ZDHHC2-Mediated AGK Palmitoylation Activates AKT–mTOR Signaling to Reduce Sunitinib Sensitivity in Renal Cell Carcinoma

ZDHHC2 confers sunitinib resistance to clear cell renal cell carcinoma by catalyzing AGK palmitoylation to activate the AKT–mTOR pathway.

Abstract

Tyrosine kinase inhibitors (TKI) that can suppress the VEGF signaling pathway and angiogenesis have been developed to impede the progression of malignant tumors and have been approved as first-line targeted agents for clear cell renal cell carcinoma (ccRCC). Dysregulation of lipid metabolism is a major driver of TKI resistance in renal cancer. In this study, we showed that the palmitoyl acyltransferase ZDHHC2 is abnormally upregulated in tissues and cell lines resistant to TKIs, such as sunitinib. Upregulation of ZDHHC2 contributed to sunitinib resistance in cells and mice, and ZDHHC2 regulated angiogenesis and cell proliferation in ccRCC. Mechanistically, ZDHHC2 mediated AGK S-palmitoylation to promote translocation of AGK into the plasma membrane and activation of the PI3K–AKT–mTOR signaling pathway in ccRCC, which modulated sunitinib sensitivity. In conclusion, these results identify a ZDHHC2–AGK signaling axis and suggest that ZDHHC2 is a targetable candidate for improving the antitumor efficacy of sunitinib in ccRCC.

Significance:

ZDHHC2 confers sunitinib resistance to clear cell renal cell carcinoma by catalyzing AGK palmitoylation to activate the AKT–mTOR pathway.

Graphical Abstract

Introduction

Clear cell renal cell carcinoma (ccRCC) is the main subtype of renal cell carcinoma (RCC) and is characterized by increased hypoxia and upregulation of angiogenesis-associated genes due to early inactivation of Von Hippel–Lindau (VHL; refs. 1, 2). As the main effector pathway of angiogenesis, the VEGF pathway is involved in promoting tumor initiation and progression (3). Tyrosine kinase inhibitors (TKI), such as sunitinib, suppress the VEGF signaling pathway and angiogenesis of tumors and have been recognized as first-line agents for advanced ccRCC (4). Primary resistance or acquired resistance after long-term use of TKIs has become an obstacle for prolonging the survival time of patients with ccRCC (5). Understanding the underlying mechanisms regulating the sensitivity of ccRCC cells to TKIs would help identify novel therapeutic candidates to improve the antitumor effect of TKIs in ccRCC.

Excessive activation of the PI3K signaling pathway was recognized as an important factor in tumor progression that facilitated AKT phosphorylation and activation by phosphorylating phosphatidylinositol 4,5-bisphosphate (PIP2) to phosphatidylinositol 3,4,5-trisphosphate (PIP3; refs. 6–8). Then, AKT activates mTOR signaling via the mTOR downstream effector, P70S6 kinase (S6K), to regulate cell-cycle progression, cell proliferation, and metabolism (9). Multiple studies have shown that abnormal activation of the PI3K–AKT–mTOR axis is a vital cause of TKI resistance in tumors, especially lung adenocarcinoma (10, 11) and ccRCC (12–15). Blocking the activation of the AKT–mTOR signaling pathway could promote the sensitivity of ccRCC to sunitinib (14, 16). Thus, further exploration of the regulatory mechanism of the AKT–mTOR signaling pathway could provide novel therapeutic candidates for ccRCC treatment.

Acylglycerol kinase (AGK), a mitochondrial inner membrane protein, not only phosphorylates monoacylglycerol and diacylglycerol to form lysophosphatidic acid and phosphatidic acid (17), but also participates in regulating the EGF signaling pathway in prostate cancer (18). In addition, Hu and colleagues also found that plasma membrane (PM)-located AGK inhibited the phosphatase activity of PTEN by phosphorylating PTEN at Ser380, Thr382, and Thr383, thereby activating the PI3K–AKT–mTOR signaling pathway, facilitating CD8⁺ T-cell proliferation and enhancing antitumor immunity (19). Moreover, AGK was reported to promote the angiogenesis of hepatocellular carcinoma by activating of the NF-κB pathway (20). Given that the activation of EGF (21, 22), PI3K–AKT–mTOR (6, 12), and NF-κB signaling (23) contributes to the resistance of TKIs, AGK might also have an effect on TKI sensitivity in tumors, but the regulatory mechanism is currently unclear and needs to be explored.

S-Palmitoylation of cysteine residues, a posttranslational modification of proteins containing the lipid palmitate, is a key regulator of protein targeting and trafficking (24). S-Palmitoylation is a reversible process and is crucial to the function of proteins as it regulates protein subcellular localization, function, and stability (25). Palmitoyl acyltransferases (PAT), first identified in 2002, are responsible for the palmitoylation of target proteins (26). In human cells, more than 20 PATs with a conserved zinc-finger DHHC (Asp–His–His–Cys) motif, namely, ZDHHCs, have been identified (27). PAT dysregulation and abnormal palmitoylation of oncogenes or tumor suppressor proteins play key roles in tumorigenesis (25). Recently, it has been reported that aberrant lipid metabolism can result in antiangiogenic drug resistance in cancer (25). However, the role of protein lipidation by palmitoylation in antiangiogenic drug resistance remains unclear. In this study, we aimed to explore whether protein palmitoylation participates in the process of TKI resistance and to elucidate the underlying mechanisms in ccRCC. We demonstrated that ZDHHC2 was closely associated with sunitinib resistance in ccRCC cells. Further study indicated that ZDHHC2 catalyzes AGK palmitoylation at Cys72 to enhance the localization of the AGK in the PM and subsequently activate the AKT–mTOR signaling, leading to sunitinib resistance in ccRCC cells.

Materials and Methods

Collection of clinical specimens

Clinical specimens of ccRCC were collected from the Department of Urology, The Second Xiangya Hospital, Central South University. Ethical approval for the use of human tissues (patients with renal cancer with or without sunitinib resistance) was obtained by the local ethics committee (The Second Xiangya Hospital, China; approval no. 2021068). Written informed consent was acquired from all patients before surgery. We collected the specimens from patients with ccRCC that were diagnosed at the late stage and underwent a palliative resection of the tumor following sunitinib therapy. Tissue specimens are collected from the primary cancer tissue. Postoperative imaging examination, such as computed tomography, was used to evaluate the therapeutic effect of sunitinib therapy. According to the RECIST version 1.1 (28, 29), we defined those patients achieved complete remission or partial remission as sunitinib sensitive, and patients with progressive disease as sunitinib resistance.

Cell lines, antibodies, and chemicals

CcRCC cell lines, A498 (CL-0254) were obtained from Procell Life Science & Technology, 786-O (SC0154) and ACHN (SC0253) cells were purchased from the Yuchicell Biology Technology. ACHN cell line was assumed, and published, to be a ccRCC but subsequent genomic analysis has shown that it lacks the VHL mutation and chromosome 3p loss associated with ccRCC and does have the gains of chromosomes 7 and 17 associated with type 1 papillary RCC (30). Because we did not explore the mechanism associated with hypoxia, ACHN was still available for the following experiment. All cell lines were identified by short tandem repeat profiling in Yuchicell Biology Technology. 786-O cells were cultured in RPMI1640 medium (Gibco) and A498 cells were cultured in in MEM (PM150410, Procell Life Science & Technology), supplemented with 10% FBS (AC03L055, Shanghai Life-iLab Biotech) and incubated at 37°C in 5% CO₂. Sunitinib-resistant 786-O cells (786-O R) were generated as previously reported (14).

Mycoplasma contamination was regularly tested every 3 months using PlasmoTest - Mycoplasma Detection Kit (# rep-pt1, InvivoGen, distributors MingRui Biotech Co., LTD). Plasmocin - Mycoplasma Elimination Reagent (#ant-mpp, InvivoGen) was routinely added to the cell culture medium to prevent Mycoplasma contamination.

