Role of ATP citrate lyase and its complementary partner on fatty acid synthesis in gastric cancer

Abstract

ATP citrate lyase (ACLY) and acyl-CoA short-chain synthetases 2 (ACSS2) are key enzymes in lipid metabolism. We explored the role of ACLY in gastric cancer (GC) and the effect of ACLY and ACSS2 compensation on GC growth. We used immunohistochemistry to verify the expression level of ACLY in GC, shRNA to stably knock down the expression level of ACLY in GC cells. The expression levels of lipid metabolizing enzymes were verified by qPCR and WB, and targeted lipidomics and quantification of lipid metabolism-related indicators helped us to understand the changes in lipid metabolism. Finally, subcutaneous graft tumors validate our findings from in vitro experiments. ACLY is upregulated in GC tissues, downregulation of ACLY reduced lipid accumulation and inhibited GC proliferation, migration, and invasion in vitro. ACSS2 maintains cell growth by compensatory elevation to maintain fatty acid synthesis activity in ACLY-depleted GC cells. Inhibition of ACSS2 enhanced the inhibitory effect of downregulation of ACLY on the growth of transplanted tumors in nude mice. Downregulation of ACLY inhibited GC cell growth in vitro and in vivo. ACSS2 was compensated to increase to maintain cell growth in ACLY-depleted GC cells.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-024-81448-1.

Introduction

Gastric cancer (GC), a highly heterogeneous and aggressive disease, is the second leading cause of cancer death worldwide¹. In several countries, the incidence of stomach cancer among young individuals has significantly risen in recent years^2,3. Despite the availability of systemic chemotherapy, radiation, surgery, immunotherapy, and targeted therapy, GC continues to have high morbidity and fatality rates⁴. There is an urgent need to understand the molecular mechanisms of GC in depth and to offer novel effective therapy options to supplement or augment existing therapeutic approaches.

Lipid metabolism is a biological process essential for the maintenance of normal life activities in living creatures^5,6. It occurs alongside the processes of several energy and chemical conversions. Previous research has established that cancer alters the components of lipid metabolism and this reprogramming of lipid metabolism has become a hallmark of cancer⁷. The reprogramming of lipid metabolism not only supplies the energy required for the rapid proliferation of cancer cells but also promotes carcinogenesis and cancer development⁸. Fatty acid (FA) metabolism is a critical component of lipid metabolism. FAs are the major components of a wide range of lipids (including phospholipids, sphingolipids, and TGs) that are made up of carboxylic acid groups and hydrocarbon chains with varying carbon lengths and desaturation⁹. FAs are found in several cellular structures and they regulate numerous biochemical processes in normal cells, including the formation and regulation of biofilm fluidity and signaling pathways. They operate as secondary messengers in signaling pathways and as an energy store in animals^10,11. Current investigations on lipid metabolism in GC have concentrated on FA metabolism, but their contribution to guiding GC treatment is lacking to date.

FA metabolism primarily includes the synthesis, desaturation, and oxidation of FAs. The FA synthesis in living organisms is performed via the cytosolic enzyme synthesis and mitochondrial enzyme synthesis pathways, of which the cytosolic enzyme synthesis pathway is the primary one. The main end product of this pathway is palmitic acid (PA); hence, it is also called the PA synthesis pathway^12,13. The cytosolic enzyme synthesis pathway is essentially involved in the ab initio synthesis of FAs, and the mitochondrial enzyme synthesis pathway is primarily involved in the extension of FAs. Sterolregulatory element binding protein 1 (SREBP1), a transcription factor of FA synthesis-related genes, regulates FA synthesis by mediating the expression of ATP citrate lyase (ACLY), acetyl-coenzyme A short-chain synthetase 2 (ACSS2), acetyl-coenzyme A carboxylase 1 (ACC1), and fatty acid synthase (FASN)⁹.

ACLY is a 1101-residue polypeptide forming a functional 0.5-MDa tetramer¹⁴. ACLY is found in the endoplasmic reticulum and it largely catalyzes the synthesis of the nucleus and plasma acetyl-coenzyme A (acetyl-CoA), which connects glucose and FA metabolism¹⁵. The specific job of ACLY as the first major enzyme in ab initio FA anabolism is to catalyze the conversion of citric acid, which is produced via the mitochondrial tricarboxylic acid (TCA) cycle and transported to the cytoplasm, into acetyl-CoA and oxaloacetate (OAA)⁹. Acetyl-CoA is implicated in lipid synthesis as a substrate for FA synthesis. Because ACLY is involved in the conversion of acetyl groups, it is also implicated in vivo activities such as histone acetylation. ACLY increases in several cancers and it plays a crucial role in malignant processes such as tumor growth, invasion, and apoptosis^16–27.

ACSS2, a member of the acyl-CoA short-chain synthetase family, is a nuclear-cytoplasmic enzyme that transforms intracellular free acetate to acetyl-CoA. ACSS2 creates a parallel but distinct pathway for ACLY in cytoplasmic and nuclear acetyl-CoA production because they both yield the same result product²⁸. This reduces the need for acetyl-CoA generated from mitochondrial metabolism²⁹. ACSS2 is overexpressed in hypoxic tumor areas and is necessary for in vivo cancer cell proliferation and survival³⁰. It can be increased at the mRNA level in ACLY-deficient cancer cells, possibly compensating for the loss of ACLY³¹.

To investigate whether ACLY influences GC progression by regulating lipid metabolism, we used the Gene Expression Omnibus (GEO) and Cancer Cell Line Encyclopedia (CCLE) databases, as well as a series of in vivo and in vitro experiments, and targeted metabolomics. Our findings indicate that ACLY promotes the expression of lipid metabolism genes and lipid accumulation in GC, which in turn enhances the proliferation, migration, and invasion of GC. In conclusion, the results of the current study demonstrate that ACLY enhances GC proliferation, migration, and invasion, and ACSS2 is compensatorily raised in ACLY-depleted GC cells. Furthermore, our study reveals that the simultaneous inhibition of ACLY and ACSS2 may be a viable therapeutic method for arresting GC growth.

