Acetyl-CoA is a vitally important and versatile metabolite used for many cellular processes including fatty acid synthesis, ATP production, and protein acetylation. Recent studies have shown that cancer cells upregulate acetyl-CoA synthetase 2 (ACSS2), an enzyme that converts acetate to acetyl-CoA, in response to stresses such as low nutrient availability and hypoxia. Stressed cancer cells use ACSS2 as a means to exploit acetate as an alternative nutrient source. Genetic depletion of ACSS2 in tumors inhibits the growth of a wide variety of cancers. However, there are no studies on the use of an ACSS2 inhibitor to block tumor growth. In this study, we synthesized a small-molecule inhibitor that acts as a transition-state mimetic to block ACSS2 activity in vitro and in vivo. Pharmacologic inhibition of ACSS2 as a single agent impaired breast tumor growth. Collectively, our findings suggest that targeting ACSS2 may be an effective therapeutic approach for the treatment of patients with breast cancer.

Significance:

These findings suggest that targeting acetate metabolism through ACSS2 inhibitors has the potential to safely and effectively treat a wide range of patients with cancer.

Cancer cells within a tumor often experience situations in which oxygen availability becomes severely limited due to inadequate and aberrant vascularization. Poor vascularization additionally causes insufficient nutrient provision for the high metabolic demands of cancer cells. It is widely believed that cancer cells continuously cycle between nutrient replete conditions and hypoxic nutrient stressed conditions over the course of tumor growth (1). Over time, cancer cells adapt and evolve to the harsh conditions within the tumor microenvironment and emerge significantly more radio-resistant, less responsive to chemotherapy and immunotherapy, and prone to malignant progression (2, 3). Thus, the changes that accompany the response to tumor hypoxia represent a critical barrier to treating cancer (4, 5).

We previously performed a functional genomic screen that identified acyl-coenzyme A synthetase short-chain family member 2 (ACSS2) as a crucial enzyme for growth during hypoxic stress (6). ACSS2 is a nucleocytosolic enzyme that converts acetate to acetyl-CoA (AcCoA) and its expression is induced by hypoxia and low nutrient stresses, such as low glucose or low lipid availability (6, 7). Stress responsive transcription factors, such as hypoxia-inducible factors and sterol regulatory element binding factors, drive ACSS2 expression (6, 8, 9). The upregulation of ACSS2 during metabolic stress then promotes uptake and utilization of acetate for AcCoA synthesis (6, 10, 11).

It is well established that hypoxic cells block conversion of pyruvate to AcCoA in the mitochondria, and instead convert pyruvate to lactate, which is then secreted from the cell (12). Interestingly, aerobic glycolysis, also known as the Warburg effect, is a hallmark of many cancers and is characterized by a high rate of glucose uptake and glycolytic flux to lactate, albeit in the presence of oxygen. The consequence of both aerobic glycolysis and hypoxia-induced redirection of pyruvate to lactate is the creation of a “substrate gap” for the synthesis of cytosolic AcCoA, a critical intermediate in many metabolic pathways (13, 14). In this manner, the capture of acetate by ACSS2 fills the substrate gap for cytosolic AcCoA created by the metabolic changes associated with aerobic glycolysis and hypoxia. Multiple independent studies have now shown that genetic depletion of ACSS2 inhibits tumor growth in a wide variety of cancers, including breast (6), prostate (6), liver (15), glioblastoma (7), soft-tissue (16, 17), and skin cancers (18). The sheer breadth of cancer types that are sensitive to loss of ACSS2 speaks to the significance of this enzyme for tumor growth, the universality of acetate metabolism in cancer, and the promise of ACSS2 inhibitors for treating cancer.

In this study, we aimed to test the hypothesis that pharmacologic inhibition of ACSS2 can inhibit tumor growth. We show that in vivo inhibition of ACSS2 with a previously uncharacterized small-molecule inhibitor strongly impedes tumor growth, and in some instances can induce tumor regression. Stable isotope tracer studies verified the on-target activity of the ACSS2 inhibitor in tumors by blocking acetate-dependent fatty acid synthesis. Overall, we describe a novel ACSS2 inhibitor that will be a useful tool for future in vitro and in vivo work examining acetate metabolism in cancer. Furthermore, our study creates a strong rationale for the development of first-in-class ACSS2 inhibitors as a novel therapeutic modality for the improved treatment of patients with cancer.

Reagents, cell lines, and cell culture

BT474 cells were the gift of Dr. Jose Baselga (AstraZeneca). Brpkp110 and A7C11 cells were the gift of Dr. Jose Conejo-Garcia and previously described in refs. 19–22. WHIM12 TNBC PDX cells were gifted from Dr. Jeffrey M. Rosen (Baylor College of Medicine, Houston, TX). The identity of all cell lines was routinely confirmed by short tandem repeat profiling done at the University of Pennsylvania (human lines) and ATCC (mouse lines), and we never use any lines for more than six months after thawing. All cells are routinely screened for Mycoplasma using a PCR Mycoplasma Detection Kit from Applied Biological Materials (catalog no. G238). All cells were cultured in DMEM/F-12 50/50 (Life Technologies) supplemented with 10% FBS (Life Technologies) and 1× penicillin–streptomycin except Hs578t (ATCC) cells, which were additionally supplemented with 0.01 mg/mL bovine insulin (Sigma), 10% FBS, and 1× penicillin–streptomycin. VY-3-249 was purchased from Ark Pharmaceuticals (Sigma) and VY-3-135 was synthesized by the Wistar Molecular Screening Facility. The chemical synthesis strategy for VY-3-135 and quality analyses are available upon request. 13C2-acetate and D3-acetate were purchased from Cambridge Isotopes, Inc.

Lentiviral transduction

CRISPR-Cas9 pools of Brpkp110 and A7C11 were generated using single guide RNAs against exon 1 in mouse Acss2. Cas9 and guide RNAs were introduced into cells by lentiviral infection. Briefly, HEK293T cells were transfected (Lipofectamine 2000) with psPAX2, pVSV-G, and pLentiCRISPRV2-blast containing a single guide RNA cloned into the BsmBI site (GE Healthcare). Transduced pools of Brkp110 and A7C11 cells were selected using blasticidin S.

Acetyl-CoA synthetase biochemical assays

ACSS2 enzyme activity was measured using the TranScreener TRF AMP/GMP assay (Bellbrook Labs). Recombinant ACSS2 was purchased from Origene. The assay was performed in white, opaque, low volume 384-well plates. Test compounds were diluted in 100% DMSO, then 100 nL of each dilution was transferred using the Janus MDT Nanohead to plates containing 3 μL ACSS1 (2 nmol/L), ACSS2 (3 nmol/L), or ACSS3 (25 nmol/L) in assay buffer (30 mmol/L HEPES, pH 7.4, 140 mmol/L NaCl, 2 mmol/L MgCl2, 5 mmol/L sodium acetate, 2 mmol/L DTT, 0.005% Brij35). For propionate assay, ACSS1 (2 nmol/L), ACSS2 (3 nmol/L) or ACSS3 (25 nmol/L) were added to assay buffer (25 mmol/L HEPES, pH 7.4, 100 mmol/L KCl, 3 mmol/L MgCl2, 5 mmol/L DTT, 0.05% CHAPS). Three microliters of substrate mix containing 100 μmol/L ATP and 10 μmol/L CoA was added to followed by a 120-minute incubation. Final substrate concentrations were 5 mmol/L acetate, 50 μmol/L ATP, and 5 μmol/L CoA in the acetate reaction mix was and 5 mmol/L propionate, 10 μmol/L ATP, and 100 μmol/L CoA in the propionate reaction mix. After incubation, 3 μL of terbium conjugated AMP antibody and AMP tracer was added to according to the methods described by BellBrook Labs. After 30 minutes, the HTRF signal was measured using an Envision Plate reader. Data were normalized to percent inhibition, where 100% inhibition equals the counts obtained in absence of ACSS2, and 0% inhibition equals the counts obtained in the complete reaction including a DMSO control.

Water solubility

One milliliter of PBS buffer (pH 7.4) was added to 10 mg of test compound. The solutions were sonicated for 30 minutes and then vortexed at low speed for 30 minutes. The solutions then sat at room temperature for 16–24 hours before filtering through a 0.22-μm filter followed by 10-fold serial dilutions in triplicate in ACN:Water (1:1, v/v). The final sample solutions were mixed with internal standard followed by LC/MS-MS analysis. Solubility was determined according to the calibration curve generated from five concentration standards.

Microsomal stability assays

Test compounds and controls at final concentrations of 0.5 μmol/L were incubated with 0.5 mg/mL of liver microsomes and an NADPH-regenerating system in potassium phosphate buffer (pH 7.4). Aliquots were taken at specific times and the reactions were quenched with acetonitrile containing an internal standard. Controls were measured that lacked the cofactor solution. Samples were analyzed by LC/MS-MS. Results were reported as peak area ratios of each analyte to internal standard. The intrinsic clearance (CLint) was determined from the first-order elimination constant by nonlinear regression. The half-life (t1/2) was calculated using following equation: k = 0.693/t1/2. Intrinsic clearance was calculated using following equation: CLint = k/min*mL/0.5 mg*mg protein/g liver; where rate = k/min; 0.5 mg protein/mL incubation; 48 mg protein/g liver for mouse; 39.7 mg protein/g liver for human.

Pharmacokinetic analysis

VY-3-135 was dissolved in 10% DMSO, 10% Solutol and 80% PBS. CD-1 mice (Charles River) were administered VY-3-135 by oral gavage (30 mg/kg), intravenous (2 mg/kg), or intraperitoneal (10 mg/kg) based on an average weight of 28 grams. Three animals per time point had ≥ 0.2 mL drawn from the trunk into K2-EDTA tubes. Whole blood samples were kept on wet ice until centrifugation. Plasma was drawn off and extracted in acetonitrile and water with an internal standard (tolbutamide). Samples were centrifuged again and the supernatant collected. LC/MS analysis was performed on a Triple Quad 5500 mass spectrometer (Sciex) coupled to a Nexera X2 (Shimadzu) liquid chromatography system.