As previously described (31), The antibodies used as follows: β-Actin (#66009–1-Ig, Proteintech, 1:5,000 dilution), ZDHHC2 (#DF4688, Affinity Bioscience, 1:1,000 dilution), ZDHHC5 (#21324–1-AP, Proteintech, 1:1,000 dilution), ZDHHC15 (#Ag16740, Proteintech, 1:1,000 dilution), ZDHHC22 (#HPA072213, Atlas Antibodies, 1:100 dilution), AGK (#ab137616, Abcam, 1:2,000 dilution), S6K1 (#2708, Cell Signaling Technology, 1:1,000 dilution), pS6k1-T398 (#9209, Cell Signaling Technology, 1:1,000 dilution), cleaved caspase-3 (#9661, Cell Signaling Technology, 1:2,000 dilution), caspase-3 (#19677–1-AP, Proteintech, 1:1,000 dilution), pAKT-S473 (#4060S, Cell Signaling Technology, 1:1,000 dilution), pAKT-T308 (#13038, Cell Signaling Technology, 1:1,000 dilution), cleaved caspase-3 (#9661, Cell Signaling Technology, 1:2,000 dilution), Calnexin (#10427–2-AP, Proteintech, 1:2,000 dilution), Flag (#20543–1-AP, Proteintech, 1:2000 dilution), ATP1A1 (#14418–1-AP, Proteintech, 1:2,000 dilution). Sunitinib (#S7781), everolimus (#S1120), tivozanib (#S1207), 2-bromohexadecanoic acid (#E0120) were purchased from Selleckchem. The sources of the remaining reagents are marked clearly in each section.

As previously described (31), Flag-ZDHHC2 and HA-AGK were constructed by cloning the cDNA of ZDHHC2 and AGK into the OmicLinkTM Expression Clone (CMV Promoter; GeneCopoeia, EX-V0006-M14). The point mutations of ZDHHC2 and AGK were synthesized by using the Hieff Mut Multi Site-Directed Mutagenesis Kit (#11004ES10, Yeasen Biotechnology). The siRNAs were purchased from RiboBio. The sequences of the siRNAs and single-guide RNAs (sgRNA) are provided in Supplementary Table S1. A caspase-3 activity assay kit (#ab39401) was purchased from Abcam. Cells were transfected with the indicated plasmids or siRNAs using Lipofectamine 2000 (Thermo Fisher Scientific) according to the manufacturer's instructions.

Transient transfection of siRNA and plasmids

Cells were cultured in plates or dishes to undergo starvation treatment with 1 mL serum-free Opti-MEM medium (Gibco) for 12 hours. Then the indicated siRNA, plasmids and Lipofectamine 2000 (Thermo Fisher Scientific) were incubated together for 20 minutes in 1 mL serum-free Opti-MEM medium. The mixture was transferred to plates or dishes from centrifuge tubes. After the cells were transfected for 6 hours, serum-free Opti-MEM medium was replaced with complete DMEM medium for another 72 hours.

Cell proliferation assay

The Cell Counting Kit-8 (CCK-8) assay was used for the in vitro cell proliferation assay as previously reported (32). In brief, CCK-8 reagent (#C0037, Beyotime) was added to each cell well and the absorbance at 450 nm was measured with a microplate reader.

Mouse xenograft assay

All animal experiments were approved by the Institutional Animal Care and Use Committee of the Second Xiangya Hospital, Central South University (animal license number: 2022589). BALB/c nude mice (6 weeks old and 22–24 g) were purchased from Vital River. These mice were housed in standard conditions (60% ± 3% humidity at 22°C ± 0.5°C) with a 12-hour light/dark cycle in the animal center of the Second Xiangya Hospital. They can easily obtain the food and water ad libitum. ccRCC cell lines infected with sgControl or sgZDHHC2 were subcutaneously injected into the left side of the backs of the mice (1 × 10⁷ cells per mouse). After the tumor volume reached to 50 mm³, these mice were treated with or without sunitinib (oral administration, 25 mg/kg, once a day for 8 days). Tumor volume was calculated using the formula (L × W²)/2. Once the mice were euthanized, the tumors were excised and weighed. The animal experiment complied with the NIH Guide for the Care and Use of Laboratory Animals (NIH Publications No. 8023, revised 1978).

Western blot and immunoprecipitation

The details of immunoprecipitation and Western blot analysis were reported previously (33). Briefly, the RIPA buffer (#P0013, Beyotime) was used to extract the proteins from cells. The protein concentration was measured with a Micro BCA Protein Assay Kit (Sigma-Aldrich). Then, proteins were resolved by SDS-PAGE after adding loading buffer and transferred onto polyvinylidene difluoride (PVDF) membranes (Thermo Fisher Scientific). The PVDF membranes were blocked with 5% skimmed milk and incubated with the corresponding primary antibodies at 4°C overnight. The next day, PVDF membranes were incubated with the secondary antibodies and developed with ECL reagent (Sigma-Aldrich). For immunoprecipitation, protein A+G beads (#P2029, Beyotime) and IgG (#A7007, Beyotime) or a primary antibody were cocultured with the proteins for more than 24 hours. The beads were collected and washed with PBS buffer for six times. Then the beads were boiled for 10 minutes and subjected to Western blotting analysis on SDS-PAGE gels. The secondary antibodies for immunoprecipitation were mouse anti-rabbit IgG (Conformation Specific, L27A9) mAb (HRP Conjugate; Cell Signaling Technology, 5127S), which did not bind denatured and reduced rabbit IgG heavy or light chains. Protein expression levels were measured by using ImageJ software (NIH). Protein identification via mass spectrometry (MS) was performed by SpecAlly Life Technology Co., Ltd.

qRT-PCR

RNA was extracted through using TRIzol reagent (Thermo Fisher Scientific). RNA concentration and quality were assessed using a Nano-Drop 2000 (Thermo Fisher Scientific). RNA was reverse-transcribed with a Prime Script RT Kit (Takara), and TB Green Fast qPCR Mix (Takara) was used to amplify cDNA. Relative gene expression levels were determined using the 2^-ΔC_t method after normalizing to β-actin levels. All primer sequences are provided in Supplementary Table S2.

Tissue microarray and IHC

CcRCC tissue microarray slides (# U081ki01) were purchased from Bioaitech. The tissue microarray slides were only used to examine correlations between protein expression, and did not explore whether there was a correlation with treatment response to TKIs. The tissue microarray specimens were immunostained with pAKT S473, pS6K1 T389, VEGFA, and ZDHHC2. The method of scoring of staining intensity was mentioned previously (32).

Apoptosis assay

The caspase-3 activity assay and Annexin V-FITC/7-AAD assay were employed to measure the apoptosis of cells. For the caspase-3 activity assay, a caspase-3 assay kit (ab39401) obtained from Abcam was used following the manufacturer's protocol. For the Annexin v-FITC/7-AAD assay, cells were stained with Annexin V-FITC and 7-AAD following the manufacturer's instructions for the Annexin V-FITC Apoptosis Detection Kit (Procell). Cells were incubated for 15 minutes at room temperature and analyzed on a flow cytometer. Data were analyzed with FlowJo analysis software.

Human umbilical vein endothelial cell tubule formation assay

The 24-well plates were polymerized with Matrigel for 30 minutes in the incubator. The cell medium of 786-O cells (conditioned medium) was collected to culture human umbilical vein endothelial cells (HUVEC). HUVECs (2 × 10⁴) resuspended in 200 μL of conditioned medium were added to the 24-well plates and cultured for 12 hours. Then, the 24-well plates were photographed with a microscope.

