Targeting Acetyl-CoA Carboxylase Suppresses De Novo Lipogenesis and Tumor Cell Growth in Multiple Myeloma

Abstract

Purpose:

In multiple myeloma (MM), tumor cells reprogram metabolic pathways to sustain growth and monoclonal immunoglobulin production. This study examines acetyl-CoA carboxylase 1 (ACC1), the enzyme driving the rate-limiting step in de novo lipogenesis (DNL), in MM metabolic reprogramming, particularly in c-MYC (MYC)-driven subtypes.

Experimental design:

ACC1 expression was evaluated across MM genetic subgroups, focusing on MYC translocations. Functional studies using ACC1 inhibitors and genetic knockdown assessed MM cell growth, lipid synthesis, and metabolic homeostasis in vitro and in vivo. The role of MYC overexpression in ACC1 sensitivity was examined, with palmitate rescue experiments. Lipidomic analysis and assessments of ER stress, protein translation, and oxidative damage elucidated underlying mechanisms.

Results:

ACC1 was overexpressed in MYC-translocated MM. Its inhibition or knockdown reduced MM cell growth in vitro and in vivo, particularly in MYC-overexpressing cells. ACC1 knockdown suppressed de novo lipid synthesis, partially rescued by palmitate. Lipidomic disruptions increased cholesterol ester desaturation and altered phospholipid ratios, inducing ER stress, impaired translation, protein carbonylation, oxidative damage, and apoptosis.

Conclusions:

ACC1 is a metabolic vulnerability in MYC-driven MM. Inhibiting ACC1 disrupts lipid homeostasis, induces ER stress, and causes oxidative damage, impairing cell survival. Targeting lipid synthesis pathways, especially in MYC-dependent subtypes, offers a promising therapeutic strategy for MM.

INTRODUCTION

Multiple myeloma (MM) is a hematologic malignancy characterized by the proliferation of clonal plasma cells (PCs) within the bone marrow. One of the reported risk factors for the onset of MM or its precursor conditions is having a lipid disorder such as obesity^1–3 or Gaucher’s disease⁴. Furthermore, bone marrow adipocytes sustain the growth of MM PCs⁵, and lipids – such as pristane – were implicated in cancer growth in the earliest models of murine plasmacytoma⁶. Lipid metabolism is required for MM PCs to meet the high bioenergetic and biosynthetic demand of malignant cell growth coupled with the unceasing production of monoclonal immunoglobulin. For example, in the MM genetic subgroup t(4;14), the cells are highly dependent on the mevalonate pathway, which synthesizes cholesterol and other isoprenoids such as geranylgeranyl pyrophosphate – a critical component for anchoring signaling proteins to the cell surface⁷. Another group found enhanced expression of fatty acid synthase (FASN) in the bone marrow of MM patients, and an inhibitor of FASN induces apoptosis in MM cells⁸.

The de novo lipogenesis (DNL) pathway synthetizes saturated fatty acids, in primis the 16-carbon palmitate. Palmitate is the building block of fatty acids in mammalian cells; it can be desaturated and elongated by specific enzymes to produce a variety of saturated and monounsaturated fatty acids (SFA and MUFA). Fatty acids are fundamental for cell homeostasis, as they can be used for cell membrane and organelle synthesis, as an energy source, and for signal transduction through post-translational modification of ligands and receptors^9–13. Upregulation of DNL is often observed in different cancer types, such as prostate cancer^14,15, colorectal cancer^16,17, ovarian cancer¹⁸, and breast cancer^19,20, and is often correlated with aggressive disease and poor prognosis. Diffuse large B-cell lymphoma has high levels of FASN, the enzyme catalyzing the last DNL step, and inhibiting lipid synthesis induces apoptosis²¹. While FASN is the last step in the DNL pathway, the rate-limiting enzyme for the whole pathway is acetyl-CoA carboxylase (ACC)^22,23. Recently, we found in MM that transcription of one of the genes encoding acetyl-CoA carboxylase, ACC1, was controlled by the combined action of the transcription factor c-MYC (MYC) and the pro-tumorigenic long noncoding RNA lnc-17–92 (aka RNA Regulator of Lipogenesis or RROL)²⁴. Inhibiting MYC or lnc-17–92 reduced the activity of the DNL pathway and subsequent MM cell proliferation.

We hypothesized that ACC1 in MYC-driven MM represents a metabolic vulnerability that can be exploited therapeutically. Here, we provide the mechanistic underpinning of growth suppression of MM via DNL inhibition as well as the rationale for using ACC inhibitors along with the identification of biomarkers of response that can be leveraged in future clinical studies.

MATERIALS AND METHODS

Cells

Cell lines (CLs) were grown at 37°C at 5% CO2 and cultured in RPMI-1640 medium (Gibco^® Life Technologies, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (Lonza Group Ltd., Basel, Switzerland) and 1% penicillin/streptomycin (Gibco^®, Life Technologies). AMO1 (RRID:CVCL_1806), NCI-H929 (RRID:CVCL_1600), SK-MM-1 (RRID:CVCL_A478), U266 (RRID:CVCL_U266), JJN3 (RRID:CVCL_2078), MOLP8 (RRID:CVCL_2124), EJM (RRID:CVCL_2030), KMS-12-PE (RRID:CVCL_1333), and KMS-12-BM (RRID:CVCL_1334) were purchased from DSMZ (Braunschweig, Germany). MM.1S (RRID:CVCL_8792) and RPMI-8226 (RRID:CVCL_0014) were purchased from ATCC (Manassas, VA, USA). AMO1 bortezomib-resistant (ABZB) were kindly provided by Dr. Christoph Driessen (Eberhand Karls University, Tübingen, Germany). KMS-11 (RRID:CVCL_2989) were purchased from Accegen Biotechnology. OPM1 (RRID:CVCL_5210), OCI-MY5 (RRID:CVCL_E332), KMS-20 (RRID:CVCL_2990), KMS-26 (RRID:CVCL_2992), KMS-27 (RRID:CVCL_2993), KMS-28-PE (RRID:CVCL_2995), KMS-28BM (RRID:CVCL_2994), and KMS-34 (RRID:CVCL_2996) are part of DFCI MMCL collection. Lenti-X^™ 293T (human embryonic kidney, purchased from Takara (cat. no. 632180)) were cultured in DMEM (Dulbecco’s modified Eagle’s medium) (Gibco^®, Life Technologies) supplemented with 10% fetal bovine serum (Lonza Group Ltd.) and 1% penicillin/streptomycin (Gibco^®, Life Technologies). Cells were periodically tested to exclude mycoplasma contamination. Cells were STR (short tandem repeats) authenticated.