Inhibitor docking simulations

Swissdock was used to carry out VY-3-135 and acetyl-CoA synthetase docking simulations. Swissdock is an EADock DSS (23) software, which can identify multiple binding sites using blind docking. The software generates one or more binding sites and the ΔG (CHARMM/Chemistry at Harvard Macromolecular Mechanics; ref. 24) is estimated. Each binding mode is evaluated with fast analytical continuum treatment of solvation (FACTS) and ranked on the basis of favorable energies. In the absence of the human acetyl-CoA PDB coordinates, we used the 1.75 Å crystal structure of Salmonella enterica, acetyl-CoA synthetase bound to adenosine-5′-propylphosphate and coenzyme A (CoA; PDB ID: 1PG4; ref. 25). The coordinates of VY-3-135 were generated in ChemDraw, energy minimized, hydrogens added, and saved in MOL2 format in Phenix (26). Prior to initiating the docking process, we removed the Adenosine-5′-propylphosphate and CoA coordinates from the acetyl-CoA synthetase PDB file. In doing so, we allowed Swissdock to sample all possible pockets of acetyl-CoA synthase for VY-3-135 binding.

Antibodies and Western blotting

Cells were lysed in 1× Laemmli buffer (Bio-Rad) supplemented with 40 mmol/L dithiothreitol (DTT). Lysates were heated and resolved using Mini-PROTEAN precast polyacrylamide gels (Bio-Rad) and blotted onto nitrocellulose membranes using the Mini Blot Module transfer system (Life Technologies). Blots were blocked using 5% milk in Tris-buffered saline solution with Tween-20 (TBST) at room temperature. Blots were incubated with primary antibodies overnight at 4°C. Primary antibodies were diluted in 1% BSA and 0.05% sodium azide in TBST. Antibodies were purchased from the following vendors: Cell Signaling Technology (ACSS2 #3658; EGFR #3658; HER2 #2165; ERK1/2 #9107; pERK1/2 #4370; AKT #4691; pAKTT308 #9275; pAKTS473 #9271; and p53 #2524; Sigma (ACSS1 #HPA041014); ProteinTech (TUBB #66240; ACTB #60008); Abcam (histone H4 #16483; GAPDH #9485); LI-COR Biosciences (goat anti-mouse #926-32210; donkey anti-rabbit #926-68073. Blots were imaged using a LI-COR Odyssey infrared imager.

IHC

Harvested tumors were fixed in 10% phosphate-buffered formalin and then embedded in paraffin blocks followed by sectioning. Sections were deparaffinized and antigen retrieval was performed using steam in the presence of citrate buffer. Endogenous peroxidase activity was blocked with H2O2 and slides were stained with the indicated antibody from Cell Signaling Technology (ACSS2 #3658; Ki67 #9027). Staining was visualized with DAB and sections were counterstained with hematoxylin.

Nuclear fractionation

A total of 2 × 107 cells were trypsinized, quenched with complete DMEM, and washed in PBS. Ten percent of the pellet was collected as total extract in buffer BC-500 (50 mmol/L Tris pH 7.6, 2 mmol/L EDTA, 500 mmol/L KCl, 10% glycerol) and incubated on ice before sonication. Debris was pelleted and supernatant collected as total extract. Five volumes of buffer A (10 mmol/L HEPES pH 7.9, 5 mmol/L MgCl2, 0.25 M sucrose, 0.1% NP-40) were added to the remaining pellet and incubated. Nuclei were pelleted and supernatant collected as cytosolic fraction. Two volumes of buffer B (10 mmol/L HEPES pH 7.9, 25% glycerol, 1.5 mmol/L MgCl2, 0.1 mmol/L EDTA, 300 mmol/L NaCl) was added to pelleted nuclei. Soluble nuclear proteins were collected from the supernatant following centrifugation. One volume of buffer BC-1000 (50 mmol/L Tris pH 7.6, 2 mmol/L EDTA, 1000 mmol/L KCl, 10% glycerol) was added to the chromatin pellet. One volume of buffer BC-0 (50 mmol/L Tris pH 7.6, 2 mmol/L EDTA, 10% glycerol) was added and the sample was sonicated. Debris was pelleted and supernatant was collected as chromatin-bound protein fraction. Individual fractions were diluted with buffer BC-0 and equal concentrations of protein loaded onto SDS-PAGE gel.

QuantSeq 3′ mRNA sequencing

RNA was extracted using TRIzol and purified using an RNeasy Plus Kit (Qiagen). Quantity was determined using the Qubit 2.0 Fluorometer (Thermo Fisher Scientific) and quality was validated using TapeStation RNA ScreenTape or Agilent Bioanalyzer Total RNA Nano chip (Agilent). Two-hundred nanograms of DNAse I–treated, total RNA was used to prepare library for Illumina Sequencing using the Quant-Seq 3′mRNA-Seq Library Preparation Kit (Lexogen). Library quantity was determined using qPCR (KAPA Biosystems). Library size was determined using Agilent TapeStation or Bioanalyzer and DNA High Sensitivity D5000 ScreenTape or High Sensitivity DNA chip (Agilent). Equimolar amounts of each sample library were pooled, denatured and run using Hi-Output, Single-read, 75 bp cycle sequencing kit. Next Generation Sequencing was done on a NextSeq 500 (Illumina).

Bioinformatics and data mining

Raw RNA-sequencing sequencing reads were aligned using bowtie2 (27) algorithm against hg19 human genome version and RSEM v1.2.12 software (28) was used to estimate raw read using gene information from Ensemble transcriptome version GRCh19.p7. Raw counts were normalized and used to estimate significance of differential expression difference between experimental groups using DESeq2 (29). Gene expression changes were considered significant if passed FDR<5% thresholds unless stated otherwise. Gene-set enrichment analysis was done using QIAGEN’s Ingenuity Pathway Analysis software (IPA, QIAGEN Redwood City, www.qiagen.com/ingenuity) using “Canonical pathways” option. Top 10 pathways that passed FDR < 5% thresholds were reported.

TCGA and METABRIC databases were mined using cbioportal.org (30–32). ACSS2 mRNA expression was analyzed in HER2+ breast cancer versus all other types. Putative HER2 and EGFR copy-number alterations are generated by GISTIC algorithm. EGFR receptor status was analyzed only in TNBC patient samples. For mRNA expression of ACSS2 in breast cancer cell lines, we used the Cancer Cell Line Encyclopedia (31, 32).

Tumor xenograft studies

All animal experiments were approved by the Institutional Animal Care and Use Committee and were performed in an Association for the Assessment and Accreditation of Laboratory Animal Care accredited facility. For BT474, mouse xenograft studies' mice were anesthetized and one 17β-estradiol 60-day release pellet (Innovative Research of America, catalog no. SE-121) was injected subcutaneously via 10 gauge precision trochar into the lateral side of the neck. A total of 1 × 106 BT474 cells, 5 × 106 MDA-MB-468 cells, 2 × 106 WHIM12 cells, 5 × 105 A7C11 cells, and 5 × 105 Brpkp110 cells in 100 μL PBS:Matrigel (growth factor–reduced; Corning, catalog no. 356231) were subcutaneously injected into 5- to 6-week-old female NSG mice (Wistar). For BT474 orthografts, mice were injected with 1 × 106 cells into the fifth mammary fat pad (50 μL) of NSG mice with estrogen implants. Following tumor establishment in all groups, mice were randomized and treated daily by intraperitoneally injection or oral gavage with vehicle (10% DMSO, 20% solutol, 70% water containing 0.5% Tween20) or 100 mg/kg VY-3-135. Tumors were measured thrice weekly via caliper measurement and tumor volume calculated as (L × W2)/2 (where L is the longer of the two measurements). Lab members were not blinded to the treatment groups. At the study, conclusion all tumors were resected and processed fresh or frozen on dry ice and stored at -80°C for downstream analysis. No significant differences in body weight were noted between groups.

LC/MS-based metabolomics

All metabolomic experiments were performed in serum-like modified Eagle medium (SMEM), containing 54 different nutrients that are found in the bloodstream at concentrations that are physiologically relevant to humans, the formula of which has been described previously (6, 33, 34). SMEM was supplemented with 10% or 1% dialyzed FBS (Life Technologies). Cells were grown in normoxia (atmospheric oxygen) or hypoxia (1% oxygen) in uniformly labeled 13C2-acetate (0.1 mmol/L; Cambridge Isotope Laboratories) for 24 hours. For glucose tracing, SMEM was supplemented with 10% dialyzed FBS and grown in normoxia in uniformly labeled 13C6-glucose (5.5 mmol/L; Cambridge Isotope Laboratories) for 4 hours. For extraction of metabolites from cultured cells, cells were washed once in ice-cold PBS and extracted by adding LC/MS-grade methanol/acetonitrile/water (5:3:2). Plates were incubated at 4°C for 5 minutes on a rocker and the extraction solution was collected. Metabolite extracts were cleared by centrifugation. Supernatants were transferred to LC/MS silanized glass vials with PTFE caps and stored at -80°C until LC/MS analysis. For extraction from tumors, tumor-bearing mice were sacrificed by exsanguination under anesthesia and tumors were immediately excised and flash frozen in liquid nitrogen. Frozen tumors were weighed and then extracted at 40 mg/mL in extraction solution using a tissue homogenizer (Bullet Blender) and stainless steel beads. The metabolite extract was cleared twice by centrifugation and transferred to glass vials.

