YAP/TAZ and ATF4 drive resistance to Sorafenib in hepatocellular carcinoma by preventing ferroptosis

Abstract Understanding the mechanisms underlying evasive resistance in cancer is an unmet medical need to improve the efficacy of current therapies. In this study, a combination of shRNA‐mediated synthetic lethality screening and transcriptomic analysis revealed the transcription factors YAP/TAZ as key drivers of Sorafenib resistance in hepatocellular carcinoma (HCC) by repressing Sorafenib‐induced ferroptosis. Mechanistically, in a TEAD‐dependent manner, YAP/TAZ induce the expression of SLC7A11, a key transporter maintaining intracellular glutathione homeostasis, thus enabling HCC cells to overcome Sorafenib‐induced ferroptosis. At the same time, YAP/TAZ sustain the protein stability, nuclear localization, and transcriptional activity of ATF4 which in turn cooperates to induce SLC7A11 expression. Our study uncovers a critical role of YAP/TAZ in the repression of ferroptosis and thus in the establishment of Sorafenib resistance in HCC, highlighting YAP/TAZ‐based rewiring strategies as potential approaches to overcome HCC therapy resistance.

(h) Intracellular GSH levels declined with the depletion of YAP/TAZ either with or without Sorafenib treatment. HLE-shLuc and HLE-shY/T cells were cultured with DMSO or 6µM Sorafenib for 18 hours, and intracellular GSH levels were measured using the GSH-Glo Glutathione Assay kit. Data are shown as mean ± standard deviation (SD). Statistical significance was calculated using one-way ANOVA. Results represent three independent experiments. (k) Knockdown efficiency of siYAP/TAZ was assessed by quantitative RT-PCR analysis.
RNA was extracted from the cells described in (i) and analyzed by quantitative RT-PCR. CTGF as a direct transcriptional target of YAP/TAZ served as positive control to confirm the knockdown efficiency of siYAP/TAZ. Data are shown as mean ± standard deviation (SD).
Statistical significance was calculated using unpaired t-test. Results represent three independent experiments. Appendix Figure S2. Ferroptosis is repressed in Sorafenib-resistant HCC cells.
(a) Cell viability assay showing that Sorafenib-resistant Huh7 cells are resistant to various ferroptosis inducers, including Erastin, RSL3, FIN56, and FINO2. Cells were treated with either 1µM Erastin for 24 hours, 0.1µM RSL3 for 12 hours, 5µM FIN56 for 18 hours, or 5µM FINO2 for 18 hours. Cell viability was measured using the Promega CellTiter-Glo 2.0 kit and normalized to the respective DMSO treatments. Data are shown as mean ± standard deviation (SD). Statistical significance was calculated using two-way ANOVA. Results represent three independent experiments.
(b) YAP/TAZ-deficiency induced death of HLE cells in response to Sorafenib treatment could be rescued by Ferrostatin-1 but not by GSK872 or Z-VAD-FMK. HLE cells transfected with either siCtrl or siYAP/TAZ were treated with 6µM Sorafenib (Srf) with or without 5µM Ferrostatin-1, 10µM GSK872 or 10µM Z-VAD-FMK for 20 hours before harvest. Cell viability was measured using Promega CellTiter-Glo 2.0 kit and normalized to siCtrl-DMSO.
Data are shown as mean ± standard deviation (SD). Statistical significance was calculated using two-way ANOVA. Results represent three independent experiments. Appendix Figure S3. YAP/TAZ promote resistance to ferroptosis in Sorafenib-sensitive cells.
(a) Cell viability assay showing that the loss of YAP/TAZ increased death of Sorafenibsensitive (parental) Huh7 and Hep3B cells in response to Sorafenib treatment which could not be reversed by treatment with Ferrostatin-1. Huh7 and Hep3B cells were transfected with siYAP/TAZ and then treated with 3µM Sorafenib (Srf) with or without 5µM Ferrostatin-1 (Fer) for 12 hours before harvest. Cell viability was measured with Promega CellTiter-Glo 2.0 kit and normalized to siCtrl-DMSO. Data are shown as mean ± standard deviation (SD). Statistical significance was calculated using two-way ANOVA. Results represent three independent experiments.
(b) Quantitative RT-PCR analysis confirmed the knockdown efficiency of YAP/TAZ as determined from the experiment shown above. RNA was extracted and analyzed by quantitative RT-PCR. Expression of CTGF as a direct transcriptional target gene of YAP/TAZ was used to assess the knockdown efficiency of siYAP/TAZ. Data are shown as mean ± standard deviation (SD). Statistical significance was calculated using unpaired t-test. Results represent three independent experiments.
Huh7 and Hep3B cells transfected with empty vector (EV) or with a cDNA construct coding for YAP-5SA were cultured with either DMSO or 6µM Sorafenib (Srf) for 12 hours before harvest. Cell viability was measured with Promega CellTiter-Glo 2.0 kit and normalized to Huh7 and Hep3B transfected with empty vector (EV) and treated with DMSO solvent. Data are shown as mean ± standard deviation (SD). Statistical significance was calculated using two-way ANOVA. Results represent three independent experiments.
(d) Quantitative RT-PCR analysis confirmed the overexpression of activated YAP. RNA was extracted from the cells described in (c) and YAP mRNA expression was analyzed by quantitative RT-PCR. Data are shown as mean ± standard deviation (SD). Statistical significance was calculated using unpaired t-test. Results represent three independent experiments.
The intracellular GSH levels were measured in Huh7 and Hep3B cells described in (c) using the GSH-Glo Glutathione Assay kit and normalized to Huh7-EV and Hep3B-EV, respectively.
Data are shown as mean ± standard deviation (SD). Statistical significance was calculated using unpaired t-test. Results represent three independent experiments. (g) Immunohistochemical staining of SLC7A11 in HCC and adjacent non-neoplastic areas from patients. Tumor tissues showed a higher expression of SLC7A11. Scale bar, 100µm.

