AKAP1 enhances glycogen accumulation and hepatocarcinogenesis through YTHDF2-mediated G6PC mRNA decay Authors: Tao Yang, Jiahao Zhang, Zifeng Zhao, Shixia Cheng, Zeyu Yan, Weifang Wang, Gang Wang, Renyu Zhang, Zhang Zhang, Peng Yuan, Xuyang Zheng, Hongxin Zhang, Jibin Li, Jinliang Xing Published: 2026 ABSTRACT During recent years, a growing body of evidence has revealed that metabolic rewiring of glycogen is a hallmark of cancer that plays a significant role in tumor development and progression. Nonetheless, the molecular mechanism underlying the dysregulation of glycogen remains largely unknown. In this study, using liver-specific AKAP1 depletion and overexpression mouse models, we demonstrated that AKAP1 deficiency markedly suppressed both chemical diethylnitrosamine/carbon tetrachloride (DEN/CCl₄)-induced and Akt/β-catenin oncogene-driven spontaneous hepatocellular carcinoma (HCC) by reducing hepatic glycogen content. Conversely, AKAP1 overexpression promoted glycogen accumulation and accelerated spontaneous hepatocarcinogenesis. Mechanically, m6A-dependent mRNA decay of glucose 6 phosphatase (G6PC) by YTHDF2, which was identified as a direct phosphorylation substrate at serine 289 and 359 sites by AKAP1 in a PKA-dependent manner, is important for AKAP1-caused glycogen accumulation and consequent hepatocarcinogenesis. Additionally, we found that AKAP1 expression is transcriptionally upregulated by Myc-associated zinc-finger protein (MAZ) in HCC cells. Importantly, treatment with AP-21, a competitive peptide inhibitor that disrupts mitochondrial localization of AKAP1, significantly reduced glycogen content and suppressed hepatocarcinogenesis without observable toxicity, highlighting its translational potential as a targeted therapeutic strategy for HCC. Collectively, our findings uncover a critical role for AKAP1 in driving HCC through metabolic reprogramming of glycogen, establishing AKAP1 as a promising therapeutic target for this malignancy. Introduction Glycogen, a highly branched cytosolic glucose polymer, is predominantly stored in the liver, muscle, and heart tissues, where it serves as a readily mobilizable energy reservoir to meet acute cellular energy demands.1 In recent years, a growing body of evidence has revealed that glycogen metabolism undergoes profound reprogramming in various human malignancies. Notably, the tumor initiation of hepatocellular carcinoma (HCC) exhibits particularly high levels of glycogen accumulation, a phenomenon historically recognized but only recently appreciated for its functional significance.2,3 Malignant cells exploit glycogen stores as an alternative energy source to survive under adverse microenvironmental conditions, such as hypoxia, nutrient deprivation, and oxidative stress, and to sustain metastatic dissemination.4 Moreover, emerging evidence indicates that glycogen is dynamically exchanged between cancer-associated fibroblasts (CAFs) and cancer cells within the tumor microenvironment, fueling tumor proliferation and metastasis through metabolic symbiosis.5 These findings collectively underscore the significance of glycogen metabolic reprogramming as a critical determinant of cancer progression, yet the molecular drivers of this metabolic shift remain poorly understood. Despite the growing recognition of glycogen’s role in malignancy, the molecular mechanisms governing its dysregulation in cancer cells have remained largely elusive. A-kinase anchoring protein 1 (AKAP1) is a multifunctional scaffold protein localized to the mitochondrial outer membrane (OMM), where it recruits protein kinase A (PKA) and other signaling molecules to coordinate mitochondrial dynamics, bioenergetics, calcium handling, and cellular stress responses.6 Beyond its well-characterized functions in mitochondrial homeostasis, AKAP1 has been implicated in the pathogenesis of metabolic disorders, including diabetic cardiomyopathy,7 diabetic kidney disease,8 obesity, and insulin resistance.9 Importantly, accumulating evidence points to a potential oncogenic role for AKAP1. It activates the mTOR pathway and promotes glioblastoma growth, and elevated AKAP1 expression has been detected in lung, prostate, and liver cancers,10–12, implying that AKAP1 may play a key role in these malignancies. However, the precise biological functions of AKAP1 in HCC, particularly its role in metabolic reprogramming and the underlying molecular mechanisms, have yet to be fully elucidated. Given the established connection between mitochondrial dysfunction and metabolic rewiring in cancer, we hypothesized that AKAP1 might serve as a critical node linking mitochondrial signaling to glycogen metabolism in hepatocarcinogenesis. To comprehensively investigate the role of AKAP1 in hepatocarcinogenesis and dissect its underlying mechanisms, we generated liver-specific AKAP1 knockout (AKAP1LKO) and hepatocyte-specific AKAP1 overexpression mouse models using CRISPR-Cas9 technology and adeno-associated virus serotype 8 (AAV8)-mediated gene delivery, respectively. We subjected these mice to both chemical-induced (diethylnitrosamine/carbon tetrachloride, DEN/CCl4) and oncogene-driven (Akt/β-catenin) HCC models to comprehensively evaluate the impact of AKAP1 on tumor initiation and progression. Gross and histopathological analyses were performed to assess tumor number, size, and burden, while Periodic Acid-Schiff (PAS) staining and immunohistochemistry for Ki67 and other markers were employed to evaluate glycogen accumulation and cellular proliferation. To elucidate the molecular mechanisms, we conducted unbiased immunoprecipitation (IP) coupled with liquid chromatography-tandem mass spectrometry (LC-MS/MS) to identify AKAP1-interacting proteins in HCC cells. Through this screening, we identified YTH domain-containing family protein 2 (YTHDF2), a well-characterized m6A reader, as a direct binding partner of AKAP1. The interaction and functional relationship between AKAP1 and YTHDF2 were rigorously validated using co-immunoprecipitation (Co-IP), immunofluorescence colocalization, fluorescence resonance energy transfer (FRET) assays, and in vitro kinase assays. RNA immunoprecipitation (RIP), methylated RNA immunoprecipitation (meRIP)-qPCR, and mRNA decay assays were employed to investigate how YTHDF2 regulates glucose-6-phosphatase (G6PC), a rate-limiting enzyme in glycogenolysis. Given the translational potential of targeting AKAP1, we also evaluated the therapeutic efficacy of AP-21, a competitive peptide inhibitor that disrupts AKAP1’s mitochondrial localization, in both in vitro and in vivo HCC models, assessing its effects on glycogen metabolism, tumor growth, and systemic toxicity. Our findings reveal that AKAP1 functions as a critical driver of hepatocarcinogenesis by promoting glycogen accumulation in hepatocytes. Mechanistically, AKAP1 phosphorylates YTHDF2 at serine residues 289 and 359 in a PKA-dependent manner, thereby enhancing YTHDF2 protein stability. The stabilized YTHDF2 subsequently binds to m6A-modified G6PC mRNA, accelerating its decay and leading to G6PC downregulation. This AKAP1/YTHDF2/G6PC axis ultimately results in glycogen accumulation, which fuels tumor initiation and progression. Additionally, we demonstrate that AKAP1 itself is transcriptionally upregulated by Myc-associated zinc-finger protein (MAZ) in HCC cells, establishing an upstream regulatory loop that amplifies this oncogenic signaling. Importantly, treatment with AP-21, a peptide that blocks AKAP1 mitochondrial localization, markedly reduces glycogen content, suppresses hepatocarcinogenesis, and prolongs survival in HCC-bearing mice without observable toxicity. This study provides the first direct evidence linking AKAP1-mediated mitochondrial signaling to glycogen metabolic reprogramming in HCC. It identifies a previously unrecognized post-translational regulatory mechanism that controls YTHDF2 stability and function. Furthermore, it establishes the AKAP1/YTHDF2/G6PC axis as a critical and targetable pathway in HCC pathogenesis. Collectively, these insights not only advance our fundamental understanding of glycogen metabolic rewiring in cancer but also position AKAP1 as a promising therapeutic target for HCC intervention, with AP-21 representing a lead candidate for further translational development. Results AKAP1 promotes both chemical-induced and oncogene-driven hepatocarcinogenesis To investigate the role of AKAP1 in hepatocarcinogenesis, we established a DEN/CCl4-induced liver carcinogenesis model using liver-specific AKAP1 knockout mice (Fig. 1a). Successful knockout of AKAP1 in the liver was confirmed by genotyping, qRT-PCR, and western blotting (Supplementary Fig. 1). At the endpoint of approximately eight months, fewer tumor nodules were observed in AKAP1 liver-specific knockout (LKO) mice, whereas tumor nodules were clearly observed in wild-type (WT) mice (Fig. 1b). Significantly decreased liver/body weight ratio (Fig. 1c), prolonged overall survival (OS) (Fig. 1d), diminished tumor number and maximum tumor size (Supplementary Fig. 2a, b) and reduced ALT and AST (Supplementary Fig. 2c, d) were also observed in AKAP1-deficient mice as compared to their WT littermates. H&E and immunohistochemical (IHC) staining assays further supported these observed phenotypes (Fig. 1e, f).Fig. 1Liver-specific AKAP1 knockout delays both chemical-induced and oncogene-driven hepatocarcinogenesis. a Scheme for the establishment of DEN/CCl4-induced liver carcinogenesis in a liver-specific AKAP1 knockout mouse model. b Representative images of tumor nodules in the livers of AKAP1 WT (AKAP1fl/fl) and LKO (AKAP1LKO) mice treated with DEN/CCl4. c, d Liver/body weight ratio (c) and