Materials and methods

Cell culture

Human GC cell lines HGC-27 and AGS were obtained from the Kunming Cell Bank, Chinese Academy of Sciences. HGC-27 cell lines were cultured in 1640 medium (G4534, Servicebio, Wuhan, China) containing 10% fetal bovine serum (FBS) (mu001SR, QmSuero, Wuhan, China) and 100 U/mL penicillin and 100 μg/mL streptomycin (BL505A, Biosharp, Wuhan, China). AGS cells were cultured in F12 medium (CM-0022, Procell, Wuhan, China) containing 10% FBS and 100 U/mL penicillin and 100 μg/mL streptomycin. All cell lines were maintained in an incubator containing 5% CO₂ at 37 °C.

Stable knockdown cell lines

8 × 10⁵ GC cells were grown in 6-well plates, and when the cells were plastered and reached about 50% density the following day, lentivirus (Genechem Co., Ltd., Shanghai, China) containing target gene-specific shRNA was added to culture wells along with transfection reagent (REVG005, Genechem Co., Ltd., Shanghai, China). The amount of lentivirus was calculated based on the optimal MOI (10) for the pre-experiment. After 12 h of viral transduction, the medium was changed back to normal and the culture was continued until 48–72 h. When the cells were observed to express adequate fluorescence under the fluorescence microscope (IX71, Olympus, Tokyo, Japan), the GC cells were screened for 48 h using puromycin (BL528A, Biosharp, Wuhan, China) at a concentration of 2 μg/mL. The knockdown effect was confirmed using real-time quantitative polymerase chain reaction (RT-qPCR) and western blot (WB). Table S1 displays all shRNA sequences used to construct the stable-transformation cell lines.

CCK8 assay

CCK8 assay was used to assess the short-term proliferative viability of the cells. 5 × 10³ GC cells were seeded in 96-well plates and five replicate wells were set up for each group. In this study, cell viability was measured at 24, 48, and 72 h: 100 μL of the medium was replaced and 10 μL of CCK8 reagent (BS350B, Biosharp, Wuhan, China) was added to each well. The cells were incubated for 1 h in the incubator and the optical density (OD) at 450 nm was measured using an enzyme marker (PerkinElmer, Waltham, MA, USA). The experiment was repeated thrice.

Colony formation assay

The colony formation assay was used to assess the long-term proliferative viability of the cells. 500 GC cells were uniformly distributed into 6-well plates and grown in normal media for 1–2 weeks. When the cell clusters containing more than 30 cells were visible under a microscope, the cells were fixed using 4% paraformaldehyde (BL539A, Biosharp, Wuhan, China) for 30 min, followed by staining with 0.1% crystal violet solution (G1014, Servicebio, Wuhan, China) for 20 min. Finally, a general view of each well was imaged with a cell phone against a white background. ImageJ software (Version 1.53 k, Wayne Rasband, National Institutes of Health, USA) was used to quantify the number of cell colonies. The experiment was repeated thrice.

Wound healing assay

The wound healing assay was used to assess lateral cell migration. 1.5 × 10⁶ GC cells were evenly seeded in 6-well plates, and the next day the cells were adherent and at 100% density. A straight line was drawn along a straightedge with a 200 μL pipette tip on the median side of the cells, followed by washing away the free cells with phosphate-buffered saline (PBS). Finally, 2 mL of medium containing 2% FBS was added and images were recorded using an inverted microscope (IX51, Olympus, Tokyo, Japan) at 40 × magnification of the same position at 0 and 48 h after drawing a line. The experiment was repeated thrice.

Transwell assay

Cell migration and invasion were assessed using transwell assays. The first step of the experiment was to confirm the addition of matrix gel (BD Biosciences, San Jose, CA, USA), which was added to detect cell invasion ability and without matrix gel to detect cell migration. The matrix gel was uniformly added to the upper layer of the 24-well chambers with 8 μm pore size (Corning, NY, USA) at a dilution ratio of 1 to 8 and the matrix gel was solidified 30 min in advance at 37 °C. Then, 100 μL of serum-free FBS medium containing 2 × 10⁴ cells to the upper layer of the chamber and 600 μL of medium containing 10% FBS to the lower layer. After 96 h in the cell incubator, the chambers containing cells were fixed with 4% paraformaldehyde for 30 min, followed by staining with 0.1% crystal violet solution for 20 min. Thereafter, the cells were photographed with an inverted microscope (IX71, Olympus, Tokyo, Japan) at 200 × magnification. ImageJ software was used to count the number of cells. The experiment was repeated thrice.

RNA extraction, reverse transcription, and RT-qPCR analysis

The total RNA solution was obtained from cells by lysing in RNA extraction reagent (G3013, Servicebio, Wuhan, China), followed by precipitation with chloroform and isopropanol, and subsequent washing with 75% ethanol. The concentration and purity of the extracted RNA using NanoDrop one (Thermo Fisher Scientific, Waltham, MA, USA). To acquire cDNA, a 20 μL reverse transcription reaction was performed using a reverse transcription kit (G3337, Servicebio, Wuhan, China): 4 μL reverse transcription mix + 1 μL gDNA remover + 2 μg total RNA + DEPC water. Reverse transcription experiments were performed using a C1000 PCR apparatus (Bio-Rad, Hercules, CA, USA). CFX96TM real-time system (Bio-Rad, Hercules, CA, USA) for performing qPCR (G3327, Servicebio, Wuhan, China) procedure with a reaction volume of 20 µL (10 µL qPCR mix + 0.4 µL 10 μM forward primer + 0.4 µL 10 μM reverse primer + 1 µg cDNA + DEPC water). The comparative cycle threshold (Ct) approach (2^-ΔΔCt) was used to assess mRNA expression levels. The experiment was repeated thrice. Table S2 displays all primer sequences.

Protein extraction and WB analysis

Total cellular protein was extracted using RIPA lysis buffer (G2002, Servicebio, Wuhan, China) and the obtained protein was quantified using the BCA method (G2026, Servicebio, Wuhan, China). SDS-PAGE was used to separate the individual proteins from the mixture. Thereafter, the separated proteins were then transferred onto nitrocellulose (NC) membranes (Millipore, MA, USA). The membranes were then dipped in a protein-free rapid closure solution (G2052, Servicebio, Wuhan, China) for 1 h before being treated with particular primary antibodies for 16 h at 4 ℃. Thereafter, the membranes were treated with HRP-labeled antibodies for 1 h at 20–25 ℃. Then, the Bio-Rad GelDoc system was used to acquire WB images and identify antibody-bound proteins using ECL luminous chemistry reagent (PMK003, Bioprimacy, Wuhan, China). ImageJ software was used for semiquantitative analysis of the WB images. The experiment was repeated thrice and Table S3 lists the antibody information for ACLY, SREBP1, ACSS2,  ACC1, FASN, and β-actin.