Glutathione S-transferase pull-down assay and purification of recombinant proteins

As previously described (31), cells were lysed with 1 × RIPA lysis buffer (P0013B, Beyotime) for 30 minutes at 4°C. Glutathione S-transferase (GST) fusion proteins were immobilized on BeyoMag Anti-GST Magnetic Beads (P2138, Beyotime). After washing with 1 × RIPA lysis buffer, the beads were incubated with cell lysates for 4 hours. The beads were then washed four times with 1 × RIPA lysis buffer and resuspended in loading buffer. The bound proteins were subjected to SDS/PAGE and Western blotting.

Escherichia coli BL21 was used to express GST-recombinant proteins and His-recombinant proteins after induction with IPTG (Beyotime). Then Escherichia coli BL21 was lysed with muramidase and sonication. Bacterial debris was removed by centrifugation for 10 minutes. Glutathione-Sepharose beads (GE Healthcare Life Sciences) were added to the liquid supernatant to purify GST-fusion proteins at 4°C overnight. The beads were then collected and washed for six times with binding buffer to remove the contaminating proteins. SDS-PAGE and high-sensitivity colloidal Coomassie blue staining were performed to examine the purification efficiency. For purification of His-tagged proteins, HisSep Ni-NTA MagBeads (Yeasen, 20561ES03) were used to isolate and purify His-tagged proteins. The remaining steps were similar to those for glutathione-Sepharose beads.

PM protein extraction and subcellular fractionation assay

PM and intracellular membrane (ICM) fractions were isolated with the Plasma Membrane Protein Extraction Kit (Abcam, ab65400). Briefly, cells were incubated with 2 mL of homogenized buffer at 4°C after being washed three times with PBS buffer. The cells were then harvested with a cell scraper. Following this process, the cells were homogenized with a homogenizer for 50 times and spun at 700 × g for 5 minutes. Then the supernatants were collected and centrifuged at 10,000×g for 30 minutes at 4°C. Total cellular membrane proteins and cytosolic proteins were located in the pellet and supernatant, respectively. The next step was to separate the PM and ICM. The pellet was resuspended in 200 μL upper phase solution. Then, the upper phase solution containing the pellet was mixed with isopyknic lower phase solution. The mixture was centrifuged at 1,000 × g for 5 minutes at 4°C in an ice bath for 5 minutes. PM proteins and ICM proteins were located in the upper phase solution and lower phase solution, respectively. The upper phase solution and lower phase solution were then mixed with 5× volume of water, respectively, followed by collecting PM proteins and ICM proteins by centrifugation at 15,000 × g for 30 minutes at 4°C.

Acyl-biotinyl exchange assay

Cell lysates were incubated with 20 mmol/L of methyl methanethiosulfonate (Sigma-Aldrich) and 1 mmol/L PMSF (Beyotime, China) at 50°C for 30 minutes to exhaustively block free thiols. Proteins were precipitated with acetone and resuspended in 1 mol/L hydroxylamine pH 7.4 (Sigma-Aldrich) to promote depalmitoylation. The proteins were incubated with 0.2 mmol/L biotin-HPDP (Top Science) for 1 hour at room temperature. Streptavidin (Yeasen) was used to purify biotinylated proteins, and biotinylated proteins were analyzed by immunoblotting.

Click-iT pull down

For click-iT identification of palmitoylation, click chemistry and streptavidin pulldown were performed according to the published procedure with slight modifications (34). RCC cells were transfected with the instructed plasmids for 48 hours. Subsequently, the cells were incubated with 100 μmol/L of Click-iT palmitic acid-azide (Thermo Fisher Scientific) for 6 hours. After incubation, the cells were lysed to extract proteins. A Click-iT Protein Reaction Buffer Kit (# C10276; Thermo Fisher Scientific) was used to catalyze the reaction of protein samples with biotin-alkyne. The biotin alkyne-azide–palmitic-protein complex was precipitated with streptavidin (Yeasen, China), and the bound proteins were eluted by boiling with SDS-PAGE sample buffer without DTT for 10 minutes at 95°C, and then analyzed through immunoblotting.

In vitro protein acyltransferase assay

The purified recombinant GST-ZDHHC2 was incubated with purified recombinant His-AGK in the reaction buffer containing 50 mmol/L Tris-HCl [pH 7.4], 10 μmol/L palmitoyl alkyne-CoA (Cayman, #15968) and 1 μmol/L palmostatin B (Sigma, #178501) at 25°C for 1 hour, followed by the Click-iT reaction with palmitic acid-azide (Thermo Fisher Scientific) and Click-iT Protein Reaction Buffer Kit (# C10276; Thermo Fisher Scientific) to biotinylate the proteins with the incorporation of biotin-alkyne. The biotin alkyne-azide–palmitic-protein complex was precipitated with streptavidin (Yeasen). The bound proteins were eluted by boiling with SDS-PAGE sample buffer without DTT for 10 minutes at 95°C, and then analyzed by immunoblotting.

Immunofluorescence assay

Cells were incubated with 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindodicarbocyanine,4-chlorobenzenesulfonate salt (DiD, Far-red Plasma Membrane Fluorescent Probe, Beyotime, C1039) for 20 minutes to label the cell membrane. Then, the cells were fixed with paraformaldehyde for 15 minutes and permeabilized with 0.2% Triton X-100 for 10 minutes. Then, the cells were incubated with anti-ZDHHC2 and anti-AGK antibodies at 4°C overnight. The next day, the cells were incubated with fluorescent secondary antibodies (Beyotime) for 1 hour, followed by washing with PBS buffer for 3 times. Then DAPI (Beyotime) was used to stain cell nuclei for 10 minutes. After washing with PBS buffer for 3 times, the samples were analyzed through confocal microscopy (Andor, Dragonfly, 63× objective lens).

CRISPR/Cas9 technique

Various sgRNAs of ZDHHCs and AGK were designed through https://www.synthego.com. Then sgRNAs were cloned into the lentiCRISPR v2 vector (Addgene, #52961). The sequence of sgRNAs is shown in Supplementary Table S1.

Statistical analysis

As previously described (14), the experimental data are presented as the mean ± SEM. The sample size (n) for each statistical analysis is provided in the figure legends. GraphPad Prism 9 software was used to calculate the P value using the unpaired two-sided Student t test for comparison of difference between two groups, or one-way ANOVA followed by Tukey multiple comparisons post hoc test for comparison of differences between more than two groups. Differences were considered statistically significant at P values less than 0.05. Statistical differences were considered at: *, P < 0.05; **, P < 0.01; ***, P < 0.001; not significant (ns), P > 0.05.

Ethics approval and consent to participate

The study was conducted in accordance with the principles of the Declaration of Helsinki (Approved no. 2021068). It was approved by the Animal Use and Care Committees at the Second Xiangya Hospital, Central South University.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding authors on reasonable request. The data generated in this study are publicly available in Gene Expression Omnibus (GEO) at GSE203485. The data analyzed in this study were obtained from GEO at GSE76068.

Other methods are provided in the Supplementary Data file.