Kaplan–Meier plot (d) of AKAP1LKO and WT mice treated with DEN/CCl4. e Representative H&E and IHC staining of Ki67 in liver tissues from AKAP1LKO and WT mice treated with DEN/CCl4. (Scale bar = 100 μm). f Statistical bar chart of Ki67 staining results in the lower panel of (e). g Scheme for the establishment of oncogene-induced (Akt and β-catenin) liver carcinogenesis in a liver-specific AKAP1 knockout mouse model. h Representative images of tumor nodules in the livers of AKAP1LKO and WT mice injected with Akt and β-catenin. i, j Liver/body weight ratio (i) and Kaplan–Meier plot (j) of AKAP1LKO and WT mice injected with Akt and β-catenin. k Representative H&E and IHC staining of Ki67 in liver tissues from AKAP1LKO and WT mice injected with Akt and β-catenin (Scale bar = 100 μm). l Statistical bar chart of Ki67 staining results for the lower panel (k) Liver-specific AKAP1 knockout delays both chemical-induced and oncogene-driven hepatocarcinogenesis. a Scheme for the establishment of DEN/CCl4-induced liver carcinogenesis in a liver-specific AKAP1 knockout mouse model. b Representative images of tumor nodules in the livers of AKAP1 WT (AKAP1fl/fl) and LKO (AKAP1LKO) mice treated with DEN/CCl4. c, d Liver/body weight ratio (c) and Kaplan–Meier plot (d) of AKAP1LKO and WT mice treated with DEN/CCl4. e Representative H&E and IHC staining of Ki67 in liver tissues from AKAP1LKO and WT mice treated with DEN/CCl4. (Scale bar = 100 μm). f Statistical bar chart of Ki67 staining results in the lower panel of (e). g Scheme for the establishment of oncogene-induced (Akt and β-catenin) liver carcinogenesis in a liver-specific AKAP1 knockout mouse model. h Representative images of tumor nodules in the livers of AKAP1LKO and WT mice injected with Akt and β-catenin. i, j Liver/body weight ratio (i) and Kaplan–Meier plot (j) of AKAP1LKO and WT mice injected with Akt and β-catenin. k Representative H&E and IHC staining of Ki67 in liver tissues from AKAP1LKO and WT mice injected with Akt and β-catenin (Scale bar = 100 μm). l Statistical bar chart of Ki67 staining results for the lower panel (k) To further substantiate the role of AKAP1 in HCC pathogenesis, a transposon-based mouse liver cancer model involving intrahepatic delivery of oncogenes (Akt and β-catenin) through hydrodynamic tail vein injection (HTVi) was also used (Fig. 1g). We observed a marked decrease in the incidence of HCC in AKAP1LKO mice compared to that in WT littermates (Fig. 1h). In line with this, a decreased liver/body weight ratio (Fig. 1i), prolonged survival (Fig. 1j), diminished tumor number and maximum tumor size (Supplementary Fig. 2e, f), reduced ALT and AST (Supplementary Fig. 2g, h) and reduced Ki67 expression (Fig. 1k, l) were also observed in AKAP1LKO mice compared to their WT littermates. Furthermore, changes in AKAP1 mRNA and protein levels among the indicated groups support the role of AKAP1 in hepatocarcinogenesis (Supplementary Fig. 3). Together, these data indicate that AKAP1 depletion in the mouse liver attenuates both chemically induced and oncogene-driven hepatocarcinogenesis. To further determine whether AKAP1 contributes to HCC development, AAV8 vector-mediated AKAP1 overexpression was performed in the livers of both chemical-induced and oncogene-driven spontaneous HCC mouse models (Supplementary Fig. 4a). QRT-PCR and western blot assays showed that the in vivo administration of AAV8-AKAP1 successfully elevated AKAP1 expression in the liver (Supplementary Fig. 5). Notably, AKAP1 overexpression in mice markedly accelerated tumor formation, as evidenced by the increased number and size of tumor nodules in the liver (Supplementary Fig. 4b). In addition, mice with AKAP1 overexpression also exhibited an elevated liver/body weight ratio (Supplementary Fig. 4c), shortened survival (Supplementary Fig. 4d), growing tumor number and maximum tumor size (Supplementary Fig. 6a, b, e, f), raised ALT and AST (Supplementary Fig. 6c, d, g, h) and increased expression of Ki67 in HCC (Supplementary Fig. 4e, f) compared to the control littermates. In addition, the differences in AKAP1 mRNA and protein levels among the indicated groups supported the role of AKAP1 in HCC tumorigenesis (Supplementary Fig. 7). To demonstrate the essential role of AKAP1 in hepatocarcinogenesis, we further explored whether AKAP1 also contributes to HCC progression. As demonstrated in Supplementary Fig. 8a, b, the proliferation of HCC cells was significantly inhibited by AKAP1 knockdown in HLF cells but accelerated by AKAP1 overexpression in SNU-739 cells. Additionally, the migration and invasion capacities were decreased by AKAP1 downregulation but increased by AKAP1 overexpression (Supplementary Fig. 8c, d). These results suggest that AKAP1 may promote hepatocarcinogenesis by enhancing both tumor initiation and progression. AKAP1 promotes glycogen accumulation in HCC cells both in vitro and in vivo Notably, we observed a clear cytoplasm with indistinct vacuoles that did not displace the nucleus in neoplastic cells overexpressing AKAP1 in spontaneous HCC mouse models, suggesting that AKAP1 might be associated with glycogen metabolism. Quantitative analysis of glycogen levels in the liver sections of mice treated with either DEN/CCl4 or Akt/β-catenin injection revealed that AKAP1 knockout significantly decreased glycogen content (Fig. 2a), whereas AKAP1 overexpression increased glycogen content compared to that in the controls (Fig. 2b). Glycogen levels in the human HCC samples were also examined. A positive correlation was found between the expression level of AKAP1 and glycogen content, which was in concordance with the findings in mouse models (Fig. 2c, d). To provide further supporting evidence, the role of AKAP1 in glycogen metabolism was analyzed in human HCC cell lines. A positive correlation was also observed between AKAP1 expression and glycogen content (Fig. 2e–h). In addition, AKAP1 knockdown significantly decreased glycogen levels in HLF cells, whereas AKAP1 overexpression dramatically increased glycogen levels in SNU-739 cells (Fig. 2i–k), as evidenced by PAS and 2-NBDG (a fluorescent D-glucose derivative) staining. In addition, considering that 2-NBDG can simultaneously reflect glucose uptake and thus affect intracellular glycogen metabolism, we detected the expression of several major glucose transporters at different AKAP1 expression levels. The results showed nonsignificant changes in GLUT1, GLUT2, GLUT3, and GLUT4 among HCC cells with different AKAP1 expression levels, suggesting that AKAP1 endogenously regulates glycogen accumulation in HCC cells (Supplementary Fig. 9).Fig. 2AKAP1 promotes glycogen accumulation in HCC cells both in vitro and in vivo. a, b Glycogen levels were detected by PAS staining in liver sections of mice treated with either DEN/CCl4 or Akt/β-catenin injection upon AKAP1 knockout (a) or overexpression (b) (scale bar = 100 μm). c Representative PAS staining and IHC staining of AKAP1 in human HCC tissue sections. d Correlation between the expression levels of AKAP1 and glycogen content in HCC tissue sections. e, f AKAP1 expression was measured by qRT-PCR (e) and western blotting (f) assays in one normal liver cell line and seven human HCC cell lines. g Glycogen levels were determined in THLE-2 and seven human HCC cell lines by colorimetric assay according to the instructions. h Corelation between glycogen content and AKAP1 expression levels in HCC cell lines. i, j Knockdown or overexpression of AKAP1 was verified by qRT-PCR (i) and western blot assays (j) in HLF and SNU-739 cells. k Glycogen levels in HLF and SNU-739 cells with AKAP1 knockdown or overexpression were determined by PAS and 2-NBDG staining assays (Scale bar = 50 μm) AKAP1 promotes glycogen accumulation in HCC cells both in vitro and in vivo. a, b Glycogen levels were detected by PAS staining in liver sections of mice treated with either DEN/CCl4 or Akt/β-catenin injection upon AKAP1 knockout (a) or overexpression (b) (scale bar = 100 μm). c Representative PAS staining and IHC staining of AKAP1 in human HCC tissue sections. d Correlation between the expression levels of AKAP1 and glycogen content in HCC tissue sections. e, f AKAP1 expression was measured by qRT-PCR (e) and western blotting (f) assays in one normal liver cell line and seven human HCC cell lines. g Glycogen levels were determined in THLE-2 and seven human HCC cell lines by colorimetric assay according to the instructions. h Corelation between glycogen content and AKAP1 expression levels in HCC cell lines. i, j Knockdown or overexpression of AKAP1 was verified by qRT-PCR (i) and western blot assays (j) in HLF and SNU-739 cells. k Glycogen levels in HLF and SNU-739 cells with AKAP1 knockdown or overexpression were determined by PAS and 2-NBDG staining assays (Scale bar = 50 μm) AKAP1 enhances glycogen accumulation by downregulating G6PC expression in HCC cells To investigate the mechanisms underlying glycogen accumulation caused by AKAP1 overexpression, the expression levels of key enzymes involved in glycogen synthesis and breakdown (Fig. 3a) were determined using qRT-PCR and western blot analysis in HLF cells with AKAP1 knockdown and SNU-739 cells with AKAP1 overexpression. We did not observe any significant changes in the expression of PGM1, PGM2, UGP2, GYS, GBE1, GAA, AGL, and PYGL following knockdown or overexpression of AKAP1 in either HLF or SNU-739 cells; however, we did observe a significant increase or decrease in G6PC expression at both the mRNA and protein levels following AKAP1 knockdown or overexpression (Fig. 3b, c). In addition, the results from targeted metabolomic analysis showed that the content of glucose-6-phosphate (G-6-P), G6PC enzyme substrate, and glucose-1-phosphate (G-1-P), upstream intermediates of the glycogen catabolism pathway, were elevated in AKAP1-overexpressing HCC cells compared with the control group (Fig. 3d, e). Further results of colorimetric assays using a commercial kit showed decreased G-6-P and G-1-P content in AKAP1-downregulated HCC cells and increased G-6-P and G-1-P content in AKAP1-upregulated HCC cells (Fig. 3f), suggesting that the downregulation of G6PC may contribute to AKAP1-induced glycogen accumulation. To test this hypothesis, we knocked down G6PC in HLF cells using AKAP1 knockdown and overexpressed G6PC in SNU-739 cells overexpressing AKAP1. The decrease in glycogen content caused by AKAP1 knockdown was significantly restored by G6PC knockdown in HLF cells, whereas glycogen accumulation induced by AKAP1 overexpression was markedly attenuated by G6PC overexpression in SNU-739 cells (Fig. 3g), indicating that AKAP1 induces glycogen accumulation by downregulating G6PC in HCC cells. Next, a significant negative correlation was observed between the protein expression levels of AKAP1 and G6PC in tumor tissues from 272 early-stage HCC patients (r = −0.336, P < 0.001) (Fig. 3h, i). To elucidate the clinical-translational implications of AKAP1/G6PC dysregulation, we integrated HCC patients’ clinical characteristics into survival analysis. Clinicopathological parameters were summarized in Supplementary Table 9. Univariate and multivariate Cox regression analyses revealed that AKAP1 upregulation and G6PC downregulation emerged as independent prognostic factors for OS in HCC patients (Supplementary Tables 10 and 11). Strikingly, patients exhibiting high AKAP1 and low G6PC had the poorest survival (Fig. 3j), establishing a synergistic prognostic value of the AKAP1/G6PC axis in HCC.Fig. 3AKAP1 enhances glycogen accumulation via downregulating G6PC expression in HCC cells. a A schematic diagram of key enzymes involved in the regulation of glycogen synthesis and breakdown. b, c The expression levels of key enzymes involved in glycogen synthesis and breakdown were determined by qRT-PCR and western blot analysis in HLF and SNU-739 cells with AKAP1 knockdown or overexpression. d, e Targeted metabolomics analysis was performed to evaluate the involvement of G6PC in AKAP1-mediated glycogen accumulation. f Colorimetric determination of G-6-P and G-1-P content in HCC cells with different AKAP1 expression levels. g Glycogen levels were determined by PAS and 2-NBDG staining assays in HLF and SNU-739 cells treated as indicated (Scale bar = 50 μm). h, i G6PC expression in tumor tissues from HCC patients was determined by IHC staining, and the correlation between the expression of AKAP1 and G6PC was analyzed (n = 272). j Kaplan‑Meier curves showing overall survival of HCC patients stratified by combined AKAP1 and G6PC expression levels AKAP1 enhances glycogen accumulation via downregulating G6PC expression in HCC cells. a A schematic diagram of key enzymes involved in the regulation of glycogen synthesis and breakdown. b, c The expression levels of key enzymes involved in glycogen synthesis and breakdown were determined by qRT-PCR and western blot analysis in HLF and SNU-739 cells with AKAP1 knockdown or overexpression. d, e Targeted metabolomics analysis was performed to evaluate the involvement of G6PC in AKAP1-mediated glycogen accumulation. f Colorimetric determination of G-6-P and G-1-P content in HCC cells with different AKAP1 expression levels. g Glycogen levels were determined by PAS and 2-NBDG staining assays in HLF and SNU-739 cells treated as indicated (Scale bar = 50 μm). h, i G6PC expression in tumor tissues from HCC patients was determined by IHC staining, and the correlation between the expression of AKAP1 and G6PC was analyzed (n = 272). j Kaplan‑Meier curves showing overall survival of HCC patients stratified by combined AKAP1 and G6PC expression levels Taking into account AKAP1’s role in regulating mitochondrial dynamics and its context-dependent modulation of cellular survival and energy metabolism,13,14 we investigated the impact of AKAP1-mediated glycogen accumulation on mitochondrial respiration. Our results show that AKAP1 silencing decreased the oxygen consumption rate (OCR) in HLF cells, and this decrease was not notably altered by G6PC overexpression or knockdown (Supplementary Fig. 10a). In contrast, AKAP1 overexpression enhanced mitochondrial respiration in SNU-739 cells, and this enhancement was not significantly affected by G6PC silencing or overexpression (Supplementary Fig. 10b). These results suggest that AKAP1-mediated glycogen accumulation may not affect mitochondrial respiration. Taken together, these results suggest that AKAP1 may enhance glycogen accumulation by downregulating G6PC expression in HCC cells. G6PC downregulation-mediated glycogen accumulation plays a crucial role in the promotion of hepatocarcinogenesis by AKAP1 Next, we examined whether G6PC downregulation-mediated glycogen accumulation plays a role in AKAP1-promoted hepatocarcinogenesis in both chemical-induced and oncogene-driven spontaneous HCC mouse models (Fig. 4a and Supplementary Fig. 11a). We noted that lentiviral vector-mediated G6PC silencing in the livers of mice dramatically restored HCC development suppressed by AKAP1 knockout (Fig. 4b–f, Supplementary Figs. 11b–f, and 12a–d). In contrast, lentiviral vector-mediated overexpression of G6PC in the livers of mice significantly attenuated HCC development promoted by AKAP1 overexpression (Fig. 4g–k, Supplementary Figs. 11g–k, and 12e–h), indicating that AKAP1 promotes hepatocarcinogenesis through G6PC downregulation-mediated glycogen accumulation. Meanwhile, the results also showed that hepatocarcinogenesis was markedly promoted by G6PC knockdown in AKAP1(LKO) mice and suppressed by G6PC overexpression in AKAP1-overexpressing mice. This finding aligns with a previous report demonstrating that glycogen accumulation caused by G6PC deficiency drives the initiation of liver cancer.3 In addition, the differences in the mRNA and protein levels of AKAP1 and G6PC among the indicated groups supported the role of AKAP1 in HCC tumorigenesis (Supplementary Fig. 13).Fig. 4G6PC downregulation-mediated glycogen accumulation plays a crucial role in the promotion of hepatocarcinogenesis by AKAP1. a Scheme for the establishment of G6PC knockdown or overexpression in the liver of AKAP1LKO or AAV8-AKAP1 mice treated with DEN/CCl4. b Representative images of tumor nodules in the livers of mice injected with DEN/CCl4. c, d Liver/body weight ratios (c) and Kaplan–Meier plots (d) of mice treated with DEN/CCl4. e Representative H&E, Ki67, and PAS staining images of the liver tissues from mice injected with DEN/CCl4 injection. (Scale bar = 100 μm). f Statistical bar chart of Ki67 and PAS staining results is shown in Fig. 4e. g Representative images of tumor nodules in the livers of mice injected with DEN/CCl4. h, i Liver/body weight ratio (h) and Kaplan–Meier plot (i) of the indicated mice treated with DEN/CCl4. j Representative H&E, Ki67, and PAS staining images of liver tissues from mice treated with DEN/CCl4. (Scale bar = 100 μm). k Statistical bar chart of the Ki67 and PAS staining results in Fig. 4j G6PC downregulation-mediated glycogen accumulation plays a crucial role in the promotion of hepatocarcinogenesis by AKAP1. a Scheme for the establishment of G6PC knockdown or overexpression in the liver of AKAP1LKO or AAV8-AKAP1 mice treated with DEN/CCl4. b Representative images of tumor nodules in the livers of mice injected with DEN/CCl4. c, d Liver/body weight ratios (c) and Kaplan–Meier plots (d) of mice treated with DEN/CCl4. e Representative H&E, Ki67, and PAS staining images of the liver tissues from mice injected with DEN/CCl4 injection. (Scale bar = 100 μm). f Statistical bar chart of Ki67 and PAS staining results is shown in Fig. 4e. g Representative images of tumor nodules in the livers of mice injected with DEN/CCl4. h, i Liver/body weight ratio (h) and Kaplan–Meier plot (i) of the indicated mice treated with DEN/CCl4. j Representative H&E, Ki67, and PAS staining images of liver tissues from mice treated with DEN/CCl4. (Scale bar = 100 μm). k Statistical bar chart of the Ki67 and PAS staining results in Fig. 4j AKAP1 stabilizes YTHDF2 via phosphorylating its serine 289 and 359 sites in HCC cells AKAP1 is a key scaffold protein for PKA that facilitates the phosphorylation of target proteins.6 Moreover, to provide insights into the underlying mechanism by which AKAP1 downregulates G6PC expression, IP and MS were conducted to identify the potential phosphorylation substrates of AKAP1 in both HCC-LM3 and HLF cells. IP-MS revealed 358 proteins in the anti-AKAP1-IP sample of HCC-LM3 cells and 824 proteins in the anti-AKAP1-IP sample of HLF cells, of which 87 were common in HCC-LM3 and HLF cells (Fig. 5a). Interestingly, Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis suggested that the 87 proteins were mainly enriched in pathways related to RNA processing and mRNA stability (Fig. 5b), and YTHDF2, the first identified m6A reader, was identified as a potential interacting partner with AKAP1 (Fig. 5c). The colocalization and interaction of AKAP1 and YTHDF2 in HCC cells were verified by immunofluorescence staining and FRET assays (Fig. 5d, e). Interestingly, we found that AKAP1 increased the expression of the YTHDF2 protein without affecting its mRNA levels (Supplementary Fig. 14a, b), implying posttranscriptional regulation of YTHDF2 by AKAP1. Moreover, both AKAP1 knockdown and PKA suppression by inhibitor H89 treatment accelerated the degradation of YTHDF2 protein (Fig. 5f, g), suggesting an