Protein phosphatase 2A-B55β mediated mitochondrial p-GPX4 dephosphorylation promoted sorafenib-induced ferroptosis in hepatocellular carcinoma via regulating p53 retrograde signaling

Abstract

Rationale: As a key endogenous negative regulator of ferroptosis, glutathione peroxidase 4 (GPX4) can regulate its antioxidant function through multiple post-translational modification pathways. However, the effects of the phosphorylation/dephosphorylation status of GPX4 on the regulation of inducible ferroptosis in hepatocellular carcinoma (HCC) remain unclear. Methods: To investigate the effects and molecular mechanism of GPX4 phosphorylation/dephosphorylation modification on ferroptosis in HCC cells. Sorafenib (Sora) was used to establish the ferroptosis model in HCC cells in vitro. Using the site-directed mutagenesis method, we generated the mimic GPX4 phosphorylation or dephosphorylation HCC cell lines at specific serine sites of GPX4. The effects of GPX4 phosphorylation/dephosphorylation modification on ferroptosis in HCC cells were examined. The interrelationships among GPX4, p53, and protein phosphatase 2A-B55β subunit (PP2A-B55β) were also explored. To explore the synergistic anti-tumor effects of PP2A activation on Sora-administered HCC, we established PP2A-B55β overexpression xenograft tumors in a nude mice model in vivo. Results: In the Sora-induced ferroptosis model of HCC in vitro, decreased levels of cytoplasmic and mitochondrial GPX4, mitochondrial dysfunction, and enhanced p53 retrograde signaling occurred under Sora treatment. Further, we found that mitochondrial p53 retrograded remarkably into the nucleus and aggravated Sora-induced ferroptosis. The phosphorylation status of GPX4 at the serine 2 site (GPX4^Ser2) revealed that mitochondrial p-GPX4^Ser2 dephosphorylation was positively associated with ferroptosis, and the mechanism might be related to mitochondrial p53 retrograding into the nucleus. In HCC cells overexpressing PP2A-B55β, it was found that PP2A-B55β directly interacted with mitochondrial GPX4 and promoted Sora-induced ferroptosis in HCC. Further, PP2A-B55β reduced the interaction between mitochondrial GPX4 and p53, leading to mitochondrial p53 retrograding into the nucleus. Moreover, it was confirmed that PP2A-B55β enhanced the ferroptosis-mediated tumor growth inhibition and mitochondrial p53 retrograde signaling in the Sora-treated HCC xenograft tumors. Conclusion: Our data uncovered that the PP2A-B55β/p-GPX4^Ser2/p53 axis was a novel regulatory pathway of Sora-induced ferroptosis. Mitochondrial p-GPX4^Ser2 dephosphorylation triggered ferroptosis via inducing mitochondrial p53 retrograding into the nucleus, and PP2A-B55β was an upstream signal modulator responsible for mitochondrial p-GPX4^Ser2 dephosphorylation. Our findings might serve as a potential theranostic strategy to enhance the efficacy of Sora in HCC treatment through the targeted intervention of p-GPX4 dephosphorylation via PP2A-B55β activation.

Keywords: Ferroptosis; Glutathione peroxidase 4; Hepatocellular carcinoma; Protein phosphatase 2A-B55β subunit; Retrograding p53 signal.

Figure 1

Ferroptosis resistance and high GPX4 expression were associated with HCC. A. The expression of ferroptosis-related genes in tumor tissues (n = 152) and peritumor tissues (n = 91) of HCC patients from the GEO database (GSE102079). B. The expression of ferroptosis-related genes in tumor tissues (n = 373) and peritumor tissues (n = 50) of HCC patients from the TCGA-LIHC database. C. KEGG analysis of DEGs between tumor tissues and peritumor tissues in HCC patients from TCGA-LIHC database. D. GSEA analysis of the ferroptosis pathways in tumor tissues and peritumor tissues of HCC patients from TCGA-LIHC database. E. GPX4 expression in the paired tumor tissues (n = 40) and the paired peritumor tissues (n = 40) from the TCGA-LIHC database. F. GPX4 expression in two cohorts of HCC patients from GEPIA. G. Kaplan-Meier analysis showed the overall survival of HCC patients from GEPIA (TCGA-LIHC) with different levels of GPX4 expression. GPX4 expression was a binary variable divided into high or low expression according to the quartile. H. Protein levels of p53 and GPX4 in the paired peritumor tissues (P) and the paired tumor tissues (T) from the HCC patients, n = 6. I. Representative IHC images of p53 and GPX4 in the paired peritumor tissues and paired tumor tissues from the HCC patients. *, P < 0.05.