Immunohistochemistry staining

Tumor and para-cancer tissues were collected from 20 patients with stomach adenocarcinoma at the People’s Hospital of Wuhan University in Wuhan, China. Six different transplanted tumor tissues were collected from each group of nude mice in the animal experiments. In brief, paraffin sections were baked, dewaxed, and hydrated, and then the antigen was recovered using sodium citrate buffer. After cooling, the slides were treated with 3% hydrogen peroxide solution for 15 min, followed by a 60-min treatment with 3% Albumin Bovine V (BSA) solution. Primary antibodies were incubated with the sections overnight at 4 °C, followed by secondary antibodies for 15 min at 20–25 ℃. Using 3,5-diaminobenzidine, the staining outcomes were visualized. Three random field images of each slide were obtained using a biological microscope (BX53, Olympus, Tokyo, Japan) with 200 × magnifications. Using ImageJ software, the immunostained slices were quantified as integrated optical density/area (AOD). Table S3 lists the antibody information for ACLY and Ki-67.

Quantification of triglycerides(TG)

TG levels in the GC cell samples were determined using a TG assay kit (single reagent GPO-PAP technique) (A110-1–1, Jiancheng Technology Co., Nanjing, China). According to the manufacturer’s instructions. 2.5 μL of double-distilled water or the standard or the sample was added either to be measured to 250 μL of working solution. Both were mixed thoroughly and then incubated for 10 min at 37 ℃. Thereafter, the OD value at 510 nm was measured with an enzyme marker. The TG content = [(sample value – blank value)/(calibration value – blank value)] × calibration product concentration/sample protein concentration to be tested. Three replicates of each sample in three wells were tested.

Oil red O staining

Cells in 6-well plates were fixed with 4% paraformaldehyde for 30 min, rinsed with 60% isopropanol for 5 min, and then stained for 2 h at 37 °C with a freshly made 0.3% Oil Red O working solution. The Oil Red O saturated stock solution was obtained from Solarbio, Beijing, China. The resultant cells were restained with hematoxylin (G1004, Servicebio, Wuhan, China) for 1 s, followed by rinsing with water. One milliliter of double distilled water was retained in each well. Random images (400 ×) were recorded using an Olympus IX71 inverted microscope, and quantified using Image-Pro Plus software (Version 6.0, Media Cybernetics, Silver Springs, MD, USA). The quantification metric is the number of lipid droplets per field of view divided by the number of cells.

Quantification of acetyl coenzyme A (acetyl-CoA) and malonyl coenzyme A (malonyl-CoA)

An acetyl-CoA ELISA kit (ED12805, LunChangShuoBiotech, Xiamen, China) and a malonyl-CoA ELISA kit (ED13193, LunChangShuoBiotech, Xiamen, China) were used to determine the acetyl-CoA and malonyl-CoA content of GC cell samples, respectively. The experimental procedure was conducted following the manufacturer’s protocol. Briefly, GC cells were repeatedly frozen and thawed with cooled PBS and the supernatant was extracted by centrifugation as the assay sample. During the experiment, acetyl-CoA or malonyl-CoA in the sample or standard bound to the antibody coated on the enzyme plate. After adding the HRP-labeled antibody, the antibody-antigen-enzyme-labeled antibody complex was formed. On adding the chromogenic substrate TMB, TMB appeared blue owing to the catalysis of HRP and turned yellow after the termination solution was added. Simultaneously, the OD value at 450 nm was measured by the enzyme standardization instrument, and the concentration of acetyl-CoA or malonyl-CoA of the sample were calculated from the standard curve.

Targeted lipidomics

Bioprofile Biotechnology Co., Ltd (Shanghai, China) performed absolute quantification of targeted medium and long-chain FAs on HGC-27 cell samples.

1. Preparation of standard solutions: A mixed standard stock solutions of 51 FAs (4000 μg/mL) were diluted into ten points. The calibration curve was made by adding the N-hexane covering a range from 1 to 2000 μg/mL.

2. Sample preparation³²: Chloroform: methanol (2:1) solution and glass beads were added to the sample tube, followed by thorough lysis using a high-throughput tissue grinder. The supernatant was collected and subjected to esterification with an 80℃ water bath and extraction with n-hexane after shaking with 1% methanol sulfuric acid solution. The required sample was obtained after washing with 4℃ water and drying with anhydrous sodium sulfate powder. After diluting the sample and adding methyl salicylate as an internal standard, the appropriate amount of supernatant was mixed and transferred into a detection vial.

3. Gas chromatography and mass spectrum conditions: The Gas chromatography analysis was performed on trace 1310 gas chromatograph (Thermo Fisher Scientific, USA). The GC was fitted with a capillary column Thermo TG-FAME (50 m*0.25 mm ID*0.20 μm) and helium was used as the carrier gas at 0.63 mL/min. THE injection was made in split mode at 8:1 with an injection volume of 1 μl and an injector temperature of 250 °C. The temperature of the ion source and interface were 300 °C and 280 °C, respectively. Mass spectrometric detection of metabolites was performed on ISQ 7000 (Thermo Fisher Scientific, USA) with electron impact ionization mode. Single ion monitoring (SIM) mode was used with the electron energy of 70 eV. 4 replicate samples per group were provided for analysis^32,33.