Primary patient cells

Following written informed consent approved by Institutional Review Board at the Dana-Farber Cancer Institute or University of Catania Hospital, CD138+ cells were isolated from the BM aspirates of MM patients by Ficoll-Hypaque (Lonza Group, Basel, Switzerland) density gradient sedimentation, followed by antibody-mediated positive selection using anti-CD138 magnetic-activated cell separation microbeads (Miltenyi Biotech, Gladbach, Germany; Cat. Number: 130–097-614). The purity of immunoselected cells was assessed by flow-cytometry analysis using a phycoerythrin-conjugated CD138 monoclonal antibody by standard procedures. For long-term culture (6 days), CD138+ cells were cultured and physically separated from HS-5 cells by means of Falcon Cell Culture Inserts (Corning, New York, NY, USA), according to manufacturer’s instructions, as previously described^24,25.

Peripheral blood mononuclear cells

Peripheral blood mononuclear cells (PBMCs) were isolated from healthy adult donors, after informed consent approved by our Institutional Review Board of the Dana-Farber Cancer Institute. Cells were separated using the Ficoll-hypaque method (Lonza Group Ltd.). PBMCs were cultured in RPMI-1640 medium (Gibco^®, Life Technologies) supplemented with 10% fetal bovine serum (Lonza Group Ltd.) and 1% penicillin/streptomycin (Gibco^®, Life Technologies)^24,25.

RNA-seq and microarray-based gene expression analysis of MM patients and cell lines.

Analyses were performed as previously described^24,25. Briefly:

RNA-seq:

As a preliminary dataset, we used previously published RNAseq data from CD138+ MM cells from 360 MM patients from the IFM/DFCI 2019 clinical trial (NCT01191060)²⁶. We used this dataset to assess expression of ACACA in newly diagnosed MM patients. Unstranded paired-end RNA sequencing data were quantified using quasi-mapping with Salmon. Reference transcripts for GRCh38 transcripts were downloaded from Gencode v24. After QC controls, TPM values for genes were generated from isoform level TPMs with tximport. All figures were created with R and ggpubr. De-novo assembly for the RNAseq data on the IFM cohort was done using TopHat. Gencode v24 GTF file was used as the reference and new isoforms annotated by TopHat were identified from the output files. Only samples collected from CD138+ selected BM samples at diagnosis were used for analysis.

Microarray-based gene expression analysis:

ACACA expression levels were evaluated in CD138+ cells purified from 170 newly diagnosed and 12 relapsed MM patients from two publicly available datasets (GSE66293, GSE47552), that were profiled by Affymetrix Human Gene 1.0 ST Array and were batch-corrected and Brainarray annotated (v19). ACACA expression levels were also evaluated in CD138+ cells purified from an independent cohort of 50 newly diagnosed and 6 relapsed MM patients, that were profiled on Affymetrix Human Gene 2.0 ST Array (GSE116294). Differential expression between the two groups was assessed by Wilcoxon rank sum test with continuity correction in the R environment (version 4.0.4).

Correlation analysis:

The Spearman correlation was used to evaluate the correlation between ACACA and MYC mRNAs.

Survival analysis:

Survival analysis was performed using the survival package in R, and the log rank test was used to compare groups.

Immunohistochemistry

Immunohistochemistry slides were scanned with Aperio (GT 450, Leica Biosystems, IL, USA) with a 40X objective. The image analysis was performed using the module “CytoNuclear” from the software Halo (v3.6, Indica Labs, NM, USA). Briefly, for each sample one or more areas suitable for the analysis were defined annotating each sample. Areas with staining problems or other kind of artifacts were not included in the annotations. If a sample was not suitable for analysis an annotation on a blank area was done. After defining the dyes for the chromogen and the counterstaining, the intensity of signal was measured in the nuclear compartment for MYC and the cytoplasmic area for ACACA protein, on all the annotated areas. The threshold for positivity was manually determined on each slide for both MYC and ACACA and the data exported as percentage of DAB positive cells for each sample. The statistical analysis was performed using R and the packages “ggplot2”²⁷, and data visualization was performed using ggplot2 package ((RRID:SCR_014601) version 3.5.1, available on the Comprehensive R Archive Network - CRAN: https://cran.r-project.org/web/packages/ggplot2/index.html) and ggpubr package ((RRID:SCR_021139) version 0.6.0, available on the Comprehensive R Archive Network - CRAN: https://cran.r-project.org/web/packages/ggpubr/). Data manipulation was conducted with the dplyr package ((RRID:SCR_016708) version 1.1.4, CRAN: https://cran.r-project.org/web/packages/dplyr/index.html), while categorical variable handling was performed using the forcats package ((RRID:N/A)version 1.0.0, CRAN: https://cran.r-project.org/web/packages/forcats/index.html). Briefly, after importing and wrangling the data, the correlation between percentage of positive cells for MYC and ACACA on multiple myeloma common samples was correlated. Furthermore, for ACACA the value of samples of multiple myeloma (38 samples) vs normal control (2 samples) was plotted and the fold difference calculated.

