Article Text

Original research
High-fat diet promotes liver tumorigenesis via palmitoylation and activation of AKT
Free
  1. Lang Bu1,2,
  2. Zhengkun Zhang2,
  3. Jianwen Chen2,
  4. Yizeng Fan3,
  5. Jinhe Guo2,
  6. Yaqing Su2,
  7. Huan Wang2,
  8. Xiaomei Zhang2,
  9. Xueji Wu2,
  10. Qiwei Jiang2,
  11. Bing Gao2,
  12. Lei Wang2,
  13. Kunpeng Hu4,
  14. Xiang Zhang5,
  15. Wei Xie2,
  16. Wenyi Wei3,
  17. Ming Kuang1,
  18. Jianping Guo2
  1. 1 Center of Hepato-Pancreate-Biliary Surgery, the First Affiliated Hospital, Sun Yat-sen University, Guangzhou, Guangdong, China
  2. 2 Institute of Precision Medicine, the First Affiliated Hospital, Sun Yat-sen University, Guangzhou, Guangdong, China
  3. 3 Department of Pathology, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts, USA
  4. 4 Division of General Surgery, the Third Affiliated Hospital of Sun Yat-Sen University, Guangzhou, Guangdong, China
  5. 5 State Key Laboratory of Digestive Disease, Institute of Digestive Disease and the Department of Medicine and Therapeutics, Li Ka Shing Institute of Health Sciences, the Chinese University of Hong Kong, Hong Kong, China
  1. Correspondence to Professor Jianping Guo; guojp6{at}mail.sysu.edu.cn; Professor Ming Kuang; kuangm{at}mail.sysu.edu.cn

Abstract

Objective Whether and how the PI3K-AKT pathway, a central node of metabolic homeostasis, is responsible for high-fat-induced non-alcoholic steatohepatitis (NASH) and hepatocellular carcinoma (HCC) remain a mystery. Characterisation of AKT regulation in this setting will provide new strategies to combat HCC.

Design Metabolite library screening disclosed that palmitic acid (PA) could activate AKT. In vivo and in vitro palmitoylation assay were employed to detect AKT palmitoylation. Diverse cell and mouse models, including generation of AKT1C77S and AKT1C224S knock-in cells, Zdhhc17 and Zdhhc24 knockout mice and Akt1C224S knock-in mice were employed. Human liver tissues from patients with NASH and HCC, hydrodynamic transfection mouse model, high-fat/high-cholesterol diet (HFHCD)-induced NASH/HCC mouse model and high-fat and methionine/choline-deficient diet (HFMCD)-induced NASH mouse model were also further explored for our mechanism studies.

Results By screening a metabolite library, PA has been defined to activate AKT by promoting its palmitoyl modification, an essential step for growth factor-induced AKT activation. Biologically, a high-fat diet could promote AKT kinase activity, thereby promoting NASH and liver cancer. Mechanistically, palmitoyl binding anchors AKT to the cell membrane in a PIP3-independent manner, in part by preventing AKT from assembling into an inactive polymer. The palmitoyltransferases ZDHHC17/24 were characterised to palmitoylate AKT to exert oncogenic effects. Interestingly, the anti-obesity drug orlistat or specific penetrating peptides can effectively attenuate AKT palmitoylation and activation by restricting PA synthesis or repressing AKT modification, respectively, thereby antagonising liver tumorigenesis.

Conclusions Our findings elucidate a novel fine-tuned regulation of AKT by PA-ZDHHC17/24-mediated palmitoylation, and highlight tumour therapeutic strategies by taking PA-restricted diets, limiting PA synthesis, or directly targeting AKT palmitoylation.

  • CELL SIGNALLING
  • DRUG DEVELOPMENT
  • FATTY ACID SUPPLEMENTATION
  • HEPATOBILIARY CANCER
  • MOLECULAR ONCOLOGY

Data availability statement

Data are available upon reasonable request.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

WHAT IS ALREADY KNOWN ON THIS TOPIC

  • AKT plays a leading role in physiological metabolic homeostasis and is closely related to liver metabolic disorders.

  • High-fat diets play a key role in liver metabolic disorders, the detailed mechanisms of different fatty acids such as palmitic acid (PA) in these diseases remain unclear.

  • AKT-mTOR activation induces lipogenesis and hepatocellular carcinoma (HCC) progression.

WHAT THIS STUDY ADDS

  • AKT undergoes palmitoylation modification, which would contribute to high-fat diets (PA uptake)-induced AKT activation in cells, mice and human liver tissues (NASH and HCC).

  • The palmitoyltransferases ZDHHC17/24 palmitoylate AKT to exert pathological effects in NASH formation and liver tumorigenesis.

  • The anti-obesity drug orlistat or specific penetrating peptides can effectively attenuate AKT palmitoylation and activation, thereby antagonising liver tumorigenesis.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

  • This study encloses a fine-tuned regulation of AKT activity by PA-ZDHHC-mediated palmitoylation, providing the strategies for HCC therapy by limiting PA uptake, restricting PA synthesis, and targeting ZDHHC17/24.

One sentence statement

Adoption of a low-fat (palmitic acid (PA)-restricted) diet, restriction of PA synthesis, or direct targeting of ZDHHC17/24 would provide effective strategies for hepatocellular carcinoma (HCC) therapy by attenuating AKT palmitoylation and activation.

Introduction

Diet not only provides nutrition and building blocks for the uncontrolled growth of tumour cells, but also promotes signal transduction to affect various aspects of tumours, so diet-changing therapies have recently been realised, including calorie restriction and low-calorie-fat diet.1 It is worth noting that high-fat diets (HFDs) are primarily associated with metabolic diseases, including obesity and cancer, through diverse and unclear mechanisms.2 3 Among them, palmitic oil is one of the most source of dietary oil and major cause of obesity, which has been considered as a metabolic disease and preferred to inducing tumours.4 Of note, PA as a major component of palmitic and animal fat oil has been tightly linked to oral tumour metastasis with its metabolic function to shape cell plasma membrane.2 Meanwhile, an active product of PA, palmitoyl-CoA, has been shown to induce palmitoyl modification in a variety of proteins, such as MC1R in response to UV-induced melanoma,5 STAT3 in response to IL6-induced Th17 differentiation and antiviral infection,6 and N-RAS for colon cancers.7 However, whether and how PA accumulation from synthesis or food uptake is involved in tumorigenesis, especially in organs involved in metabolic homeostasis, such as liver, is unclear.