Results

ZDHHC2-mediated S-palmitoylation plays a key role in modulating the sensitivity of ccRCC to sunitinib

We first explored whether palmitoylation regulates the response of ccRCC to sunitinib by using an inhibitor of palmitoylation (2-BP). The CCK-8 assay indicated that 2-BP treatment enhanced the antiproliferation effect of sunitinib on 786-O, 786-O R, and A498 cells (Supplementary Fig. S1A). In addition, we showed that 2-BP treatment also reduced the IC₅₀ values of sunitinib in these ccRCC cells (Supplementary Fig. S1B). Because palmitoylation was involved in modulating the sensitivity of ccRCC cells to sunitinib, we were curious about the PATs that specifically participated in this process. Gene-specific sgRNAs were transfected into 786-O cells, and an unbiased screen was performed by knocking out all 24 PATs individually using the clustered regularly interspaced short palindromic repeats (CRISPR)/Cas9 system (Fig. 1A; Supplementary Fig. S1C and S1D). To avoid the sgZDHHC knockouts (KO) having no effect due to no expression of the target, we analyzed the count values of ZDHHCs in the RNA sequencing (RNA-seq) data of 786-O cells (GSE203485; Supplementary Fig. S1C). ZDHHC15 and ZDHHC22 had the lowest expression levels among these 24 ZDHHCs (Supplementary Fig. S1C). Because we could not buy antibodies for all ZDHHCs, we only examined the expression levels of ZDHHC15, ZDHHC22, and ZDHHC5 in 786-O cells (Supplementary Fig. S1D). We showed that ZDHHC2, ZDHHC5, ZDHHC15, and ZDHHC22 were knocked out in 786-O cells (Supplementary Fig. S1D; Supplementary Fig. S2A). Then, the IC₅₀ values of sunitinib were measured in each group of cells and compared with those in the control group. We found that some PATs contributed to enhancing the sensitivity of sunitinib, but a portion of PATs were related to insensitivity to sunitinib (Fig. 1A). We noticed that KO of ZDHHC2 resulted in the greatest decrease in the IC₅₀ value of sunitinib among the 24 PATs (Fig. 1A). Clustering analysis (using the GSE76068 dataset) also indicated that ZDHHC2 was the most upregulated PAT in a TKI-resistant patient-derived tumor xenograft (PDX) mouse model (Fig. 1B). In addition, we also found that ZDHHC2 was upregulated in sunitinib-resistant ccRCC tissues compared with sunitinib-sensitive ccRCC tissues derived from patients, as well as in sunitinib-resistant 786-O cells, as reported previously (14), compared with wild-type (WT) cells (Fig. 1C and D; Supplementary Fig. S1E and S1F). Subsequently, KO of ZDHHC2 was found to promote apoptosis after sunitinib treatment (Fig. 1E–G). At the same time, we found that relatively low-dose sunitinib (2 μmol/L) treatment had less of an effect on cell proliferation (Supplementary Fig. S1A). Thus, we believed that ZDHHC2 knockdown synergistically enhanced the antitumor effect of sunitinib in 786-O R cells. Moreover, we also demonstrated that downregulation of ZDHHC2 increased the sensitivity of ccRCC cells and mouse models to sunitinib (Fig. 1H–K, Supplementary Fig. S2A). In contrast, overexpression of ZDHHC2 decreased apoptosis after treatment with sunitinib and resulted in sunitinib resistance in 786-O and A498 cells (Fig. 1L–N; Supplementary Fig. S2B). Furthermore, our data also demonstrated that 2-BP treatment attenuated the sunitinib resistance-promoting effect induced by overexpression of ZDHHC2 (Supplementary Fig. S2C). Moreover, we showed that overexpression of a ZDHHC2 enzymatically dead mutant (C129A; refs. 35, 36) had no effect on the sensitivity of ccRCC cells to sunitinib, which indicated that the resistance of sunitinib mediated by ZDHHC2 is closely related to ZDHHC2 enzyme activity (Supplementary Fig. S2D). Together, these data suggest that ZDHHC2 is crucial for regulating the sensitivity of ccRCC cells to sunitinib.

Figure 1.

ZDHHC2-mediated S-palmitoylation plays a key role in modulating the sensitivity of ccRCC to sunitinib. A, 786-O cells were transfected with indicated constructs for 72 hours. These cells were harvested and treated with a serial dose of sunitinib for 24 hours. The CCK-8 assay was applied to measure the IC₅₀ values of sunitinib in each group. Data presents as mean ± SEM with three replicates. B, Expression levels of ZDHHCs during sunitinib pretreatment (n = 4), response (n = 4), and resistance (escape; n = 4) phases. The heatmap shows the log-fold change (LogFC) of escape versus pretreatment and log-fold change of escape versus response. C, The protein level of ZDHHC2 from patients with RCC with (n = 6) or without (n = 6) sunitinib resistance was examined by Western blotting analysis, and the protein level of ZDHHC2 was quantified by ImageJ software. P values as indicated. D, The protein level of ZDHHC2 from 786-O cells with or without sunitinib resistance (786-O R) was examined by Western blotting analysis. Three replicates were performed. E–G, 786-O cells were transfected with indicated constructs for 48 hours. After 24-hour puromycin selection, cells were treated with or without sunitinib (2 μmol/L) for another 24 hours. Cells were collected for Western blot analysis (E), caspase-3 activity assay (F), and Annexin V-FITC/7-AAD assay (G). Data presented as mean ± SEM with three replicates. H, 786-O, A498, and 786-O R cells were transfected with indicated constructs for 72 hours. After puromycin selection, these cells were treated with a serial dose of sunitinib for 24 hours and subjected to CCK-8 assay. The IC₅₀ values of sunitinib in each group are indicated. I–K, 786-O cells were transfected with indicated constructs for 72 hours. After puromycin selection, these cells were subcutaneously injected into the nude mice. These mice were treated with or without sunitinib (oral administration, 25 mg/kg, once a day for 8 days). Tumor image (I); tumor mass (J); tumor growth curve (K). Data presented as mean ± SEM with five replicates. L and M, 786-O cells were transfected with indicated constructs for 24 hours. Cells were treated with or without sunitinib (2 μmol/L) for another 24 hours. Cells were collected for Western blot analysis (L) and caspase-3 activity assay (M). Data presented as mean ± SEM with three replicates. N, 786-O and A498 cells were transfected with indicated constructs for 24 hours. These cells were treated with a serial dose of sunitinib for 24 hours and subjected to CCK-8 assay. The IC₅₀ values of sunitinib in each group are indicated. ns, not significant; *, P < 0.05; **, P < 0.01; ***, P < 0.001. EV, empty vector.

ZDHHC2 contributes to ccRCC proliferation and angiogenesis

Next, we employed the CancerSEA dataset (cancer single-cell state atlas) to explore the function of ZDHHC2 in various types of malignant tumors (Fig. 2A). ZDHHC2 was involved in promoting the epithelial–mesenchymal transition (EMT) and proliferation of tumor cells (Fig. 2A). In addition, ZDHHC2 was positively correlated with hypoxia and angiogenesis, which might be the reason why ZDHHC2 was associated with TKI (sunitinib) resistance in ccRCC (Fig. 2A; Supplementary Fig. S2E). Consistently, we showed that downregulation of ZDHHC2 by transfection with sgRNAs inhibited the proliferation and invasion of 786-O and A498 cells (Fig. 2B and C; Supplementary Fig. S2F). In addition, the proliferation of ccRCC cells was more or less suppressed after four additional PATs (ZDHHC9, ZDHHC11b, ZDHHC13, and ZDHHC20) affecting sunitinib sensitivity mentioned in Fig. 1A were knocked out (Supplementary Fig. S2G). In addition, inhibiting the expression of ZDHHC2 also reduced angiogenic capability (Fig. 2D and E). In contrast, overexpression of WT ZDHHC2 but not the enzymatically dead mutant promoted the proliferation ability of ccRCC cells (Fig. 2F and G). Unsurprisingly, overexpression of ZDHHC2 promoted angiogenesis in a manner dependent on its enzymatic activity (Fig. 2H and I). In addition, ZDHHC2 was knocked out and overexpressed in sunitinib-resistant 786-O cells (Fig. 2J and K) to further demonstrate the effect of ZDHHC2 in sunitinib-resistant cells. The CCK-8 assay showed that ZDHHC2 promoted the proliferation of sunitinib-resistant 786-O cells, but the enzymatically dead mutant lost this effect (Fig. 2L and M). Furthermore, we demonstrated that ZDHHC2 KO decreased the angiogenic capability of 786-O R cells (Fig. 2N). The HUVEC tubule formation assay also showed consistent results that ZDHHC2-WT, but not the enzymatically dead mutant induced angiogenesis (Fig. 2O). Thus, our data indicate that ZDHHC2 plays an essential role in ccRCC proliferation and angiogenesis.