LC/MS analysis was performed on a Q Exactive HF-X mass spectrometer (Thermo Fisher Scientific) equipped with a HESI II probe and coupled to a Vanquish Horizon UHPLC system (Thermo Fisher Scientific). 0.002 mL of sample is injected and separated by HILIC chromatography on a ZIC-pHILIC 2.1-mm. Peak analysis was performed using an annotated compound library and TraceFinder 4.1 software.

LC/MS analysis of fatty acid extracts

For stable isotope tracer studies, mice were provided 2% heavy labeled acetate (either D3-acetate or 13C2-acetate) in the drinking water for two days. 90 minutes prior to sacrifice mice were given an intraperitoneal bolus of 2 g/kg heavy acetate. Tumor tissue was homogenized in LC/MS grade methanol (40 mg/mL) using gentleMACS M Tubes (Miltenyi Biotec). Cultured cells were washed in PBS and scraped into methanol. Chloroform and deoxygenated ice-cold PBS were added to samples then vortexed and centrifuged. The lower phase was collected and dried under nitrogen. Samples were redissolved in 0.3 mol/L potassium hydroxide in 90% methanol and incubated at 70°C for 1 hour to saponify the lipids. Formic acid and hexane were added, samples were vortexed, and briefly centrifuged to phase separate. The upper phase was dried under a nitrogen before resuspension in methanol. LC/MS analysis was performed using C18 reverse-phase chromatography and Thermo Q-Exactive HF-X mass spectrometer. Raw data analysis was performed using TraceFinder 4.1 software.

Development of an in vitro biochemical assay for ACSS2 inhibitors

Since ACSS2 catalyzes the conversion of acetate, coenzyme A and ATP into acetyl-CoA, AMP and PPi, we optimized a TranScreener AMP fluorescence polarization readout assay as a biochemical measurement of ACSS2 activity in vitro (Fig. 1A). Since salt can affect protein stability and enzyme kinetics, we tested different concentrations of recombinant human ACSS2 in the presence and absence of physiological-like levels of sodium chloride (Supplementary Fig. S1A). Sodium chloride boosted ACSS2 activity by five- to seven-fold and the activity of ACSS2 was linear at lower concentrations of enzyme. We next generated an AMP standard curve and dose response curves for coenzyme A (CoA) and ATP (Supplementary Fig. S1B–S1D). The optimal assay concentrations were determined to be 5 μmol/L CoA (Km = 1.7 nmol/L/hour; Vmax = 14.6) and 10 μmol/L ATP (Km = 35.2 nmol/L/hour; Vmax = 147.4).

Figure 1.

VY-3-135 is a potent, stable ACSS2 inhibitor with good bioavailability. A, Schematic of the forward reaction catalyzed by ACSS2. B and C, Chemical structures of VY-3-249 and VY-3-135 and IC50 determinations for inhibitors against ACSS1 (blue squares) and ACSS2 (red circles). Data points represent mean (filled shapes) and replicates (empty shapes). D and E, Mouse and human microsomal stability assay for VY-3-135 and VY-3-249. n = 1. F, Pharmacokinetic analysis of VY-3-135 by oral gavage, intraperitoneal, and intravenous injection. Data represent mean ± SD, n = 3 mice/group/time point. Table describes the calculations of pharmacokinetic analysis.

Figure 1.

VY-3-135 is a potent, stable ACSS2 inhibitor with good bioavailability. A, Schematic of the forward reaction catalyzed by ACSS2. B and C, Chemical structures of VY-3-249 and VY-3-135 and IC50 determinations for inhibitors against ACSS1 (blue squares) and ACSS2 (red circles). Data points represent mean (filled shapes) and replicates (empty shapes). D and E, Mouse and human microsomal stability assay for VY-3-135 and VY-3-249. n = 1. F, Pharmacokinetic analysis of VY-3-135 by oral gavage, intraperitoneal, and intravenous injection. Data represent mean ± SD, n = 3 mice/group/time point. Table describes the calculations of pharmacokinetic analysis.

Close modal

We obtained a previously described quinoxaline compound 1-(2,3-di(thiophen-2-yl)quinoxalin-6-yl)-3-(2-methoxyethyl)urea, hereafter referred to as VY-3-249, which has been shown to inhibit ACSS2 activity in vitro (Fig. 1B) (15, 35, 36). We additionally created a chemical synthesis scheme for another compound (R)-1-ethyl-2-(hydroxydiphenylmethyl)-N-(2-hydroxypropyl)-1H-benzo[d]imidazole- 6-carboxamide, referred to as VY-3-135, that was identified in the same screen as VY-3-249, but was not further characterized at that time (Fig. 1C; ref. 15). Using our in vitro biochemical TranScreener assay, we determined that VY-3-249 has an IC50 value of 1,214 ± 128 nmol/L, which is in the range of a previous report (Fig. 1B and C; ref. 15). Remarkably, VY-3-135 is nearly 30-fold more potent than VY-3-249 and has an IC50 value of 44 ± 3.85 nmol/L (Fig. 1C).

Stability and pharmacokinetic analysis of VY-3-135

The aqueous solubility of VY-3-135 was 21.7 μmol/L, while VY-3-249 was poorly soluble (Supplementary Table S1). In addition, VY-3-135 had superior stability in both mouse and human liver microsomal assays (Fig. 1D and E). Given the approximately 30-fold higher activity of VY-3-135 over VY-3-249 and its better solubility and stability, we decided to test the pharmacokinetic parameters of VY-3-135 by oral gavage, intraperitoneal and intravenous administration in mice. VY-3-135 was completely absorbed by all routes of administration and demonstrated good exposure with well-behaved pharmacokinetics (Fig. 1F).

VY-3-135 does not inhibit ACSS1 or ACSS3 enzymatic activity

The AcCoA synthetase family consists of three members: ACSS1, ACSS2, and ACSS3. Human ACSS1 and ACSS3 have 48.56% and 36.24% protein identity with ACSS2. ACSS1 is a bona fide AcCoA synthetase (37), while ACSS3 has been suggested to be a propionyl-CoA synthetase (38). We therefore tested the ability of all three family members to catalyze the conversion of acetate to AcCoA and propionate to propionyl-CoA (Supplementary Fig. S1E). ACSS1 has high specific activity for both acetate and propionate as substrates. ACSS2 greatly prefers acetate, but also has measurable activity with propionate as a substrate (Supplementary Fig. S1E). ACSS3 strongly prefers propionate with virtually no activity with acetate. Importantly, neither VY-3-249 nor VY-3-135 displayed any inhibitory activity toward recombinant human ACSS1 or ACSS3 regardless of the substrate (Fig. 1B and C; Supplementary Fig. S1F and S1G). These results suggest that VY-3-135 is specific to ACSS2 among the AcCoA synthetase family of enzymes.

VY-3-135 is predicted to be a transition-state mimetic of ACSS2

There are currently no published crystal structures of human ACSS2, but there is an X-ray crystal structure of Salmonella typhimurium AcCoA synthetase complexed with adenosine-5′-propylphosphate and CoA (25). This structure is suggested to be highly relevant because the nucleoside binding site between human, mouse, and salmonella ACSS2 is almost identical, where the only difference is an aromatic tryptophan residue present in salmonella versus an aromatic phenylalanine residue in human and mouse (Fig. 2A). Supplementary Figure S2A depicts the geometry of the transition state of the ACSS2-catalyzed reaction where the sulfur of the CoA attacks the carbonyl carbon of the acetate of acetyl-AMP to form a tetrahedral intermediate. VY-3-135 resembles this transition state, and suggests that the benzimidazole ring in the compound may mimic the adenine moiety of the acetyl-AMP intermediate, while the tetrahedral carbon bears the hydroxy group potentially mimicking the oxyanion in the transition state (Supplementary Fig. S2A).

Figure 2.

Model of Salmonella enterica acetyl-CoA synthetase in complex with VY-3-135. A, Amino acid sequence alignment of yeast, salmonella, human, and mouse ACSS2. Key residues that create the nucleotide binding pocket of ACSS2 are highlighted in yellow. B, Docking of VY-3-135 into acetyl-CoA synthetase shows that the inhibitor has a preference for the acetyl-AMP binding site of the protein but also occupies a portion of the CoA site. C, Superposition of acetyl-CoA synthetase in complex with VY-3-135 with the published structure of acetyl-CoA synthetase in complex with adenosine-5′-propylphosphate and CoA. The overlay shows how VY-3-135 interferes with both acetyl-AMP and CoA binding. D, The surface (acetyl-CoA synthetase) and stick (VY-3-135) representation shows that the inhibitor fits into the site with no clashes with the protein.

Figure 2.

Model of Salmonella enterica acetyl-CoA synthetase in complex with VY-3-135. A, Amino acid sequence alignment of yeast, salmonella, human, and mouse ACSS2. Key residues that create the nucleotide binding pocket of ACSS2 are highlighted in yellow. B, Docking of VY-3-135 into acetyl-CoA synthetase shows that the inhibitor has a preference for the acetyl-AMP binding site of the protein but also occupies a portion of the CoA site. C, Superposition of acetyl-CoA synthetase in complex with VY-3-135 with the published structure of acetyl-CoA synthetase in complex with adenosine-5′-propylphosphate and CoA. The overlay shows how VY-3-135 interferes with both acetyl-AMP and CoA binding. D, The surface (acetyl-CoA synthetase) and stick (VY-3-135) representation shows that the inhibitor fits into the site with no clashes with the protein.