Drugs

ND-646 and ND-630, for in vitro studies, were purchased from MedChemExpress (NJ 08852, USA). ND-646, for in vivo application, was provided by Gilead Sciences. Both compounds were dissolved in DMSO according to the manufacturer’s protocol.

Transient and stable transfection of cells

Cell transfection and transduction were performed as previously described²⁸. Briefly. Cells were transfected (electroporation) by the Neon Transfection System (Invitrogen, CA, US) (2 pulses at 1150, 30ms). siRNAs were used at 25nM. The transfection efficiency evaluated by flow-cytometric analysis relative to a FAM dye–labeled anti-miR–negative control reached 85% to 90%.

Stable expression using lentiviral plasmids

To generate cells stably expressing c-MYC, U266 was transduced with the Lenti ORF clone of Human v-myc myelocytomatosis viral oncogene homolog (avian) (MYC), Myc-DDK-tagged (RC201611L3) (Origene Technologies, Rockville, Maryland, MD). To knockdown ACC1, MM cells were infected with lentiviral particles carrying the expression of shRNAs against ACC1, under constitutive (pLKO.1) or isopropyl β- d-1-thiogalactopyranoside (IPTG)-inducible (pLKO_IPTG_3xLacO) promoters. Vector were obtained from Millipore Sigma (Bedford, MA). IPTG was used at 1mM, added every other day to culture medium.

Cell viability, apoptosis, cell cycle, and colony-forming assays

Analyses were performed as previously described^24,25. Briefly, cell viability was evaluated by Cell Counting Kit-8 (CCK-8) assay (Dojindo Molecular Technologies) and 7-Aminoactinomycin (7-AAD) flow cytometry assays (BD Biosciences), according to the manufacturer’s instructions. Flow cytometry analysis was performed either by FACS CANTO II (BD Biosciences) or by Attune NxT Flow cytometer (Thermo Fisher Scientific). Apoptosis was investigated by an Annexin V/7-AAD flow cytometry assay (BD Biosciences) and by electronic microscopy. For the colony-forming assay, 200 cells treated with DMSO or ND-646 were plated in triplicate in 1 ml of a mixture composed of 1.1% methylcellulose (MethoCultTM STEMCELL Technologies) in RPMI-1640 + 10% FBS. Crystal violet-stained colonies were scored after 2 weeks under an inverted microscope (Leica DM IL LED) at ×5 magnification.

Western blot analysis.

Protein extraction and western blot analysis were performed as previously described^24,25,29. Briefly, cells were lysed in 1x RIPA buffer (Cell Signaling Technology) supplemented with Halt Protease Inhibitor Single-Use cocktail (100X, Thermo Scientific). Whole cell lysates (~20 μg per lane) were separated using 4–12% Novex Bis-Tris SDS-acrylamide gels (Invitrogen) and electro-transferred onto Nitrocellulose membranes (Bio-Rad). Extraction of nuclear proteins was performed using the NE-PER^™ Nuclear and Cytoplasmic Extraction Reagents (Thermo Fisher, #78833), according to the manufacturer’s instructions. After electrophoresis, the nitrocellulose membranes were blocked and probed over-night with primary antibodies at 4°C, then the membranes were washed 3 times in PBS-Tween and then incubated with a secondary antibody conjugated with horseradish peroxidase for 2 hours at room temperature. Chemiluminescence was detected using the Western Blotting Luminol Reagent (sc-2048, Santa Cruz, Dallas, TX, USA).

Primary antibodies: Anti-ACC1 (Cell Signaling Technology Cat# 4190, RRID: AB_10547752), Anti-ACC2 [D5B9] (Cell Signaling Technology Cat# 8578, RRID:AB_10949898), Anti-phospho-ACC-Ser79 [D7D11] (Cell Signaling Technology Cat# 11818, RRID:AB_2687505), Anti-phosphorylated eIF2α (Cell Signaling Technology Cat# 9721, RRID:AB_330951), Anti-PERK (Cell Signaling Technology Cat# 5683, RRID:AB_10841299), BIP [C50B12] (Cell Signaling Technology Cat# 3177, RRID:AB_2119845), yH2AX [S139][20E3] (Cell Signaling Technology Cat# 9718, RRID:AB_2118009) and Anti-IRE1α (Cell Signaling Technology Cat# 3294, RRID:AB_823545), were purchased from Cell Signaling Biotechnology (Danvers, MA). Anti-Vinculin (Sigma-Aldrich Cat# V9131, RRID:AB_477629) was purchased from Sigma Aldrich.

Secondary antibodies: Anti-rabbit IgG, HRP-linked Antibody (Cell Signaling Technology Cat# 7074, RRID:AB_2099233), Anti-mouse IgG, and HRP-linked Antibody (Cell Signaling Technology Cat# 7076, RRID:AB_330924) were purchased from Cell Signaling Biotechnology (Danvers, MA).

De novo lipogenesis assay

De novo lipogenesis assay was performed as previously described²⁹. Briefly, cells were seeded at 5×10⁵ cells per well in 6-well plates and incubated for 3 days in the presence of treatments (ND-646, IPTG or respective controls). Twenty-four hours before the end of treatment, 1 μCi of ¹⁴C-labeled glucose (ARC-0122D) was added to each well. Cells were harvested, washed with cold PBS and collected in glass tubes. Purified lipid extract was obtained by a chloroform-methanol based extraction³⁰. Glucose incorporation into the cellular lipids was quantitated by photon emission through scintillation counting and normalized to total protein content.

Protein synthesis

The synthesis of protein was measured by puromycin incorporation, as previously described for the surface sensing of translation (SUnSET) assay. Briefly, cells were incubated with 1 μg/mL of puromycin aminonucleoside (Sigma Aldrich) after treatment with ND-646. The incorporation of puromycin into neosynthesized protein was detected by immunoblotting with an Anti-puromycin antibody (Sigma Aldrich, MABE343).