Subcutaneous xenograft tumor model

BALB/c nude mice were ordered from the Animal Experiment Center of the People’s Hospital of Wuhan University and housed in the SPF-level barrier environment of the Animal Experiment Center of the People’s Hospital of Wuhan University. After one week of acclimatization, 5 × 10⁶ HGC-27 NC and HGC-27 shACLY cells were injected subcutaneously into the right back of the right flank of BALB/c nude mice. Following one week of observation of transplanted tumors, the nude mice grown with different cells were equally subdivided into two groups according to the randomization principle. The groups were NC, NC + VY-3–249, shACLY, and shACLY + VY-3–249. 25 mg/kg of VY-3–249 was intraperitoneally administered into nude mice in the NC + VY-3–249 and shACLY + VY-3–249 groups daily, and the same volume of vehicle was injected daily into nude mice in the NC and shACLY groups for three weeks. At the end of the fourth week, all BALB/c nude mice were executed by CO₂ euthanasia apparatus (LC500, Yuyan Scientific Instruments Co., Shanghai, China) and the transplanted tumors were removed. All experimental procedures involving animals used in this study were reviewed and approved by the Experimental Animal Welfare Ethics Committee of the Animal Experimentation Center, Wuhan University People’s Hospital. Animal protocols followed the National Institutes of Health and ARRIVE guidelines³⁴ and were approved by the animal ethics committee of our hospital (Approval No. WDRM20231205E).

Preparation of drug solutions

ACSS2 inhibitor VY-3–249, 1-(2,3-di(thiophen-2-yl)quinoxalin-6-yl)-3-(2-methoxyethyl)urea, was purchased from Selleck Chemicals (S8588, Houston, TX, USA). The concentration used in animals was 25 mg/mL³⁵, the formulation consisted of 5% DMSO + 40% PEG300 + 5% Tween 80 + 50% ddH2O. PA (P9767, Sigma-Alrdich, MO, USA) was prepared as 2 mM stock solutions by dissolving in ddH₂O at 70 °C, filtering through 0.4 μm and storing at 4 °C. The stock solutions were then added to 2% bovine serum albumin (BSA, GC305010, Servicebio, Wuhan, China) medium made by dissolving BSA in serum-free medium to a final concentration of 2% (w/v) and supplementing with 5 μg/mL penicillin/streptomycin to achieve the desired final FA concentration³⁶.

Data collection and analysis in bioinformatics

CCLE data for ACLY gene expression in tumor cells can be downloaded from the DepMap site. GEO and the International Cancer Genome Consortium Data Portal were used to obtain the GSE26253 and GSE84437 datasets. Using the R packages ‘survival’ and ‘survminer’, the Kaplan–Meier technique was employed to assess survival disparities, and ‘survminer’ was used to compare survival rates.

Statistical analysis

Statistical analyses were performed using the Statistical Product Service Solutions (SPSS) software package (Version 26.0, IBM Corp., Armonk, NY, USA). Comparisons between the two groups were made using Student’s t-test for independent samples. p < 0.05 was considered statistically significant. The data were presented as mean ± SD. The plots were constructed using GraphPad Prism software (Version 9.3, La Jolla, CA, USA). Each experiment was conducted at least thrice.

Result

ACLY is upregulated in GC tissues, and its overexpression predicts poor prognosis for GC patients

To verify the ACLY expression level in GC, we performed immunohistochemical staining on paraffin sections of 20 human GC tissues (Fig. 1A). The findings, which indicated that ACLY was highly expressed in human GC tissues compared with para-cancerous tissues (Fig. 1B). Thereafter, we searched the GEO database for GC patient data and selected the largest sample size datasets, GSE26253 and GSE84437, to explore the relationship between ACLY and prognosis. Subsequent analysis suggested that GC patients with high ACLY expression had a shorter overall survival (Fig. 1C and Fig. 1D). Our study validated the expression level of ACLY in GC and confirmed the poor prognosis of patients with high ACLY levels.

Fig. 1.

Expression and prognosis of ACLY in GC. (A) Representative immunohistochemical images of ACLY expression in GC and paraneoplastic tissue. Scale bar is 50 μm. (B) Immunohistochemical analysis of ACLY expression in GC and paracancerous tissues. (C and D) Prognosis of ACLY expression levels in GC in the Gene Expression Omnibus (GEO) database (GSE26253 and GSE84437). Data shown as mean ± SD. ***p < 0.001 (Student’s t-test).

Downregulation of ACLY reduces lipid accumulation and inhibits GC proliferation, migration, and invasion in vitro

To find suitable GC cell lines, we used the DepMap portal to download ACLY expression data from the CCLE database in tumor cell lines. HGC-27 and AGS, two GC cell lines with relatively high ACLY expression, were analyzed and screened (Figure S1A). Stable knockdown cell lines were created using shRNA. We stably knocked down ACLY in HGC-27 and AGS cells to investigate its function in GC. ACLY knockdown was confirmed using qPCR and WB (Figure S1B and Fig. 2A), and the results indicated that ACLY was successfully knocked down in HGC-27 and AGS cells. Furthermore, because ACLY is the first key enzyme in FA ab initio synthesis, we studied the expression levels of key molecules in HGC-27 and AGS cells (Figure S1B). After ACLY was knocked down, the relative expression of ACSS2 significantly increased in HGC-27 and AGS cells, while that of ACC1 with FASN considerably decreased. Thereafter, we used the TG kit to examine the in vitro TG levels, and we found that downregulating ACLY notably reduced the in vitro TG levels when compared to the negative control (Fig. 2B). To demonstrate intracellular neutral lipid levels, oil red O staining was used (Fig. 2C). In HGC-27 and AGS cells, the number of intracellular lipid droplets in the shACLY group was significantly lower than in the NC group (Fig. 2D). According to these findings, ACLY downregulation may affect intracellular lipid deposition by influencing the expression of FA ab initio metabolism genes.

Fig. 2.

Effect of knockdown of ACLY in GC cells on intracellular lipid levels and cell proliferation. (A) Representative western blot images of stable knockdown of ACLY using shRNA in HGC-27 and AGS cells. (B) Intracellular triglyceride (TG) levels in HGC-27 and AGS cells. (C) Representative oil red O staining images of HGC-27 and AGS cells. Scale bar is 20 μm. (D) Quantitative analysis of oil red O staining in HGC-27 and AGS cells. (E and F) Cell viability of HGC-27 and AGS cells at 24, 48 and 72 h as determined by CCK8 assay. (G) Representative images of clone formation of HGC-27 and AGS cells. (H) Quantitative analysis of clone formation of HGC-27 and AGS cells. Data shown as mean ± SD. *p < 0.05, **p < 0.01 and ***p < 0.001 (Student’s t-test).