Close modal

Because VY-3-135 resembles a transition-state mimetic of the acetyl-AMP intermediate we expected it to occupy the acetyl-AMP binding site of the salmonella AcCoA synthetase (Fig. 2AD). The Swissdock results revealed several possible binding sites on AcCoA synthetase (Supplementary Fig. S2B). The sites are ranked based on ΔG with the most favorable ΔG being ranked as the top hit. The results indicate that VY-3-135 indeed binds the acetyl-AMP site with the most favorable energy (ΔG = −9.45). VY-3-135 also potentially occupies a pocket on the surface of the protein formed by residues E256, P258, L275, R475, T476, F478, and H481, but with less favorable energy (ΔG = −8.46) than the acetyl-AMP site, so we focused on site 1 (Supplementary Fig. S2B). Interestingly, although VY-3-135 occupies the pocket of the acetyl-AMP, the inhibitor is shifted toward the core of the protein also occupying a portion of the CoA binding site (Fig. 2B and D). The approximately 3 Å shift of VY-3-135 away from the position of acetyl-AMP is due to its bulky diphenyl moiety, which is too large to fit the acetyl-AMP pocket. According to the model, the adenosine-like moiety makes primarily hydrophobic interactions with the large side chains of W413 and W414 and the aliphatic portion of the side chains of G387, Q415, R515, R526. Hydrogen bonds are also mediated between this portion of the inhibitor and the guanine and carboxylate of the side chains of R515 and D500 (Fig. 2A). The diphenyl portion of VY-3-135 stacks against the large hydrophobic side chain of W309. VY-3-135 occupation of the acetyl-AMP and the CoA binding pockets of AcCoA synthetase prevents substrate binding and AcCoA synthesis.

VY-3-135 is a potent low nanomolar inhibitor of ACSS2 in cancer cells in vitro

We next wanted to test the effectiveness of VY-3-135 in cell-based assays. Previous studies showed that high ACSS2 expression correlates with high acetate uptake (6, 11), we therefore screened a panel of 12 breast cancer cell lines to identify those with high ACSS2 expression. BT474, MDA-MB-468, and SKBr3 cells all have high basal expression of ACSS2 (Fig. 3A). Our previous studies also showed that acetate uptake and its relative contribution to de novo fatty acid synthesis is highest under nutrient and oxygen stressed conditions (6), conditions that also induce expression of ACSS2 (6, 11). We therefore incubated SKBr3 cells for 24 hours in 13C2-acetate in normoxic, nutrient-replete conditions (N10) and in hypoxic, low lipid conditions (H1). Lipid extracts were saponified and cellular fatty acids were analyzed by LC/MS for carbon-13 labeling from 13C2-acetate (Supplementary Fig. S3A). It is important to note that all in vitro stable isotope tracing experiments are performed in cell culture medium that contains 59 different nutrients at serum-like relevant concentrations (6, 33, 34). 13C2-acetate tracing into SKBr3 cells showed that hypoxia and low lipid stress caused a significant increase in carbon-13 labeling of the saturated fatty acid palmitate (Fig. 3B), suggesting increased ACSS2 activity in metabolically stressed cancer cells. Importantly, addition of VY-3-135 completely blocked fatty acid synthesis from acetate in SKBr3 cells (Fig. 3B). Similar results were obtained in BT474 and MDA-MB-468 cells (Fig. 3C; Supplementary Fig. S3B).

Figure 3.

VY-3-135 is a potent inhibitor of ACSS2 in cells. A, Immunoblot for ACSS2, EGFR, HER2 expression in a panel of human breast cancer cell lines. GAPDH was used as the loading control for ACSS2 and EGFR. ACTB was used as the loading control for HER2. B, Enrichment of 100 μmol/L 13C2-acetate in the intracellular palmitate pool in SKBr3 cells treated with vehicle or VY-3-135 and cultured in normoxia and SMEM + 10% serum (N10) or hypoxia and SMEM+1% serum (H1) over a 24-hour period. Data represent mean ± SD, n = 3. C, Enrichment of 100 μmol/L 13C2-acetate into palmitate in BT474 cells cultured in H1 conditions over a 24-hour period using a 10-fold dilution series of VY-3-135. Data represent mean ± SD, n = 3. D, Growth of BT474 and SKBr3 cells in 10 μmol/L VY-3-135 for 72 hours in H1 culture conditions supplemented with 200 μmol/L sodium acetate. Data are mean ± SD, n = 2 performed in triplicate. P values are Student t tests. E, Enrichment of 13C2-acetate into the intracellular citrate pool. Experimental parameters were identical to C. For all metabolomic data Padj values are reported on the graphs. P values were generated by two-way ANOVA with Tukey multiple comparison testing of mole percent enrichment of carbon-13.

Figure 3.

VY-3-135 is a potent inhibitor of ACSS2 in cells. A, Immunoblot for ACSS2, EGFR, HER2 expression in a panel of human breast cancer cell lines. GAPDH was used as the loading control for ACSS2 and EGFR. ACTB was used as the loading control for HER2. B, Enrichment of 100 μmol/L 13C2-acetate in the intracellular palmitate pool in SKBr3 cells treated with vehicle or VY-3-135 and cultured in normoxia and SMEM + 10% serum (N10) or hypoxia and SMEM+1% serum (H1) over a 24-hour period. Data represent mean ± SD, n = 3. C, Enrichment of 100 μmol/L 13C2-acetate into palmitate in BT474 cells cultured in H1 conditions over a 24-hour period using a 10-fold dilution series of VY-3-135. Data represent mean ± SD, n = 3. D, Growth of BT474 and SKBr3 cells in 10 μmol/L VY-3-135 for 72 hours in H1 culture conditions supplemented with 200 μmol/L sodium acetate. Data are mean ± SD, n = 2 performed in triplicate. P values are Student t tests. E, Enrichment of 13C2-acetate into the intracellular citrate pool. Experimental parameters were identical to C. For all metabolomic data Padj values are reported on the graphs. P values were generated by two-way ANOVA with Tukey multiple comparison testing of mole percent enrichment of carbon-13.

Close modal

We next isolated polar metabolites in order to analyze carbon-13 labeling of the cytosolic metabolite UDP-N-acetylglucosamine (UDP-GlcNAc). UDP-GlcNAc is synthesized from glucosamine 6-P and AcCoA by the cytosolic enzyme glucosamine-phosphate N-acetyl-transferase (GNPNAT1) and is an alternative readout of cytosolic AcCoA metabolism. VY-3-135 blocked incorporation of 13C2-acetate into UDP-GlcNAc (Supplementary Fig. S3C and S3D). Finally, we followed the growth of BT474 and SKBr3 cells treated with VY-3-135. Inhibition of ACSS2 in hypoxia and low lipid conditions caused modest growth inhibition in vitro over a 72-hour period (Fig. 3D).

To test for an off-target effect on ACSS1 in cells, we used citrate as a readout of ACSS1 activity because it is a mitochondrial AcCoA synthetase (Supplementary Fig. S3A). VY-3-135 did not decrease 13C2-acetate labeling of citrate in BT474 cells (Fig. 3E). Because citrate is strictly synthesized in the mitochondria, this result agrees with the biochemical assay results in Fig. 1C and further suggests that VY-3-135 has little to no activity against ACSS1. We did note a small increase in acetate contribution to the TCA cycle as evidenced by increased M+2 citrate in hypoxia and low serum-stressed cells treated with VY-3-135 (Fig. 3E). This is likely due to increased availability of acetate due to loss of ACSS2 activity that is then available for use by the mitochondria. In sum, the cell-based assays suggested that VY-3-135 is potent on-target ACSS2 inhibitor in cells.

Targeting ACSS2 inhibits mouse TNBC tumor growth in ACSS2high tumors, but not ACSS2low tumors

Having established that VY-3-135 is a potent ACSS2 inhibitor in vitro and in cells, we now wanted to test the inhibitor in preclinical models. We predicted that tumors with high ACSS2 expression (ACSS2high) would be more sensitive to ACSS2 inhibition compared with low ACSS2–expressing tumors (ACSS2low). To address this, we tested two mouse triple-negative breast cancer (TNBC) models, ACSS2high Brpkp110 (p53−/−/KrasG12D/+/Pik3ca-myr) and ACSS2low A7C11 (p53−/−/KrasG12D/+; Supplementary Fig. S4A). Brpkp110 and A7C11 are derived from mice with loss of p53 and activation of the KRAS (Supplementary Fig. S4B and S4C). TP53 is the most frequently mutated gene in breast cancer, while KRAS is mutated in less than 1% of breast cancer. However, MAPK signaling involving KRAS is one of the most important downstream mediators of EGFR and human EGFR 2 (ERBB2; also known as HER2) signaling, which are also two commonly activated oncogenes in breast cancer. Indeed, inhibition of EGFR or HER2 signaling by the small-molecule inhibitor lapatinib decreases phosphorylation of ERK in MDA-MB-468, BT474, and SKBr3 cells, suggesting a robust activation of the MAPK pathway by EGFR/HER2 signaling in breast cancer (Supplementary Fig. S4D). Compared with A7C11 cells, Brpkp110 cells additionally express constitutively active PI3K (Supplementary Fig. S4E). PI3K is one of the most frequently mutated enzymes in breast cancer and, like the MAPK pathway, the PI3K–AKT pathway is also activated downstream of EGFR and HER2 receptors in breast cancer (Supplementary Fig. S4D). Although A7C11 and Brpkp110 cells lack EGFR and HER2, they do maintain downstream signaling pathways through activation of KRAS and PI3K (Supplementary Fig. S4F).

We compared the labeling of palmitate by 13C2-acetate in ACSS2low A7C11 and ACSS2high Brpkp110 cells. Brpkp110 cells displayed higher incorporation of acetate into palmitate compared with A7C11 cells and the level of incorporation was further enhanced by hypoxia and low serum stress (Fig. 4A). Acetate-dependent labeling of palmitate by 13C2-acetate was efficiently blocked by VY-3-135 treatment in both cell lines (Fig. 4A). We also found that VY-3-135 did not affect 13C6-glucose–dependent palmitate synthesis, suggesting that the inhibition of fatty acid synthesis in VY-3-135–treated Brpkp110 cells was specific to ACSS2 and that other fatty acid synthesis enzymes were not affected by VY-3-135 (Supplementary Fig. S4G).

Figure 4.