HFD has been considered as an unabated risk factor of non-alcoholic steatohepatitis (NASH), an advanced form of non-alcoholic fatty liver disease (NAFLD), now has also been termed as metabolic dysfunction-associated steatohepatitis. NASH involves lobular inflammation, fibrosis and hepatocyte ballooning, which can progress to HCC.8 Although multiple potential mechanisms such as autophagy and immune responses are associated with NASH,9 10 whether and how the PI3K-AKT pathway, a central node of metabolic homeostasis and playing potent role in lipogenesis and HCC progression,11 is responsible for high-fat-induced NASH and HCC remain a mystery.

As a central signal in response to extracellular stimuli, the PI3K-AKT pathway plays pivotal roles in cell growth, survival and metabolic homeostasis physiologically, as well as in diabetes and tumorigenesis pathologically.12 13 Therefore, the regulation of AKT, especially at the post-translational level, has been extensively established.14 15 Notably, the binding of AKT PH domain to PI3K-generated PIP3 is thought to be a critical step in the full activation of AKT by PDK1-mediated T308 and mTORC2-mediated S473 phosphorylation.16 17 For example, ubiquitination, acetylation and methylation all contribute to the binding of AKT to PIP3 for its further activation.18–20 Thus, genetic alterations of PIP3-associated enzymes, including its kinase PIK3CA, phosphatase PTEN, and their upstream regulatory proteins, such as RAS and EGFR, have been observed in around 50% of diverse types of cancer and largely contribute to PIP3 accumulation and subsequent AKT activation.13 To this end, specific inhibitors against the PIK3CA-H1047R mutant have been approved for breast cancer intervention.21 However, besides canonical PI3K-PIP3 dependent AKT activation, whether and how AKT activation in a PIP3-independent manner remain unclear. In this study, we report that accumulating PA from HFDs or aberrant synthesis promotes ZDHHC17/24-catalysing palmitoylation and activation of AKT in a PIP3-independent manner, leading to NASH and liver tumorigenesis, and highlight the strategy to target PA-ZDHHC-AKT axis for combating HCC.

Results

HFD activates AKT and promotes NASH and liver tumorigenesis

To systematically explore the regulatory role of metabolites on AKT kinase, we established an in vitro AKT activity screening approach based on and optimised from previous reports.22 23 Briefly, using AKT recognised peptide (PRPRSCTWPDPRPEF) as a substrate, phosphorylation of that can be recognised by a specific FHA1 domain, leading to the assembly of a functional Renilla luciferase (figure 1A). This reporter system was initially validated with an insulin-based positive control (online supplemental figure S1A), and then challenged with a metabolite library (figure 1B). Among them, several metabolites were identified and validated to positively activate AKT, including equol, menadione, PA, propylparaben, xanthosine dihydrate, isohomovanillic acid, and 5'-adenylic acid (figure 1B–C and online supplemental figure S1B). Notably, PA was shown on be the top-ranked metabolites for activating AKT (figure 1C and online supplemental figure S1B), which was effectively antagonised by the palmitoyl-transferase inhibitor 2-bromopalmitate (2BP) (online supplemental figure S1C). Consistent with these findings, PA promoted, whereas 2BP attenuated AKT kinase activity (online supplemental figure S1D-G), its interaction with PDK1 and PIP3 (online supplemental figure S1H-I), and its plasma membrane translocation (online supplemental figure S1J). RNA-sequencing results also demonstrated that PA could strongly enhance the activity of PI3K-AKT pathway (online supplemental figure S1K-N), thus, collectively suggesting PA as a potent activator of AKT.

Supplemental material

Figure 1

Palmitic acid (PA) activates AKT and promotes liver tumorigenesis. (A) Mode diagram of AKT activity detection system. (B) HEK293 cells were transfected with AKT-Sub-TK-FHA1 plasmid, serum-starved for 10 hours, then treated with more than 500 metabolites for 8 hours and subjected to luciferase assay. (C) HEK293 cells were transfected with AKT-Sub-TK-FHA1, serum-starved for 10 hours, then treated with screened metabolites for 8 hours and subjected to luciferase assay. Data are presented as the mean±SD of three independent experiments. **P<0.01, ***p<0.001, ****p<0.0001 (Student’s t test). (D) Different hepatocarcinogenesis mouse models and the degree of tumorigenic capacity. (E) Representative image of livers harvested from mice which were hydrodynamic transfected with N‐RasV12, SB transposase, and AKT1 or myr-mAkt1 plasmids, then intraperitoneally injected or intragastric administration with or without PA (10 mg/kg) every day and fed on a normal diet (ND), or a high-fat diet (HFD) for 1 or 2 months. (F) The percentage of liver/body weight was shown. Data are presented as the mean±SD (n=6 mice per group). ***P<0.001, ****p<0.0001 (Student’s t test). (G) Liver sections were stained with H&E and analysed by immunohistochemistry (IHC) for pS6 (pS240/244), AFP, Ki67, and Cleaved-Caspase3 levels. (H) Liver tissues from above group were subjected to immunoblotting (IB) analysis. (I) Liver sections of C57BL/6 mice with HFHC diet for different months were stained with H&E, Masson, and analysed by IHC for pS6, Pan-CK, and CPS1 levels. (J) Liver tissues from above groups were subjected to IB analysis.

Despite recent reports that PA promotes metastasis of oral cancer in mice,2 we sought to determine whether PA promotes tumorigenesis through activation of AKT. To this end, a well-established tail-vein hydrodynamic transfection liver tumour mouse model (transfection with hyperactive mouse Akt1 (myr-mAkt1), mutated N-Ras (N-RasV12) and Sleeping Beauty transposon) was adopted and further optimised. Consistent with previous findings, cocktail constructs containing myr-mAkt1, but not wild-type (WT) AKT1 could promote liver tumorigenesis (figure 1D).11 24 Interestingly, in our optimised model, HFD or administration of PA (I.G. or I.P.), both significantly promoted WT-AKT1-induced NASH (figure 1E–F), coupled with increased AKT activity and hepatocyte proliferation (figure 1G–H). Consistent with these findings, we obtained liver tissues from a reported HFLC-feeding induced NAFLD/NASH/HCC murine model,25 and observed that AKT activity was increasingly elevated on progression of NASH to HCC (figure 1I–J). These results suggest that HFD or PA administration could activate AKT, thereby potentially promoting NASH formation and liver tumorigenesis.