To determine if ACLY affects the biological function of GC cells, we first investigated the viability of GC cell proliferation. CCK8 assay was used to determine the viability of short-term cell proliferation. In the CCK8 assays conducted at 24, 48, and 72 h, the downregulation of ACLY significantly reduced the cell viability of HGC-27 cells, as indicated by both the OD values and the rates of increase in OD values (Fig. 2E). For AGS cells, the growth rates of cell viability between 24 to 48 h were comparable between the two groups. However, during the period from 48 to 72 h, the downregulation of ACLY significantly inhibited the growth rate of cell viability (Fig. 2F). These findings suggest that ACLY may suppress the short-term cell viability of GC cells and might also inhibit their long-term cell viability. We then tested the long-term proliferation viability of HGC-27 and AGS cells using clonogenic assays (Fig. 2G), and the results indicated that knocking down ACLY significantly reduced the number of colonies of cells (Fig. 2H). The downregulation of ACLY affected on the short- and long-term viabilities of GC cells in this study. In the following study, we used a wound healing assay to assess GC cell migration ability (Fig. 3A). The results demonstrated that compared to negative control cells, cells with knockdown ACLY had a lower migration rate in HGC-27 and AGS cells (Fig. 3B). Then, transwell experiments with and without matrix gel were performed (Fig. 3C). Our quantitative data revealed that the downregulation of ACLY significantly reduced the number of cells crossing the chambers, as well as the cells’ ability to migrate and invade (Fig. 3D). To further verify the effect of ACLY on gastric cancer cells, rescue experiments were conducted on HGC-27 cells with ACLY knockdown, and the results indicated that the neutral lipid content of HGC-27 cells with ACLY knockdown increased upon overexpression of ACLY (Fig. 4A–C), and the proliferation, migration, and invasion abilities of HGC-27 cells were significantly restored (Fig. 4D–H). In conclusion, the current study confirmed that the downregulation of ACLY inhibits GC proliferation, migration, and invasive ability in vitro.

Fig. 3.

Effect of knockdown of ACLY in GC cells on cell migration and invasion. (A) Representative wound healing images of HGC-27 and AGS cells. Scale bar is 200 μm. (B) Wound healing migration rate of HGC-27 and AGS cells. (C) Representative transwell images of HGC-27 and AGS cells. Scale bar is 50 μm. (D) Quantitative analysis of transwell migration and invasion of HGC-27 and AGS cells. Data shown as mean ± SD. **p < 0.01 and ***p < 0.001 (Student’s t-test).

Fig. 4.

Rescue experiments of ACLY. (A) Expression levels of ACLY in HGC-27 cells. (B) Representative oil red O staining images of HGC-27 cells. Scale bar is 20 μm. (C) Quantitative analysis of oil red O staining in HGC-27 cells. (D) Representative wound healing images of HGC-27 cells. Scale bar is 200 μm. (E) Wound healing migration rate of HGC-27 cells. (F) Cell viability of HGC-27 cells at 24, 48, and 72 h as determined by CCK8 assay. (G) Representative transwell images of HGC-27 cells. Scale bar is 50 μm. (H) Quantitative analysis of transwell migration and invasion of HGC-27 cells. **p < 0.01 and ***p < 0.001 (Student’s t-test).

ACSS2 maintains the growth of cells after the downregulation of ACLY by compensatory elevation

Because previous results have indicated that ACSS2 mRNA levels were significantly increased after ACLY knockdown (Figure S1B), we performed WB experiments to confirm the protein levels of ACSS2 in HGC-27 and AGS cells. The findings revealed that the protein levels of ACSS2 were significantly higher after ACLY knockdown in both cell lines ( Fig. 5A–C).To further investigate the roles of ACLY and ACSS2 in lipid metabolism, we established stable knockdown cell lines of SREBP1 and ACSS2 in HGC-27 and AGS cells, respectively. Subsequently, we assessed the levels of lipid metabolism-related genes in these cells. Our results demonstrated that the protein levels of various molecules do not appear to be significantly altered upon knockdown of SREBP1 in both cells. Furthermore, after downregulating ACSS2, we observed a significant decrease in the protein levels of ACC1 and FASN in HGC-27 cells, whereas a significant increase in AGS cells ( Fig. 5A–C). The qPCR results were consistent with the WB findings. Specifically, we observed a significant increase in ACSS2 mRNA levels in both HGC-27 and AGS cells after ACLY knockdown, along with downregulation of ACC1 and FASN mRNA levels in both cell lines ( Fig. 5D and E). We additionally found that the knockdown of SREBP1 led to a significant decrease in ACLY, ACSS2, ACC1, and FASN expression in both HGC-27 and AGS cells. Interestingly, consistent with the observations in Fig. 5A, we discovered that the downregulation of ACSS2 in AGS cells resulted in an upregulation of ACC1 and FASN levels, whereas the opposite effect was observed in HGC-27 cells ( Fig. 5D and E). Our findings suggest that SREBP1 regulates both ACLY and ACSS2, with ACSS2 upregulation possibly being employed to correct an imbalance in lipid metabolism caused by ACLY deficiency.

Fig. 5.

Effect of knocking down ACLY on ab initio synthesis of FAs in GC cells. (A) Representative western blot images of protein expression levels involved in the ab initio synthesis of fatty acids in HGC-27 and AGS cells. (B and C) Expression levels of fatty acid ab initio anabolic proteins in HGC-27 and AGS cells relative to the internal reference protein β-actin. (D and E) Expression levels of fatty acid ab initio anabolic genes in HGC-27 and AGS cells relative to the internal reference gene β-actin. Data shown as mean ± SD. *p < 0.05, **p < 0.01 and ***p < 0.001 (Student’s t-test).