Knockout or VY-3-135 inhibition of ACSS2 inhibits tumor growth. A, Enrichment of 100 μmol/L 13C2-acetate into palmitate in A7C11 and Brpkp110 cells cultured in N10 and H1 conditions over a 24-hour period in the presence and absence of VY-3-135. Data represent mean ± SD, n = 3. Padj values are reported on the graph. Two-way ANOVA Tukey multiple comparison testing of mole percent enrichment of carbon-13. B, Immunoblot for ACSS2 in A7C11 and Brpkp110 pools after CRISPR-Cas9 targeting of Acss2. Lysates were prepared from cells grown in N10 or H1 over a 24-hour period. sgNTC, single guide RNA against nontargeting control. sgACSS2, single guide RNA against Acss2. C, CRISPR-Cas9 knockout of Acss2 in A7C11 cells has a modest effect on tumor growth. Data represent mean ± SEM with ANOVA P value displayed, n = 5. D, CRISPR-Cas9 knockout of Acss2 in Brpkp110 cells causes a significant decrease in tumor growth. Data represent mean ± SEM with ANOVA P value displayed, n = 5. E, VY-3-135 treatment (100 mpk daily i.p.) causes a significant decrease in Brpkp110 tumor growth. Data represent mean ± SEM with ANOVA P value displayed, n = 5.

Figure 4.

Knockout or VY-3-135 inhibition of ACSS2 inhibits tumor growth. A, Enrichment of 100 μmol/L 13C2-acetate into palmitate in A7C11 and Brpkp110 cells cultured in N10 and H1 conditions over a 24-hour period in the presence and absence of VY-3-135. Data represent mean ± SD, n = 3. Padj values are reported on the graph. Two-way ANOVA Tukey multiple comparison testing of mole percent enrichment of carbon-13. B, Immunoblot for ACSS2 in A7C11 and Brpkp110 pools after CRISPR-Cas9 targeting of Acss2. Lysates were prepared from cells grown in N10 or H1 over a 24-hour period. sgNTC, single guide RNA against nontargeting control. sgACSS2, single guide RNA against Acss2. C, CRISPR-Cas9 knockout of Acss2 in A7C11 cells has a modest effect on tumor growth. Data represent mean ± SEM with ANOVA P value displayed, n = 5. D, CRISPR-Cas9 knockout of Acss2 in Brpkp110 cells causes a significant decrease in tumor growth. Data represent mean ± SEM with ANOVA P value displayed, n = 5. E, VY-3-135 treatment (100 mpk daily i.p.) causes a significant decrease in Brpkp110 tumor growth. Data represent mean ± SEM with ANOVA P value displayed, n = 5.

Close modal

We used CRISPR-Cas9 technology and multiple different guide RNAs to delete Acss2 in Brpkp110 and A7C11 mouse breast cancer cells (Fig. 4B; Supplementary Fig. S4A). The guide RNAs are specific to ACSS2 and do not match any sequence in ACSS1 or impact ACSS1 expression (Supplementary Fig. S4H). We then tested the impact of Acss2 knockout on the growth of ACSS2low A7C11 tumors and ACSS2high Brpkp110 tumors. CRISPR-mediated depletion of ACSS2 had a modest effect on A7C11 tumor growth (Fig. 4C). In contrast, Brpkp110 ACSS2high tumor growth was strongly affected by knockout of ACSS2 (Fig. 4D).

Because we observed significant reduction in tumor growth in sgACSS2 Brpkp110 tumors, we next wanted to test if VY-3-135 could repress Brpkp110 tumor growth. In agreement with the CRISPR knockout studies, VY-3-135 treatment caused marked inhibition of Brpkp110 tumor growth (Fig. 4E). To validate that tumor growth inhibition was due to the on-target effect of VY-3-135 on ACSS2, we treated sgACSS2 Brpkp110 tumor-bearing mice with VY-3-135. Although ACSS2 knockout tumors were smaller and grew more slowly, there was no further effect on tumor growth by VY-3-135, suggesting that the antitumor growth properties of VY-3-135 are ACSS2 specific (Supplementary Fig. S4I). We additionally stained the tumors for Ki67 and found that VY-3-135–treated tumors were less proliferative, suggesting that targeting ACSS2 may cause inhibition of tumor growth by decreasing proliferation (Supplementary Fig. S4J).

VY-3-135 inhibits human TNBC tumor growth in ACSS2high tumors, but not ACSS2low tumors

We next wanted to test VY-3-135 in human models of breast cancer. Figure 3A showed that the highest expression of ACSS2 tended to occur in human breast cancer cell lines that also had high expression of EGFR and HER2, such as BT474, BT20, MDA-MB-468, and SKBr3 cells. Similar correlations in gene expression were also observed in human breast cancer databases (Supplementary Fig. S5A–S5C). EGFR is the requisite binding partner of HER2 and is most commonly elevated in TNBC where >50% of patient tumors have elevated EGFR expression (39, 40).

We compared VY-3-135 sensitivity in two different human TNBC models, one with high ACSS2 expression (MDA-MB-468) and one with low ACSS2 expression (WHIM12; Fig. 5A and B). WHIM12 is a stable patient-derived xenograft line derived from a TNBC patient (41). VY-3-135 repressed MDA-MB-468 (ACSS2high) tumor growth, but was mostly ineffective at blocking WHIM12 (ACSS2low) growth (Fig. 5C and D). These results further suggest that ACSS2 expression itself could be a useful marker for sensitivity to ACSS2 inhibitors. We therefore also tested the efficacy of VY-3-135 on BT474 tumors that have the highest expression of ACSS2 (Figs. 3A and 5A; ref. 6). Strikingly, VY-3-135 treatment of BT474 tumor–bearing mice completely abrogated tumor growth over two weeks of treatment (P = 0.0002; Fig. 5E). All VY-3-135–treated mice had palpable BT474 tumors at day 14; however, two of the tumors in VY-3-135–treated group were not palpable by the end of the regimen, with one tumor fully regressing (Supplementary Fig. S5D). Conversely, all vehicle-treated mice had palpable tumors that progressed over the same period. We repeated the BT474 tumor studies as orthografts in mammary tissue and obtained similar results upon VY-3-135 treatment (Supplementary Fig. S5E). The orthotopic experiments, though promising, represent only a small cohort, and future orthotopic experiments are necessary to fully address the still outstanding issue of treating breast tumors that reside in the mammary fat pad. Altogether, these results demonstrate for the first time that pharmacologic targeting of ACSS2 can inhibit human tumor growth in preclinical models.

Figure 5.

VY-3-135 inhibits growth of ACSS2high but not ACSS2low human breast tumors. A, Immunoblots for ACSS2, ACSS1, and HER2 expression. B, Immunoblot for EGFR expression. C, WHIM12 tumor growth ± VY-3-135. Data represent mean ± SEM with ANOVA P value displayed, n = 6 per group. Black arrow, start of treatment (100 mpk daily orally). D, MDA-MB-468 tumor growth ± VY-3-135. Data represent mean ± SEM with ANOVA P value displayed, n = 8 per group. Black arrow, start of treatment (100 mpk daily orally). E, BT474 tumor growth ± VY-3-135. Black arrow, start of treatment (100 mpk daily i.p.). ANOVA P value is displayed, n = 5 mice per group. F and G, D3-acetate dependent labeling of palmitate in WHIM12 tumors (F) and MDA-MB-468 tumors (G). N.D., not detected. n ≥ 5. H and I, D3-acetate dependent labeling of UDP-GlcNAc in WHIM12 tumors (H) and MDA-MB-468 tumors (I). J and K, D3-acetate–dependent labeling of citrate in WHIM12 tumors (J) and MDA-MB-468 tumors (K). Data represent mean ± SD, n ≥ 5. For all metabolomic data, Padj values are reported on the graphs. P values were generated from two-way ANOVA Tukey multiple comparison testing of mole percent enrichment of carbon-13.

Figure 5.

VY-3-135 inhibits growth of ACSS2high but not ACSS2low human breast tumors. A, Immunoblots for ACSS2, ACSS1, and HER2 expression. B, Immunoblot for EGFR expression. C, WHIM12 tumor growth ± VY-3-135. Data represent mean ± SEM with ANOVA P value displayed, n = 6 per group. Black arrow, start of treatment (100 mpk daily orally). D, MDA-MB-468 tumor growth ± VY-3-135. Data represent mean ± SEM with ANOVA P value displayed, n = 8 per group. Black arrow, start of treatment (100 mpk daily orally). E, BT474 tumor growth ± VY-3-135. Black arrow, start of treatment (100 mpk daily i.p.). ANOVA P value is displayed, n = 5 mice per group. F and G, D3-acetate dependent labeling of palmitate in WHIM12 tumors (F) and MDA-MB-468 tumors (G). N.D., not detected. n ≥ 5. H and I, D3-acetate dependent labeling of UDP-GlcNAc in WHIM12 tumors (H) and MDA-MB-468 tumors (I). J and K, D3-acetate–dependent labeling of citrate in WHIM12 tumors (J) and MDA-MB-468 tumors (K). Data represent mean ± SD, n ≥ 5. For all metabolomic data, Padj values are reported on the graphs. P values were generated from two-way ANOVA Tukey multiple comparison testing of mole percent enrichment of carbon-13.

