A RIPK3-independent role of MLKL in suppressing parthanatos promotes immune evasion in hepatocellular carcinoma

Abstract

Mixed lineage kinase domain-like (MLKL) is widely accepted as an executioner of necroptosis, in which MLKL mediates necroptotic signaling and triggers cell death in a receptor-interacting protein kinase 3 (RIPK3)-dependent manner. Recently, it is increasingly noted that RIPK3 is intrinsically silenced in hepatocytes, raising a question about the role of MLKL in hepatocellular carcinoma (HCC). This study reports a previously unrecognized role of MLKL in regulating parthanatos, a programmed cell death distinct from necroptosis. In HCC cells with intrinsic RIPK3 deficiency, knockout of MLKL impedes the orthotopic tumor growth, activates the anti-tumor immune response and enhances the therapeutic effect of immune checkpoint blockade in syngeneic HCC tumor models. Mechanistically, MLKL is required for maintaining the endoplasmic reticulum (ER)-mitochondrial Mg²⁺ dynamics in HCC cells. MLKL deficiency restricts ER Mg²⁺ release and mitochondrial Mg²⁺ uptake, leading to ER dysfunction and mitochondrial oxidative stress, which together confer increased susceptibility to metabolic stress-induced parthanatos. Importantly, pharmacological inhibition of poly(ADP-ribose) polymerase to block parthanatos restores the tumor growth and immune evasion in MLKL-knockout HCC tumors. Together, our data demonstrate a new RIPK3-independent role of MLKL in regulating parthanatos and highlight the role of MLKL in facilitating immune evasion in HCC.

Fig. 1. MLKL deficiency impairs the orthotopic growth of HCC tumors.

a RIPK3 expression in patient samples across different types of cancer. Data were extracted from TCGA database. b Reconstitution of RIPK3 restores the necroptotic signaling in human HCC cells. HepG2 cells were transfected with the empty vector or pcDNA3.1-hRIPK3-Flag for 24 h and necroptotic signaling was stimulated by 50 μM Z-VAD-FMK followed by 50 ng/mL TNFα plus 20 μM Birinapant (T/B/Z). c MLKL expression in cancer tissues and adjacent normal tissues in two independent cohorts of HCC patients. d Survival analysis of MLKL high- and low-expression patients. The same set of data as in c were used for the analysis. e Reconstitution of RIPK3 restores the necroptotic signaling in murine HCC cells. Hepa 1–6 cells were treated and analyzed as described in b. f Growth curves of MLKL-KO and control murine HCC cells in vitro. Hepa 1–6-luc cells were infected with lentivirus-delivered sgMLKL or negative control (sgNC). KO efficiency was assessed by immunoblotting and cell growth was monitored by the live-cell analysis system IncuCyte (n = 3). g, h Orthotopic tumor growth of MLKL-KO and control HCC cells in a syngeneic model. Cells as described in f were orthotopically implanted in the liver of C57BL/6 mice (n = 13 for sgNC; n = 10 for sgMLKL #1 and sgMLKL #2). Tumor growth was monitored by whole-animal imaging. Endpoint tumor weight was shown in g. Representative luminescent images were shown in h. Data are represented as means ± SEM. Two-tailed Student’s t-test was used for statistical analysis. ns, not significant; *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. 2. MLKL deficiency activates anti-tumor immunity in HCC tumors.

a Endpoint tumor weight of MLKL-KO and control murine HCC cells in nude mice. MLKL-KO Hepa 1–6-luc (sgMLKL) or control (sgNC) cells were orthotopically implanted in the liver of nude mice (n = 8 for sgNC; n = 9 for sgMLKL #1 and sgMLKL #2). b, c Percentage of CD8⁺ T cells and TNFα⁺ CD8⁺ T cells in orthotopic tumors. MLKL-KO and control cells as described in a were orthotopically implanted in the liver of C57BL/6 mice and tumor-infiltrating immune cells were analyzed. d, e Relative expression of Granzyme B and TNFα in tumor-infiltrating immune cells as described in b. f, g Percentage of CD8⁺ T cells and TNFα⁺ CD8⁺ T cells in subcutaneous tumors. MLKL-KO and control cells as described in a were subcutaneously implanted in the right flank of C57BL/6 mice and tumor-infiltrating immune cells were analyzed. h Endpoint tumor weight upon treatment of anti-PD-1. Mice carrying MLKL-KO or control tumors as described in b were treated with anti-PD-1 antibody or isotype control antibodies every 3 days for 21 days (n = 7). i Correlation between checkpoint expression and MLKL expression in HCC patients. Data were from TCGA public database. Data are represented as means ± SEM. Two-tailed Student’s t-test was used for statistical analysis. ns, not significant; *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. 3. MLKL deficiency enhances metabolic stress-induced parthanatos.