VY-3–249 is a selective inhibitor of ACSS2. By inhibiting this enzyme, VY-3–249 can affect metabolic pathways involving acetate and acetyl-CoA, potentially impacting processes such as cell growth and proliferation in cancer³⁷. We exploited the ACSS2 inhibitor, VY-3–249, to confirm whether upregulation of ACSS2 is required to maintain GC growth in vitro. In addition, VY-3–249 drug concentration-cell viability curves were derived using the CCK8 assay in HGC-27 and AGS cells. In HGC-27 and AGS cells, the IC50 was 21.78 and 34.43 μM, respectively (Fig. 6A and B). Then, we determined different drug concentrations in the two cells based on the IC50 of VY-3–249 and assessed cell viability in each group of cells at 24, 48, and 72 h time points. VY-3–249 had a significant inhibitory effect on cell viability in HGC-27 and AGS cells at different concentrations and time points, and knocking down ACLY rendered both cells more sensitive to VY-3–249. (Fig. 6C and D). Experiments on clone formation produced identical results (Figure S2A–C). In addition, no significant difference was observed between the shACLY group and the NC group AGS cells when VY-3–249 = 15 μM (48 h and 72 h). Therefore, it is likely that VY-3–249 = 15 μM caused AGS cell death (Fig. 6D). Consequently, we concluded that ACSS2 was compensated for elevated and maintained GC cell growth after ACLY downregulation in GC cells.

Fig. 6.

Effect of ACSS2 inhibition on GC cells. (A and B) Inhibition curves of HGC-27 and AGS cell viability by VY-3–249 (an ACSS2 inhibitor). (C) Cell viability of HGC-27 cells at concentrations of 0, 2.5 μΜ, 5 μM, and 10 μM of VY-3–249 at 24, 48, and 72 h. (D) Cell viability of AGS cells at concentrations of 0, 3.75 μΜ, 7.5 μM, and 15 μM of VY-3–249 at 24, 48, and 72 h. Control cells were treated with the same volume of solvent (DMSO) as the drug-treated group, and all groups had identical volumes of DMSO (at a concentration of 0.08%). The cell viability of NC cells or shACLY cells treated with different drugs was calculated with respect to the cell viability of the control cells set at 100%. Comparisons were made based on the same treatment duration. The reduction in cell viability relative to the control cells was expressed as “(100—cell viability)%”. Data shown as mean ± SD. *p < 0.05, **p < 0.01 and ***p < 0.001 (Student’s t-test).

Ab initio synthesized FA content is further reduced in ACLY-depleted GC cells after ACSS2 inhibition

Acetyl-CoA and malonyl-CoA are substrates for FA ab initio metabolism. We used the corresponding ELISA kits to examine the acetyl-CoA and malonyl-CoA levels in pairs of GC cells to understand the effect of ACLY downregulation versus ACSS2 inhibition. Figure 7A shows that either the downregulation of ACLY, or inhibition of ACSS2, or downregulation of ACLY followed by inhibitor application significantly reduced acetyl-CoA levels in HGC-27 cells, whereas malonyl-CoA levels did not significantly change under different treatment conditions (Figure S2E). In addition, we quantified medium- and long-chain FAs in HGC-27 cell samples to determine if the downregulation of ACLY followed by the application of an ACSS2 inhibitor further reduced intracellular FA content. The resultant heat maps displayed our quantitative data, in which odd-chain FAs not involved in the FA synthesis process were excluded. The data was normalized using the z-score method for rows and subjected to clustering analysis for rows and columns (Fig. 7B). We intuitively observed a potential difference in the distribution of medium- and long-chain FA content between the NC group and the shACLY/shACLY + VY-3–249 groups. We conducted an analysis of the actual data from the heatmap and observed that knockdown of ACLY alone and knockdown of ACLY followed by inhibition of ACSS2 significantly reduced the total fatty acid (TFAs) content in HGC-27 cells compared to the control group. Furthermore, inhibition of ACSS2 resulted in a further decrease in cellular TFAs content compared to the shACLY group (Fig. 7C). FAs can be categorized as saturated fatty acids (SFAs) and unsaturated fatty acids (UFAs). All directly synthesized FAs are SFAs, with PA accounting for more than 50% of SFAs¹³. As shown in Figure S2D, both the shACLY group and the shACLY + VY-3–249 group demonstrated significant reductions in SFAs and UFAs content compared to the NC group. Additionally, the decline in SFAs was greater than that of UFAs in both the shACLY group and the shACLY + VY-3–249 group, suggesting that they are more likely to affect FA synthesis rather than desaturation. The analysis of ab initio synthesized FAs revealed a significant reduction in the content of Caproate (C6:0), Caprylate (C8:0), Caprate (C10:0), Laurate (C12:0), Myristate (C14:0), and PA (C16:0) in ACLY-knockdown cells and ACLY-knockdown and ACSS2 inhibition cells compared to the NC group (Fig. 7D and E). The content and changes of various ab initio synthesized FAs are presented in Fig. 7E. As the number of carbon atoms increases, the content of FAs also increases, with PA accounting for the majority (> 90%) of ab initio synthesized FAs. Compared to the control group, the levels of C6:0, C12:0, C14:0, and C16:0 were significantly downregulated after ACLY knockdown, and ACSS2 inhibition further reduced the content of these FAs compared to the ACLY knockdown group. In conclusion, ab initio FA content was further reduced in ACLY-depleted GC cells post ACSS2 inhibition.

Fig. 7.

Effect of ACSS2 inhibition on ab initio fatty acid synthesis in HGC-27 cells. (A) Levels of acetyl-CoA in HGC-27 cells. Cells were treated with 10 μM of VY-3–249 or vehicle for 24 h. (B) Heat map of data for absolute quantification of 39 medium and long chain fatty acids. The heat map data were normalized by the z-score method for rows. (C) Quantification of total fatty acids (TFAs) in HGC-27 cells. (D) Quantification of ab initio synthesized FAs (C ≤ 16) in HGC-27 cells (E) Quantification of various de novo synthesized FAs in HGC-27 cells. (F and G) Cell viability of HGC-27 and AGS cells at 24, 48, and 72 h as determined by CCK8 assay. The concentration of PA is 100 μM. Data shown as mean ± SD. *p < 0.05, **p < 0.01 and ***p < 0.001 (Student’s t-test). Abbreviations: CoA, Coenzyme A; C6:0, Caproate; C8:0, Caprylate; C10:0, Caprate; C12:0, Laurate; C14:0, Myristate; C16:0, Palmitate.

Next, we investigated the effect of FA on the proliferation of GC cells. In vitro addition of 100 μM PA promoted the proliferation of HGC-27 and AGS cells (Fig. 7F and G). After elucidating the compensatory mechanism of ACLY and ACSS2 in ab initio FA synthesis, we reasonably proposed that ACLY and ACSS2 affect the proliferative capacity of GC cells by influencing the ab initio synthesis of FAs.