AKT undergoes palmitoyl-modification

Although PA has been shown to factilitate PI3K-AKT signalling,26 we observed that 2BP, which has the function of antagonising palmitoyltransferase, readily inhibited PA-induced AKT activation (online supplemental figure S1C-J). Therefore, we hypothesised that AKT undergoes palmitoyl-modification. To this end, two palmitoyl-conjugation assays were employed, the click chemistry reaction to biotin-azide assay and acyl-biotin exchange (ABE) palmitoylation assay. Both ectopically expressed and endogenous AKT1 were proved to undergo palmitoyl-modification (figure 2A–B and online supplemental figure S2A-D). Similar results were observed in AKT2 and AKT3 subtypes (online supplemental figure S2E). Importantly, AKT1 palmitoylation was significantly induced by PA stimulation, which was antagonised by 2BP (figure 2C and online supplemental figure S2F). In contrast, challenges with insulin, PI3K inhibitors, or enforced AKT membrane translocation (myr-AKT1) had mild effects on AKT palmitoyl-conjugation (online supplemental figure S2G-J).

Figure 2

AKT1 is palmitoylated at C77/C224 to enhance its kinase activity and oncogenic function. (A) Huh7 cells were labelled with dimethyl sulfoxide (DMSO) or alk-C16 (50 µM) for 4 hours, followed by click chemistry reaction to biotin-azide assay. (B) Whole cell lysates of Huh7 cells were incubated with α-AKT1 antibody and processed for acyl-biotin exchange (ABE) palmitoylation assay. (C) HepG2 cells were treated with palmitic acid (PA) (100 µM) with or without 2BP (50 µM) for 4 hours, incubated with α-AKT1 antibody and processed for ABE palmitoylation assay. (D) Mouse liver tissues obtained from figure 1E were subjected for ABE palmitoylation assay. (E) Four non-alcoholic steatohepatitis (NASH)-induced hepatocellular carcinoma (HCC) with their adjacent normal tissues were subjected for ABE palmitoylation assay. N, indicates adjacent normal tissue; T, indicates liver tumour; Mock, indicates no hydroxylamine (HAM) treatment. (F) Mouse liver tissues obtained from figure 1I were subjected for ABE palmitoylation assay. (G) Human liver sections with no steatosis, NASH and HCC were stained with H&E, Masson, and analysed by IHC for pS6 levels. (H) Liver tissues from above groups were subjected to ABE palmitoylation assay, and the relative palmitoylation of AKT was quantified with total AKT1 and normalised with no steatosis tissues. Data are presented as the mean±SD (n=7 for no steatosis tissues, n=8 for NASH, n=8 for HCC). **P<0.01, ***p<0.001, ****p<0.0001 (Student’s t test). (I) AKT1WT , AKT1C77S and AKT1C224S HEK293 cells were treated with or without PA for 4 hours, followed by ABE palmitoylation assay. (J) AKT1WT , AKT1C77S and AKT1C224S HEK293 cells were treated with or without PA and alk-C16 for 4 hours, subjected to click chemistry reaction to biotin-azide assay. (K) DLD1-AKT1/2-/--EV, DLD1-AKT1/2-/--Flag-AKT1-WT, C77S, C224S, and C77/224S cells were subjected to colony formation (I) and soft agar assays (J). Data are presented as the mean±SD of three independent experiments. **P<0.01, ***p<0.001, ****p<0.0001 (Student’s t test). (L) Representative image of livers harvested from hydrodynamic transfection of EV, myr-mAkt1, or myr-mAkt1-C77/224S and N‐RasV12, SB plasmids groups, then fed on a normal diet (ND) for 1 month. The percentage of liver/body weight was shown. Data are presented as the mean±SD (n=6 mice per group). **P<0.01, ***p<0.001 (Student’s t test). (M) Liver tissues from above group were subjected to IB analysis. (N) Representative image of livers harvested from mice which were hydrodynamic transfected with N‐RasV12, SB transposase, and Flag-AKT1-WT, C77S, C224S, or C77/224S plasmids, then fed on ND or a high fat diet (HFD) for 2 months. The percentage of liver/body weight was shown. Data are presented as the mean±SD (n=8 mice per group). ***P<0.001, ****p<0.0001 (Student’s t test). (O) Liver tissues from above group were subjected to IB analysis. (P) Representative images of 4-week-old Akt1WT and Akt1C224S mice. Representative liver images and the percentage of liver/body weight were shown. Data are presented as the mean±SD (n=5 mice per group). ***P<0.001 (Student’s t test). (Q) Representative images and body weight of male mice fed with HFD for 11 weeks. Data are presented as the mean±SD (n=6 mice per group). ****p<0.0001 (two-way analysis of variance analysis). (R) Weights of liver, visceral adipose tissue (VAT), and brown adipose tissue (BAT) were shown as percentage of body weight. Data are presented as the mean±SD (n=6 mice per group). **P<0.01, ***p<0.001 (Student’s t test). (S) Liver tissues, VAT and BAT in (R) were stained with H&E, Oil Red, or analysed by IHC for pS6 (pS240/S244) levels. (T) Liver tissues in (R) were subjected to IB analysis. (U) Akt1WT and Akt1C224S mice fed with ND or HFD, then subjected to ABE palmitoylation assay.

To investigate AKT palmitoylation in vivo, we harvested liver tissues from PA or HFD-feeding mice as we mentioned above (figure 1D), and observed that PA and HFD markedly enhanced AKT palmitoylation (figure 2D). More importantly, compared with adjacent normal liver tissues, NASH-induced HCC displayed higher AKT palmitoylation levels (figure 2E). Consisnt with the association of AKT activity with NASH/HCC progression (figure 1I), we observed a progressive increase of AKT palmitoylation following HFLC-feeding time (figure 2F). Consistently, NASH tissue specimens also displayed increased AKT palmitoylation, accompanied with elevated AKT activity, by contrast, HCC tissues exhibited more stronger AKT palmitoylation and activation (figure 2G–H and online supplemental figure S2K-L). These observations together suggest that AKT undergoes palmitoylation modification both in NASH and HCC.

To point out potential palmitoyl-residues, the amino acid sequence of AKT was analysed (online supplemental figure S3A). Of note, mutation of C77 or C224, but not other cystines, including previously reported C344,27 significantly abolished AKT phosphorylation (online supplemental figure S3B), and reduced their binding to PDK1 (online supplemental figure S3C). Interestingly, C77S or/and C224S mutations clearly abolished baseline or PA-induced AKT palmitoylation (online supplemental figure S3D-H). To further investigate the physiological role of palmitoylation on AKT, we generated C77S and C224S-AKT1 knock-in HEK293 cells (termed as AKT1C77S and AKT1C224S ) (online supplemental figure S3I), which showed readily reduced PA-induced AKT palmitoylation (figure 2I–J), indicating that AKT undergoes palmitoyl-modification at both C77 and C224 residues.