Lipidomics

Quantitative lipid profiling analysis was performed by Lipometrix (KU Leuven, Belgium). The analysis used liquid chromatography-electrospray ionization tandem mass spectrometry (LC-ESI/MS/MS) on a Nexera X2 UHPLC system (Shimadzu) coupled with a hybrid triple quadrupole/linear ion trap mass spectrometer (6500+ QTRAP system; AB SCIEX). Peak integration was done with MultiQuantTM 3.0.3, and lipid species signals were corrected for isotopic contributions with Python Molmass 2019.1.1 and quantified based on internal standard signals. Counts were normalized to DNA content and lipid levels are shown as a sum notation of indicated classes.

Protein carbonylation

Protein carbonylation levels were assessed using total protein lysate prepared in RIPA buffer. Derivatization of protein carbonyl groups was done with 2,4-Dinitrophenylhydrazine and determined with anti-DNP antibody (A6430, Molecular Probes) through ELISA³¹.

Membrane fluidity

Changes in membrane fluidity were assessed using the membrane fluidity kit according to the manufacturer’s instructions (Abcam, ab189819) following 6 days of treatment with ND-646 in the presence of 50 μM BSA-palmitate or BSA only. Thereafter, cells were washed and counted, then loaded with pyrenedecanoic acid (5 μM, PDA) supplemented with Pluronic F127 (0.08%) for 1h at room temperature. Cells were then washed and resuspended in fresh media. After excitation at 350nm, emission signals for PDA monomers (400nm) and excimers (470 nm) were determined. Data are expressed as normalized ratios ([excimer/monomer]/viable cells).

Animal study

6-week-old female immunodeficient NOD.CB17-Prkdcscid/NCrCrl (NOD/SCID; RRID:IMSR_CRL:394) mice (Charles River) were housed in our animal facility at Dana-Farber Cancer Institute (DFCI). All experiments were performed after approval by the Animal Ethics Committee of the DFCI and performed using institutional guidelines and following the declaration of Helsinki. Briefly, 5*10⁶ H929 cells were subcutaneously injected in NOD SCID mice. As the tumor became palpable (~50mm), mice were randomized to receive ND-646 or vehicle (−) as control (5 mice/group). Treatments were administered via I.P. injection, B.I.D., at 50mg/kg. Tumor sizes were measured by electronic caliper. For the xenograft inducible mouse-model, we subcutaneously implanted 5 × 10⁶ H929 cells transduced with IPTG-inducible shNC or shACC1–3 vectors. IPTG (Sigma, 10mM) was added to drinking water every other day. Early engraftment model: IPTG treatment started 2 days after injection of tumor cells (4 mice/group). Late engraftment model: IPTG treatment started after tumors became palpable (100–200 mm3) and mice randomized (5 mice/group). Tumor sizes were measured by electronic caliper.

Statistical Analysis

All in vitro experiments were repeated at least three times and performed in triplicate. A representative experiment was shown in the figures. Statistical significances of differences were determined using a Student’s t-test (unless otherwise specified), with the minimal level of significance specified as p<0.05. Kaplan-Meier survival curves were compared by log-rank test. Statistical analyses were determined using GraphPad software (http://www.graphpad.com). Graphs were obtained using GraphPad software (unless otherwise specified).

Data availability

The authors declare that all data supporting the findings of this study are available within the article and its Supplementary Information. Files or reagents are available from the corresponding authors on request. The transcriptome data analyzed in this study were publicly available and obtained from Gene Expression Omnibus (GEO) at GSE66293, GSE47552 and GSE116294. Raw data is available upon request. Normalized lipidomics data is provided in supplementary material (Data file S1).

RESULTS

Higher ACC1 expression is associated with c-MYC abnormalities and predicts poor clinical outcomes in MM.

To assess ACC1 (aka ACACA) expression and its association with any specific MM subtype, we analyzed RNA sequencing (RNA-seq) data from purified CD138+ plasma cells (PCs) from 319 newly diagnosed MM patients (NDMM) who were enrolled in the IFM/DFCI clinical trial 2009 (NCT01191060). We observed that higher expression of ACC1 mRNA (top quartile) predicted shorter event-free survival (Fig. 1A) (p=0.016) and overall survival (Fig. 1A) (p= 0.07). We also found a significant upregulation of ACC1 mRNA at the refractory/relapsed MM (RRMM) stage of the disease compared to NDMM in two independent cohorts of patients (Fig. 1B) (p=0.0015 and p=0.012).

Figure 1. Higher ACC1 expression predicts worse clinical outcomes in MM patients and is associated with c-MYC abnormalities.

A) Prognostic significance (EFS and OS, respectively) of high ACC1 expression (top quartile) in the IFM-DFCI cohort of NDMM patients.

B) Expression of ACC1 mRNA in NDMM (n=170) vs RRMM (n=12) patients profiled on Affymetrix Human Gene 1.0 ST Array from two publicly available datasets (GSE66293, GSE47552) that were batch-corrected and Brainarray annotated (v19) and expression of ACC1 mRNA in NDMM (n=50) vs RRMM (n=6) patients profiled on Affymetrix Human Gene 2.0 ST Array (GSE116294).

C) Expression of ACC1 mRNA in NDMM patients with or without translocations of MYC locus, from the IFM-DFCI cohort.

D) Correlation analysis of MYC and ACC1 mRNA expression in 170 NDMM patients, and in an independent cohort of 50 NDMM patients

E) IHC analysis ACC1 protein in normal bone marrow (upper panel) and MM (middle panel), and MYC in MM (bottom panel).