essential role for PKA in AKAP1-regulated YTHDF2 expression. Next, we investigated whether YTHDF2 is phosphorylated by AKAP1/PKA and found that both AKAP1 knockdown and PKA suppression significantly impaired the phosphorylation of YTHDF2, whereas the activation of PKA by FSK treatment markedly enhanced the phosphorylation of YTHDF2 (Fig. 5h, i). Additionally, GPS 5.0 predicted that the serine and threonine sites in YTHDF2 might be phosphorylated by PKA (Supplementary Fig. 14c), of which S289 and S359 are conserved across different mammalian species (Supplementary Fig. 14d), implying that these two sites are most likely AKAP/PKA phosphorylation sites. An in vitro kinase assay showed that WT-YTHDF2 could be directly phosphorylated by AKAP1/PKA, whereas both the S289A and S359A mutants exhibited compromised phosphorylation. In particular, the S289A/S359A double mutation completely abrogated the phosphorylation of YTHDF2 by PKA (Fig. 5j). Next, we determined whether phosphorylation of YTHDF2 at serine 289 and 359 by AKAP1/PKA affects the stability of the YTHDF2 protein. Compared with the WT, the S289A/S359A mutant displayed significantly accelerated degradation of YTHDF2 (Fig. 5k). Moreover, overexpression of the S289A/S359A mutant had minimal effects on the promotion of glycogen accumulation in HLF cells compared to WT YTHDF2 (Fig. 5l), suggesting the importance of the AKAP1/PKA/YTHDF2 cascade in glycogen accumulation in HCC cells. Finally, a positive correlation between the protein expression levels of AKAP1 and YTHDF2 was detected in HCC tumor tissues from the patients (r = 0.407, P < 0.001) (Supplementary Fig. 14e, f).Fig. 5AKAP1 stabilizes YTHDF2 via phosphorylation at serine residues 289 and 359 in HCC cells. a Co-IP and MS were conducted to identify potential phosphorylation substrates of AKAP1 in both HCC-LM3 and HLF cells with high expression levels of AKAP1. b Enriched KEGG pathways by 87 proteins, which were identified as potential phosphorylation substrates of AKAP1 in the intersection between HCC-LM3 and HLF cells. c co-IP of YTHDF2 with AKAP1, or co-IP of AKAP1 with YTHDF2, in whole-cell extracts from HLF cells. d, e Immunofluorescence staining (d) and FRET (e) assays for colocalization of AKAP1 and YTHDF2 in HLF cells (scale bar = 10 μm). f Stability of YTHDF2 was measured by CHX chase assay in HLF cells with AKAP1 knockdown. g The stability of YTHDF2 was measured by CHX chase assay in HLF cells treated with H89 at a dose of 10 μM. h, i Phosphorylation of YTHDF2 was determined in HLF cells with the indicated treatment (FSK, 50 μM; H89, 10 μM). j In vitro kinase assay of YTHDF2. Recombinant active PKA and wild type (WT), S289A, S359A, or S289A/S359A YTHDF2 were incubated in the presence of [γ-32P] ATP. Autoradiography was performed after the proteins separated by SDS-PAGE were transferred to PVDF membranes. MBP served as a positive control, and coincubation with H89 served as a negative control. Coomassie blue staining served as a loading control. k The stability of YTHDF2 was measured by CHX chase assay in HLF cells with the indicated treatment. l Glycogen levels were determined by PAS and 2-NBDG staining assays in HLF and SNU-739 cells treated as indicated AKAP1 stabilizes YTHDF2 via phosphorylation at serine residues 289 and 359 in HCC cells. a Co-IP and MS were conducted to identify potential phosphorylation substrates of AKAP1 in both HCC-LM3 and HLF cells with high expression levels of AKAP1. b Enriched KEGG pathways by 87 proteins, which were identified as potential phosphorylation substrates of AKAP1 in the intersection between HCC-LM3 and HLF cells. c co-IP of YTHDF2 with AKAP1, or co-IP of AKAP1 with YTHDF2, in whole-cell extracts from HLF cells. d, e Immunofluorescence staining (d) and FRET (e) assays for colocalization of AKAP1 and YTHDF2 in HLF cells (scale bar = 10 μm). f Stability of YTHDF2 was measured by CHX chase assay in HLF cells with AKAP1 knockdown. g The stability of YTHDF2 was measured by CHX chase assay in HLF cells treated with H89 at a dose of 10 μM. h, i Phosphorylation of YTHDF2 was determined in HLF cells with the indicated treatment (FSK, 50 μM; H89, 10 μM). j In vitro kinase assay of YTHDF2. Recombinant active PKA and wild type (WT), S289A, S359A, or S289A/S359A YTHDF2 were incubated in the presence of [γ-32P] ATP. Autoradiography was performed after the proteins separated by SDS-PAGE were transferred to PVDF membranes. MBP served as a positive control, and coincubation with H89 served as a negative control. Coomassie blue staining served as a loading control. k The stability of YTHDF2 was measured by CHX chase assay in HLF cells with the indicated treatment. l Glycogen levels were determined by PAS and 2-NBDG staining assays in HLF and SNU-739 cells treated as indicated YTHDF2 downregulates G6PC through m6A-dependent mRNA decay YTHDF2 is a cytoplasmic m6A reader that executes its function by destabilizing its target mRNAs.15 Given that G6PC expression is negatively regulated by AKAP1 at the mRNA level, we investigated whether G6PC is a direct target of YTHDF2. The results showed that G6PC expression at both the mRNA and protein levels was significantly increased upon YTHDF2 knockdown in HLF cells, but markedly decreased upon YTHDF2 overexpression in SNU-739 cells (Fig. 6a, b). In fact, several YTHDF2 binding sites in G6PC were identified by meRIP-qPCR (Fig. 6c). In addition, RNA binding protein immunoprecipitation (RIP) and cross-linking immunoprecipitation (CLIP) assays revealed a direct interaction between G6PC mRNA and YTHDF2 in both HLF and SNU-739 cells (Fig. 6d, e). Next, we explored whether YTHDF2 regulates G6PC expression via m6A methylation-mediated mRNA decay by evaluating the mRNA half-life of G6PC in HLF and SNU-739 cells. The mRNA decay of G6PC was significantly slowed or accelerated when YTHDF2 was either silenced in HLF cells or overexpressed in SNU-739 cells (Fig. 6f). In addition, we explored whether G6PC expression regulated by YTHDF2 depends on N6 adenylate methylation using an in vivo RNA precipitation assay. The streptavidin aptamer-based capture assay showed that the mRNA with mutated m6A sites for G6PC abrogated its association with YTHDF2 (Fig. 6g), indicating that m6A modification of G6PC mRNA is vital for its association with YTHDF2. In addition, the luciferase reporter assay showed that ectopic YTHDF2 induced a significant decrease in the luciferase activity of the WT reporter but not the m6A site mutant reporter in SNU-739 and SNU-398 cells (Fig. 6h). Moreover, tryptophan residues at positions 432, 486, and 491 in the YTH domain are critical for m6A recognition by YTHDF. Therefore, we generated m6A-recognition defective YTHDF2 (YTHDF2 mutant) using the W432A, W486A, and W491A triple mutations. RIP and CLIP assays showed the interaction of G6PC mRNA with WT YTHDF2 but not with the YTHDF2 mutant (Fig. 6i). Compared to WT YTHDF2, the YTHDF2 mutant failed to decrease the level of G6PC at both the mRNA and protein levels, thus accelerating the mRNA decay of G6PC (Fig. 6j–l). Our results also showed that overexpression of YTHDF2 impaired the upregulation of G6PC caused by AKAP1 knockdown, whereas knockdown of YTHDF2 rescued the downregulation of G6PC caused by AKAP1 overexpression (Fig. 6m). Consistent with the effect of YTHDF2, activation of PKA by FSK treatment also significantly attenuated AKAP1 knockdown-upregulated G6PC expression, whereas suppression of PKA by H89 treatment markedly rescued AKAP1 overexpression-downregulated G6PC expression (Fig. 6n). Together, these results indicated that AKAP1 downregulates G6PC through YTHDF2-mediated m6A-dependent mRNA decay in HCC cells.Fig. 6YTHDF2 downregulates G6PC through m6A-dependent mRNA decay. a, b QRT-PCR (a) and western blotting (b) assays for mRNA and protein expression levels of YTHDF2 and G6PC in HLF and SNU-739 cells with YTHDF2 either silenced or overexpressed. c m6A enrichment of G6PC in HLF cells with YTHDF2 knockdown. d, e YTHDF2 CLIP (d) and RIP (e) assays for the interaction between the YTHDF2 protein and G6PC mRNA in HLF and SNU-739 cells. f Lifetime of G6PC mRNA in HLF and SNU-739 cells with YTHDF2 either silenced or overexpressed, treated with 5 μM actinomycin D for different times. g Top, schematic representation of the S1m‑mediated in vivo RNA precipitation assay. Bottom, immunoblot for YTHDF2 in HCC cells with WT or mutant S1m tagged G6PC mRNA was pulled down by streptavidin beads. h Luciferase activity of WT (G6PC‑WT) or mutated (G6PC‑Mut) G6PC reporters in SNU-739 and SNU-398 cells with ectopically expressed YTHDF2. i CLIP and RIP assays for the interaction of G6PC with WT or m6A recognition-defective YTHDF2. j qRT-PCR analysis of the mRNA expression levels of YTHDF2 and G6PC in SNU-739 and SNU-398 cells expressing WT or m6A recognition-defective YTHDF2. k Western blotting assay for protein expression levels of YTHDF2 and G6PC in SNU-739 and SNU-398 cells expressing WT or m6A recognition-defective YTHDF2. l Lifetime of G6PC mRNA in SNU-739 and SNU-398 cells expressing WT or m6A recognition defective YTHDF2 treated with 5 μM actinomycin D. m, n Relative mRNA levels of G6PC were detected in SNU-739 and HLF cells YTHDF2 downregulates G6PC through m6A-dependent mRNA decay. a, b QRT-PCR (a) and western blotting (b) assays for mRNA and protein expression levels of YTHDF2 and G6PC in HLF and SNU-739 cells with YTHDF2 either silenced or overexpressed. c m6A enrichment of G6PC in HLF cells with YTHDF2 knockdown. d, e YTHDF2 CLIP (d) and RIP (e) assays for the interaction between the YTHDF2 protein and G6PC mRNA in HLF and SNU-739 cells. f Lifetime of G6PC mRNA in HLF and SNU-739 cells with YTHDF2 either silenced