Palmitoylation deficiency represses AKT kinase activity and oncogenic functions

Next, we examined whether PA activates AKT by enhancing its palmitoylation at C77 and C224. The results showed that C77S or/and C224S mutant significantly inhibited PA-induced or insulin-induced AKT activity (online supplemental figure S4A), phosphorylation and palmitoylation in HEK293 (online supplemental figure S4B-D) and AKT1/2 double knockout DLD1 (DLD1-AKT1/2-/- ) cells (online supplemental figure S4E-G). In support of this observation, ectopic expression of C77S or/and C224S-AKT1 mutants dramatically reduced colony formation and anchorage growth (figure 2K and online supplemental figure S4H-I), as well as tumour growth in vivo (online supplemental figure S4J-L), coupled with decreased AKT phosphorylation (online supplemental figure S4M) in DLD1-AKT1/2-/- cells. Similarly, HEK293-AKT1C77S and AKT1C224S cells also exhibited decreased AKT phosphorylation on insulin or PA stimulation (online supplemental figure S4N-R), accompanied with attenuated its interaction with PDK1 (online supplemental figure S4S-T), decreased cell colony formation (online supplemental figure S4U). Interestingly, RNA-sequencing results showed that PA could strongly enhance insulin signalling, NAFLD, and activate PI3K-AKT signalling in HEK293-AKT1WT cells compared with AKT1C224S cells (online supplemental figure S4V-W), indicating the potential link of PA to AKT activation and fatty liver disease. Notably, the C77S and/or C224S mutants reduced AKT kinase activity measured by in vitro kinase assays (online supplemental figure S4X), collectively suggesting that PA partially activates AKT by promoting AKT palmitoylation at C77 and C224.

Due to the critical role of fat deposition in NASH and its-induced liver tumorigenesis, we sought to investigate whether palmitoylation deficiency defects AKT-induced tumorigenesis in vivo. To this end, we employed the canonical hydrodynamic transfection mouse model, in comparation with intact myr-mAkt1, palmitoylation deficient mutations (myr-mAkt1-C77/224S) could significantly attenuate HCC formation (figure 2L), coupled with decreased Akt phosphorylation and activation (figure 2M and online supplemental figure S5A). Furthermore, to determine whether HFD-induced NASH and HCC through palmitoylation of AKT, we used the optimal murine model in combination with HFD-feeding (figure 1D), and observed that injected AKT1 construct bearing different mutants (C77S and/or C224S) could obsviously alleviate HFD-induced NASH (figure 2N), accompanied with decreased AKT phosphorylation, palmitoylation and downstream activation (figure 2O and online supplemental figure S5B-C), indicating that palmitoylation of AKT at C77 and C224 is critical for NASH and HCC.

To further explore the physiological role of AKT palmitoylation in vivo, we generated Akt1-C224S knock-in mice (Akt1C224S ) (online supplemental figure S6A-B). Similar to the phenotype of Akt1-KO mice (Akt1-/- ),28 29 Akt1C224S mice were born with expected Mendelian ratio (online supplemental figure S6C), and displayed lower growth rate and decreased body size/weight, organ size/weight and Akt1 activation compared with intact mice (figure 2P and online supplemental figure S6D-K). A similar phenomenon was also observed in mice fed a HFD (figure 2Q–T), accompanied with promoting Akt1 palmitoylation (figure 2U). Notably, insulin or PA stimulation had a slight effect on pT308-AKT in liver tissues or MEFs from Akt1C224S mice compared with the counterpart mice (online supplemental figure S6L-O). These findings suggest that defective palmitoylation impairs AKT kinase activity and physiological function.

Palmitoylation promotes AKT membrane localisation in a PI3K-PIP3 independent manner

It is previously reported that palmitoyl-conjugation prefers to anchoring target proteins, such as MC1R, STAT3, N-RAS and NOD1/2, to membrane.5 30–32 Therefore, we tended to define whether palmitoyl-conjugation could promote AKT membrane anchoring to further enhance its interaction with PDK1 (online supplemental figure S1H-J), which mainly occurs in the plasma membrane in a PIP3-dependent manner.16 Expectedly, similar to insulin, PA also apparently enhanced AKT membrane localisation (figure 3A and online supplemental figure S7A), whereas 2BP administration (online supplemental figure S1J) or C77S and/or C224S mutant (figure 3A and online supplemental figure S7A) markedly impaired insulin or PA-induced AKT membrane localisation. In addition, cell fractionations also demonstrated that HEK293-AKT1C77S and AKT1C224S , as well as MEFs-Akt1C224S exhibited a marked decrease in AKT membrane localisation induced by insulin or PA (figure 3B and online supplemental figure S7B-D), concurrently decreased AKT binding with PIP3 (figure 3C–D and online supplemental figure S7E-F). These findings suggest that the canonical function of insulin-activated AKT depends at least in part on AKT palmitoyl-conjugation.

Figure 3

Palmitoylation conjugation enhances AKT membrane translocation independent of PIP3. (A) AKT1WT , AKT1C77S , and AKT1C224S HEK293 cells were serum-starved for 18 hours before treated with or without insulin (1 µg/mL) for 10 min, palmitic acid (PA) (100 µM) for 4 hours, or wortmannin (1 µM) for 2 hours, then subjected to endogenous IF analysis. Scale bars, 10 µm. (B) AKT1WT , AKT1C77S and AKT1C224S HEK293 cells were serum-starved for 18 hours before treated with or without PA for 4 hours, then subjected to cell membrane separation assay. (C and D) AKT1WT , AKT1C77S and AKT1C224S HEK293 cells were serum-starved for 18 hours before treated with or without insulin for 10 min (C) or PA for 4 hours (D), followed by PIP3 pull-down assay. (E) HEK293 cells were transfected with Flag-AKT1-PH or Flag-AKT1-PH-C77S, serum-starved for 18 hours before treated with or without insulin, then subjected to PIP3 pull-down assay. (F) HEK293 cells were transfected with Flag-AKT1-PH or Flag-AKT1-PH-C77S, serum-starved for 18 hours before treated with or without insulin or PA, then subjected to IF analysis. Scale bars, 10 µm. (G) HEK293 cells were serum-starved for 18 hours before treated with or without insulin for 10 min, PA for 4 hours, or wortmannin for 2 hours, followed by endogenous IF analysis. Scale bars, 10 µm.