A comprehensive analysis of ACC1 expression in NDMM patients with specific chromosomal abnormalities revealed upregulated ACC1 in patients carrying MYC-associated translocations t(8;14), t(8;2), and t(8;22) (Fig. 1C and Supplementary Fig. 1A) (p=0.009), as well as in those with 17p deletion (del17p) (Supplementary Fig. 1B–C) (p=0.047 and p=0.0051), which is linked to more aggressive disease. ACC1 upregulation was also observed in patients carrying the t(14;16) translocation but not in other subgroups (Supplementary Fig. 1D). Notably, MYC directly regulates ACC1 expression MM cells^24,32, and we found a positive correlation between MYC and ACC1 expression in two independent cohorts of 129 (R=0.25, p=0.0047) (Fig. 1D) and 170 NDMM patients (R=0.28, p=0.00016) (Fig. 1D). Despite observing no significant difference in ACC1 mRNA expression between NDMM patients (n=360) and normal donor PCs (n=16) (Supplementary Fig. 1E), a higher proportion of cells expressing ACC1 protein was noted in bone marrow biopsies from MM patients (n=38) compared to healthy donors (n=2) (Fig. 1E). This finding underscores the complexity of ACC1 regulation, potentially involving post-transcriptional and/or translational mechanisms. Additionally, we observed a positive correlation between ACC1 and MYC proteins in MM patient biopsies (Spearman R = 0.29) (Supplementary Fig. 1F). Finally, pathway analysis of NDMM patients with higher ACC1 expression revealed activation of MYC-related transcriptional signatures and fatty acid metabolism, along with other metabolic pathways such as glycolysis, MTORC1 signaling, and cholesterol homeostasis (Supplementary Fig. 1G).

Targeting ACC1 inhibits MM cell growth in vitro and in vivo.

We investigated the impact of pharmacological inhibition of ACC1 on MM cell growth and survival. We utilized allosteric inhibitors ND-646³³ and ND-630³⁴, which prevent the dimerization of ACC1 necessary for its biological function. We found that treating AMO1 cells with ND-630 or ND-646 reduced the phosphorylation of ACC1 (Arg172) and ACC2 (Arg277), and is also known to decrease phosphorylation on Ser79, here referred to as pACC, in a dose-dependent manner (Fig. 2A). The effects of ND-646 were more pronounced, with a strong reduction of pACC starting at a 100 nM concentration (Fig. 2A), which guided its prioritization in our study. In addition, ND-646 is orally available with a favorable biodistribution to non-liver tissues³³. Therefore, we focused further investigation on this compound. Reduced phosphorylation of these residues has been reported as an effective marker of ND-646 and ND-630 activity³³.

Figure 2. Targeting ACC1 inhibits MM cell growth in vitro and in vivo.

A) WB analysis of pACC and ACC1 in AMO1 cells treated for 3 days with the indicated concentrations of ND-630 or ND-646, or DMSO as control. GAPDH was used a protein loading control.

B) CCK-8 viability assay in 20 MM cell lines treated for 6 days with the indicated concentrations of ND-646. DMSO (1%) was used as the control. Cell lines are listed from most sensitive (top) to least sensitive (bottom) based on their response to ND-646.

C) Annexin V / 7-AAD staining of AMO1 cells treated for 6 days with ND-646 (100nM) or DMSO (1%) as the control.

D) Colony forming assay of AMO1 and H929 cells treated with ND-646 (0.25uM) or DMSO as the control.

E) 7-AAD staining of CD138+ cells purified from 4 MM patients and treated with ND-646 (1uM) or DMSO as the control.

F) Subcutaneous in vivo tumor growth of H929 cells in NOD SCID mice, after treatment with ND-646 (n=5) or VEHICLE (n=5).

G) CCK-8 proliferation assay of AMO1 and H929 cells expressing the shACC1–3 under an IPTG-inducible promoter. CCK-8 analysis was performed at the indicated time points after IPTG induction. Effects are represented compared to uninduced conditions. WB shows ACC1 expression in the different experimental settings. Y-tubulin is used as a protein loading control.

H) In vivo engraftment of subcutaneously injected H929 expressing shACC1–3, in the absence (−) or presence (+) of IPTG via drinking water. Results are represented as the percentage of engraftment after the indicated observation time. IPTG treatment started 2 days after the injection of MM cells.

I) In vivo tumor growth of subcutaneously injected H929 expressing shACC1–3, in the absence (−) or presence (+) of IPTG via drinking water. IPTG treatment started when tumors became palpable, after mice randomization.

In a panel of 20 MM cell lines with heterogeneous ACC1 expression (shown in Supplementary Fig. 2 for ACC1 mRNA and Supplementary Fig. 3 for ACC1 protein in 11 cell lines), both ND-646 (Fig. 2B) and ND-630 (Supplementary Fig. 4A) antagonized growth in a dose-dependent manner, with ND-646 being effective already at 100nM concentration, consistently with pACC reduction (Fig. 2B). ND-646 also induced apoptosis (Fig. 2C) and prevented tumor colony formation (Fig. 2D). It was equally effective against AMO1 cells resistant to the proteasome inhibitor bortezomib (Supplementary Fig. 4B). Interestingly, in MM cell lines, sensitivity to these inhibitors did not appear to correlate with ACC1 mRNA or protein expression (Supplementary Figs. 2 and 3), nor with common genetic abnormalities such as Ig translocations or mutations. However, ectopic expression of MYC significantly enhanced the growth-inhibitory effect of ND-646 in U266 cells (Supplementary Fig. 4C), which express very low levels of this transcription factor. Ectopic expression of MYC consistently promoted the DNL pathway in these cells, as indicated by the increased incorporation of radiolabeled glucose (Glucose D-[14C(U)]) into lipids (Supplementary Fig. 4D).

Importantly, ND-646 antagonized the growth of CD138+ PCs isolated from three out of four MM patients, even if cultured in the presence of protective bone marrow stromal cells (patients #1 and #2) (Fig. 2E), while sparing PBMCs isolated from five healthy donors (HDs) (Supplementary Fig. 4E). Moreover, in an H929 xenograft model we observed that ND-646 (50mg/kg, BID) significantly inhibited tumor growth in vivo (Fig. 2F) (p<0.05).