Close modal

VY-3-135 potently inhibits ACSS2-dependent fatty acid metabolism but has no effect on gene expression in tumors

We next checked the on-target activity of VY-3-135 in tumors. ACSS2 has been shown to affect histone and transcription factor acetylation in the nucleus of cells, which then affects gene transcription (7, 11, 15–17, 35, 36, 42). For instance, liver cancer cells exposed to hypoxia upregulate fatty acid metabolism genes, such as FASN, in an ACSS2-dependent manner associated with changes in histone acetylation (43), while a more recent study showed that Acss2−/− mice fed a high-fat diet were less obese than wild-type mice and that this was regulated, in part, by an inability to regulate the expression of lipid metabolism genes in the liver (44). On the basis of the literature, we tested whether inhibition of ACSS2 affects gene transcription in vivo. QuantSeq 3′ mRNA sequencing of BT474 tumor tissue from vehicle and VY-3-135–treated mice showed that the effects of VY-3-135 on gene transcription in vivo were minimal. 248 genes were differentially regulated at nominal significance of P < 0.05, but none of those genes passed an FDR < 5%. Furthermore, of the 248 differentially regulated genes only 7 changed by more than 2-fold and none by more than 2.5-fold (Supplementary Fig. S6A and S6B). As such, an IPA of the RNA-seq data did not predict significant alteration of any canonical pathways and the activity of only five transcriptional regulators (P < 0.05; Z > 2) were modestly altered, with inactivation of estrogen receptor signaling and activation of p53 as the most significantly altered pathways (Supplementary Fig. S6C). These results suggested that inhibition of ACSS2 by VY-3-135 in BT474 tumors did not substantially affect gene transcription.

The degree of nuclear ACSS2 accumulation in cells is often regulated by metabolic stress (7, 11, 16, 17, 42). IHC staining for ACSS2 indicated localization to both the nucleus and the cytosol in BT474 tumor cells (Supplementary Fig. S6D). To further probe the localization of ACSS2 in the nucleus, nuclear extracts were subfractionated into soluble and chromatin-bound fractions. Overall, ACSS2 was mostly localized to the soluble fraction of the nucleus and the cytosol with very little bound to the chromatin fraction (Supplementary Fig. S6E). The lack of ACSS2 bound to chromatin could possibly explain the lack of gene regulation after pharmacologic inhibition of ACSS2 in BT474 cells.

The lack of transcriptional effects by VY-3-135 suggests that tumor growth inhibition was more likely to be due to direct effects on tumor acetate metabolism. To measure tumor acetate metabolism, we supplemented the drinking water of BT474, MDA-MB-468, and WHIM12 tumor–bearing mice with heavy labeled acetate (D3-acetate or13C2-acetate) for 48 hours and additionally gave a single intraperitoneal bolus of heavy acetate 90 minutes prior to tumor harvesting. Heavy acetate provision to the mice did not alter plasma acetate levels (Supplementary Fig. S7A). Lipids and polar metabolites were extracted from matching tumor samples and analyzed by LC/MS. Profiling of fatty acids and polar metabolites showed that TNBC WHIM12 ACSS2low tumors had less palmitate labeling and less UDP-GlcNAc labeling than TNBC MDA-MB-468 ACSS2high tumors (Fig. 5FI). More importantly, VY-3-135 caused a marked decrease in both palmitate synthesis and UDP-GlcNAc synthesis from heavy labeled acetate, with acetate labeling often becoming undetectable at higher order isotopologs in VY-3-135–treated mice (Fig. 5FI; Supplementary Fig. S7B). These results clearly demonstrate that VY-3-135 is able to inhibit acetate metabolism in tumors in vivo and that measurement of palmitate and/or UDP-GlcNAc labeling by a heavy labeled acetate can be used as a biomarker of ACSS2 inhibitor activity in vivo.

In accordance with our in vitro findings from Fig. 3E, we did not observe any inhibitory effect on D3-acetate or 13C2-acetate labeling of citrate in tumors, further validating that VY-3-135 does not inhibit ACSS1 (Fig. 5J and K; Supplementary Fig. S7C). We did observe an increase in D3-acetate labeling of citrate in MDA-MB-468 tumors similar to what we observed in Fig. 3E for BT474 cells. However, the data suggest that an increase in acetate contribution to the TCA cycle does not confer a growth advantage because MDA-MB-468 tumor growth was still inhibited.

In our study, we describe a previously uncharacterized inhibitor of ACSS2. We set out by optimizing biochemical assays for assessing ACSS1, ACSS2, and ACSS3 activity in vitro. We also developed methods for routine chemical synthesis of VY-3-135. We find that the properties of VY-3-135 are a dramatic improvement over the only other currently known ACSS2 inhibitor, VY-3-249. Compared with VY-3-249, VY-3-135 is a markedly superior compound that is more soluble, more stable, and more potent. Docking simulations show that the effectiveness of VY-3-135 may be due to its ability to span the acetyl-AMP pocket as well as the CoA pocket. Future cocrystallization studies of VY-3-135 with human ACSS2 will help to determine the binding site and exact mechanism of inhibition and help guide development of more potent inhibitors.

Our in vivo inhibitor studies are the first to show that pharmacologic inhibition of ACSS2 causes tumor growth inhibition and regression. Given the diverse role of AcCoA, it is likely that acetate-fueled production of AcCoA fills many roles for the cancer cell, none of which are mutually exclusive (45). One major consumer of cytosolic AcCoA in a proliferating cancer cell is fatty acid biosynthesis. Indeed, many cancers undergo a “lipogenic switch” wherein they switch from taking up lipids to using de novo lipid biosynthesis pathways as their main source for membrane biogenesis (46, 47). The lipogenic switch can occur in response to metabolic stress, such as low nutrient availability and hypoxia, as well as the activation of oncogenes and/or loss of tumor suppressors (6, 48, 49). Our previous in vitro studies in cultured cells showed that AcCoA generated from acetate can contribute to fatty acid biosynthesis during metabolic stress (6, 50). Many other studies have also traced acetate into cancer cells in vitro (6, 7, 11, 50, 51), but very few studies have traced acetate into tumors in vivo (14, 52), and no studies have looked at acetate metabolism in tumors after inhibition of ACSS2. Our in vivo stable isotope tracer studies are the first to show that targeting ACSS2 is an effective means to substantially reduce the use of acetate for fatty acid biosynthesis in tumors. This finding is most relevant to tumors that are dependent on acetate metabolism for AcCoA, such as ACSS2high tumors, and is likely the main driver behind VY-3-135–mediated tumor growth inhibition and regression.

Successful translation of ACSS2 inhibitors to the clinic will require methods to identify patients that are likely to respond to treatment. The growth of BT474, MDA-MB-468, and Brpkp110 tumors, which have high expression of ACSS2 was affected by VY-3-135, but not WHIM12 and A7C11 tumors, which have low ACSS2 expression. This data supports the notion that high ACSS2 expression predicts acetate metabolism dependency and could be used to identify patients with acetate avid tumors.

ACSS2 expression is mainly controlled by SREBP transcription factors (6, 8). As such, its expression is intimately linked with activation of transcriptional programs associated with fatty acid and cholesterol biosynthesis and may therefore be a marker of more “lipogenic” tumors. Mining patient data suggested that HER2+ breast cancer and TNBC, particularly EGFR+ TNBC, have relatively high ACSS2 expression. Interestingly, EGFR mutation is also commonly found in glioblastoma (53), which is another cancer type that readily consumes acetate (14) and has been shown to be sensitive to inhibition of ACSS2 activity (7). EGFR and HER2 stimulate PI3K-AKT-mTORC signaling, which is a known activator of SREBP and lipid biosynthesis (54–56). Future studies will be needed to further investigate a potential link between EGFR/HER2 signaling, the PI3K–AKT–mTOR pathway, ACSS2 expression, and acetate metabolism. For now, the main predictor of sensitivity to acetate metabolism inhibition seems to be ACSS2 expression.

We propose that ACSS2 represents a novel therapeutic target for treating any tumor that has high ACSS2 expression and readily consumes acetate as an alternative nutrient source. Fortunately, 11C-acetate PET is an established imaging tool in the clinic that has been extensively used for cancer detection (57–62) and has also been used to study acetate-dependent fatty acid synthesis in preclinical models (15, 52). 11C-acetate PET imaging could easily be applied to help identify patients with cancer whose tumors are “hot” for acetate uptake. It could also serve as a pharmacodynamic biomarker to longitudinally monitor tumor responses in patients treated with ACSS2 inhibitors.

A number of publications have suggested a critical role for ACSS2 in the regulation of gene expression (7, 35, 36, 43, 63). However, we observed little to no effect of VY-3-135 on gene transcription in vivo. Indeed, only one gene had an FDR < 5% in VY-3-135–treated tumors, strongly suggesting that acetate metabolism is not a critical determinant of transcriptional responses in certain contexts. It is important to highlight that this result may be specific to BT474 tumors. Further studies in other tumor models are warranted to elucidate the intricate interplay between acetate metabolism and gene transcription, including the involvement of other factors such as acetyltransferases and transcription factors as well as the metabolic conditions under which ACSS2 influences gene transcription in cancer. One caveat of our study is that stress within the tumor microenvironment is likely to be highly dynamic and therefore bulk tumor sequencing could overlook more subtle regionalized increases in ACSS2 expression or the ability of ACSS2 to regulate gene expression in those regions.

Antimetabolite drugs have been a cornerstone of chemotherapy for decades. The dramatic increase in the number of clinical trials involving inhibitors of metabolic enzymes, such a mutant isocitrate dehydrogenase and glutaminase, attest to the potential of targeting metabolic pathways in cancer (64). We put forth that targeting acetate metabolism represents an unrealized opportunity for improving cancer treatment. Acetate is consumed by many different cancer types, even more so under conditions of metabolic stress (45). Because most, if not all, solid tumors experience some degree of hypoxia and nutrient stress during their evolution, targeting acetate metabolism has the potential to be broadly applicable to any acetate consuming cancer type. In addition, because most normal tissues barely consume acetate as a major nutrient source, targeting acetate metabolism exploits a unique metabolic vulnerability that is both cancer specific and likely to be safe for the patient. In summation, we performed a preclinical evaluation of a novel tool compound for use in future investigations into tumor acetate metabolism. Our results highlight the promise of ACSS2 inhibitors as a novel therapeutic modality in cancer and the potential to synergize ACSS2 inhibitors with other therapies.

No disclosures were reported.