Figure 2.

ZDHHC2 contributes to ccRCC proliferation and angiogenesis. A, The cancerSEA dataset (http://biocc.hrbmu.edu.cn/CancerSEA/) was used to analysis the cancer-related function of ZDHHC2 in various types of malignant tumor. B–E, 786-O and A498 cells were transfected with indicated constructs for 72 hours. After puromycin selection, cells were collected for Western blot analysis (B), CCK-8 assay (C), and in vitro angiogenesis assay (D and E). Data presented as mean ± SEM with three replicates. F–I, 786-O and A498 cells were transfected with indicated constructs for 24 hours. Cells were collected for Western blot analysis (F), CCK-8 assay (G), and in vitro angiogenesis assay (H and I). Data presented as mean ± SEM with three replicates. J–M, 786-O R cells were transfected with the indicated constructs and plasmids. Cells were harvested for Western blotting analysis (J and K) and CCK-8 assay (L and M). Data presented as mean ± SEM with three replicates. N and O, 786-O R cells were transfected with the indicated constructs and plasmids. Then, the cell medium of transfecting 786-O R cells was collected to culture HUVECs and to perform the in vitro angiogenesis assay. Data presented as mean ± SEM with three replicates. ns, not significant; *, P < 0.05; **, P < 0.01; ***, P < 0.001. EV, empty vector.

ZDHHC2 activates the AKT–mTOR signaling pathways in ccRCC

To further elucidate the underlying mechanism of how ZDHHC2 regulates the sensitivity of ccRCC to sunitinib, 786-O cells were transfected with ZDHHC2-specific siRNAs or the corresponding control. Cells were subjected to transcriptome analysis (Fig. 3A). Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis and gene set enrichment analysis (GSEA) of the RNA-seq data indicated that ZDHHC2 silencing modulated a number of signaling pathways in 786-O cells, such as the HIF1, PI3K–AKT, mTOR, PD-L1 expression and PD-1 checkpoint, and FOXO signaling pathways (Fig. 3B–D; Supplementary Fig. S3A–S3C). KEGG enrichment analysis and GSEA of The Cancer Genome Atlas-Kidney Renal Clear Cell Carcinoma (TCGA-KIRC) dataset also showed that ZDHHC2 was involved in regulating the PI3K–AKT and mTOR signaling pathways (Fig. 3E–G). Because dysregulation of the PI3K–AKT–mTOR axis is one of the major causes of sunitinib resistance in ccRCC (13–15), we wondered whether ZDHHC2 regulates the sensitivity of sunitinib through the PI3K–AKT–mTOR axis. Notably, overexpression of ZDHHC2 increased the phosphorylation of AKT S473 and T308 sites and phosphorylation of the S6K1-T389 site in 786-O, A498 and ACHN cells (Fig. 3H). However, knockdown of ZDHHC2 inactivated the PI3K–AKT–mTOR axis, which was reactivated after rescuing the expression of ZDHHC2 in ccRCC cells (Fig. 3I and J). Immunoblotting showed decreased pAKT S473, pAKT T308 and pS6K1 T389 after knocking out ZDHHC2 and increased pAKT S473, pAKT T308, and pS6K1 T389 after reintroducing ZDHHC2 in 786-O R cells (Fig. 3K). Moreover, there was a positive correlation between ZDHHC2 and pAKT S473 (Spearman r = 0.7482; n = 40; P < 0.001), or ZDHHC2 and pS6K1 T389 (Spearman r = 0.7482; n = 40; P < 0.001) after staining ZDHHC2, pAKT S473, and pS6K1 T389 in the tissue microarray of ccRCC (Fig. 3L and M). Together, these data suggest that ZDHHC2 contributes to the activation of the AKT–mTOR pathway in ccRCC.

Figure 3.

ZDHHC2 activates the AKT–mTOR signaling pathways in ccRCC. A, The transcriptome analysis of 786-O cells after transfection with siNC or siZDHHC2 for 48 hours. B, KEGG enrichment analysis of the RNA-seq data after knockdown of ZDHHC2 in 786-O cells. P values as indicated. C and D, GSEA analysis of the RNA-seq data after knockdown of ZDHHC2 in 786-O cells, P values as indicated. E, KEGG enrichment analysis of the TCGA-KIRC dataset after dividing the patient specimens into ZDHHC2 high expression and low expression group. F and G, GSEA analysis of the TCGA-KIRC dataset after dividing the patient specimens into ZDHHC2 high expression and low expression group. H, Western blot analysis of the whole cell lysates of 786-O, A498, and ACHN cells after transfection with indicated plasmids for 24 hours. I, Western blot analysis of the whole cell lysates of 786-O, A498, and ACHN cells after transfection with indicated constructs for 48 hours. J, Western blot analysis of the whole cell lysates of 786-O, A498, and ACHN cells after transfection with indicated constructs for 48 hours. SR, siRNA resistance. K, 786-O R cells were transfected with the indicated constructs for 48 hours. After puromycin selection for 24 hours, cells were harvested for Western blotting analysis. L and M, The tissue microarray of renal cancer was stained with ZDHHC2, pAKT S473, or pS6K1 T389 by using the IHC assay. The typical images of IHC are shown in L. The expression and correlation of pAKT S473 and ZDHHC2, or pS6K1 T389 and ZDHHC2 are shown in M. P values as indicated. EV, empty vector.

Everolimus treatment overcomes ZDHHC2-induced sunitinib resistance in ccRCC

It has been reported that the AKT–mTOR signaling pathway modulates angiogenesis by increasing the expression of VEGF or other angiogenic factors, including nitric oxide and angiopoietins (37). The mTOR inhibitors (everolimus or temsirolimus) are effective in the treatment of metastatic RCC (38, 39). Inhibition of the AKT–mTOR pathway has already been considered for treating metastatic RCC either in combination with anti-angiogenesis agents, including TKIs, or as a second therapy following resistance (38, 40–43). Because we have mentioned that ZDHHC2 activated the AKT–mTOR pathway and regulated the sensitivity of sunitinib in ccRCC, we wondered whether everolimus could reverse the sunitinib resistance induced by ZDHHC2. First of all, we found that overexpression of ZDHHC2 increased the IC₅₀ values of everolimus, while ZDHHC2 silencing reduced the IC₅₀ value of everolimus in 786-O cells (Supplementary Fig. S4A and S4B). Sunitinib-resistant 786-O cells were used to repeat the above experiments. The sensitivity of 786-O R cells to everolimus also increased or decreased along with ZDHHC2 KO or overexpression (Supplementary Fig. S4C). Then, we revealed that everolimus treatment diminished the changes in angiogenic capability and VEGFA expression induced by knockdown or overexpression of ZDHHC2 in A498 and ACHN cells (Fig. 4A–D; Supplementary Fig. S4D and S4E). Furthermore, we showed there was a positive correlation between ZDHHC2 and VEGFA (Spearman r = 0.743; n = 40; P < 0.001) after staining ZDHHC2 and VEGFA in the tissue microarray of ccRCC (Supplementary Fig. S4F and S4G). We also demonstrated that depletion of ZDHHC2 increased the sensitivity of A498 cells to tivozanib, a more VEGF selective TKI than sunitinib (Supplementary Fig. S4H; ref. 44). In addition, we demonstrated that treatment with everolimus attenuated the changes in sunitinib IC₅₀ values induced by silencing or overexpressing ZDHHC2 in ccRCC cells (Fig. 4E and F). In addition, KO or overexpression of ZDHHC2 failed to change the IC₅₀ values of sunitinib after mTOR was inhibited with everolimus in 786-O R cells, which indicated that ZDHHC2 also regulated sunitinib resistance in 786-O R cells via the AKT–mTOR signaling pathway (Fig. 4G and H). Moreover, the nude mouse xenograft assay indicated that the combination of everolimus enhanced the anti-growth effect of sunitinib, similar to the effect seen in Fig. 1L after KO of ZDHHC2 (Fig. 4I–K). We also noted that depletion of ZDHHC2 attenuated combined antitumor effect of everolimus and sunitinib compared with everolimus and sunitinib treatment alone (Fig. 4I–K). Therefore, these data suggest that the AKT–mTOR signaling pathway mediates ZDHHC2-induced sunitinib resistance in ccRCC.