(h) Quantification of SLC7A11-positive cells in tumor and non-tumor samples showed that
HCC tumors present higher SLC7A11 levels. Data are shown as mean ± standard deviation (SD). Statistical significance was calculated using unpaired t-test. Data are shown as mean ± standard deviation (SD). Statistical significance was calculated using unpaired t-test. (g) ATF4 deficiency induced cell death in response to Sorafenib which was overcome by treatment with Ferrostatin-1 (Fer). HLE cells were transfected with siCtrl, siATF4-1 or siATF4-2 and treated with 6µM Sorafenib with or without 5µM Ferrostatin-1 for 12 hours before harvest. Cell viability was measured with Promega CellTiter-Glo 2.0 kit and normalized to HLE-siCtrl DMSO. Data are shown as mean ± standard deviation (SD). Statistical significance was calculated using two-way ANOVA. Results represent three independent experiments.
(h) Quantitative RT-PCR analysis confirmed the knockdown efficiency of siATF4-1 and siATF4-2. RNA from the cells described in (g) was extracted and analyzed by quantitative RT-PCR. mRNA levels of SLC7A11 and CHAC1 decreased significantly with the deficiency of ATF4. Data are shown as mean ± standard deviation (SD). Statistical significance was calculated using two-way ANOVA. Results represent three independent experiments. Appendix Figure S7. YAP/TAZ and TEADs stabilize ATF4 protein.
(a) YAP/TAZ protect ATF4 from proteasome-mediated degradation. HLE cells were transfected with siCtrl or siYAP/TAZ (siY/T) and treated with Sorafenib and the autophagy inhibitor chloroquine (CQ) or the proteasome inhibitor MG132 for 18 hours, as indicated.
Immunoblotting showed that YAP/TAZ deficiency declined Sorafenib-induced ATF4 protein levels which was blocked by MG132 but not by CQ. GAPDH served as loading control. Results represent three independent experiments.
(b) YAP/TAZ deficiency promoted ubiquitylation of ATF4. HLE cells were treated with Sorafenib and MG132 for 18 hours, before ATF4 was immunoprecipitated, and the immunoprecipitates were analyzed by immunoblotting for ATF4 and ubiquitylation of ATF4 (Ub) and for ATF4 and YAP/TAZ in the input samples. Results represent three independent experiments.
Immunoblotting visualized the levels of ATF4. GAPDH served as loading control. Results represent three independent experiments.
(d) TEADs direct the nuclear localization of ATF4. HLE cell were transfected with siCtrl or siTEADs, and the nuclear localization of ATF4 was analyzed by immunofluorescence microscopy. DAPI visualized nuclei. Scale bars, 50µm. Results represent three independent experiments.
(e) Validation of the knockdown efficiency of siTEAD-transfected HLE cells as determined by quantitative RT-PCR for TEAD1, TEAD2, TEAD3 and TEAD4 mRNA levels. Data are shown as mean ± standard deviation (SD). Statistical significance was calculated using oneway ANOVA. Results represent three independent experiments.