To elucidate how palmitoyl-conjugation affects AKT membrane translocation, we observed that depletion of PIP3 by PI3K inhibitors attenuated insulin but not PA-induced AKT membrane localisation (figure 3A, online supplemental figure S7D,G). On the other hand, PI3K inhibitors readily disrupted PA-induced AKT interaction with PIP3 and PDK1 (online supplemental figure S7H-I) but not their membrane location (online supplemental figure S7J-K). Due to the important role of AKT-PH domain in mediating its interaction with PIP3 and further membrane location,13 both insulin-induced and PA-induced AKT-PH domain interaction with PIP3 and membrane translocation could be markedly abolished by C77S mutation (figure 3E–F, and online supplemental figure S7L). To directly detect whether PA-induced AKT membrane location through binding with PIP3, specific staining for PIP3 was performed, and only partial colocalisation of PIP3 with PA-induced AKT membrane localisation was observed (figure 3G and online supplemental figure S7M). Insulin-induced colocalisation of AKT and PIP3 could be attenuated by PI3K inhibitors, while PI3K inhibitor did not impair PA-induced AKT membrane localisation (figure 3G and online supplemental figure S7M). Consistent with the property of palmitoylation to anchor proteins directly at the membrane, our results indicate that palmitoylation enhances AKT membrane localisation in a PI3K-PIP3 independent manner.

Palmitoylation deficiency facilitates AKT inactive polymer formation

Although palmitoylation at C77 of AKT-PH domain has been shown to promote AKT membrane anchoring, the function of palmitoylation at C224 remain unclear. As protein dimerisation/polymerisation plays a key role in its biological function, this has been shown to be regulated by palmitoyl-modification, including STAT3.30 Therefore, we sought to detect whether palmitoylation at C77 or C224 affects AKT dimer/polymerisation. To this end, split synthetic renilla luciferase, GST pull-down, fluorescent resonance energy transfer (FRET), and disuccinimidyl suberate (DSS) crosslinking assays shown that AKT formed dimer/polymer both in vitro and in cells and C224S mutant enhanced AKT1 dimerisation (figure 4A–F and online supplemental figure S8A-B). Interestingly, PA stimulation reduced AKT dimer/polymer, which was antagonised by 2BP (figure 4G and online supplemental figure S8C-E). Meanwhile, the C224S but not the C77S mutant enhanced AKT dimerisation compared with WT-AKT (figure 4H and online supplemental figure S8F-G). In support of this finding, C224S (but not C77S) strongly enhanced AKT aggregate formation detected by gel filtration (figure 4I–J and online supplemental figure S8H-I), with a large complex containing AKT inactive species (online supplemental figure S8H-I). To further validate the effect of palmitoyl-modification for AKT dimer/polymer, we employed both insect cell and bacteria purified systems, in which palmitoyl-transferases was very low or evolutionarily lost, respectively. We observed that, similar to C224S, purified WT-AKT protein displayed large-sized probable aggregates in vitro (figure 4K–L and S8J). These observations imply that palmitoylation at C224 antagonises the formation of AKT aggregates.

Figure 4

Palmitoylation defects AKT polymerisation. (A) Split synthetic Renilla luciferase system. (B) HEK293 cells were transfected with EV, Renilla-N, Renilla-C, AKT1-Renilla-N, AKT1-Renilla-C, or AKT1-C224S-Renilla-C, then subjected to luciferase assay. Data are presented as the mean±SD of three independent experiments. ***P<0.001, ****p<0.0001 (Student’s t test and two-way analysis of variance analysis). (C) Bacterially purified His-AKT1 protein and glutathione S-transferase (GST) or GST-AKT1 proteins were subjected to GST pull-down assay. (D) HEK293 cells were transfected with cyan fluorescent protein (CFP)-AKT1 or CFP-AKT1-C224S and yellow fluorescent protein (YFP)-AKT1 plasmids, then subjected to fluorescent resonance energy transfer (FRET) assay (the red boxes were bleached, the green boxes were not bleached, and the blue boxes were negative control). Scale bars, 10 µm. (E) The quantifications of CFP changes after YFP photobleaching. Data are presented as the mean±SD of three independent experiments. **P<0.01, ***p<0.001, ****p<0.0001 (Student’s t test and two-way analysis of variance analysis). (F) Huh7 and Jhh7 were subjected to disuccinimidyl suberate (DSS) crosslinking assay. (G) HEK293 cells were transfected with EV or Flag-AKT1, treated with PA or 2BP for 4 hours, the whole cell extract was immunoprecipitated with anti-Flag agarose and eluted with 3×Flag peptide, then subjected to Native PAGE assay. (H) HEK293 cells were transfected with EV, Flag-AKT1-WT, C77S, C224S, or C77/224S, the whole cell extract was immunoprecipitated with anti-Flag agarose and eluted with 3×Flag peptide, then subjected to Native PAGE assay. (I and J) HEK293 cells were transfected with Flag-AKT1-WT, C77S, C224S, or C77/224S, then the whole cell extract was separated by gel filtration and subjected to IB analysis (I). The quantitative trendline results of protein expression levels were quantified by Image J software (J). (K) Bacterially purified His-AKT1 protein was subjected to gel filtration and IB analysis. (L) Bacterially purified His-AKT1 was subject to in vitro palmitoylation assays with purified GST-ZDHHC24. The reaction was analysed with Native PAGE. SE, shorter exposure; LE, longer exposure.

ZDHHC17/24 interact with and palmitoylate AKT

Next, we screened a panel of palmitoyl-transferases (ZDHHCs) to point out potential enzymes for AKT palmitoylation. Notably, certain ZDHHCs readily associated with AKT, among that ZDHHC17 and ZDHHC24 were further shown to colocalise with AKT and promote AKT palmitoylation (online supplemental figure S9A-C). Thereafter, we focused on investigating the role of ZDHHC17/24 in AKT activation. First, the PH domain of AKT was narrowed to interact with ZDHHC17, while both the PH and kinase domain bound to ZDHHC24 (online supplemental figure S9D-E). Furthermore, direct interaction of AKT with ZDHHC17/24 was demonstrated both in vitro and in cells (figure 5A and online supplemental figure S9F-G). Interestingly, the binding of AKT to ZDHHC17/24 was significantly increased after PA stimulation and attenuated after 2BP administration (figure 5B and online supplemental figure S9H-I), whereas insulin and PI3K inhibitors did not affect their interaction (online supplemental figure S9J-M). Moreover, WT, but not the catalytically inactive ZDHHC17/24 mutant (C467S for ZDHHC17 and C124S for ZDHHC24), could promote AKT palmitoylation in cells (online supplemental figure S10A-H). Similar results were observed in the in vitro palmitoylation assay (online supplemental figure S10I-M). Notably, ZDHHC17 and ZDHHC24 promoted AKT palmitoylation at C77 and C224, respectively, in vitro (figure 5C–D and online supplemental figure S10N-O). However, the C224S mutant also decreased ZDHHC17-mediated AKT palmitoylation in cells (online supplemental figure S10P-Q), which may be due to its role in promoting the assembly of AKT polymers to evade recognition by ZDHHC17. To this end, coexpression of ZDHHC17 and ZDHHC24 further enhanced AKT palmitoylation, PIP3 binding and membrane localisation (online supplemental figure S10R-T). Interestingly, consistent with the finding above that lack of palmitoyl-transferases leading to bacterially purified AKT forming polymer, ZDHHC24-mediated in vitro palmitoylation significantly reduced purified AKT aggregates (figure 4L). These results suggest that ZDHHC17 and ZDHHC24-mediated palmitoylation of AKT may perform different functions (online supplemental figure S10U).