Figure 2

Downregulation of GPX4 and mitochondrial dysfunction in sorafenib-induced ferroptosis of HCC cells. A. HCC cells were treated with different doses of sorafenib (Sora, 5, 10, 20, and 40 µM, for 12 h and 24 h) alone or combined with ferroptosis inhibitor Fer-1 (1 µM) for 24 h. Cell viability was measured with the MTS assay. B. HCC cells were treated with Sora (10 µM) individually or combined with inhibitors of different cell death patterns, including Fer-1 (1 µM), Z-VAD (10 µM), Nec-1 (10 µM), or CQ (5 µM) for 24 h. Cell viability was measured with the MTS assay. C. WB detection of GPX4 protein expression in HCC cells after treatment with different doses of Sora (5, 10, 20 µM, 24 h) alone or combined with Fer-1 (1 µM, 24 h). D. TEM observation of mitochondrial morphological characteristics (red arrows) of ferroptosis in HCC cells treated with Sora (10 µM, 24 h). N, nucleus; M, mitochondria. Scale bar, 0.5 μm. E. FCM detection of LPO levels in HCC cells treated with Sora (10 μM, 24 h) individual or combined with Fer-1 (1 μM, 24 h). F. IF detection of mitochondrial LPO in HCC cells treated with Sora (10 µM, 24 h). Scale bars, 20 μm. G. FCM detection of mitochondrial ROS in HCC cells treated with Sora (10 µM, 24 h) individual or combined with Mito-Tempo (20 µM, pretreatment for 2 h). H. FCM detection of mitochondrial membrane potential in HCC cells treated with CCCP (10 µM, 6 h) and Sora (10 µM, 24 h) individual or combined with Fer-1 (1 μM, 24 h) in HCC cells. *, P < 0.05.

Figure 3

Mitochondrial p53 retrograde into the nucleus was involved in Sora-induced ferroptosis. A. Protein levels of p53 in cytosolic (Cyto) or mitochondrial (Mito) fraction of HCC cells treated with Sora (10 µM, 24 h). B. Protein levels of p53 in cytosolic (Cyto) or nucleus (Nu) fraction of HCC cells treated with Sora (10 µM, 24 h). C. IF assay of the p53 distribution in mitochondria and nucleus of HCC cells treated with Sora individual (10 µM, 24 h) or combined with Fer-1 (1 µM, 24 h). Cells were subjected to IF staining of p53 (green), Mitotracker (red), and DAPI (blue). Scale bars, 20 μm. D. FCM detection of LPO in Hep3B cells. Staining cells with C11 BODIPY probe. E. FCM detection of LPO in Hep3B cells transfected with a p53-expressing plasmid or p53 nuclear localization sequence deleted (ΔNLS) plasmid. Staining cells with C11 BODIPY probe. F. Hep3B cells transfected with a p53-expressing plasmid or p53 ΔNLS plasmid were treated with Sora (10 µM, 24 h) individually or combined with Fer-1 (1 µM, 24 h). Cell viability was measured with the MTS assay. G-H. HepG2 cells transfected with a p53-expressing plasmid or p53 ΔNLS plasmid were treated individually with Sora (10 µM, 24 h) or combined with Fer-1 (1 µM, 24 h). Staining cells with a C11 BODIPY probe and LPO levels were measured with the FCM assay (G). Cell viability was measured with the MTS assay (H). *, P < 0.05.

Figure 4

Mitochondrial p-GPX4^Ser2 dephosphorylation triggered ferroptosis and induced mitochondrial p53 translocation. A. Mitochondrial-IP (Mito-IP) detection of the mitochondrial p-GPX4 level in HCC cells upon Sora treatment (10 µM, 24 h). B. Mito-IP detection of the interaction between GPX4 and PINK1 in mitochondria of HCC cells treated with Sora (10 µM, 24 h). C. Venn diagram showed the shared and unique phosphorylation sites of GPX4 predicated by DISPHOS 1.3, PhosphoSVM, MusiteDeep, Gps 6.0, and NetPhon3.1. D. IP detection of the phosphorylation levels of GPX4 in the constructed S2A, S40A, S45A, and S112A cells of HepG2. E. Sequences alignment of the conserved serine residues on GPX4. F. Mito-IP detection of the mitochondrial p-GPX4 levels in WT cells and the constructed S2D and S2A cells of HepG2. G. FCM detection of the LPO levels in WT, S2D, S112D, S2A, and S112A cells of HepG2. Staining cells with C11 BODIPY probe. H. WT, S2D, S112D, S2A, and S112A cells of HepG2 were treated with Sora (10 µM) individually or combined with Fer-1 (1 µM, 24 h). Cell viability was measured with the MTS assay. I. PLA detection of the interaction between p53 and GPX4 in HCC cells upon Sora treatment (10 µM, 24 h). J. PLA detection of the interaction between p53 and GPX4 in S2A and S112A cells of HepG2. K. Mito-IP detection of the interaction between GPX4 and p53 in S2D and S2A cells of HepG2 treated with Sora (10 µM, 24 h). L. Protein levels of p53 in cytosolic (Cyto), mitochondrial (Mito), and nucleus (Nu) fraction of WT, S2D, and S2A cells of HepG2. M. IF assay of the distribution of p53 in mitochondria and nucleus of S2D and S2A cells of HepG2 upon Sora treatment (10 µM, 24 h). Cells were subjected to IF staining of p53 (green), mitotracker (red), and DAPI (blue). Scale bars, 20 μm. *, P < 0.05.