or overexpressed, treated with 5 μM actinomycin D for different times. g Top, schematic representation of the S1m‑mediated in vivo RNA precipitation assay. Bottom, immunoblot for YTHDF2 in HCC cells with WT or mutant S1m tagged G6PC mRNA was pulled down by streptavidin beads. h Luciferase activity of WT (G6PC‑WT) or mutated (G6PC‑Mut) G6PC reporters in SNU-739 and SNU-398 cells with ectopically expressed YTHDF2. i CLIP and RIP assays for the interaction of G6PC with WT or m6A recognition-defective YTHDF2. j qRT-PCR analysis of the mRNA expression levels of YTHDF2 and G6PC in SNU-739 and SNU-398 cells expressing WT or m6A recognition-defective YTHDF2. k Western blotting assay for protein expression levels of YTHDF2 and G6PC in SNU-739 and SNU-398 cells expressing WT or m6A recognition-defective YTHDF2. l Lifetime of G6PC mRNA in SNU-739 and SNU-398 cells expressing WT or m6A recognition defective YTHDF2 treated with 5 μM actinomycin D. m, n Relative mRNA levels of G6PC were detected in SNU-739 and HLF cells Given the close association between metabolic dysfunction-associated steatohepatitis (MASH) and HCC development,16 we investigated the role of the AKAP1-YTHDF2-G6PC axis in MASH progression using a murine model of MASH induced in AKAP1 LKO mice.17 In line with the previous regulatory pattern, our analysis revealed that AKAP1 and YTHDF2 expression levels were significantly downregulated and hepatic glycogen content was markedly depleted in AKAP1LKO mice compared to AKAP1fl/fl controls fed a fast-food diet, whereas the G6PC expression level was elevated in these models (Supplementary Fig. 15). These results indicated that the AKAP1/YTHDF2/G6PC pathway might be enrolled in MASH-HCC initiation or progression. AKAP1 is transcriptionally upregulated by MAZ in HCC cells It has been reported that the protein expression level of AKAP1 was increased in HCC compared with normal livers.12 Using the UALCAN database,18 we found that AKAP1 expression was significantly upregulated at the mRNA level in HCC cells (Supplementary Fig. 16a), suggesting that the upregulation of AKAP1 may occur at the epigenetic or transcriptional level. UALCAN-based analysis indicated no significant difference in the promoter methylation level of AKAP1 between tumor and non-tumor HCC tissues (Supplementary Fig. 16b), implying that AKAP1 upregulation may occur at the transcriptional level. Based on an overlay of the PROMO, GeneCards, hTFtarget, and GRAND databases, two transcription factors, YY1 and MAZ, were predicted as potential upstream transcription factors of AKAP1 (Supplementary Fig. 17a). To examine the role of YY1 and MAZ in the regulation of AKAP1 expression in HCC cells, changes in AKAP1 expression were evaluated by qRT-PCR and western blotting following knockdown of YY1 or MAZ in HLF cells. Unlike YY1, MAZ knockdown markedly decreased AKAP1 expression at both the mRNA and protein levels (Supplementary Fig. 17b). As expected, MAZ knockdown significantly elevated G6PC expression at both the mRNA and protein levels while decreasing YTHDF2 expression (Supplementary Fig. 17c). Next, a series of truncated AKAP1 promoter constructs was developed to test whether AKAP1 is a direct transcriptional target of MAZ. The luciferase reporter assay revealed that unlike AKAP1 promoter constructs truncated from −2000 to −300, the construct truncated from −300 to +24 completely abolished the transcriptional activity of MAZ in HLF and SNU-739 cells (Supplementary Fig. 17d). Site-directed mutagenesis further identified putative DNA-binding sites of AKAP1 is located between nt-404 and nt-394 (Supplementary Fig. 17e). Additionally, the ChIP-PCR assay verified that MAZ directly bound to the WT AKAP1 promoter (Supplementary Fig. 17f). However, when the DNA-binding site was mutated, MAZ binding was significantly impaired (Supplementary Fig. 17g). Clinically, cBioPortal database analysis showed a significant positive correlation between the mRNA expression levels of MAZ and AKAP1 in the tumor tissues of HCC patients (Supplementary Fig. 16c). Consistently, a positive correlation was observed between the protein expression levels of MAZ and AKAP1 in tumor tissues from 272 patients with HCC (r = 0.365; P < 0.001) (Supplementary Fig 16d). Furthermore, we clearly observed decreased expression of AKAP1 and YTHDF2 but increased expression of G6PC upon MAZ knockdown in HLF cells, which was markedly reversed by forced AKAP1 expression. In contrast, overexpression of MAZ elevated AKAP1 and YTHDF2 while decreasing G6PC in SNU-739 cells, which was rescued by AKAP1 silencing (Supplementary Fig. 17h). Consistent with these findings, the levels of glycogen were significantly down- or upregulated when MAZ was either knocked down or overexpressed in HLF and SNU-739 cells, which was significantly rescued by either overexpression or knockdown of AKAP1 (Supplementary Fig. 17i). Besides, the effects of MAZ on HCC cell proliferation, migration, and invasion were determined. We found that the proliferation, migration, and invasion of HCC cells were significantly suppressed by MAZ knockdown and restored by forced AKAP1 expression. By contrast, MAZ overexpression promoted the proliferation, migration, and invasion of HCC cells, which were abolished by AKAP1 knockdown (Supplementary Fig. 18a–d). Regarding the effect of MAZ on the expression of EMT markers, we found that MAZ knockdown decreased mesenchymal markers (N-cadherin and ZEB1) and increased epithelial markers (E-cadherin), which was reversed by forced AKAP1 expression. In contrast, MAZ overexpression increased mesenchymal markers and decreased epithelial markers, which was reversed by AKAP1 knockdown (Supplementary Fig. 18e, f). Moreover, the univariate and multivariate Cox regression analyses (Supplementary Tables 10 and 11) identified aberrant AKAP1, YTHDF2, and G6PC expression levels as independent prognostic factors for OS in HCC patients, whereas MAZ showed a nonsignificant trend toward being an independent prognostic factor (HR = 1.209, 95% CI: 0.895–1.633, P = 0.217), which is in line with the study by Li et al.19 To clinically translate these findings, we developed a nomogram integrating the gene signature with key clinicopathological parameters (Supplementary Fig. 19a). The nomogram exhibited superior predictive accuracy compared to conventional staging systems, with time-dependent ROC curves demonstrating AUC values of 0.810 (1-year), 0.783 (3-year), and 0.738 (5-year) (Supplementary Fig. 19b–h). These results establish a novel prognostic biomarker panel for HCC outcome prediction. AP-21 inhibited HCC tumorigenesis To explore the therapeutic potential of targeting AKAP1 in HCC, AP-21, a competitive peptide inhibitor that blocks AKAP1 mitochondrial localization,9 was utilized. A non-targeting mutant AP-21 served as a control. AP-21 treatment significantly reduced glycogen content, decreased YTHDF2 expression, and accelerated YTHDF2 protein degradation in HLF cells (Fig. 7a–c). In addition, AP-21 significantly slowed G6PC mRNA decay in HLF cells (Fig. 7d). We next investigated the role of AP-21 in the suppression of hepatocarcinogenesis in vivo (Fig. 7e). Mice treated with AP-21 (2 mg/kg, twice weekly for 4 weeks) exhibited fewer tumors, a lower liver-to-body weight ratio, and reduced tumor burden versus the control/mutant groups (Fig. 7f, g and Supplementary Fig. 20a, b). Survival analysis confirmed that AP-21 extended the survival of HCC-bearing mice (Fig. 7h). Histological evaluation revealed that AP-21-treated mice had dramatically fewer Ki67⁺ tumor cells and reduced intratumoral glycogen (Fig. 7i–j). Tumor samples from AP-21-treated mice showed decreased YTHDF2 expression and increased G6PC expression relative to the controls (Supplementary Fig. 20c). Importantly, AP-21 caused no obvious toxicity, as assessed by organ histology and blood biochemical markers (Supplementary Fig. 21). Collectively, as shown in Fig. 7k, these data support AKAP1 targeting via AP-21 as a promising therapeutic strategy for HCC.Fig. 7AP-21 inhibited HCC tumorigenesis. a Glycogen content determined by PAS staining and 2-NBDG staining assays in HLF cells treated as indicated. b The expression levels of YTHDF2 and G6PC in HLF cells were treated as indicated. c Stability of YTHDF2 was measured by the CHX chase assay in HLF cells with AP-21 and mutant AP-21 treatments. d Lifetime of G6PC mRNA in HLF cells given AP-21 and mutant AP-21 peptides after 24 h was evaluated by sampling at different times of 5 μM actinomycin D treatment. e Scheme for the establishment of AP-21 treatment in the livers of mice treated with Akt/β-catenin. f Representative images of tumor nodules in the liver of Akt/β-catenin plasmid hydrodynamically injected mice treated with AP-21. g, h Liver/body weight ratios (g) and Kaplan–Meier plot (h) of Akt/β-catenin plasmid hydrodynamic injection mice treated with AP-21. i Representative H&E, Ki67, and PAS staining images of the liver tissues from Akt/β-catenin plasmid hydrodynamically injected mice treated with AP-21. (Scale bar = 100 μm). j Statistical bar chart of Ki67 and PAS staining results is shown in (i). k Graphic summary of the mechanisms through which AKAP1 regulates glycogen accumulation to promote hepatocarcinogenesis AP-21 inhibited HCC tumorigenesis. a Glycogen content determined by PAS staining and 2-NBDG staining assays in HLF cells treated as indicated. b The expression levels of YTHDF2 and G6PC in HLF cells were treated as indicated. c Stability of YTHDF2 was measured by the CHX chase assay in HLF cells with AP-21 and mutant AP-21 treatments. d Lifetime of G6PC mRNA in HLF cells given AP-21 and mutant AP-21 peptides after 24 h was evaluated