Figure 5

ZDHHC24 promotes AKT palmitoylation and performs oncogenic functions. (A) Whole cell lysates of HepG2 cells were immunoprecipitated with α-ZDHHC24 antibody and subjected to IB analysis. (B) HepG2 cells were treated with or without PA or 2BP for 4 hours, then subjected to endogenous IP analysis. (C and D) Bacterially purified His-AKT1-WT, C77S, C224S, or C77/224S and GST, GST-ZDHHC24 (C), or GST-ZDHHC17 (D) proteins were subjected to in vitro palmitoylation modification assay. (E) Liver tissues of Zdhhc17+/+ , Zdhhc17-/- , Zdhhc24+/+ , Zdhhc24+/- and Zdhhc24-/- mice were subjected to IB analysis. (F) Liver tissues from above group were stained with H&E and analysed by IHC for pS6 levels. (G) Gene Set Enrichment Analysis (GSEA) analysis of non-alcoholic fatty liver disease and PI3K-AKT signalling pathway in liver tissues from Zdhhc24+/+ or Zdhhc24-/- mice that were profiled by RNA-Seq. (H) Representative image of livers harvested from Zdhhc24+/+ , Zdhhc24-/- , Zdhhc17+/+ , or Zdhhc17-/- mice which were hydrodynamic transfected with N‐RasV12, SB transposase, and Flag-AKT1 plasmids, then fed on a normal diet (ND) or a high-fat diet for 2 months. The percentage of liver/body weight was shown. Data are presented as the mean±SD (n=8 mice per group). ***P<0.001, ****p<0.0001 (Student’s t test). (I) Liver sections were stained with H&E and analysed by IHC for pS6, AFP, Ki67, and Cleaved-Caspase3 levels. (J) Liver tissues from above mice were subjected to ABE palmitoylation assay. (K) Representative image of livers harvested from Zdhhc24+/+ or Zdhhc24-/- mice which were fed with a ND or a HFMCD diet. The percentage of liver/body weight was shown. Data are presented as the mean±SD (n=6 mice per group). **P<0.01, ***p<0.001, ****p<0.0001 (Student’s t test and two-way analysis of variance analysis). (L) Liver sections from above mice were stained with H&E, Masson and analysed by IHC for pS6, AFP, F4/80 and Ki67 levels. (M) Liver tissues from above mice were subjected to ABE palmitoylation assay. (N) Liver tissues from above mice were subjected to IB analysis.

ZDHHC17/24 play oncogenic roles by activating AKT

Next, we tended to explore the biological functions of ZDHHC17/24 in activating AKT. To this end, ectopic expression of both ZDHHC17 and ZDHHC24 enhanced, whereas depletion of ZDHHC17 or ZDHHC24 reduced AKT phosphorylation on insulin or PA stimulation (online supplemental figure S11A-G). Consistently, ZDHHC17/24 significantly enhanced the interaction of AKT and PDK1 (online supplemental figure S11H-I), whereas expression of WT but not catalytically inactive ZDHHC24 reduced AKT dimerisation (online supplemental figure S11J). Furthermore, depletion of ZDHHC24 largely enhanced AKT polymerisation compared with counterpart cells (online supplemental figure S11K). To investigate the biological functions of ZDHHC17/24, ectopic expression of WT but not catalytically inactive ZDHHC17/24 enhanced the growth and colony formation of normal hepatic cell THLE3 (online supplemental figure S12A-E). Furthermore, depletion of ZDHHC17 or ZDHHC24 significantly reduced malignancy and tumour growth in HCC cells (online supplemental figure S12F-R), indicating the oncogenic roles of ZDHHC17 and ZDHHC24. To examine whether ZDHHC17/24 exert oncogenic effects through palmitoylation and activation of AKT, we found that the AKT inhibitor MK2206 significantly inhibited WT but only slightly affected ZDHHC17/24-induced colony formation (online supplemental figure S13A-B). Furthermore, ectopic expression of ZDHHC17/24 markedly promoted cell growth in WT, but not in C77S or C224S mutant AKT1, rescued DLD1-AKT1/2-/- cells (online supplemental figure S13C-D).

To explore the physiological role of ZDHHC17/24 in vivo, Zdhhc17 and Zdhhc24 knockout mice (termed Zdhhc17-/ - and Zdhhc24-/- ) were generated. Of note, both Zdhhc17-/ - and Zdhhc24-/ - mice exhibited reduced Akt phosphorylation (figure 5E–F and online supplemental figure S14A-M), along with reduced Akt palmitoylation (online supplemental figure S14N). Furthermore, RNA-sequencing-based transcriptomic analysis of liver tissues from Zdhhc24-/- and counterpart mice showed that depletion of Zdhhc24 severely impaired the metabolic functions in liver tissues, such as reducing NAFLD, and repressing PI3K-AKT signalling pathway (figure 5G and online supplemental figure S15A-E), suggesting a potent role of ZDHHC24 in AKT activation and liver diseases. Furthermore, in the optimised HFD-fed hydrodynamics transfection NASH/HCC mouse model, ablation of Zdhhc17 and Zdhhc24 resisted NASH induced by Akt-transfection (figure 5H), coupled with reduced Akt phosphorylation, palmitoylation and its downstream activation (figure 5I–J and online supplemental figure S15F-G). To confirm the potential role of ZDHHC17/24 in NASH, we employed HFMCD-based NASH mouse model.33 34 The results showed that deletion of Zdhhc17 or Zdhhc24 could protect HFMCD-induced NASH (figure 5K–L and online supplemental figure S15H-I), coupled with decreased AKT palmitoylation and activation (figure 5M–N and online supplemental figure S15J-K). Together, these results suggest that ZDHHC17/24-mediated palmitoylation and activation of AKT is important for their function in promoting NASH and HCC.