Inhibition of ACSS2 enhances the inhibitory effect of down-regulation of ACLY on the growth of transplanted tumors in nude mice

To investigate the function of ACLY and ACSS2 in GC in vivo, a nude mouse subcutaneous xenograft tumor model was established. In the right dorsal aspect of five-week-old BALB/c nude mice, we subcutaneously injected HGC-27 human GC cells with stably downregulated ACLY and negative control HGC-27 human GC cells. After seven days of normal feeding, subcutaneous transplanted tumors were observed, while nude mice were randomly divided into four groups of equal numbers (n = 6): NC, NC + VY-3-249, shACLY, and shACLY + VY-3-249. Nude mice in the NC + VY-3-249 and shACLY + VY-3-249 groups were intraperitoneally injected with 25 mg/kg of VY-3-249 daily, while the NC and shACLY groups received the same volume of vehicle for three weeks, and subcutaneous tumors were collected at week 4 (Fig. 8A). Figure 8B shows the photographs of tumor tissue. We found that compared with the NC group, both intraperitoneal injection of VY-3–249 and ACLY knockdown resulted in significant inhibition of transplanted tumor growth and significant reduction of tumor weight. And the subcutaneous xenograft tumor weight of mice in the combined treatment group of intraperitoneal injection of VY-3-249 and ACLY knockdown was the lightest (Fig. 8C). Furthermore, immunohistochemical staining of xenograft tumors revealed that Ki-67 expression was lower in transplanted tumors with downregulated ACLY compared to that in the controls, and Ki-67 levels were significantly lower after three weeks of VY-3-249 intraperitoneal administration (Fig. 8D and E). In addition, Ki-67 expression was the lowest in transplanted tumors in the shACLY + VY-3–249 group. Our findings from a subcutaneous transplantation tumorigenesis study of GC cells in nude mice revealed that downregulation of ACLY significantly inhibited GC growth in vivo, and ACSS2 inhibition enhanced the inhibitory effect of ACLY depletion on GC growth.

Fig. 8.

Effect of simultaneous inhibition of ACLY and ACSS2 on xenograft tumor formation by HGC-27 cells in BALE/c nude mice. (A) Schematic diagram depicting the construction of a subcutaneous transplantation tumor model of HGC-27 cells in nude mice and the intraperitoneal administration of VY-3–249 (25 mg/kg). The Figure was partly generated using Servier Medical Art, provided by Servier, licensed under a Creative Commons Attribution 3.0 unported license. (B) Pictures of transplanted tumors formed by subcutaneous injection of HGC-27 cells in nude mice. (C) Quantification of transplanted tumor weight in nude mice. (D) Representative images of immunohistochemical staining of Ki-67 in tissue sections of nude mice transplanted with tumors. Scale bar is 50 μm. (E) Quantification of immunohistochemical staining of Ki-67 in tissue sections of nude mice transplanted with tumors. Data shown as mean ± SD. *p < 0.05, **p < 0.01 and ***p < 0.001 (Student’s t-test).

Discussion

ACLY catalyzes the Mg-ATP-dependent conversion of citric acid and CoA to OAA and acetyl-CoA to maintain lipid and sterol synthesis, isoprenylation, and protein glycosylation³⁸. In vivo, ACLY inhibition decreases tumor cell proliferation and migration,  induces apoptosis and differentiation, and inhibits tumor growth^23,24. In acute myeloid leukemia, low ACLY expression was related to a favorable outcome³⁹. ACLY promotes the expression of lipid metabolism-related genes, significantly promotes lipid metabolism in GC cells, and increases cancer cell proliferation¹⁶. Furthermore, ACLY knockdown decreases acetyl-CoA, free FA content, and total cholesterol in hepatocellular carcinoma cells while increasing TCA metabolic activity⁴⁰. Our findings were similar: ACLY knockdown led to the lower expression of lipid metabolism genes and lower fat buildup. Meanwhile, cell proliferation, migration, and invasion were considerably reduced, which could be attributed to FA ab initio metabolism reduction. However, in prostate and lung cancer investigations, although ACLY depletion stops cell growth in vivo and in vitro and inhibits adipogenesis from the head, lipid buildup occurs in the cells, primarily TG^27,41. ACLY interacts with snail proteins and regulates snail transcriptional activity to retain lung cancer stem cell features. It also stabilizes β-catenin and thus, regulates the traditional Wnt pathway by promoting tumor cell migration and invasion^18,22,42. AMPK is activated in ACLY-deficient cells, which leads to p53 phosphorylation and activation, ultimately causing DNA damage and subsequent apoptosis or cellular senescence in both normal and cancer cells^43,44. ACLY inhibitors reduce monolayer cell proliferation in a dose- and time-dependent manner, produce ER stress, and activate the p-eIF2/ATF4/CHOP axis in HCC cells to promote apoptosis, and synergize with sorafenib to improve tumor treatment efficacy^20,45,46. ACLY knockdown can exhibit anticancer effects with reduced cisplatin resistance by suppressing the PI3K-AKT pathway and activating the AMPK-ROS pathway^19,47,48. Furthermore, ACLY targeting can boost the therapeutic impact of irinotecan⁴⁹. While normal mammalian cells obtain FAs primarily through exogenous uptake, the main source of FAs in cancer cells is from ab initio synthesis rather than the microenvironment. Metabolomic analysis has demonstrated enhanced FA synthesis and PA production in cancer, which plays a crucial role in cancer growth⁵⁰, and FA can promote cell proliferation by activating different signals^51–53. Because ACLY can affect tumor growth and invasive status through influencing lipid metabolism^54–56, we concluded that in our study ACLY promoted GC progression by regulating the FA ab initio synthesis pathway.