As ND-646 and ND-630 also bind ACC2, in addition to ACC1, we investigated their individual effects using genetic manipulation. We observed that stable conditional knockdown of ACC1, but not ACC2, antagonized the in vitro growth of MM cells (Supplementary Fig. 4F–G). We validated the anti-MM activity of ACC1 depletion using a conditional knockdown system in vitro (Fig. 2G) and in vivo in early- and late-engraftment H929 xenograft models (Fig. 2H–I).

Targeting ACC1 impairs DNL and alters the MM cell lipidome.

We investigated the impact of pharmacologic and genetic blockade of ACC1 on DNL rates in AMO1 and H929 cells—two highly sensitive cell lines to ND-646 and primary models used to demonstrate RNA-mediated MYC regulation of ACC1²⁴. We observed a significant reduction in the incorporation of radiolabeled glucose (Glucose D-[14C(U)]) into lipids, indicating decreased lipogenesis rates, following ACC1 knockdown with an inducible shRNA (Fig. 3A) or ND-646 treatment (Fig. 3B). These effects were comparable to those achieved through FASN inhibition with IPI-9119 (Fig. 3B) (p<0.01). Notably, supplementing cell cultures with palmitate, the end-product of DNL, reversed the growth suppression caused by ND-646 treatment, rescuing MM cell proliferation (Fig. 3C).

Figure 3. ACC inhibition diminishes MM cell growth and DNL rates.

A) DNL rates in AMO1 and H929 cells expressing ACC1 shRNA induced by IPTG. DNL measured as [14C]-glucose incorporation into lipids. Results are expressed as cpm/ug (n=9, mean ± SEM). Unpaired t-tests, **p<0.01, ***p<0.001.

B) DNL rates in AMO1 and H929 cells treated for 6 days with 100 nM of ND-646 or 100 nM of IPI-9119. Results are expressed as cpm/ug (n=9, mean ± SEM). One-way ANOVA with Dunnett’s test, **p<0.01, ****p<0.0001.

C) Cell growth assessment of AMO1 and H929 cells co-treated with ND-646 at 100 nM and 50 μM of palmitate-BSA or 50 μM BSA for 6 days. Data plotted as % relative to control of viable cells (n=9, Mean ± SEM). One-way ANOVA with Tukey’s test, ****p<0.0001.

D) Total triacylglycerides (TG) levels after ACC inhibition with ND-646 at 100nM for 6 days. Lipid levels are shown as a sum notation of TG and expressed as nmol/mg of DNA (n = 3, mean ± SEM). Multiple unpaired t-test, *q<0.05, ****q<0.0001.

E)Total triacylglyceride (TG) species containing palmitate (16:0) after 6 days of ACC inhibition with 100 nM ND-646. Data are represented as the sum of all relevant TG species and expressed in nmol per mg DNA (n=3, Mean ± SEM). Multiple unpaired t-test, p < 0.005 and ****p < 0.0001.

F) Acyl chain unsaturation in TG in MM cells treated with 100 nM of ND-646 for 6 days. TG containing fatty acids with the same number of unsaturations (0 or 1) were summed together. Data expressed as nmol/mg of DNA (n = 3, mean ± SEM). Two-way ANOVA with Šídák’s test, ****p<0.0001.

G) Unsaturated cholesterol ester after ACC inhibition with ND-646 at 100nM for 6 days. Lipid levels are shown as a sum notation of CE with 3 or more double bonds and expressed as nmol/mg of DNA (n = 3, mean ± SEM). Multiple unpaired t-test, *q<0.05.

Targeted HILIC LC-MS/MS lipidomics analysis confirmed the effect of pharmacologic blockade of ACC on the MM cell lipidome. Treatment with ND-646 reduced the levels of total triacylglycerides (TG) in both AMO1 and H929 cells (Fig. 3D), decreased the levels of tripalmitin (TG 16:0), a palmitate-derived TG that has been used as a biomarker of response in clinical trials with DNL inhibitor TVB-2640³⁵, was significantly reduced with ND-646 treatment (Fig. 3E), and specifically reduced the levels of fatty acids with saturated and monounsaturated acyl chains (Fig. 3F), consistent with the de novo lipogenesis blockade. Interestingly, gene set enrichment analysis (GSEA) using RNA-seq data revealed an upregulation of gene sets associated with the cholesterol pathway (Supplementary Figure 5A) following ACC inhibition with ND-646, leading to increased levels of unsaturated cholesterol esters in AMO1 and H929 cells (Fig. 3G). This suggests that MM cells utilize the increased availability of acetyl-CoA (since ACC can no longer convert acetyl-CoA to malonyl-CoA) to fuel the mevalonate pathway as an alternative to the DNL pathway for lipid creation. We hypothesize that this increase in cholesterol ester levels could be targeted through HMGCR inhibition to potentiate the effect of ACC blockade on cell growth. This was confirmed by combining ND-646 with Simvastatin, which showed an enhanced cell growth reduction effect than either drug alone (Supplementary Figure 5B).

DNL suppression induces ER stress and diminishes protein synthesis in MM

Endoplasmic reticulum (ER) stress is observed in response to cell perturbations, including oxidative stress and fatty acid dysregulation³⁶. Treatment with ND-646 induced ER stress in AMO1 cells, as observed by increased phosphorylation of PERK, increased levels of IRE1α, and increased phosphorylation of eIF2α (Fig. 4A). Although the effects on p-eIF2α were less pronounced in H929 cells, ACC inhibition still led to increased phosphorylation of PERK and levels of IRE1α. Both MM cell lines exhibited downregulation of BIP expression following ACC inhibition. BIP, also known as GRP78, is a direct sensor of ER stress as it activates IRE1^37,38. In cancer cells, BIP is often overexpressed, facilitating tumor cell survival by activating pathways that suppress apoptosis and promote metabolic plasticity ^39,40. Its downregulation leads to the accumulation of misfolded protein, triggering ER stress⁴¹. Importantly, the exogenous addition of palmitate partially rescued the ER stress induced by ND-646 treatment (Fig. 4A).