Bioinformatically, ZDHHC17/24 were observed to be amplified in various cancers, including about 2% of HCC (online supplemental figure S16A), which mutually excluded with AKT activation pathways, such as alterations in PIK3CA, PTEN, AKT1 and EGFR (online supplemental figure S16B). To test this finding, expression of ZDHHC17/24 and activation of AKT (indicated by pS6) were detected in samples of patient with HCC. As a result, ZDHHC17/24 were relatively highly expressed in HCC compared with adjacent normal tissues (figure 6A and online supplemental figure S16C), and positively correlated with AKT activation (figure 6B and online supplemental figure S16D). More importantly, both the expression of ZDHHC17 and ZDHHC24 were significantly associated with the poor survival rate of patients with HCC (online supplemental figure S16E), suggesting that amplification of ZDHHC17/24 may promote HCC through activation of AKT.

Figure 6

Targeting AKT palmitoylation to repress AKT activity and oncogenic roles in hepatocellular carcinoma (HCC). (A) IHC staining of ZDHHC17, ZDHHC24 and pS6 (pS240/S244) were performed in liver tumour and adjacent tissues. (B) The correlation of pS6 (pS240/S244) and ZDHHC17 or ZDHHC24 protein levels was calculated in liver carcinoma cirrhosis microarray. ****P<0.0001 (χ2 test). (C) Schematic of AKT1-C224 and AKT1-S224 peptides. The different residues are shown in red. CPPtat, CPP from HIV-1 Tat protein. (D) HEK293 cells were transfected with Flag-AKT1 or Flag-AKT1-C224S, treated with or without AKT1-C224 and AKT1-S224 peptides (10 µM) for 8 hours, then subjected to acyl-biotin exchange (ABE) palmitoylation assay. (E) Huh7 cells were serum-starved for 10 hours, treated with AKT1-C224 and AKT1-S224 peptides (10 µM) for 8 hours, before stimulated with or without palmitic acid (PA), then subjected to IB analysis. (F–H) Mice bearing Huh7 xenografts were treated with AKT1-C224 and AKT1-S224 peptides (10 mg/kg/day, I.P.), and tumour sizes were monitored (F). Data are presented as the mean±SD (n=8 mice per group). N.s, no significant; ***P<0.001 (two-way analysis of variance analysis). Tumours were dissected (G) and weighed (H). Data are presented as the mean±SD (n=8 mice per group). ***P<0.001 (Student’s t test). (I) Tumours tissues in (G) were subjected to ABE palmitoylation assay. (J) Illustration of the flow for treating Cebpb-Tta-TetO-Myc mice with AKT competing peptides. Briefly, female mice were terminated doxycycline water (100 μg/mL) at age 4 weeks, and then injected intraperitoneally with PBS, AKT1-C224, or AKT1-S224 peptides (10 mg/kg/day) for 10 days. The livers were harvest for further analysis. (K and L) Macroscopic liver images of Tet-on C-Myc mice with different treatments (K). And the ration of liver and bodyweight was calculated (L). (n=7 mice per group). ***P<0.001 (Student’s t test). (M) Cebpb-Tta-TetO-Myc female mice were terminated doxycycline water (100 µg/mL) at age 4 weeks, and then injected intraperitoneally with phosphate buffered saline (PBS), AKT1-C224, or AKT1-S224 peptides (10 mg/kg/day) until death, and then survival analysis of these mice was performed. (n=8 mice per group). ***P<0.001 (Log-rank Mantel-Cox test). (N–P) Liver sections were stained with H&E, Masson, and analysed by IHC for pS6 (pS240/244), pGSK3β, AFP, Ki67, Cleaved-Caspase3, Pan-CK, and CPS1 levels (N). The dissected tissues were harvested for IB analysis (O) and palmitoylation assay (P).

APT2 reduces AKT palmitoylation and oncogenic functions

Palmitoylation is the only reversible fatty modification that can be converted by acyl-protein thioesterases such as APT1 and APT2.35 36 To this end, APT2, but not APT1, was identified to interact with AKT mainly in its PH domain (online supplemental figure S17A-D). As a result, PA-induced AKT palmitoylation could be attenuated by APT2, which was further observed as a main eraser of AKT palmitoylation (online supplemental figure S17E-F). Consistent with this finding, ectopic expression of APT2 reduced AKT phosphorylation, membrane translocation, and PDK1 interaction on insulin or PA stimulation (online supplemental figure S17G-K). However, depletion of APT2 markedly elevated cell proliferation while enhancing AKT palmitoylation and phosphorylation (online supplemental figure S17L-N). Moreover, ectopic expression of APT2 could enhance AKT polymerisation (online supplemental figure S17O). These findings suggest that APT2 plays an opposite role to ZDHHC17/24 in regulating AKT palmitoylation, activity and oncogenic function.

Targeting AKT palmitoylation to repress HCC tumour growth

Since PA can be synthesised by the tricarboxylic acid (TCA) cycle with citrate as the source and fatty acid synthase (FASN) as the restrict enzyme,37 abnormal homeostasis of this process also leads to PA accumulation, especially in tumours. Previous reports have shown that, in fact, liposynthetic enzymes including FASN were elevated in NASH and HCC.38 Furthermore, amplification of FASN tended to be mutually exclusive with AKT activation and ZDHHC17/24 amplification in HCC (online supplemental figure S18A), indicating that abnormal FASN would lead to synthesis of PA, in turn promoting NASH and HCC. Consistent with previous findings, the approved anti-obesity drug FASN inhibitor orlistat can fight NASH, and synergise with chemotherapeutics to treat cancer, possibly by activating AKT.39 Therefore, we hypnotised that the elevation of FASN may palmitoylate and activate AKT through the synthesis of PA. To this end, we observed that orlistat attenuated AKT palmitoylation, activation and membrane location (online supplemental figure S18B-D). Meanwhile, orlistat could significantly attenuate AKT palmitoylation, activation and the proliferation and colony formation of cells bearing WT but not C77/224S-AKT (online supplemental figure S18E-F). Consistent with these findings, deletion of ZDHHC24 was resistant to orlistat treatment both in cells (online supplemental figure S18G) and in vivo (online supplemental figure S18H-K), indicating that aberrant synthesis of PA can be eliminated by FASN inhibitors, resulting in reduced AKT palmitoylation and activation against according cancers.