by sampling at different times of 5 μM actinomycin D treatment. e Scheme for the establishment of AP-21 treatment in the livers of mice treated with Akt/β-catenin. f Representative images of tumor nodules in the liver of Akt/β-catenin plasmid hydrodynamically injected mice treated with AP-21. g, h Liver/body weight ratios (g) and Kaplan–Meier plot (h) of Akt/β-catenin plasmid hydrodynamic injection mice treated with AP-21. i Representative H&E, Ki67, and PAS staining images of the liver tissues from Akt/β-catenin plasmid hydrodynamically injected mice treated with AP-21. (Scale bar = 100 μm). j Statistical bar chart of Ki67 and PAS staining results is shown in (i). k Graphic summary of the mechanisms through which AKAP1 regulates glycogen accumulation to promote hepatocarcinogenesis Discussion Cancer cells alter their metabolism to support their rapid proliferation and invasion, which has emerged as a promising therapeutic target.20 In recent years, investigators have turned their attention to the metabolic rewiring of glycogen, which plays a critical role in tumor progression.2 However, deeper insights into several important questions regarding the regulation and role of dysregulated glycogen metabolism in cancer are urgently needed. Here, using a liver-specific AKAP1 depletion and overexpression mouse model, we found that AKAP1 prompted the development of both DEN/CCl4-induced and Akt/β-catenin-driven spontaneous liver cancers by enhancing glycogen accumulation, emphasizing the importance of AKAP1 in reprogramming glycogen metabolism during hepatocarcinogenesis. Mechanistically, we identified YTHDF2 as a direct phosphorylation target of AKAP1 and found that AKAP1 promotes glycogen accumulation via YTHDF2-mediated G6PC downregulation. Additionally, we revealed that AKAP1 is transcriptionally upregulated by MAZ, a critical transcription factor that is abnormally expressed in many human cancer types, including HCC. What’s more, AP-21, a competitive peptide that inhibits the localization of AKAP1 to mitochondria, was demonstrated to obviously suppress cellular glycogen content and HCC tumorigenesis. Thus, the MAZ/AKAP1/YTHDF2/G6PC cascade may control glycogen accumulation and hepatocarcinogenesis, and targeting AKAP1 might be a potential therapeutic treatment for HCC. High levels of glycogen have been described in various human cancers compared to their tissue of origin. However, scientists have recently turned their attention to reprogrammed glycogen metabolism in cancer.4 Glycogen accumulation has been reported to enhance the survival of lung cancer cells and DRP1-deficient colon cancer cells after chemoradiotherapy.21,22 Glycogens derived from CAFs have also been demonstrated to promote the proliferation and metastasis of ovarian cancer cells.5 Although it is well recognized that glycogen contributes to the aggressiveness and survival of cancer cells,23 the mechanisms underlying reprogrammed glycogen metabolism remain unclear. Here, by utilizing a liver-specific AKAP1 depletion or overexpression mouse model, we found that AKAP1, a crucial mitochondrion-localized protein involved in the maintenance of mitochondrial function, plays a critical role in glycogen accumulation in DEN/CCl4-induced and Akt/β-catenin-driven spontaneous HCC tissues, suggesting that in addition to exerting the functions in tumor progression, accumulated glycogen also plays crucial roles in the occurrence of human cancers. Similar to our findings, it has been reported that accumulated glycogen drives the initiation of HCC by suppressing the Hippo signaling pathway,10 which further supports the notion that glycogen not only functions as an energy source for cancer cell proliferation and metastasis but also serves as a key tumor-initiating promoter. In addition, we verified the contribution of AKAP1 to glycogen accumulation in HCC cell lines and tumor tissues of patients. As an OMM-localized protein that recruits various molecules to regulate mitochondrial functions,6 few studies have focused on the role of AKAP1 in human cancers. It has been reported that AKAP1 is upregulated in HCC and that its upregulation is closely associated with tumor size, metastasis stage, and poor patient survival,12 which further supports our findings that AKAP1 functions as a critical oncogene in the development and progression of HCC. As an OMM-localized protein, AKAP1 anchors PKA to regulate the phosphorylation of multiple mitochondrial proteins to maintain the integrity and function of mitochondria.11 Recent studies have revealed that AKAP1 contributes to metabolic diseases by phosphorylating multiple mitochondrial proteins. For example, AKAP1 has been reported to promote the phosphorylation of dynamin-related protein 1 (DRP1) in a PKA-dependent manner, which suppresses mitochondrial fragmentation and protects against cerebral ischemic stroke.13 AKAP1 also contributes to podocyte injury in diabetic kidney disease by phosphorylating La-related protein 1 (Larp1), which disrupts mtDNA replication and mitochondrial function.8 In brown adipocytes and cardiomyocytes, AKAP1 has been demonstrated to phosphorylate ACSL1 and NDUFS1 to inhibit thermogenesis and energy expenditure.7,9 Here, we identified YTHDF2 as a direct phosphorylation target of the AKAP1 complex in HCC cells and uncovered an unappreciated localization of YTHDF2 in the mitochondria. In addition, AKAP1 promotes the phosphorylation of YTHDF2 at serine 289 and 359 in a PKA-dependent manner, leading to enhanced YTHDF2 protein stability. The S289A/S359A double mutant largely dampened the role of YTHDF2 in glycogen accumulation in HCC. Moreover, a previous study demonstrated that glycogen accumulation could activate Hippo signaling to accelerate hepatocarcinogenesis.3 Notably, starch-binding domain-containing protein 1 (Stbd1), a glycogen scaffold protein, mediates glycogen dysregulation independent of mitochondrial respiration.24 Our study reveals that AKAP1-mediated glycogen accumulation does not affect mitochondrial respiration, suggesting a novelty in AKAP1’s functional roles. While AKAP1 regulates mitochondrial dynamics, it simultaneously governs cytoplasmic processes such as glycogen metabolism. Additionally, AKAP1-induced glycogen accumulation may exacerbate HCC malignancy through Hippo signaling activation, rather than mitochondrial pathways, thereby expanding the functional repertoire of AKAP1 beyond its established mitochondrial regulatory roles. As the major reader of m6A, YTHDF2-mediated m6A modifications have been shown to participate in many biological processes, especially in the development and progression of various human cancers.15,25 YTHDF2 promotes m6A modification-mediated mRNA degradation of LHPP and NKX3-1 in prostate cancer26 and SOCS2 in HCC.27 YTHDF2 has also been reported to stabilize MYC and VEGFA transcripts in glioblastoma stem cells to promote glioblastoma growth in vivo.28 While most studies have focused on the role of YTHDF2 in m6A methylation-mediated mRNA metabolism, little is known about the regulation of YTHDF2, especially posttranslational modifications. Recently, it was reported that SUMOylation of YTHDF2 at K571 increased its binding activity with m6A-RNAs, which accounts for cancer progression.29 Similarly, it has also been reported that increased O-GlcNAcylation of YTHDF2 during HBV infection promotes protein stability, and thus, the progression of HCC.30 Here, we uncovered another important posttranslational modification of YTHDF2: AKAP1/PKA-mediated phosphorylation, which promotes the stability of YTHDF2, supporting the notion that posttranslational modifications are critical for the stability and activity of YTHDF2. Glycogen synthesis and breakdown are both regulated by several enzymes. Glucose 6 phosphatase (G6PC) is a key enzyme that catalyzes the final step of glycogenolysis. We demonstrated that AKAP1 promotes glycogen accumulation by downregulating G6PC through YTHDF2-mediated m6A-dependent mRNA decay, implying the importance of the AKAP1/YTHDF2/G6PC axis in reprogramming glycogen metabolism in HCC. Notably, we found that G6PC knockdown increases tumorigenesis in both AKAP1(LKO) and AKAP1(fl/fl) mice, and G6PC overexpression suppresses tumorigenesis in both AAV8-AKAP1 and AAV8-GFP controls, suggesting a dominant role of G6PC in hepatocarcinogenesis, which is independent of AKAP1. Given the established link between MASH and HCC development,16 we investigated the regulatory role of the AKAP1/YTHDF2/G6PC axis in MASH progression using our previously reported murine MASH model.17 Our results confirmed that the AKAP1-YTHDF2-G6PC axis also exists in MASH development. Strikingly, AKAP1 deficiency exacerbated MASH progression and depleted hepatic glycogen stores. Recently, hepatic glycogenesis has been shown to antagonize lipogenesis,31 highlighting the complexity of metabolic crosstalk. These findings imply that AKAP1 may protect MASH development by promoting glycogen accumulation, which still requires further confirmation. In addition, given the protective function of AKAP1 in metabolic dysfunction-associated steatotic liver disease (MASLD),17 as well as the close association between MASH and HCC development, we propose that AKAP1 may play a stage-dependent role in the development of MASH and subsequent MASH-associated HCC. Notably, the univariate and multivariate Cox regression analyses demonstrated the significant prognostic value of the dysregulated AKAP1/YTHDF2/G6PC axis in our cohort, a finding corroborated by our clinically validated nomogram model. Strikingly, patients exhibiting high AKAP1/low G6PC expression had the shortest median OS. These results establish AKAP1/YTHDF2/G6PC