Figure 4.

Everolimus treatment overcomes ZDHHC2-induced sunitinib resistance in ccRCC. A–D, A498 cells were transfected with indicated constructs. After 48 hours, cells were treated with or without everolimus (10 μmol/L) for 24 hours. The cell medium was collected for in vitro angiogenesis assay (A and C) or detecting the expression level of VEGFA by enzyme-linked immunosorbent assay (B and D). Data presented as mean ± SEM with three replicates. E and F, 786-O and A498 cells were transfected with indicated constructs for 24 hours. Then, these cells were treated with or without everolimus (5 μmol/L) for 24 hours. Cells were treated with a serial dose of sunitinib and harvested for CCK-8 assay, and the IC₅₀ values of sunitinib were measured. G, 786-O R cells were transfected with the indicated constructs for 72 hours. After puromycin selection, these cells were treated with or without everolimus (5 μmol/L) for 24 hours. Cells were treated with a serial dose of sunitinib and harvested for CCK-8 assay to measure the IC₅₀ values of sunitinib. H, 786-O R cells were transfected with the indicated constructs for 24 hours. These cells were treated with or without everolimus (5 μmol/L) for 24 hours. Cells were treated with a serial dose of sunitinib and harvested for CCK-8 assay to measure the IC₅₀ values of sunitinib. I–K, 786-O cells were transfected with indicated constructs for 72 hours. These cells were subjected to nude mice xenograft assay. These mice were treated with vehicle, sunitinib alone, everolimus alone, or sunitinib plus everolimus. Tumor image (I); tumor mass (J); tumor growth curve (K). Data presented as mean ± SEM with six replicates. ns, not significant; ***, P < 0.001.

ZDHHC2 binds with AGK to activate the AKT–mTOR signaling pathway in ccRCC

Subsequently, we wondered how ZDHHC2 activates the AKT–mTOR signaling pathway in ccRCC. As ZDHHC2 is a PAT, it might participate in the palmitoylation of some proteins to regulate cellular processes. MS analysis using anti-ZDHHC2 or IgG antibodies was performed in 786-O cells (Fig. 5A). We found that AGK was the top ranked protein in the MS results from the anti-ZDHHC2 antibody condition (Supplementary Table S3; Fig. 5B). Coimmunoprecipitation, immunofluorescence and GST-pull down assays indicated that ectopically overexpressed AGK and ZDHHC2 interacted with each other in 293T cells, and endogenously expressed AGK and ZDHHC2 also bound with each other in 786-O, A498 and ACHN cells (Fig. 5C–F; Supplementary Fig. S5A). AGK has been reported to contributes to activating the AKT signaling pathway in renal cancer cells (45). We wondered whether AGK was the key mediator of ZDHHC2-induced activation of the AKT pathway in renal cancer cells. We analyzed the expression of AGK in ccRCC tissues and cell lines by analyzing the Gene Expression Profiling Interactive Analysis, Encyclopedia of RNA Interactomes, and Cancer Cell Line Encyclopedia web tools (Supplementary Fig. S5B–S5D). We found that there was no difference between ccRCC tumor tissue and nontumor tissues (Supplementary Fig. S5B and S5C). Moreover, we also showed that AGK was expressed in most renal cancer cell lines (Supplementary Fig. S5D). We showed that depletion of AGK decreased pAKT S473, pAKT T308, and pS6K1 T389 in A498 and ACHN cells (Fig. 5G). Then, we also demonstrated that KO of AGK diminished the effect of ZDHHC2 on the changes in the phosphorylation of AKT and S6K1 in ACHN and A498 cells (Fig. 5H and I). Thus, these data suggest that AGK might be responsible for the ZDHHC2 induced activation of the AKT–mTOR pathway in ccRCC.

Figure 5.

ZDHHC2 binds with AGK to activate the AKT–mTOR signaling pathway in ccRCC. A, The silver staining of immunoprecipitation by using the ZDHHC2/AGK or IgG antibodies. B, The peptide information of ZDHHC2 and AGK in the MS of ZDHHC2. C, Immunoprecipitation analysis of the cell lysates of 293T cell transfected with indicated plasmids, 786-O, or A498 cells by using the ZDHHC2 and AGK antibodies. D, Immunofluorescence assay by staining the ZDHHC2 and AGK in 786-O and A498 cells. DAPI, nucleus. DiD dye stain, PM. Scale bar is indicated in the panel. E, Western blotting analysis of ZDHHC2 GST-pulled down by AGK recombinant, or AGK GST-pulled down by ZDHHC2 recombinant. F, Bacterially expressed His-AGK and GST, GST-ZDHHC2 recombinant proteins were subjected to in vitro protein binding assay, followed by Western blot analysis. Input samples were analyzed by Coomassie blue staining. *, recombinant protein. G–I, Western blot analysis of the whole cell lysates of A498 and ACHN after transfection with indicated constructs for 72 hours. EV, empty vector.

AGK is responsible for ZDHHC2-induced sunitinib resistance in ccRCC

Of note, analysis of the CancerSEA dataset indicated that AGK was positively correlated with angiogenesis and hypoxia in RCC (Fig. 6A). The subsequent tubule formation assay showed that KO of AGK decreased the angiogenesis capability, but overexpression of WT AGK but not a kinase-dead mutant (AGK G126E; ref. 46) promoted the angiogenesis of HUVECs (Fig. 6B and C; Supplementary Fig. S5E). Given that EMT, angiogenesis and hypoxia are involved in promoting TKI resistance in ccRCC (47, 48), we evaluated whether AGK is the key mediator of ZDHHC2-induced TKI resistance in ccRCC. First, we showed that inhibiting the expression of AGK decreased the IC₅₀ values of sunitinib in 786-O, A498, and ACHN cells (Fig. 6D; Supplementary Fig. S5F). However, overexpression of WT AGK but not a kinase-dead mutant (AGK G126E) increased the IC₅₀ values of sunitinib in ccRCC cells (Fig. 6E). Subsequently, CCK-8 and xenograft assays indicated that AGK depletion enhanced the anti-proliferation efficacy of sunitinib in 786-O, 786-O R, and A498 cells (Fig. 6F–I). Moreover, we showed that downregulation of AGK diminished the changes in the IC₅₀ values of sunitinib seen in ccRCC cells with overexpression of ZDHHC2 or coknockdown of ZDHHC2 (Fig. 6J and K; Supplementary Fig. S5G–S5J). Thus, these data indicated that ZDHHC2 modulates sensitivity to sunitinib by binding with AGK.

Figure 6.

AGK is responsible for ZDHHC2-induced sunitinib resistance in ccRCC. A, The cancer-related role of AGK was analyzed by the CancerSEA datasets. B and C, A498 cells were transfected with indicated constructs. After 48 hours, the cell medium was collected for in vitro angiogenesis assay. Data presented as mean ± SEM with three replicates. D and E, 786-O and A498 cells were transfected with indicated constructs for 72 hours. Cells were treated with a serial dose of sunitinib and subjected to CCK-8 assay. The IC₅₀ of values of sunitinib is indicated. F, 786-O, 786-O R, and A498 cells were transfected with indicated constructs for 72 hours. After puromycin selection, cells were treated with or without sunitinib (2 μmol/L) and subjected to CCK-8 assay. Data presented as mean ± SEM with three replicates. G–I, 786-O cells were transfected with indicated constructs for 72 hours. After puromycin selection, these cells were subcutaneously injected into the nude mice and the mice were treated with or without sunitinib. Tumor mass (H); tumor growth curve (I). Data presented as mean ± SEM with six replicates. J and K, 786-O and A498 cells were transfected with indicated constructs for 48 hours. Then, these cells were treated with a serial dose of sunitinib and harvested for CCK-8 assay to measure the IC₅₀ values of sunitinib. *, P < 0.05; ***, P < 0.001. EV, empty vector.

