METTL3 confers oxaliplatin resistance through the activation of G6PD-enhanced pentose phosphate pathway in hepatocellular carcinoma

Abstract

Oxaliplatin-based therapeutics is a widely used treatment approach for hepatocellular carcinoma (HCC) patients; however, drug resistance poses a significant clinical challenge. Epigenetic modifications have been implicated in the development of drug resistance. In our study, employing siRNA library screening, we identified that silencing the m⁶A writer METTL3 significantly enhanced the sensitivity to oxaliplatin in both in vivo and in vitro HCC models. Further investigations through combined RNA-seq and non-targeted metabolomics analysis revealed that silencing METTL3 impeded the pentose phosphate pathway (PPP), leading to a reduction in NADPH and nucleotide precursors. This disruption induced DNA damage, decreased DNA synthesis, and ultimately resulted in cell cycle arrest. Mechanistically, METTL3 was found to modify E3 ligase TRIM21 near the 3'UTR with N⁶-methyladenosine, leading to reduced RNA stability upon recognition by YTHDF2. TRIM21, in turn, facilitated the degradation of the rate-limiting enzyme of PPP, G6PD, through the ubiquitination-proteasome pathway. Importantly, high expression of METTL3 was significantly associated with adverse prognosis and oxaliplatin resistance in HCC patients. Notably, treatment with the specific METTL3 inhibitor, STM2457, significantly improved the efficacy of oxaliplatin. These findings underscore the critical role of the METTL3/TRIM21/G6PD axis in driving oxaliplatin resistance and present a promising strategy to overcome chemoresistance in HCC.

Fig. 1. METTL3 is involved in oxaliplatin resistance in HCC.

A–C Relative survival rate of Huh7, HepG2, and PLC/PRF/5 cells with specific siRNA transfection following OXA treatment for 48 h. D The efficiency of METTL3 knockdown verified at protein levels via WB. E The m⁶A levels of METTL3-knockdown cells detected via dot blot. F CCK-8 assays showed the relative cell viability in shNC and shMETTL3 cells with different doses of OXA treatment for 48 h. G IC50 of shNC and shMETTL3 cells calculated according to the CCK-8 assays. H The colony formation assays performed to detect the survival of Huh7 cells exposed to oxaliplatin with or without METTL3 silencing (left panel). The relative viability curve was shown (middle panel) with IC50 values calculated (right panel). I Colony formation assays were conducted with oxaliplatin, STM2457, and their combination at various concentrations (left panel). Combination indices of STM2457 and OXA in Huh7 cells calculated with CalcuSyn software (right panel). J Combination indices of STM2457 and OXA in Huh7 and HepG2 cells based on CCK-8 assays calculated with CalcuSyn software indicated a synergistic anti-tumor effect. Data are presented as mean ± SD (*p < 0.05, **p < 0.01, and ***p < 0.001).

Fig. 2. METTL3 regulates OXA susceptibility through G6PD.

A Diagram of the oxidative phase of the phosphate pentose pathway. B Ratio of 6-phosphogluconic acid/Glucose-6-phosphate (6PG/G6P) and D-Ribulose-5-phosphate/Glucose-6-phosphate (R5P/G6P) according to the non-targeted metabolomics. C The protein levels of three key enzymes in oxidative PPP detected via WB upon METTL3 silencing. D Relative RNA levels of G6PD in Huh7 and HepG2 cells detected via real-time PCR upon METTL3 silencing. E Relative G6PD activity determined in HCC cells with METTL3 silencing. F Expression of G6PD detected via WB in 12 pairs of frozen HCC tissues. G Expression of G6PD in liver cancer and adjacent normal tissues from TCGA database. H Combination indices of DHEA and OXA in Huh7 and HepG2 cells calculated with CalcuSyn software indicated a synergistic anti-tumor effect. I WB detection of G6PD and METTL3 expression in shMETTL3 and control Huh7 cells with or without G6PD overexpression. J CCK-8 assays showed the relative cell viability in shNC and shMETTL3 cells with or without G6PD overexpression following OXA treatment for 48 h (left panel). IC50 of shNC + PLVX, shMETTL3 + PLVX, and shMETTL3 + G6PD in Huh7 cells calculated according to the CCK-8 assays (right panel). K The colony formation assays performed to detect the survival of shNC + PLVX, shMETTL3 + PLVX, and shMETTL3 + G6PD cells following different doses of OXA treatment (left panel). The relative survival rate calculated was shown (right panel). Data are presented as mean ± SD (*p < 0.05, **p < 0.01, and ***p < 0.001).