Our work also revealed that ACSS2 is compensated for upregulation in ACLY-deficient GC cells and xenograft tumors and continues to maintain GC growth in vivo and in vitro (Fig. 9). Furthermore, after the ACSS2 knockdown, the mRNA level of ACLY was considerably increased in AGS cells, yet there was no significant change in HGC-27 cells (Fig. 5D and E). And the downregulation of ACSS2 in AGS cells results in a increase in the levels of FASN and ACC1, while the opposite effect is observed in HGC-27 cells (Fig. 5A–E). This led us to speculate that some different statuses of these two cell lines may have contributed to this difference in results. HGC-27 cells are p53 mutant cells, whereas AGS is p53 wild-type cells^57,58. In addition, previous studies have shown that p53 plays a critical role in regulating intracellular lipid metabolism. p53 represses the expression of SREBP1, a key transcription factor driving several adipogenic gene expressions at the transcriptional level, and the disruption of p53 in ob/ob mice partially restored the adipose tissue expression of SREBP1 and the expression of its downstream target adipogenic enzymes such as ACLY, ACC1 and FASN^59,60. The results of our study in GC are just the opposite, in the current study, ACLY expression was considerably increased following ACSS2 knockdown in AGS cells, but not in HGC-27 cells, so we speculate that AGS cells in wild-type p53 status after ACSS2 knockdown may be through the activation of SREBP1 function, thus leading to the upregulation of ACLY, ACC1 and FASN expressions. In addition, Wang et al. reported that the NQO1/p53/SREBP1 axis promotes lipid synthesis in hepatocellular carcinoma⁶¹. ACSS2 is a member of the acetyl-CoA synthase short-chain family, along with ACSS1 and ACSS3. ACSS2 is found mostly in the nucleus and cytoplasm and catalyzes the synthesis of CoA from acetate. ACSS1 acts on the same substrate as ACSS2, but differs in that it functions within the mitochondria. ACSS3, which is also found in mitochondria, preferentially uses propionate. ACSS2 is linked with lipid synthesis and protein acetylation processes, whereas ACSS1 is involved in acetate oxidation, allowing ACSS1, ACSS2, and ACSS3 enzymes to act synergistically to store, mobilize, consume, and recycle acetate throughout the cell²⁸. ACSS2 nuclear localization and expression are increased under hypoxia or in the absence of serum, and ACSS2 may enhance tumor growth independently of ACLY^29,48,62. ACSS2 inhibition reduces tumor growth both in vivo and in vitro^30,35,62,63. Human carcinomas with high ACSS1/2 expression have enhanced FASN expression, whereas low FASN expression is related to the suppression of cell proliferation, energy production, and tumor growth^64,65. Furthermore, the expression level of ACSS2 in different cancers has a fundamentally distinct clinical significance^66,67. The elimination of ACLY improves ACSS2-dependent lipid synthesis³¹. CMA regulates lipid droplet formation by controlling ACLY and ACSS2 expressions, and the acetate and ACSS2 pathways can fully compensate for ACLY depletion^68,69. The in vivo targeted knockdown of ACLY in mouse adipocytes resulted in increased ACSS2 expression⁷⁰. Furthermore, ACLY and ACSS2 have been identified as important targets of mTORC2 control of lipid production in brown adipocytes⁷¹. ACSS2 and ACLY siRNA silencing lowered overall cellular TG levels and lipid droplet formation⁷². For the first time, we demonstrate that ACSS2 is increased at both the mRNA and protein levels in GC, and it compensates for the cell growth arrest caused by ACLY deficiency. Furthermore, acetylation is one of the functions performed by ACSS2. Rapamycin, according to Liang, Yi et al., suppresses cadmium-enhanced breast cancer growth by upregulating ACSS2⁷³. ACSS2 interacts with the oncoprotein interferon regulatory factor 4 (IRF4) to improve IRF4 stability and IRF4-mediated gene transcription³⁵. Acetyl-CoA metabolism is linked to histone acetylation and gene regulation. Furthermore, lower ACSS2 levels reduce nuclear acetyl-CoA levels and histone acetylation, while decreasing hippocampal ACSS2 expression affects long-term spatial memory⁷⁴. Katelyn D. Miller et al. recently disclosed a novel ACSS2 inhibitor (VY-3–135) that inhibits ACSS2 more effectively⁷⁵. VY-3–135 may be explored more in future studies of ACSS2. However, despite providing several important insights, our study has some limitations. We focused only on the effect of ACLY on FA metabolism, whether there is an interaction between ACLY and ACSS2, and whether the compensatory relationship between ACLY/ACSS2 is related to another lipid metabolism (cholesterol metabolism) or glucose metabolism is still unelucidated. In future studies, we must focus on understanding this aspect.

Fig. 9.

Schematic representation of compensatory elevation of ACSS2 in ACLY downregulation to maintain GC growth. In the cytoplasm, ACLY and ACSS2 produce the same product acetyl coenzyme A through different pathways to support the synthesis of ab initio fatty acids and thus promote the growth of GC cells. When ACLY is downregulated, in order to maintain the homeostasis of fatty acid synthesis activity, ACSS2 is compensatingly elevated to increase the content of acetyl coenzyme A. The fatty acid production that is diminished by the absence of ACLY is rescued. The Figure was partly generated using Servier Medical Art, provided by Servier, licensed under a Creative Commons Attribution 3.0 unported license.

Conclusions

Our findings indicate that ACLY may increase GC cell proliferation, migration, and invasion through modulating lipid metabolism, while ACLY downregulation reduces cell growth in vitro and in vivo. In ACLY-depleted cells, ACSS2 was compensated to increase to maintain cell development, and simultaneous inhibition of ACLY and ACSS2 drastically inhibited cell proliferation in vitro and in vivo. Consequently, the combined targeted inhibition of ACLY and ACSS2 may be a viable GC therapy. Therefore, more fundamental studies and clinical investigations are required to validate and advance the research.

Author contributions

DW, LC came up with the design and conception. LW, LY and LC prepared material. LC and LY conducted experiments and collected data, and LC analyzed the data. LC and LW wrote the first draft of the manuscript. WW and DW revised the manuscript and guided the subject. All authors com-mented on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by grants from National Natural Science Foundation of China (No. 82172855). The funding bodies played no role in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.

Data availability

The dataset analyzed during the current study consists of public datasets such as GEO, and CCLE, which can be downloaded from the methodology section of the manuscript or obtained from the cor-responding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Ethical approval

All subjects gave their informed consent for inclusion before they participated in the study. The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Ethics Committee of Renmin Hospital of Wuhan University (2022 K-K063(C01)). Patients involved obtained informed consent in the study. We confirmed that all methods were carried out following relevant guidelines and regulations.