GalNAc-siNC or GalNAc-siTfrc and treated with 300 mg/kg APAP for 6 h. GAPDH was used as an internal reference. Followed by quantification of relative levels of 4-HNE proteins adduct, normalized by saline control (Data are mean ± SEM of n = 3 and 6 mice/group, one-way ANOVA). PMH, primary mouse hepatocyte; APAP, acetaminophen; TFRC, transferrin receptor.