Fig. 3. METTL3 regulates protein stability of G6PD via E3 ligase TRIM21.

A METTL3-silencing and control cells were treated with CHX for the indicated time, and the protein levels of G6PD and GAPDH were determined by WB (left panel). The quantitation data are shown in the right panel. B METTL3-silencing and control cells were treated with MG132 for the indicated time, and the protein levels of G6PD and GAPDH were determined by WB. C METTL3-silencing and control cells were transfected with indicated plasmids and then incubated with MG132. The levels of different proteins were detected with WB. D The lysate extracted from G6PD-Flag-overexpressing or control Huh7 cells was pulled down by anti-Flag beads and resolved by SDS-PAGE. The silver-stained gel showed differential bands with G6PD and TRIM21 highlighted. E Protein levels of TRIM21, G6PD, and GAPDH were determined in shMETTL3 Huh7 cells with or without siTRIM21 transfection. F Huh7 cells transfected with G6PD-Flag were subjected to immunoprecipitation using anti-Flag antibody or IgG control. The protein expression was detected via WB. G Huh7 cells transfected with TRIM21-Myc were subjected to immunoprecipitation using anti-Myc antibody or IgG control. The protein expression was detected via WB. H Protein levels of TRIM21, G6PD, and GAPDH were determined in TRIM21-Myc overexpression and control cells. Data are presented as mean ± SD (*p < 0.05, **p < 0.01, and ***p < 0.001).

Fig. 4. METTL3 affects mRNA stability of TRIM21 via m⁶A-YTHDF2.

A The expression of TRIM21, METTL3, and GAPDH was determined in Huh7-shNC, shMETTL3 transfected with METTL3-WT, METTL3-mut, or control plasmids. B The expression of TRIM21, G6PD, and GAPDH was determined in Huh7 cells transfected with FTO, ALKBH5, or control plasmids. C Changes of indicated proteins with different doses of STM2457 treatment detected via WB. D Results of qPCR showed the effect of different doses of STM2457 treatment on the expression of TRIM21 in Huh7 cells. GAPDH was used as the control. E Results of qPCR showed expression of TRIM21 in Huh7-shMETTL3 or control cells. GAPDH was used as the control. F Agarose electrophoresis and real-time PCR showing the abundance of TRIM21 and GAPDH in Huh7 cells immunoprecipitated with anti-METTL3 antibody or IgG control. G m⁶A modifications of TRIM21 mRNA in Huh7 and HepG2 cells from RMVar database. shMETTL3 in HepG2 diminished m⁶A modifications of TRIM21. H MeRIP was performed in Huh7 cells treated with STM2457 or vehicle using an anti-m⁶A antibody. qPCR showed the m⁶A modification levels of different fractions of TRIM21 detected with eight pairs of primers overlapping the whole length of TRIM21 mRNA. The diagram of the primers is shown above, with the blue triangle indicating the m⁶A-modification enriched sites. I qPCR analysis of TRIM21 decay rates in Huh7 cells treated with STM2457 or vehicle was performed at specified time points following Act-D treatment. The expression of TRIM21 was normalized to U6. J WB determined the expression of TRIM21, G6PD, and GAPDH in Huh7 cells transfected with YTHDF1-Myc, YTHDF2-HA, or control plasmids. K WB determined the expression of TRIM21 in Huh7 cells incubated with STM2457 or vehicle and transfected with YTHDF2 or control vectors. L Real-time PCR showing the abundance of TRIM21 and GAPDH in Huh7 cells immunoprecipitated with anti-YTHDF2 antibody or IgG control. M Real-time PCR showing the abundance of TRIM21 in Huh7 cells immunoprecipitated with anti-YTHDF2 antibody following STM2457 or vehicle treatment. Data are presented as mean ± SD (*p < 0.05, **p < 0.01, and ***p < 0.001).

Fig. 5. Silencing METTL3 blocked cell cycle due to reduced nucleotide synthesis and increased oxidative damage.