K.D. Miller: Conceptualization, formal analysis, investigation, methodology, writing-review and editing. K. Pniewski: Formal analysis, investigation. C.E. Perry: Formal analysis, investigation. S.B. Papp: Investigation. J.D. Shaffer: Investigation. J.N. Velasco-Silva: Investigation. J.C. Casciano: Resources. T.M. Aramburu: Formal analysis, investigation, visualization. Y.V.V. Srikanth: Resources. J. Cassel: Resources, formal analysis, investigation. E. Skordalakes: Formal analysis, supervision, investigation, visualization. A.V. Kossenkov: Data curation, software, formal analysis. J.M. Salvino: Conceptualization, supervision, methodology, writing-review and editing. Z.T. Schug: Conceptualization, formal analysis, supervision, funding acquisition, validation, investigation, methodology, writing-review and editing.

This work was supported by grants from NIH NCI DP2 CA249950-01 (to Z.T. Schug), the W.W. Smith Charitable Trust (to Z.T. Schug), Susan G. Komen CCR19608782 (to Z.T. Schug), the V Foundation for Cancer Research (to Z.T. Schug) and NIH NCI T32 CA009171 (to K.D. Miller and J.C. Casciano). The Wistar Molecular Screening Facility and Genomics Facility are supported by NIH grant P30 CA010815. The Wistar Proteomic and Metabolomic Facility is supported in part by NIH grants R50 CA221838 and S10 OD023586. We would like to thank Hsin-Yao Tang, Nicole Gorman, Aaron Goldman, and Thomas Beer of the Wistar Institute Proteomic and Metabolomic core. We also would like to acknowledge the staff of the Wistar Institute Genomics Core and Animal Facility. We thank Seamus O’Connor, Michael Hulse, and Adam Cohen-Nowak for helpful insight and discussions in preparing the manuscript.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Al Tameemi
W
,
Dale
TP
,
Al-Jumaily
RMK
,
Forsyth
NR
. 
Hypoxia-modified cancer cell metabolism
.
Front Cell Dev Biol
2019
;
7
:
4
.
2.
Wilson
WR
,
Hay
MP
. 
Targeting hypoxia in cancer therapy
.
Nat Rev Cancer
2011
;
11
:
393
410
.
3.
Vaupel
P
. 
Hypoxia and aggressive tumor phenotype: implications for therapy and prognosis
.
Oncologist
2008
;
13
:
21
6
.
4.
Jing
X
,
Yang
F
,
Shao
C
,
Wei
K
,
Xie
M
,
Shen
H
, et al
Role of hypoxia in cancer therapy by regulating the tumor microenvironment
.
Mol Cancer
2019
;
18
:
157
.
5.
Muz
B
,
de la Puente
P
,
Azab
F
,
Azab
AK
. 
The role of hypoxia in cancer progression, angiogenesis, metastasis, and resistance to therapy
.
Hypoxia
2015
;
3
:
83
92
.
6.
Schug
ZT
,
Peck
B
,
Jones
DT
,
Zhang
Q
,
Grosskurth
S
,
Alam
IS
, et al
Acetyl-CoA synthetase 2 promotes acetate utilization and maintains cancer cell growth under metabolic stress
.
Cancer Cell
2015
;
27
:
57
71
.
7.
Li
X
,
Yu
W
,
Qian
X
,
Xia
Y
,
Zheng
Y
,
Lee
JH
, et al
Nucleus-translocated ACSS2 promotes gene transcription for lysosomal biogenesis and autophagy
.
Mol Cell
2017
;
66
:
684
97
.
8.
Luong
A
,
Hannah
VC
,
Brown
MS
,
Goldstein
JL
. 
Molecular characterization of human acetyl-CoA synthetase, an enzyme regulated by sterol regulatory element-binding proteins
.
J Biol Chem
2000
;
275
:
26458
66
.
9.
Xu
H
,
Luo
J
,
Ma
G
,
Zhang
X
,
Yao
D
,
Li
M
, et al
Acyl-CoA synthetase short-chain family member 2 (ACSS2) is regulated by SREBP-1 and plays a role in fatty acid synthesis in caprine mammary epithelial cells
.
J Cell Physiol
2018
;
233
:
1005
16
.
10.
Yoshii
Y
,
Furukawa
T
,
Yoshii
H
,
Mori
T
,
Kiyono
Y
,
Waki
A
, et al
Cytosolic acetyl-CoA synthetase affected tumor cell survival under hypoxia: the possible function in tumor acetyl-CoA/acetate metabolism
.
Cancer Sci
2009
;
100
:
821
7
.
11.
Bulusu
V
,
Tumanov
S
,
Michalopoulou
E
,
van den Broek
NJ
,
MacKay
G
,
Nixon
C
, et al
Acetate recapturing by nuclear acetyl-CoA synthetase 2 prevents loss of histone acetylation during oxygen and serum limitation
.
Cell Rep
2017
;
18
:
647
58
.
12.
Kim
JW
,
Tchernyshyov
I
,
Semenza
GL
,
Dang
CV
. 
HIF-1-mediated expression of pyruvate dehydrogenase kinase: a metabolic switch required for cellular adaptation to hypoxia
.
Cell Metab
2006
;
3
:
177
85
.
13.
Maher
EA
,
Marin-Valencia
I
,
Bachoo
RM
,
Mashimo
T
,
Raisanen
J
,
Hatanpaa
KJ
, et al
Metabolism of [U-13 C]glucose in human brain tumors in vivo
.
NMR Biomed
2012
;
25
:
1234
44
.
14.
Mashimo
T
,
Pichumani
K
,
Vemireddy
V
,
Hatanpaa
KJ
,
Singh
DK
,
Sirasanagandla
S
, et al
Acetate is a bioenergetic substrate for human glioblastoma and brain metastases
.
Cell
2014
;
159
:
1603
14
.
15.
Comerford
SA
,
Huang
Z
,
Du
X
,
Wang
Y
,
Cai
L
,
Witkiewicz
AK
, et al
Acetate dependence of tumors
.
Cell
2014
;
159
:
1591
602
.
16.
Chen
R
,
Xu
M
,
Nagati
J
,
Garcia
JA
. 
Coordinate regulation of stress signaling and epigenetic events by Acss2 and HIF-2 in cancer cells
.
PLoS One
2017
;
12
:
e0190241
.
17.
Chen
R
,
Xu
M
,
Nagati
JS
,
Hogg
RT
,
Das
A
,
Gerard
RD
, et al
The acetate/ACSS2 switch regulates HIF-2 stress signaling in the tumor cell microenvironment
.
PLoS One
2015
;
10
:
e0116515
.
18.
Lakhter
AJ
,
Hamilton
J
,
Konger
RL
,
Brustovetsky
N
,
Broxmeyer
HE
,
Naidu
SR
. 
Glucose-independent acetate metabolism promotes melanoma cell survival and tumor growth
.
J Biol Chem
2016
;
291
:
21869
79
.
19.
Allegrezza
MJ
,
Rutkowski
MR
,
Stephen
TL
,
Svoronos
N
,
Perales-Puchalt
A
,
Nguyen
JM
, et al
Trametinib drives T-cell-dependent control of KRAS-mutated tumors by inhibiting pathological myelopoiesis
.
Cancer Res
2016
;
76
:
6253
65
.
20.
Rutkowski
MR
,
Allegrezza
MJ
,
Svoronos
N
,
Tesone
AJ
,
Stephen
TL
,
Perales-Puchalt
A
, et al
Initiation of metastatic breast carcinoma by targeting of the ductal epithelium with adenovirus-cre: a novel transgenic mouse model of breast cancer
.
J Vis Exp
2014
;
51171
.
DOI: 10.3791/51171
.
21.
Rutkowski
MR
,
Stephen
TL
,
Svoronos
N
,
Allegrezza
MJ
,
Tesone
AJ
,
Perales-Puchalt
A
, et al
Microbially driven TLR5-dependent signaling governs distal malignant progression through tumor-promoting inflammation
.
Cancer Cell
2015
;
27
:
27
40
.
22.
Sheen
MR
,
Marotti
JD
,
Allegrezza
MJ
,
Rutkowski
M
,
Conejo-Garcia
JR
,
Fiering
S
. 
Constitutively activated PI3K accelerates tumor initiation and modifies histopathology of breast cancer
.
Oncogenesis
2016
;
5
:
e267
.
23.
Grosdidier
A
,
Zoete
V
,
Michielin
O
. 
SwissDock, a protein-small molecule docking web service based on EADock DSS
.
Nucleic Acids Res
2011
;
39
:
W270
7
.
24.
Brooks
BR
,
Brooks
CL
 III
,
Mackerell
AD
 Jr
,
Nilsson
L
,
Petrella
RJ
,
Roux
B
, et al
CHARMM: the biomolecular simulation program
.
J Comput Chem
2009
;
30
:
1545
614
.
25.
Gulick
AM
,
Starai
VJ
,
Horswill
AR
,
Homick
KM
,
Escalante-Semerena
JC
. 
The 1.75 A crystal structure of acetyl-CoA synthetase bound to adenosine-5′-propylphosphate and coenzyme A
.
Biochemistry
2003
;
42
:
2866
73
.
26.
Adams
PD
,
Afonine
PV
,
Bunkoczi
G
,
Chen
VB
,
Davis
IW
,
Echols
N
, et al
PHENIX: a comprehensive Python-based system for macromolecular structure solution
.
Acta Crystallogr D Biol Crystallogr
2010
;
66
:
213
21
.
27.
Langmead
B
,
Salzberg
SL
. 