Figure 5

PP2A-B55β regulated p-GPX4^Ser2 dephosphorylation and promoted Sora-induced ferroptosis in HCC cells. A. Protein levels of B55α (coded by PPP2R2A), B55β (coded by PPP2R2B), and B55ε (coded by PPP2R5E) in HCC cells treated with Sora (5, 10, 20 µM, 24 h). B. Mito-IP detection of the interaction between GPX4 and B55β in HCC cells treated with Sora (10 µM, 24 h). C. PLA detection of the interaction between B55β and GPX4 in HCC cells treated with Sora (10 µM, 24 h). Interaction events were shown as red dots. Scale bars, 20 μm. D. IP detection of the interaction between B55β and GPX4 in WT cells and the constructed S2A, S40A, S45A, and S112A cells. E. PLA detection of the interaction between B55β and GPX4 in WT, S2A, S40A, S45A, and S112A cells. Interaction events were shown as red dots. Scale bars, 20 μm. F. HepG2 cells were treated with si2R2B and/or OA, while the whole cell lysates were isolated and subjected to PhosTag™ gel electrophoresis. The protein levels of the phosphorylated GPX4 (p-GPX4) and non-phosphorylated GPX4 (N-GPX4) were tested. G. Mito-IP detection of the interaction between GPX4 and p53 in HCC cells transfected with the PPP2R2B-overexpression plasmid. H. Protein levels of B55β and GPX4 in HepG2 cells treated with PPP2R2B-overexpression plasmid and/or Sora (10 µM, 24 h). I. FCM detection of LPO levels in HepG2 cells transfected with the PPP2R2B-overexpression plasmid. Staining cells with C11 BODIPY probe. J. Cell viability of HCC cells treated with Sora (10 µM) alone or combined with PPP2R2B-overexpression plasmid or PP2A agonists DT-061 (20 µM) and iHAP1 (10 µM) for 24, 48, 72, and 96 h. K. EdU staining and colony formation assay detection of the cell proliferation ability of HCC cells treated with Sora (10 µM) alone or combined with PPP2R2B-overexpression plasmid or PP2A agonists DT-061 (20 µM) and iHAP1 (10 µM) for 24 h. *, P < 0.05.

Figure 6

PP2A-B55β promoted the anti-tumor effect of Sora via aggravation of ferroptosis. A. The diagram depicted the experimental design of the in vivo study. BALB/c nude mice were subcutaneously inoculated with the constructed HepG2-PPP2R2B cells with B55β-overexpression, while HepG2-pBabe cells were used as control cells. Mice harbored with HCC-xenograft tumors were randomly divided into 4 groups (n = 4 for each group): pBabe-Ctrl, pBabe-Sora, PPP2R2B-Ctrl, PPP2R2B-Sora. The Sora (10 mg/kg) was injected via the tail vein every two days. B. The growth curves of xenograft tumors in nude mice. Tumor volume = (W² × L)/2, where L is the longer dimension and W is the shorter one. C-D. Representative images (C) and weight (D) of xenograft tumors. E. Representative H&E staining images of xenograft tumor tissues. F-G. Levels of GSH (F) and MDA (G) in xenograft tumor tissues. H. TEM observation of mitochondrial morphological characteristics (red arrows) of HCC cells in xenograft tumor tissues. N, nucleus; M, mitochondria. Scale bar, 0.5 μm. I. Representative IHC staining images of GPX4 and B55β expression in xenograft tumor tissues. J. Protein levels of p53 in the nucleus and in the mitochondrial or cytoplasmic fractions of xenograft tumor tissues. *, P < 0.05.