As intake of PA from food bypasses FASN, we observed that orlistat could not efficiently block PA accumulation on PA stimulation (online supplemental figure S19A-B). As a result, orlistat displayed inefficacy to block free PA-induced AKT phosphorylation and PIP3 interaction (online supplemental figure S19C-D). Alternatively, aim to directly block AKT palmitoylation, we synthesised candidate peptides containing the intact (C224) or mutant (S224) region of AKT1 (figure 6C), wherein the CPPtat sequence was included to enable the peptides penetrating cells.40 Notably, the C224 but not the S224 peptide significantly attenuated AKT palmitoylation (figure 6D), and insulin/PA-induced AKT activation (figure 6E and online supplemental figure S19E). Furthermore, administration of C224 but not S224 peptides markedly reduced hepatoma cell growth and colony formation (online supplemental figure S19F-G) and in vivo tumour growth (figure 6F–H), coupled with decreased AKT palmitoylation, activation (figure 6I and online supplemental figure S19H) and promotion of proliferation index (Ki67), as well as enhanced cell apoptosis (Cleaved Caspase 3 and PARP) (online supplemental figure S19I). Since blocking AKT could diminish C-Myc transgenic mice HCC tumorigenesis,41 we employed this murine model to validate AKT target peptides (figure 6J). The results showed that C224 but not the S224 peptides attenuated C-Myc-induced HCC formation (figure 6K–L and online supplemental figure S19J) and mouse survival (figure 6M), coupled with decreased AKT activity and palmitoylation (figure 6N–P). These findings suggest that specific peptides against AKT palmitoylation may benefit PA-induced tumour growth.

Thereby, amplified ZDHHC17/24 or accumulated PA by HFD or amplification of FASN-mediated synthesis promotes AKT activation through S-palmitoylation, promoting NASH and further liver tumorigenesis. Therefore, adopting a low-fat (PA-restricted) diet, restricting PA synthesis (with orlistat), directly targeting AKT palmitoylation (AKT peptides) or ZDHHC17/24 (2BP or specific inhibitors) will provide effective strategies for HCC therapies (figure 7).

Figure 7

Model of palmitoylation enhancing AKT activity and oncogenic function. Schematic showing that palmitic acid (PA) uptake, and amplified ZDHHC or FASN could promote AKT activation through S-palmitoylation to promote non-alcoholic steatohepatitis (NASH) and subsequent liver cancer, which could be ameliorated by PA restriction and orlistat-mediated PA blocking, or targeting AKT palmitoylation by ZDHHC inhibitors or specific AKT peptides.

Discussion

Here, based on the HFLC mouse model of NAFLD and human liver specimens, we observed that AKT palmitoylation and activation are positively associated with the progression from NASH to HCC. Importantly, HFD (or PA stimulation) can promote the formation of NASH and HCC in a WT Akt hydrodynamically transfected mouse model. These findings were confirmed in mouse models under palmitoylation-deficient Akt-C77/224S hydrodynamic transfection or Zdhhc17/24 knockout conditions. We further established a HFMCD-based NASH mouse model, in which depletion of Zdhhc17/24 effectively alleviates HFMCD-feeding-induced NASH. These observations suggest that AKT palmitoylation and activation play an important role in promoting NASH and HCC. Although hydrodynamic transfection HCC mouse models have shown the important role of high-fat induced AKT activation in NASH/HCC, generatinon of the liver-specific Akt knockout or Akt-C224S knockin mice, using for HFLC and HFMCD models, are important to further confirm AKT roles in the progression of NAFLD (from NASH to HCC).

Although HFD-induced obesity play a key role in tumorigenesis, progression, metastasis and drug resistance mainly through manipulation of metabolic homeostasis and inflammation,42 the detailed mechanisms of different fatty acids in these malignancies remain unclear. Although cholesterol synthesis compensates for fatty acid inhibition in HCC,43 we have not observed obviously compensatory activation of the cholesterol pathway following inhibition of palmitoylation (data not shown). Alternatively, in this study, we focused on one long-chain saturated fatty acid, PA, which is a major component of palmitic oil and has been reported to play a potential role in oral tumour metastasis.2 In addition to its metabolic role in supporting lipogenesis, here we observed that PA can directly activate AKT through palmitoyl-modification. More importantly, the baseline of AKT palmitoylation is important for AKT response to insulin or other growth factor stimulation.

Since PA is mainly derived from two pathways, FASN-mediated synthesis following the TCA cycle and intake from food. We used the FASN inhibitor orlistat, which has been approved for clinical obesity treatment. We found that orlistat could largely attenuate AKT palmitoyl-modification and activation by reducing basal PA levels, especially under insulin-stimulation conditions. However, orlistat was unable to block PA-induced AKT activity, suggesting that orlistat was not effective in overcoming PA intake-induced obesity or cancer, a possibility of tumour resistance to orlistat, while potent amplification of FASN-induced PA accumulation and AKT activation may benefit from this inhibitor. Due to the important physiological roles of PA in providing lipids for neuroprotection, orlistat would cause potential side effects. Alternatively, we develop small peptides that can compete ZDHHC24-mediated AKT palmitoylation and activation, resulting in decreased its oncogenic functions, with the potent efficacy for HCC therapy. As we demonstrate above, ZDHHC17/24 are also observed amplification in HCC, indicating poor outcome for patients with HCC and acting as oncogenes through activation of AKT, which provides a promising target for HCC therapy. Thus, developing the specific inhibitors to target ZDHHC17/24 would provide a promising strategy for cancer therapy.

In sum, our findings not only enclose a fine-tuned regulation of AKT activity by PA-ZDHHC-mediated palmitoylation, but also highlight strategies for HCC therapy by limiting PA uptake, restricting PA synthesis and targeting ZDHHC17/24.

Data availability statement

Data are available upon reasonable request.

Ethics statements

Patient consent for publication

Ethics approval

Not applicable.

Acknowledgments

We thank members of the Guo laboratory for critical reading and kind suggestion of the manuscript.

References

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Footnotes

  • Correction notice This article has been corrected since it published Online First. The ORCID ID's for authors have been corrected.

  • Contributors Guarantor: JG. Conception and design: JG, LB and MK. Development of methodology: LB, ZZ, JC, YF, JG and YS. Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc): LB, ZZ, JC, YF, HW, XZ, XW. Analysis and interpretation of data (eg, statistical analysis, biostatistics, computational analysis): LB, ZZ, JC, QJ, BG. Writing the manuscript and revision: JG, LB. Administrative, technical, or material support (ie, reporting or organising data, constructing databases): LB, ZZ, JC, LW, XZ, WX, KH, WW, MK. Study supervision: JG, MK. Approved manuscript: all authors.

  • Funding This work was supported in part by National Key Research and Development Program of China (2023YFC3402100 to J.G), National Nature Science Foundation of China (32070767 to J.G, 82302911 to L.B) and Guangdong Basic and Applied Basic Research Foundation (2022A1515220004 to J.G).

  • Competing interests WW is a co-founder and consultant for the ReKindle Therapeutics.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

Linked Articles