Figure 4. ER stress and protein translation are modulated by ACC inhibition in MM.

A) Representative immunoblotting of ER stress markers in cells co-treated with ND-646 at 100nM and 50 μM of palmitate-BSA or 50 μM BSA for 6 days. Data was plotted as % relative to the control of normalized marker expression (marker/vinculin) (n=3 independent experiments, Mean±SEM). One-way ANOVA with Tukey’s test, *p<0.05, **p<0.001, ***p<0.0001.

B) NES values of pre-ranked gene sets with FDR below 0.2 for H929 cells treated with ND-646 at 100 nM for 6 days.

C) Representative immunoblotting of SUnSET assay from cells co-treated with ND-646 at 100nM and 50 μM of palmitate-BSA or 50 μM BSA for 6 days. At least 3 independent experiments were performed.

Cells respond to ER stress by inhibiting protein synthesis, a process mediated through eIF2α phosphorylation. GSEA revealed that pathways associated with protein translation and rRNA processing were significantly downregulated following ACC inhibition with ND-646 (Fig. 4B). In both cell lines, global protein synthesis was reduced, as evidenced by decreased puromycin incorporation into newly synthesized proteins. Notably, this effect was partially reversed by the addition of palmitate (Fig. 4C).

Because alterations in phospholipids can play a role in ER stress activation, we next analyzed how ND-646 treatment affects the major cellular phospholipids, phosphatidylcholine (PC) and phosphatidylethanolamine (PE). ACC inhibition increased the ratio of PC/PE in AMO1 cells but decreased it in H929 cells (Fig. 5A) (q<0.05). Additionally, membrane fluidity was increased in MM cells after DNL blockade, and this was reversed by palmitate addition (Fig. 5B). Protein oxidative damage, measured by protein carbonylation levels, was also increased in MM cells treated with the ACC inhibitor (Fig. 5C) (p<0.01). Furthermore, ACC inhibition led to increased γH2AX expression, a marker of DNA damage, in both MM cell lines (Fig. 5D). Palmitate supplementation partially rescued this effect, particularly in H929 cells (Fig. 5D).

Figure 5. Impact of ACC inhibition on cellular metabolism and oxidative stress responses in multiple myeloma cells.

A) Phosphatidylcholine (PC) to phosphatidylethanolamine (PE) ratios in AMO1 and H929 cells treated with ND-646 at 100 nM for 6 days. Data expressed as the ratio of indicated lipid classes (n = 3, mean ± SEM). Multiple unpaired t-test, *q<0.05.

B) Membrane fluidity assay from cells co-treated with ND-646 at 100nM and 50 μM of palmitate-BSA or 50 μM BSA for 6 days. Data shown are means and SEM of 3 independent experiments. One-way ANOVA with Tukey’s test,*p<0.05, ***p<0.0006 ****p<0.0001

C) Protein carbonylation levels assessed after ND-646 treatment at 100 nM combined with 50 μM of palmitate-BSA or 50 μM BSA. Analysis performed on protein lysate after 6 days of treatment. Results are expressed as absorbance at 450 nm (n=12, mean ± SEM). One-way ANOVA with Tukey’s test, **p<0.01, ****p<0.0001.

D) DNA damage evaluated following ACC inhibition from cells treated with ND-646 at 100 nM and 50 μM of palmitate-BSA or 50 μM BSA for 6 days. Data shown are means and SEM of 3 independent biological replicates. One-way ANOVA with Tukey’s test,*p<0.05, ***p<0.0002 ****p<0.0001

Taken together, these findings suggest that ACC inhibition disrupts cellular lipid homeostasis, leading to oxidative damage and triggering ER stress, which is associated with impaired protein synthesis in MM cells.

DISCUSSION

Metabolic alterations, including increased lipid synthesis, are a hallmark of cancer since highly proliferative tumor cells must adapt to meet elevated energetic and metabolic demands^42,43. Consequently, tumor cells are particularly dependent on metabolic pathways such as DNL. Previous studies have shown that metabolic alterations can induce cell death in MM. Adiponectin, a cytokine produced by adipocytes, and its receptor agonist cause apoptosis of MM cells through AMPK activation and consequent inhibition of ACC^44,45. Similarly, targeting DNL through blockade of FASN induced apoptosis in MM cells⁸. Our data corroborate these findings, supporting the therapeutic relevance of lipid metabolism in MM and providing insights into the molecular mechanism by which palmitate depletion can impair cell growth.

We have shown a positive correlation between MYC and ACC1 expression in MM. In addition, previous studies had confirmed the relevance of MYC in lipid metabolism rewiring in cancer. In MM, we previously demonstrated that MYC occupies the ACC1 promoter in an RNA-dependent manner, resulting in the transcription of ACC1⁴⁶. In a murine MYC-driven prostate cancer model, we observed that high-fat diet (HFD) enhances the MYC transcriptional program via metabolic alterations that favor histone hypomethylation at the promoter regions of MYC-regulated genes, leading to increased cellular proliferation and tumor burden⁴⁷. We have also shown via mass spectrometry-based metabolite profiling that MYC-driven tumors and cells are predominantly lipogenic, with enrichment of glycerophospholipid metabolites and citrate, known to feed lipogenesis⁴⁸. Our data supports the rationale that MM patients with MYC alterations can benefit from metabolic therapies targeting ACC and lipid synthesis.