ZDHHC2 promotes AGK S-palmitoylation in renal cancer cells

Because ZDHHC2 is a PAT, we next examined whether AGK was palmitoylated by ZDHHC2 in ccRCC cells. 786-O and A498 cells were transfected with HA-tagged AGK and subsequently analyzed with the acyl-biotinyl exchange (ABE) technique with biotin-HPDP. The streptavidin blot showed that AGK was palmitoylated in human renal clear cell carcinoma cells (Fig. 7A). Biotin alkyne was used to label palmitoylated proteins by the click-iT reaction. Overexpression of AGK increased palmitoylated proteins, but the palmitoylation levels of AGK were dramatically decreased after treatment with hydroxylamine (Fig. 7B). In addition, we detected that endogenous AGK could also be palmitoylated (Fig. 7C) and approximately 20% of AGK could be palmitoylated in 786-O and A498 cells (Supplementary Fig. S6A). The above results indicated that there was an interaction between AGK and ZDHHC2, but whether AGK can be palmitoylated by ZDHHC2 remains unknown. We constructed a set of 786-O cell lines with ZDHHC KO with a CRISPR screen. The results showed that knocking out ZDHHC2 abolished AGK palmitoylation (Fig. 7D), which indicated that ZDHHC2 functions as a PAT of AGK. Then, ZDHHC2 was knocked out or overexpressed in 786-O and A498 cells, and the palmitoylation level of AGK also decreased or increased, respectively (Fig. 7E and F). In addition, the catalytically inactive ZDHHC2-C129A mutant failed to change the palmitoylation level of AGK compared with ZDHHC2-WT (Fig. 7G). To further validate our results, we reintroduced ZDHHC2-WT and the ZDHHC2-C129A mutant into ZDHHC2 KO cells (Fig. 7H), and the results showed that AGK could be palmitoylated by ZDHHC-WT but not by the catalytically inactive ZDHHC2-C129A mutant (Fig. 7I). To determine the site of AGK palmitoylation, CSS-Palm 4.0 software was used to predict the palmitoylation site (Fig. 7J), and all potential sites were mutated from cysteine (C) to serine (S; Fig. 7K). According to the literature, the purpose of eliminating palmitoylation modification could be achieved by mutating cysteine to serine (34, 49) or alanine (A; refs. 50, 51). Therefore, cysteine was mutated to serine to eliminate the palmitoylation modification of AGK in our study. We found that only the AGK-C72S mutant prevented AGK palmitoylation (Fig. 7K), and this site was highly conserved among different species (Fig. 7L). To eliminate the effects of endogenous AGK interference on the experimental results, we reintroduced AGK-WT and the AGK-C72S mutant into AGK KO cells and showed that exogenous AGK-WT could be palmitoylated, while the AGK-C72S mutant could not (Fig. 7M and N). Moreover, we performed an in vitro palmitoylation assay in the presence of palmitoyl alkyne-CoA as the palmitate donor. The results showed that AGK-WT but not the AGK-C72S mutant was palmitoylated only by ZDHHC2-WT (catalytically active), while the catalytically inactive ZDHHC2-C129A mutant did not induce palmitoylation (Fig. 7O). In summary, ZDHHC2 promotes AGK-C72 S-palmitoylation through its enzymatic activity.

Figure 7.

ZDHHC2 promotes AGK S-palmitoylation in renal cancer cells. A, 786-O and A498 cells were transfected with empty vector (EV) or HA-AGK plasmids for 72 hours. ABE assay was used to detect the palmitoylation of AGK. B, 786-O and A498 cells were transfected with empty vector or HA-AGK plasmids for 72 hours. Click-iT pull down assay was used to detect the palmitoylation of AGK. C, ABE assay was used to detect the palmitoylation level of endogenous AGK in 786-O and A498 cells. D, 786-O cells were infected with lentiviruses containing expression cassettes of control guide RNA or guide RNAs targeting different ZDHHCs and Cas9 protein. ABE assay was used to detect the palmitoylation of AGK. E, 786-O and A498 cells were infected with lentiviruses containing expression cassettes of control guide RNA or guide RNAs targeting ZDHHC2 and Cas9 protein. Then AGK was labeled with Biotin-HPDP for streptavidin pull down to analyze the palmitoylation levels of AGK. F, 786-O and A498 cells were transfected with empty vector or Flag-ZDHHC2 plasmids for 72 hours. Then AGK was labeled with Biotin-HPDP for streptavidin pull down to analyze the palmitoylation levels of AGK. G, 786-O and A498 cells were transfected with indicated plasmids for 48 hours. Then the palmitoylation levels of AGK were detected by ABE assay. H and I, Endogenous ZDHHC2 were knocked out in 786-O and A498 cells by CRISPR/Cas9. Then 786-O and A498 cells of ZDHHC2-KO were transfected with indicated plasmids for 72 hours for immunoblotting analysis. ABE assay was used to detect the palmitoylation level of AGK. J, The potential palmitoylation sites of human AGK protein were predicted through CSS-Palm 4.0 software. K, 786-O cells were transfected with the indicated plasmids for 48 hours. ABE assay was used to detect the palmitoylation level of AGK. L, Alignment of the similarity and identity of protein sequences in different AGK vertebrate orthologs. M and N, Endogenous AGK were knocked out in 786-O and A498 cells by CRISPR/Cas9. Then 786-O and A498 cells of AGK-KO were transfected with indicated plasmids for 48h for immunoblotting analysis. ABE assay was used to detect the palmitoylation level of AGK. O,In vitro palmitoylation analysis was performed by mixing purified His-AGK-WT and His-AGK-C72S with purified GST-ZDHHC2-WT or GST-ZDHHC2-C129A in the presence of palmitoyl alkyne-CoA. ABE assay was used to detect the palmitoylation level of AGK.

The S-palmitoylation of AGK induced by ZDHHC2 promotes AGK PM localization and activates the AKT–mTOR signaling pathway in ccRCC

It has been documented that AGK leads to activation of AKT–mTOR signaling by accumulating in the PM (19). The S-palmitoylation of proteins is reported to regulate protein PM localization (34). Thus, we hypothesized that ZDHHC2 palmitoylates AGK to maintain AGK in the PM. Subcellular fractions were isolated with a PM protein extraction kit and immunoblotting analysis showed that there was no cross-contamination among every subcellular fraction (Supplementary Fig. S6B). Unsurprisingly, silencing ZDHHC2 significantly reduced AGK PM localization in both 786-O and A498 cells (Fig. 8A–C), and the same effect was also observed in 786-O R cells (Supplementary Fig. S6C and S6D). In contrast, ectopic re-expression of ZDHHC2-WT but not the C129 mutant in ZDHHC2 KO cells increased AGK PM localization (Fig. 8D and E). Moreover, we found that AGK-WT was located in the PM and cytoplasm, but AGK-C72S was less located in the PM than AGK-WT in both 786-O and A498 cells (Fig. 8F–H). These data indicate that S-palmitoylation of AGK mediated by ZDHHC2 is involved in the PM localization of AGK. In addition, we showed that overexpression of ZDHHC2 WT but not the C129A mutant increased the phosphorylation of AKT at Ser473 and S6K1 at Thr389 in 786-O and A498 cells (Fig. 8I). Our results also demonstrated that overexpression of AGK-WT but not the C72S mutant resulted in upregulation of AKT and S6K1 phosphorylation and decreased ccRCC sensitivity to sunitinib (Fig. 8J and K). Taken together, our data suggest that ZDHHC2-mediated AGK S-palmitoylation activates the AKT–mTOR signaling pathway to reduce the sensitivity of ccRCC to sunitinib (Fig. 8L).