Fast gapped-read alignment with Bowtie 2
.
Nat Methods
2012
;
9
:
357
9
.
28.
Li
B
,
Dewey
CN
. 
RSEM: accurate transcript quantification from RNA-Seq data with or without a reference genome
.
BMC Bioinformatics
2011
;
12
:
323
.
29.
Love
MI
,
Huber
W
,
Anders
S
. 
Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2
.
Genome Biol
2014
;
15
:
550
.
30.
Curtis
C
,
Shah
SP
,
Chin
SF
,
Turashvili
G
,
Rueda
OM
,
Dunning
MJ
, et al
The genomic and transcriptomic architecture of 2,000 breast tumours reveals novel subgroups
.
Nature
2012
;
486
:
346
52
.
31.
Cerami
E
,
Gao
J
,
Dogrusoz
U
,
Gross
BE
,
Sumer
SO
,
Aksoy
BA
, et al
The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data
.
Cancer Discov
2012
;
2
:
401
4
.
32.
Gao
J
,
Aksoy
BA
,
Dogrusoz
U
,
Dresdner
G
,
Gross
B
,
Sumer
SO
, et al
Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal
.
Sci Signal
2013
;
6
:
pl1
.
33.
Tardito
S
,
Oudin
A
,
Ahmed
SU
,
Fack
F
,
Keunen
O
,
Zheng
L
, et al
Glutamine synthetase activity fuels nucleotide biosynthesis and supports growth of glutamine-restricted glioblastoma
.
Nat Cell Biol
2015
;
17
:
1556
68
.
34.
Vande Voorde
J
,
Ackermann
T
,
Pfetzer
N
,
Sumpton
D
,
Mackay
G
,
Kalna
G
, et al
Improving the metabolic fidelity of cancer models with a physiological cell culture medium
.
Sci Adv
2019
;
5
:
eaau7314
.
35.
Mews
P
,
Donahue
G
,
Drake
AM
,
Luczak
V
,
Abel
T
,
Berger
SL.
Acetyl-CoA synthetase regulates histone acetylation and hippocampal memory
.
Nature
2017
;
546
:
381
6
.
36.
Mews
P
,
Egervari
G
,
Nativio
R
,
Sidoli
S
,
Donahue
G
,
Lombroso
SI
, et al
Alcohol metabolism contributes to brain histone acetylation
.
Nature
2019
;
574
:
717
21
.
37.
Fujino
T
,
Kondo
J
,
Ishikawa
M
,
Morikawa
K
,
Yamamoto
TT.
Acetyl-CoA synthetase 2, a mitochondrial matrix enzyme involved in the oxidation of acetate
.
J Biol Chem
2001
;
276
:
11420
6
.
38.
Yoshimura
Y
,
Araki
A
,
Maruta
H
,
Takahashi
Y
,
Yamashita
H.
Molecular cloning of rat acss3 and characterization of mammalian propionyl-CoA synthetase in the liver mitochondrial matrix
.
J Biochem
2017
;
161
:
279
89
.
39.
Dent
R
,
Trudeau
M
,
Pritchard
KI
,
Hanna
WM
,
Kahn
HK
,
Sawka
CA
, et al
Triple-negative breast cancer: clinical features and patterns of recurrence
.
Clin Cancer Res
2007
;
13
:
4429
34
.
40.
Masuda
H
,
Zhang
D
,
Bartholomeusz
C
,
Doihara
H
,
Hortobagyi
GN
,
Ueno
NT
. 
Role of epidermal growth factor receptor in breast cancer
.
Breast Cancer Res Treat
2012
;
136
:
331
45
.
41.
Li
S
,
Shen
D
,
Shao
J
,
Crowder
R
,
Liu
W
,
Prat
A
, et al
Endocrine-therapy-resistant ESR1 variants revealed by genomic characterization of breast-cancer-derived xenografts
.
Cell Rep
2013
;
4
:
1116
30
.
42.
Xu
M
,
Nagati
JS
,
Xie
J
,
Li
J
,
Walters
H
,
Moon
YA
, et al
An acetate switch regulates stress erythropoiesis
.
Nat Med
2014
;
20
:
1018
26
.
43.
Gao
X
,
Lin
SH
,
Ren
F
,
Li
JT
,
Chen
JJ
,
Yao
CB
, et al
Acetate functions as an epigenetic metabolite to promote lipid synthesis under hypoxia
.
Nat Commun
2016
;
7
:
11960
.
44.
Huang
Z
,
Zhang
M
,
Plec
AA
,
Estill
SJ
,
Cai
L
,
Repa
JJ
, et al
ACSS2 promotes systemic fat storage and utilization through selective regulation of genes involved in lipid metabolism
.
Proc Natl Acad Sci U S A
2018
;
115
:
E9499
E506
.
45.
Schug
ZT
,
Vande Voorde
J
,
Gottlieb
E
. 
The metabolic fate of acetate in cancer
.
Nat Rev Cancer
2016
;
16
:
708
17
.
46.
Rysman
E
,
Brusselmans
K
,
Scheys
K
,
Timmermans
L
,
Derua
R
,
Munck
S
, et al
De novo lipogenesis protects cancer cells from free radicals and chemotherapeutics by promoting membrane lipid saturation
.
Cancer Res
2010
;
70
:
8117
26
.
47.
Munir
R
,
Lisec
J
,
Swinnen
JV
,
Zaidi
N
. 
Lipid metabolism in cancer cells under metabolic stress
.
Br J Cancer
2019
;
120
:
1090
8
.
48.
Peck
B
,
Schug
ZT
,
Zhang
Q
,
Dankworth
B
,
Jones
DT
,
Smethurst
E
, et al
Inhibition of fatty acid desaturation is detrimental to cancer cell survival in metabolically compromised environments
.
Cancer Metab
2016
;
4
:
6
.
49.
Jain
IH
,
Calvo
SE
,
Markhard
AL
,
Skinner
OS
,
To
TL
,
Ast
T
, et al
Genetic screen for cell fitness in high or low oxygen highlights mitochondrial and lipid metabolism
.
Cell
2020
;
181
:
716
27
.
50.
Kamphorst
JJ
,
Chung
MK
,
Fan
J
,
Rabinowitz
JD
. 
Quantitative analysis of acetyl-CoA production in hypoxic cancer cells reveals substantial contribution from acetate
.
Cancer Metab
2014
;
2
:
23
.
51.
Liu
X
,
Cooper
DE
,
Cluntun
AA
,
Warmoes
MO
,
Zhao
S
,
Reid
MA
, et al
Acetate production from glucose and coupling to mitochondrial metabolism in mammals
.
Cell
2018
;
175
:
502
13
.
52.
Lewis
DY
,
Boren
J
,
Shaw
GL
,
Bielik
R
,
Ramos-Montoya
A
,
Larkin
TJ
, et al
Late imaging with [1-(11)C]acetate improves detection of tumor fatty acid synthesis with PET
.
J Nucl Med
2014
;
55
:
1144
9
.
53.
Furnari
FB
,
Fenton
T
,
Bachoo
RM
,
Mukasa
A
,
Stommel
JM
,
Stegh
A
, et al
Malignant astrocytic glioma: genetics, biology, and paths to treatment
.
Genes Dev
2007
;
21
:
2683
710
.
54.
Guo
D
,
Prins
RM
,
Dang
J
,
Kuga
D
,
Iwanami
A
,
Soto
H
, et al
EGFR signaling through an Akt-SREBP-1-dependent, rapamycin-resistant pathway sensitizes glioblastomas to antilipogenic therapy
.
Sci Signal
2009
;
2
:
ra82
.
55.
Duvel
K
,
Yecies
JL
,
Menon
S
,
Raman
P
,
Lipovsky
AI
,
Souza
AL
, et al
Activation of a metabolic gene regulatory network downstream of mTOR complex 1
.
Mol Cell
2010
;
39
:
171
83
.
56.
Kumar-Sinha
C
,
Ignatoski
KW
,
Lippman
ME
,
Ethier
SP
,
Chinnaiyan
AM
. 
Transcriptome analysis of HER2 reveals a molecular connection to fatty acid synthesis
.
Cancer Res
2003
;
63
:
132
9
.
57.
Spick
C
,
Herrmann
K
,
Czernin
J
. 
Evaluation of prostate cancer with 11C-acetate PET/CT
.
J Nucl Med
2016
;
57
:
30S
7S
.
58.
Spick
C
,
Polanec
SH
,
Mitterhauser
M
,
Wadsak
W
,
Anner
P
,
Reiterits
B
, et al
Detection of bone metastases using 11C-acetate PET in patients with prostate cancer with biochemical recurrence
.
Anticancer Res
2015
;
35
:
6787
91
.
59.
Kim
S
,
Kim
D
,
Kim
SH
,
Park
MA
,
Chang
JH
,
Yun
M
. 
The roles of (11)C-acetate PET/CT in predicting tumor differentiation and survival in patients with cerebral glioma
.
Eur J Nucl Med Mol Imaging
2018
;
45
:
1012
20
.
60.
Zhu
W
,
Dang
Y
,
Ma
Y
,
Li
F
,
Huo
L
. 
11C-acetate PET/CT monitoring therapy of multiple myeloma
.
Clin Nucl Med
2016
;
41
:
587
9
.
61.
Fontana
F
,
Ge
X
,
Su
X
,
Hathi
D
,
Xiang
J
,
Cenci
S
, et al
Evaluating acetate metabolism for imaging and targeting in multiple myeloma
.
Clin Cancer Res
2017
;
23
:
416
29
.
62.
Pantel
AR
,
Ackerman
D
,
Lee
SC
,
Mankoff
DA
,
Gade
TP
. 
Imaging cancer metabolism: underlying biology and emerging strategies
.
J Nucl Med
2018
;
59
:
1340
9
.
63.
Cai
L
,
Sutter
BM
,
Li
B
,
Tu
BP
. 
Acetyl-CoA induces cell growth and proliferation by promoting the acetylation of histones at growth genes
.
Mol Cell
2011
;
42
:
426
37
.
64.
Luengo
A
,
Gui
DY
,
Vander Heiden
MG
. 
Targeting metabolism for cancer therapy
.
Cell Chem Biol
2017
;
24
:
1161
80
.