Mechanistically, we propose that ACC inhibition impairs MM cell growth by blocking the DNL pathway and altering the cellular lipidome. The consequences of DNL inhibition are of course pleiotropic, affecting energy production, signal transduction via post-translational modification of key proteins, significant alterations in membrane phospholipids and membrane fluidity and more⁴⁹. One of the major consequences of this lipidome re-wiring is of course ER stress and the UPR. Notably, the observed cell death is accompanied by altered phospholipid ratio and increased membrane fluidity, which result in ER stress and oxidative damage. This conclusion is strongly supported by the rescue effect observed with exogenous palmitate supplementation. These findings align with previous studies reporting ER stress induction following DNL blockade in prostate, colon, and cervical cancer cells^29,50. A cell death modality associated with ER stress induced by ACC inhibitors, triggers autophagy and programmed cell death but has been shown to also prompt an immunogenic cell death modality⁵¹. Perturbing phospholipid composition can alter ER homeostasis and modulate Sarco/Endoplasmic Reticulum Calcium ATPase (SERCA) activity⁵² or directly activate IRE1 and PERK enzymes⁵³. As an important site for protein and lipid synthesis, the ER is sensitive to membrane alterations, which can also trigger the Lipid Bilayer Stress (LBS) pathway and activate the unfolded protein response⁵⁴. Our data shows that the lipid modulation induced by ACC inhibition can increase activation of PERK, consequently leading to phosphorylation of eIF2α, an event that impairs protein translation. It also leads to IRE1α activation. While ER stress occurs in MM cells as a result of lipid metabolism rewiring, the downstream pathways that get activated differ in the different MM lines and the complexity, which will require exploration in different molecular contexts, is substantial.

These findings suggest that inhibition of DNL may synergize with current established therapeutic regimens or with others being explored in MM. After ACC blockade, MM cells increase cholesterol levels. This is expected, as it leads to accumulation of acetyl-CoA, the substrate of ACC and the major source for cholesterol synthesis and steroid production⁵⁵. Interestingly, hypocholesterolemia is observed in MM patients, likely due to MM cells utilizing cholesterol⁵⁶. Altered cholesterol homeostasis influences cancer cell migration, invasion, and the metastatic cascade in many cancers. Cancer cells upregulate HMGCR, SREBP and LDL receptors, all aimed at increasing cholesterol synthesis. Cholesterol also modulates membrane fluidity which disrupts protein folding and interactions between lipids and proteins^57,58. Furthermore, increased membrane fluidity renders the endoplasmic reticulum more susceptible to oxidative stress^59,60. Taken together, inhibition of DNL via ACC targeting activates the unfolded protein response (UPR). It is well known that MM cells rely on the UPR. In fact, bortezomib exacerbates ER stress, leading MM cells to apoptosis^61,62. Thus, inhibition of DNL via ACC targeting may synergize with this mainstay therapeutic regimen. In addition, potentiation of MM cell growth inhibitory effects of ND-646 by the addition of HMGCR inhibition by simvastatin, may also provide a mechanism to increase the anti-MM effects of ACC1 inhibition and its clinical application. Finally, we show that ACC inhibition leads to increased γH2AX foci in MM cell lines. Increased DNA damage, as we previously showed in prostate cancer⁶³, could also be leveraged therapeutically. It has been previously shown that MM with high MYC expression is dependent on non-homologous end joining (NHEJ) DNA repair pathway⁶⁴. Thus, PARP inhibition could in theory synergize with ACC targeting.

One caveat of this study is that it does not address the consequences of ACC inhibitors on overall survival in pre-clinical models. In the subcutaneous xenograft model used in this study, animal survival is inherently linked to tumor growth, as mice were euthanized when tumors reach a predetermined size to comply with ethical guidelines. While this model demonstrates the efficacy of ACC1 inhibitors in reducing tumor burden, it does not allow for a direct assessment of overall survival. Additional studies, such as orthotopic or systemic models, would be valuable to better evaluate the long-term effects of ACC1 inhibitors on survival and disease progression.

It should be said that both ACC and FASN inhibitors have been utilized in clinical trials with only minor side effects^35,65–67. ND-630 (also known as Firsocostat) is the most advanced ACC1 inhibitor currently in clinical trials. It is undergoing a Phase 2 clinical trial for adults with cirrhosis due to nonalcoholic steatohepatitis (GS-US-454–6075), being tested both as a standalone treatment and in combination with semaglutide. This represents a significant step in evaluating the clinical utility of ACC1 inhibitors.

Finally, we observed that DNL blockade with ND-646 significantly reduced levels of TG composed of saturated fatty acids in MM cells. Interestingly, tripalmitin, the TG with three palmitic acid acyl chains, was the lipid species used to surrogate DNL inhibition in patients treated with the first FASN inhibitor to enter clinical studies, TVB-2640³⁵. Thus, circulating tripalmitin may be the ideal biomarker of response in MM patients treated with ACC inhibitors.

In conclusion, this study expands on our previous work showing that ACC1 is regulated by the transcription factor c-MYC and the pro-tumorigenic long noncoding RNA Lnc-17–92²⁴, establishing a pivotal role for ACC1 and DNL in supporting MM cell growth. Importantly, this pathway is particularly relevant in patients with MYC translocations, where MYC-driven metabolic reprogramming amplifies the dependency of tumor cells on DNL, presenting a compelling case for therapeutic targeting (Figure 6).

Figure 6. Schematic mechanism of ACC inhibition disrupting lipid metabolism and inducing ER stress.

Lnc-17–92 upregulates the transcription of ACC1 by interacting with MYC, which binds to the promoter region of ACC1, enhancing its transcription. De novo lipogenesis inhibition by ACC1 activity blockade reduces lipid synthesis, leading to cholesterol accumulation and PC/PE impairment (phosphatidylcholine/phosphatidylethanolamine impairment). These changes alter cholesterol levels and membrane fluidity, disrupting ER homeostasis and leading to ER stress, which activates UPR pathways and increases ROS production. Sustained ER stress reduces cellular proliferation and increases apoptosis (Created with BioRender.com).

Statement of translational relevance.

This study provides the rationale for the use of metabolic therapies targeting lipid synthesis in multiple myeloma, with a focus on the enzyme ACC1. Our data includes the discovery and validation of biomarkers that can be used in patient-tailored therapy, such as lipid-based biomarkers for patient response monitoring and genetic markers like c-MYC translocations to predict treatment sensitivity.