a KEGG pathway enrichment analysis of transcriptome data. MLKL-KO Hepa 1–6-luc (sgMLKL) or control (sgNC) cells were orthotopically implanted in the liver of C57BL/6 mice. Tumors collected at the endpoint of the study were subjected to RNA-seq analysis. b Cell death upon metabolic stress in HCC cells. MLKL-KO and control Hepa 1–6 cells were challenged with indicated nutrient deprivation or lipid stress (0.2 mM PA) for 24 h and cell death was measured by LDH leakage. c Cell death induced by PA. MLKL-KO and control Hepa 1–6 cells were treated with BSA or 0.2 mM PA for 24 h. Cell death was measured by SYTOX Green assay. d Dendritic cell activation. MLKL-KO Hepa 1–6 and control cells were treated with or without PA (0.2 mM, 24 h) and supernatant was collected as conditioned medium. BMDCs isolated from C57BL/6 mice were treated with conditioned medium for 24 h. Left, schematic diagram of the assay. Right, CD80 expression was analyzed using flow cytometry analysis. e PARP inhibition reverses cell death in MLKL-deficient cells. MLKL-KO and control Hepa 1–6 cells were treated with 0.2 mM PA and/or Olaparib at indicated concentrations for 24 h. f Immunoblotting analysis of PAR polymer accumulation. MLKL-KO and control Hepa 1–6 cells were treated with BSA or 0.2 mM PA for 20 h and then analyzed by immunoblotting. g AIF staining. Cells were treated as in f and AIF was stained using immunofluorescence. Left, representative images, scale bar, 10 μm. Arrows indicate AIF staining in the nucleus. Right, quantification of nuclear AIF positive cells. h Orthotopic tumor growth of MLKL-KO and control murine HCC cells. Cells were orthotopically implanted in the liver of C57BL/6 mice (n = 6). Mice were treated with or without Olaparib (50 mg/kg) daily. Tumor growth was monitored by whole-animal imaging. Shown were representative luminescent images. i Endpoint tumor weight as in h. Data are represented as means ± SEM. Two-tailed Student’s t-test was used for statistical analysis. ns, not significant; *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. 4. MLKL deficiency leads to ER dysfunction in HCC cells.

a GO pathway enrichment analysis. MLKL-KO Hepa 1–6-luc (sgMLKL) or control (sgNC) cells were treated with BSA or 0.2 mM PA for 12 h and cells were collected for RNA-seq analysis. b, c mRNA and protein level validation of genes involved in ER stress. MLKL-KO Hepa 1–6 and control cells were subjected to qPCR analysis (b). Cells treated with BSA or PA (0.2 mM, 12 h) were subjected to immunoblotting analysis (c). d Transmission electron microscopy of ER. MLKL-KO Hepa 1–6-luc and control cells were treated with BSA or 0.2 mM PA for 4 h. Shown are representative images, scale bar, 200 nm. Arrows indicate ER. e Cell death analysis. Hepa 1–6 cells were pretreated with indicated inhibitors for 2 h and then challenged with 0.2 mM PA for 24 h and cell death was analyzed by SYTOX Green assay. Ceapin-A7 as ATF6 inhibitor (ATF6i, 20 μM), GSK2606414 as PERK inhibitor (PERKi, 20 μM) and 4μ8C as IRE1α inhibitor (IRE1αi, 20 μM). Data are represented as means ± SEM. Two-tailed Student’s t-test was used for statistical analysis. ns, not significant; *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. 5. MLKL deficiency disrupts Mg²⁺ homeostasis in the ER.

a Mg²⁺ signaling in ER. MLKL-KO Hepa 1–6 and control cells were treated with BSA or 0.2 mM PA for 12 h. ER Mg²⁺ signal was visualized by co-staining the cells using ER Tracker and Mag-Fluo4-AM. Shown were representative fluorescent images, scale bar, 10 μm. b Mg²⁺ signal in mitochondria. Cells were treated as in a. Mitochondrial Mg²⁺ signal was visualized by co-staining the cells using MitoTracker and Mag-Green-AM. Shown were representative fluorescent images, scale bar, 10 μm. c Oxygen consumption rate (OCR) measurement. Cells were treated as in a. Oligomycin (Oligo, 2 μM), FCCP (1 μM) and rotenone/antimycin (ROT/AA, 0.5 μM) were added as indicated. d AIF staining. MLKL-KO Hepa 1–6 and control cells were pretreated with Mg²⁺ (MgCl₂, 10 mM) or N-acetyl-L-cysteine (NAC, 2 mM) for 1 h followed by treatment with BSA or 0.2 mM PA for 12 h. Left, representative images, scale bar, 10 μm. Arrows indicate AIF staining in the nucleus. Right, quantification of nuclear AIF positive cells. Data are represented as means ± SEM. Two-tailed Student’s t-test was used for statistical analysis. ns, not significant; *P < 0.05, **P < 0.01, ***P < 0.001.