dysregulation as a novel prognostic biomarker panel with translational potential for HCC risk stratification. Besides, a limitation of our current study is the use of only male mice in the carcinogenesis models. Future studies incorporating both sexes will be important to assess potential sex-dependent differences in the AKAP1/YTHDF2/G6PC axis. MAZ is a transcription factor with a special zinc finger structure that has been reported to be overexpressed in many types of human cancers and is thought to be an oncogene and a therapeutic target for cancer treatment.32 MAZ has been reported to transactivate a variety of oncogenes, such as c-Myc,33,34 Ras gene family,35 and vascular endothelial growth factor (VEGF).36 Our data showed that MAZ also transcriptionally upregulated AKAP1 in HCC cells by directly binding to the promoter region of AKAP1 at sites nt-521 to nt-300. In line with our data, MAZ has been reported to promote the proliferation and invasion of HCC cells via regulating METTL3-mediated ZEB1 expression,19 which further supports a significant role for MAZ in the development of different types of cancers. In addition, in this cohort, MAZ showed a non-significant trend toward being an independent prognostic factor (HR = 1.209, 95% CI: 0.895–1.633, P = 0.217), which is also in line with the study reported by Li et al.19 Previously, we demonstrated that AP-21, a competitive peptide disrupting AKAP1 mitochondrial localization, exhibits therapeutic potential for obesity treatment without observable adverse effects.9 Building on these findings, we investigated the therapeutic efficacy of AP-21 in HCC. Our results revealed that AP-21 effectively suppressed HCC progression through targeted inhibition of the AKAP1/YTHDF2/G6PC regulatory axis. Notably, 4 weeks of intraperitoneal administration of AP-21 (2 mg/kg) did not induce detectable toxicity in major organs (heart, kidney, intestine, spleen, muscle, brain, and lung), as evidenced by histopathological analysis or serum biochemical markers, while demonstrating significant antitumor activity. These findings establish AP-21 as a promising candidate for HCC therapy with dual advantages of targeted action and systemic safety. In summary, our data provide evidence for the contribution of AKAP1 to glycogen accumulation and thus the development of HCC through YTHDF2-mediated G6PC downregulation, emphasizing the importance of the AKAP1/YTHDF2/G6PC axis in glycogen metabolism reprogramming and HCC malignancy and strongly suggesting that treatment targeting AKAP1 via the AP-21 peptide may serve as a potent therapeutic strategy for HCC. Materials and methods Ethics statement All mice were routinely housed in groups of 6 to 8 in a pathogen-free animal center at Fourth Military Medical University. All animal experiments were conducted in accordance with the guidelines of the Animal Care Committee of Fourth Military Medical University (Permission Number: 20210354). In addition, 272 paired early-stage HCC samples were collected at Tangdu Hospital for this study and were approved by the Ethics Committee (Permission Number: 202103-002). Animal studies AKAP1LKO mice with a C57BL6/J background were generated utilizing CRISPR–Cas9 from the Shanghai Model Organisms Center, Inc. For AKAP1 overexpression in hepatocytes in vivo, we performed tail vein injection of AAV8-cloned AKAP1 coding sequence tagged with 3 × Flag (AAV8-TBG-mAKAP1-3 × Flag, Hanbio, Shanghai) twice at six and twenty-two weeks old. For G6PC knockdown or overexpression in vivo, we conducted tail vein injection of lentivirus-cloned G6PC shRNA or coding sequence (Hanbio, Shanghai) at six weeks of age. Chemically and oncogenically driven HCC models were used to establish a hepatocarcinogenesis model. For the chemically induced HCC model, male mice aged two weeks were treated with a single intraperitoneal injection of DEN (25 mg/kg) and subsequent weekly intraperitoneal injections of CCl4 (4 weeks after DEN, 2.5 ml/kg, 1:4 diluted with olive oil).37 For oncogene-induced HCC, HTVi of Sleeping Beauty (pCMV/SB10), pT3-EF1a-HA-myr-Akt, and pT3-EF1aH N90-beta-catenin was conducted in male mice aged 6 weeks according to the instructions previously reported by Chen Xin et al.38 PT3-EF1a plasmid-injected mice were used as the controls. In this study, animals that died during modeling were excluded. Six-week-old male C57BL6/J mice given HTVis of Sleeping Beauty (pCMV/SB10), pT3-EF1a-HA-myr-Akt, and pT3-EF1aH N90-beta-catenin were grouped into the saline group, mutant AP-21 group, and AP-21 group. The peptide drug was intraperitoneally injected twice a week for 4 weeks, starting from the fourth week after the Akt/ β-catenin-induced model construction. Then, liver specimens from mice were acquired to analyze histological changes. Peptide synthesis of AP-21 and mutant AP-21 The peptide named AP-21 (Ac–RRRRRRPLALPGMLALLGWWWFFSRKK–NH2), which was fused with a cell-penetrating peptide (RRRRRR) to the N-terminus and acetylated and amidated at the N- and C-termini, encompassed the 10–30 amino acids in the mitochondrial targeting domain of AKAP1. A mutant AP-21 peptide(Ac–RRRRRRPLALPGMLAAQGWAWFFSRKK–NH2), with no ability to target mitochondria, was synthesized as a control peptide. Fluorescein isothiocyanate (FITC) was coupled to the N-terminus of the two peptides. The peptides were synthesized and purified by Ontores Biotechnologies Co., Ltd. (Hangzhou, China). In vivo imaging Mice treated with AP-21 and mutant AP-21 peptides were conventionally anesthetized with isoflurane at the moribund time or at the end of observation. Subsequently, the MOIS HT in vivo optical imaging system was used to perform in vivo imaging of FITC fluorescence in mice according to the operation instructions. Glycogen content detection assays Glycogen content was assessed using PAS and 2-NBDG staining. Glycogen measurements were performed using a colorimetric assay kit in accordance with the manufacturer’s instructions (Solarbio, Catalog: #BC0345, China). The cell extracts were loaded onto a 96-well plate, and the absorbance was measured at 620 nm. For the PAS staining assay, tissue sections or HCC cells fixed with 4% paraformaldehyde were subjected to routine hydration, followed by the reaction of a solution of periodate and Schiff reagent (Beyotime, Catalog: #C0142M, China) for 8 min. After washing with tap water, the nuclei were stained with hematoxylin. The samples were sealed with neutral resin. The 2-NBDG staining assay was performed on paraformaldehyde-fixed HCC cells by incubating them with 500 μM 2-NBDG (Selleck, USA) in the dark three times, followed by rinsing thrice with PBS. Fluorescence was detected using a confocal microscope (Olympus FV3000, Tokyo, Japan) at an excitation wavelength of 488 nm. Finally, the PAS and 2-NBDG staining results were analyzed using ImageJ software to determine the integrated optical density (IOD) of each group. Targeted metabolomic analysis of metabolites Targeted metabolomic analysis was performed by Beijing Qinglian Biotech Co., Ltd., China, based on a Shimadzu Nexera X2 LC-30AD system equipped with an ACQUITY UPLC BEH Amide column (1.7 μm, 2.1 mm×100 mm, Waters) and a triple quadruple mass spectrometer (QTRAP 5500, AB SCIEX). Specifically, the metabolite mixtures were extracted from cell precipitates with 1 mL precooled methanol/acetonitrile/water (v/v, 2:2:1) under sonication for 1 h, and then transferred to the sampling vial for LC-MS analysis. Quality control (QC) samples were prepared and analyzed with the same procedure as that for the experimental samples in each batch. Metabolites with FC greater than 1.2 and p value less than 0.05 were considered to be statistically significant metabolites. G-1-P and G-6-P content determination A total of 5 × 106 cells with different AKAP1 expression levels were used for the content detection of G-1-P and G-6-P via commercial colorimetric methods according to the instructions of the kits (Uping Biotechnology, Hangzhou, and Beyotime Biotechnology, Shanghai). Mitochondrial respiration assessment Cellular mitochondrial respiration was assessed using a Seahorse XF Analyzer. Briefly, 1 × 104 cells were seeded in XFp plates and incubated overnight. Before the assay, the growth medium was replaced with prewarmed, unbuffered XF assay medium. The OCR was measured in real-time under basal conditions and following the sequential injection of mitochondrial modulators: oligomycin, carbonyl cyanide-4 (trifluoromethoxy) phenylhydrazone (FCCP), and a mixture of rotenone and antimycin A. All OCR measurements were normalized to the total protein content per well. H&E and immunohistochemistry The H&E assay was implemented routinely using hematoxylin and eosin successively to assess the pathological histology of xenograft tumors and lung metastasis nodules. IHC analysis was performed on 5 μm HCC tissue via a commercial IHC Kit (ZSGB Biotechnologies, Beijing, China). The primary antibodies used are listed in Supplementary Table 1. Tissue collection and cell line culture A total of 272 paired early-stage HCC samples were collected at Tangdu Hospital for this study, which was approved by the Ethics Committee. The human HCC cell lines SNU-739, SNU-398, SNU-368, HLE, HLF, HCC-LM3, and Huh7, and hepatocyte cell line THLE-2 were all obtained from COBIOER (Nanjing, China) and were routinely cultured at 37 ˚C and 5% CO2 in Dulbecco’s modified Eagle’s medium (DMEM) or Roswell Park Memorial Institute (RPMI)-1640 medium supplemented with 10% fetal bovine serum and 1% penicillin and streptomycin. Quantitative real-time PCR (qRT-PCR) Total RNA from HCC cells was extracted using TRIzol reagent (Invitrogen, United States) and reverse