Figure 8.

The S-palmitoylation of AGK induced by ZDHHC2 promotes AGK PM localization and activates the AKT–mTOR signaling pathway in ccRCC. A and B, 786-O and A498 cells were transfected with indicated siRNAs for 48 hours. The subcellular fraction was extracted and subjected to Western blot analysis (A). The protein level was quantitated by using the imageJ software (B). Data presented as mean ± SEM with three replicates. WCL, whole cell lysate. C, 786-O cells were transfected with indicated siRNAs for 48 hours. Immunofluorescence assay was used to determine the subcellular location of AGK. DiD dye, PM; DAPI, nucleus. D and E, 786-O and A498 cells were transfected with indicated plasmids for 24 hours. The subcellular fraction was extracted and subjected to Western blot analysis (D). The protein level was quantitated by using the imageJ software (E). Data presented as mean ± SEM with three replicates. F and G, 786-O and A498 cells were transfected with indicated plasmids for 24 hours. The subcellular fraction was extracted and subjected to Western blot analysis (F). The protein level was quantitated by using the imageJ software (G). Data presented as mean ± SEM with three replicates. H, 786-O and A498 cells were transfected with HA-AGK WT or HA-AGK C72S for 24 hours. Immunofluorescence assay by staining the HA in 786-O and A498 cells. Scale bar is indicated. I and J, Western blot analysis of the whole cell lysates of 786-O and A498 cells after transfection with indicated plamids for 48 hours. K, 786-O and A498 cells were transfected with indicated plasmids for 24 hours. Cells were treated with a serial dose of sunitinib and subjected to measuring the IC₅₀ values by CCK-8 assay. L, A model depicting that ZDHHC2-mediated AGK S-palmitoylation promotes AGK PM localization and activates the AKT–mTOR signaling pathway to reduce the sensitivity of sunitinib in ccRCC. ns, not significant; **, P < 0.01; ***, P < 0.001. EV, empty vector. (L, Created with BioRender.com.)

Discussion

Hyperactivation of the PI3K–AKT–mTOR signaling pathway is one of the important reasons for modulating sensitivity of cancer to TKIs (14, 32, 52–54). A previous study demonstrated that AGK phosphorylates PM-localized PTEN and activates the AKT–mTOR signaling pathway in CD8⁺ T-cells but not in CD4⁺ T cells (19). This finding indicated that the PM localization of the AGK is crucial for modulating the PI3K–AKT–mTOR signaling pathway. Here, we showed that palmitoylation of AGK Cys72 maintains AGK PM localization and subsequently activates the AKT–mTOR pathway in ccRCC cells. Our results highlight the molecular mechanism by which AGK PM localization is regulated by protein posttranslational modification.

Protein S-palmitoylation mediated by ZDHHCs in human cancer cells has been proven to modulate protein membrane localization. Recently, Zhang and colleagues reported that DHHC9 palmitoylates GLUT1 to maintain GLUT1 PM localization and promote glioblastoma glycolysis (34). DHHC7 has been reported to palmitoylate STAT3 and promote its membrane recruitment and Th17 differentiation (55). Moreover, palmitoylation of one or two cysteine residues of N-Ras and H-Ras is required for their membrane association to activate the downstream pathway (56). On the other hand, the palmitoylation of EGFR leads to its nuclear translocation and promotes tumor growth in lung cancer cells (57). These findings demonstrate that protein trafficking mediated by S-palmitoylation is important for modulating the inflammatory response and tumor growth. Consistent with previous findings, our study showed that ZDHHC2 was abnormally upregulated in sunitinib-resistant ccRCC PDX samples or cells. Subsequently, our results revealed that ZDHHC2 palmitoylated AGK to promote PM localization, which resulted in reduced sunitinib sensitivity in ccRCC cells. Thus, our findings reveal the function of ZDHHC2 in regulating TKI sensitivity, expanding the role of protein S-palmitoylation in tumors.

TKIs, including sunitinib, are approved for the treatment of patients with advanced ccRCC (58). While, almost all patients develop resistance to sunitinib in 2 years (58). Overcoming resistance to antiangiogenic drugs helps improve the prognosis of patients with ccRCC (59, 60). To explore the underlying mechanism, we studied molecular mechanisms that can collectively influence the sensitivity of ccRCC to sunitinib and the resistance of sunitinib-resistant ccRCC cells. We mentioned that ZDHHC2 led to reduced sunitinib sensitivity mainly through the AGK–AKT–mTOR axis, and this finding was also verified in 786-O R cells. Apart from this, we believe that ZDHHC2- reduced sunitinib sensitivity in ccRCC cells or induced sunitinib resistance in 786-O R cells and in patients with ccRCC is not only reliant on AGK. There must be other candidates contributing to these cellular processes. Thus, studies identifying novel target proteins of ZDHHC2 and elucidating a new role of ZDHHC2 in ccRCC are needed in the future.

Inhibition of the AKT–mTOR signaling pathway has bene proven to overcome sunitinib resistance in ccRCC (14, 16). The combinations of multitargeted TKIs with mTOR inhibitors had additive or synergistic effects on the inhibition of renal cancer cell proliferation and tumor angiogenesis (42). Clinical trials have also demonstrated that the sequential use of multitargeted TKIs and mTOR inhibitors is effective for advanced ccRCC patients (41, 61). Given that ZDHHC2 activated the AKT–mTOR pathway in a manner dependent on its PAT activity, and an unselective palmitoylation inhibitor (2-BP) was proven to enhance sunitinib activity in this study. Our data indicated that targeting the activity of ZDHHC2 could further inhibit the AKT signaling axis, thereby improving the sensitivity of ccRCC to TKIs and mTOR inhibitors. Moreover, a variety of factors confer resistance to mTOR inhibitors in renal cancer, such as HIF-mediated suppression of DEPTOR (62), aberrant expression of PTEN (63), and decreased DNMT1 enzyme activity (64). After long-term use of mTOR inhibitors leads to drug resistance, targeting ZDHHC2 can be used as a complementary and alternative therapy in ccRCC. Therefore, searching for specific inhibitors of ZDHHC2 can provide ideas for new drug development for ccRCC.

Collectively, our results demonstrate that ZDHHC2 is abnormally upregulated in sunitinib-resistant ccRCC cell lines and ccRCC tissue samples with TKI resistance. Furthermore, ZDHHC2 was found to modulate the sunitinib sensitivity of ccRCC cells. Further study showed that ZDHHC2 is associated with the regulation of cell proliferation, EMT, and angiogenesis in ccRCC. Mechanistically, we revealed that ZDHHC2 palmitoylates AGK to modulate the sensitivity of sunitinib by activating the AKT–mTOR signaling pathway in ccRCC cells. Taken together, these findings reveal a novel ZDHHC2–AGK axis that is responsible for the modulation of sunitinib sensitivity in ccRCC cells. Our results also suggest that ZDHHC2 is a targetable candidate for improving the antitumor efficacy of TKIs, especially sunitinib, in ccRCC.

Consent for publication

All subjects have written informed consent.

Supplementary Material

Supplementary Data

Supplementary Data

Authors' Disclosures

No disclosures were reported.

Authors' Contributions

Y. Sun: Methodology. L. Zhu: Methodology. P. Liu: Methodology. H. Zhang: Formal analysis. F. Guo: Conceptualization, formal analysis. X. Jin: Funding acquisition, investigation, methodology, project administration, writing–review and editing.