A Relative NADPH levels were determined in HCC cells with METTL3 silencing. B ROS levels of HCC cells with METTL3 silencing were determined via FCS using a DCFH-DA probe. C The level of γH2AX in METTL3-silencing Huh7 cells was determined via IF (left panel), and the MFI of γH2AX was counted (right panel). D Detection of proteins in ATM-CHK2 DNA damage response pathway and proteins responsible for G1/S transition in METTL3 silencing and control cells. E EdU assay images (left panel) were obtained from METTL3-silencing and control Huh7 cells with the quantification of positive cells shown in the right panel. F Cell cycle distributions of METTL3-silencing Huh7 cells were determined via FCS following PI and RNase treatment. G Cell cycle release assays showing the transition of G1/S with thymidine used for synchronization (lower panel). Statistics of G1/S transition rates were shown in the upper panel. H CCK-8 proliferation assays showing the rescue effect of NAC and deoxyribonucleosides (Nuc) in METTL3-silencing or control Huh7 cells. I The DNA platination detected via ICP-MS in Huh7 cells with different treatments. J The DNA platination detected via ICP-MS in both PLC/PRF/5 control cells and METTL3-overexpression cells. Data are presented as mean ± SD (*p < 0.05, **p < 0.01, and ***p < 0.001).

Fig. 6. Silencing METTL3 sensitizes HCC cells to oxaliplatin treatment in vivo and in organoid models.

The image (A) and the weights (B) of subcutaneous xenografts from NOD/SCID mice. Mice received OXA or vehicle ten days after Huh7-shNC or shMETTL3 cells plantation and were sacrificed on Day 32. C Tumor growth of subcutaneous xenografts in NOD/SCID mice was monitored every 5 days by measuring the length (L) and width (W) with a caliper. And the tumor volume (V) was calculated according to the formula V = (L × W²)/2. D IHC images of Ki-67 and c-caspase and HE images detected in subcutaneous xenografts from NOD/SCID mice (left panel). The statistics of expression of Ki-67 and c-caspase were shown in the right panel. Scale bar: 100 μm. E General images (upper) and HE images (lower) of orthotopic xenografts from C57BL/6 J mice. The mice received indicated medicine five days after laparotomy and were sacrificed on Day 24. Scale bar: 100 μm. F The in vivo imaging of orthotopic C57BL/6 J mice with an injection of D-luciferin i.p. (left panel). The statistics of the bioluminescence log values for each group are shown in the right panel. G WB detection of indicated protein in orthotopic tumors of C57BL/6 J mice following treatment of vehicle or STM2457. Each group contained three animals. H Representative IHC staining images of METTL3 in HCC specimens. Scale bar: 100 μm. I Representative IF staining images of METTL3 in HCC organoids. Scale bar: 100 μm. J Dose-response curves for OXA for 5 days in HCC organoids. K Average IC50 values of HCC organoids stratified by METTL3 expression level. L Combination indices of STM2457 and OXA on HCC organoids calculated with CalcuSyn software. Data are presented as mean ± SD (*p < 0.05, **p < 0.01, and ***p < 0.001).

Fig. 7. METTL3 expression correlates with G6PD expression and treatment responses to OXA-based HAIC in HCC.

A Representative images showing both low expression of METTL3 and G6PD in a biopsy of an HCC case. B Pre-treatment and post-treatment CT scan images of HCC in a patient with low expression of METTL3 and G6PD. C Representative images showing both high expression of METTL3 and G6PD in a biopsy of an HCC case. D Pre-treatment and post-treatment CT scan images of HCC in a patient with high expression of METTL3 and G6PD. E ROC curve analyses of METTL3 expression (left panel: continuous score; right panel: binary category) for predicting the treatment responses to OXA-based HAIC using the RECIST and modified RECIST criteria. F ROC curve analyses of G6PD expression (left panel: continuous score; right panel: binary category) for predicting the treatment responses to OXA-based HAIC using the RECIST and modified RECIST criteria. G Correlation between therapeutic efficacy and METTL3 expression as assessed by RESIST and mRESIST criteria. H Correlation between therapeutic efficacy and G6PD expression as assessed by RESIST and mRESIST criteria. PR Partial Response, CR Complete Response, SD Stable Disease, PD Progressive Disease. I The correlation between the expression of METTL3 and G6PD in a HAIC-treated HCC cohort. J Waterfall plots depicting the maximum response of intrahepatic target lesions treated with OXA-based HAIC in the METTL3-low and -high groups.