transcribed. Primers were obtained from Tsingke Biotech (Beijing). The qRT-PCR assay was conducted routinely, and the 2−ΔΔCT method was used to analyze relative mRNA levels. Supplementary Table 2 lists all primers used in this work. Knockdown and forced expression of target genes shRNA sequences targeting AKAP1, YTHDF2, and MAZ were synthesized and inserted into the pLKO.1 vector by Tsingke Biotechnology Ltd. To generate overexpression plasmids, the coding sequences of AKAP1, YTHDF2, and MAZ were cloned and inserted into the pLVX-Puro vector. To establish stable cell lines, lentivirus-infected HCC cells were selected with 2 μg/mL of puromycin for 2 weeks. All relevant sequences used are listed in Supplementary Table 3. Cell proliferation, migration, and invasion assays Cell viability was analyzed by MTS assay (Promega, Madison, USA) and colony formation assays following the instructions. Briefly, for the cell viability assay, 2 × 103 HCC cells with different target gene expression levels were seeded into 96-well plates and detected daily by measuring the absorbance at 490 nm. For the colony formation assay, 1 × 103 HCC cells were seeded into 6-well plates. After culturing for 2 weeks, the plates were stained with crystal violet, and the colonies were counted. A wound-healing assay was performed to determine the migration ability of HCC cells as described previously in ref.39 Briefly, cells in 6-well plates were pretreated with mitomycin C (Catalog: HY-13316, MCE) at a dose of 0.5 μM for 2 h to inhibit HCC cell proliferation. Wound healing was photographed at the time points of 0 h and 48 h after scratching. The migration rate was then calculated. For assessment of HCC cell invasion, HCC cells were seeded into the 24-well transwell chambers with 8.0 μm pore size polycarbonate membranes (Corning Costar, China) pretreated with Matrigel. Cells with serum-starvation for 12 h were resuspended in serum-free medium. A total of 5 × 104 cells was added to the upper chamber, while 600 μL of complete medium containing 10% FBS was added to the lower chamber as a chemoattractant. After culturing for 48 h, the transwell chambers were stained with crystal violet. For each well, five random fields were captured under a light microscope (Olympus, Tokyo, Japan) at 200× magnification, and the number of stained cells was counted. Each experiment was performed in triplicate. Western blot Western blot assays were conducted as previously reported.9 Briefly, approximately 60 μg of total protein was used to detect changes in the expression levels of the molecules involved. β-Actin was selected as the housekeeping gene. Three independent experiments were performed. Supplementary Table 1 lists the primary antibodies used. IP and co-IP Cell lysates acquired using IP buffer were cleared by incubation with 25 μL of Protein A/G magnetic beads (Thermo Fisher Scientific) and a suitable amount of AKAP1 antibody at 4 °C overnight. The pellets were collected and washed with lysis/rinsing buffer twice the next day. Next, the collected protein samples were washed using 100 μL of 1 × SDS-PAGE buffer. Protein gels were then silver stained by a commercial kit (Beyotime Biotechnology, Shanghai). Additionally, an LC-MS assay was conducted to analyze the interacting proteins of AKAP1 in HCC cells using the stained protein gels, followed by bioinformatics analysis. The method for co-IP is roughly the same as that of IP. A suitable amount of the indicated antibody was used. Supplementary Table 1 lists the antibodies used in the co-IP assay. Immunofluorescence For immunofluorescence analyses of YTHDF2, HCC cells in dishes were incubated with primary antibodies at 4 °C overnight. After rinsing twice in PBS, the cells were then incubated with the corresponding secondary antibodies. Afterwards, the nucleus was stained with 4’, 6-diamidino-2-phenylindole (DAPI, Sigma-Aldrich), and the photomicrographs were acquired on an Olympus confocal microscope. FRET assay The acceptor photo bleach method was used for the FRET assay. Briefly, pEGFP-N1-AKAP1-3 × Flag and pDsRed2-N1-YTHDF2 vectors as the donor and acceptor, respectively, were constructed and transfected into HCC cells. Then, we acquired the donor and acceptor images at corresponding excitation and emission wavelengths under an oil immersion condition. FRET image processing was conducted with the Life Science Analysis module (Olympus FV3000, Japan). In vitro kinase assay In E. coli, GST-tagged wild type (WT), S289A, S359A, or S289A/S359A YTHDF2 were purified using glutathione-Sepharose beads. Recombinant active PRKACA (Cat: #Ag21478) and Dephosphorylated MBP-GST (Cat: #Ag0713) were obtained from Proteintech. Recombinant proteins were mixed in the buffer supplemented with 50 mM ATP (Selleck) and 5 μCi[γ-32P] ATP (RR Scientific) for half an hour at 32 °C with or without H89 (5 μM). The reacted samples were determined by SDS-PAGE electrophoresis and subsequently autoradiographed. In addition, Coomassie Brilliant Blue R-250 was used to stain the PVDF membrane. CHX chase assay HCC cells with distinct AKAP1 expression statuses or treated with the PKA inhibitor H89 (10 μM) or activator forskolin (FSK, 50 μM) were incubated with 50 μM cycloheximide for the indicated time points. Total cellular proteins were extracted and subjected to western blot analysis. Protein quantification was performed using ImageJ software, with β-actin serving as the loading control. The remaining YTHDF2 protein levels at each time point were normalized to the initial YTHDF2/β-actin ratio of the corresponding experimental groups. Streptavidin aptamer-based capture assay For in vivo streptavidin precipitation, WT or mutant streptavidin-tagged G6PC was transfected into HCC cells seeded on T75 bottles for 48 h. The harvested cells were treated with the inhibitor of proteinase and RNase inhibitors (Roche). The supernatant was added to streptavidin beads (Invitrogen) for 15 min at 4°C. Ten percent of the supernatant was used as the input, and the remaining sample was blended with the beads at 4 °C overnight. Finally, the precipitation products were denatured using 20 μL of 5 × SDS-PAGE buffer before detection by western blot. RIP and MeRIP-qPCR assays RIP and MeRIP assays were performed using an EZ-Magna RIP™ RNA-binding protein IP kit (Catalog: #17-700, Merck, Germany). The isolated mRNA samples from HCC cells were treated with 20 min of ultrasonic fragmentation after being added to an appropriate volume of IP buffer and 2 μL of RNase inhibitor, followed by incubation with 5 μg of the specified antibodies (YTHDF2, m6A, and IgG) at 4 °C for 4 h. Then, the complexes were eluted after Protein A/G magnetic bead enrichment. Finally, reverse transcription and qPCR validation were conducted for equal volumes of RNA samples. Cross-linked immunoprecipitation (CLIP) assay In brief, 0.15 J/cm2 of ultraviolet light was used to crosslink RNA and protein complexes in vivo for 10 min, followed by harvesting cells and adding DL-dithiothreitol, a protease inhibitor. The lysates were centrifuged at 4 °C and 13,000×g for 15 min, and the supernatants were incubated with RNase T1 for 15 min in a water bath at 22 °C and immediately in an ice bath for 5 min. Five micrograms of YTHDF2 antibody were added to the supernatant mixture at 4 °C overnight. Then, the mRNA samples were extracted from complexes after Protein A/G magnetic bead enrichment and DNase I and proteinase K digestion. Finally, reverse transcription and qPCR validation were successively performed. MRNA decay assay HCC cells were incubated with actinomycin D (Selleck Chemicals, Houston, USA) at a dose of 5 μM for various times. Total RNA acquired from each sample was determined for qRT–PCR analysis. The remaining mRNA levels were normalized according to previous studies.40 The mRNA half-life was also calculated. Dual-luciferase reporter assay Different lengths of the AKAP1 promoter and site-mutant AKAP1 promoter were inserted into the pGL3-basic reporter plasmid via SacI and XhoI restriction endonucleases. HCC cells with different AKAP1 expression levels were cotransfected with the reporter plasmids and pRL-SV40 plasmid at a ratio of 10:1. Then, we detected the activity of firefly and Renilla luciferase. Besides, for YTHDF2 regulating G6PC mRNA stability analysis, the pMIR-GLO reporter and pRL-TK plasmids were cotransfected into HCC cells with the indicated treatment. The luciferase activities were assessed using a dual-luciferase reporter assay kit (Promega). Finally, the relative ratio of luciferase activity was quantified. Chromatin immunoprecipitation (ChIP)-qPCR assay The ChIP-qPCR assay was conducted using a SimpleChIP® ChIP Kit (Catalog: #56383, CST Signaling, USA). Briefly, 1% formaldehyde cross-linked HCC cells were homogenized in lysis buffer and sonicated. The collected supernatant was subjected to immunoprecipitation with an equivalent mass of anti-MAZ antibody (1:100) or rabbit IgG antibody. Purified and eluted DNA samples were amplified, followed by qPCR assessment. The sequences of the primers used are listed in Supplementary Table 2. Statistical analysis All experiments were repeated thrice. Student’s t-test or analysis of variance (ANOVA) was used to analyze quantitative data. Tukey’s honest significant difference test was used to determine whether there were significant differences among multiple groups. Pearson’s correlation analysis was used to evaluate the correlation between the two variables. SPSS 18.0 and GraphPad 7.0 software were employed, and P < 0.05 was defined as significant. Supplementary information Sigtrans_Supplementary_Materials Uncropped WB band Sigtrans_Supplementary_Materials Uncropped WB band