Itaconate alleviates cholesterol burden via ABCA1 stabilization and cholesterol efflux

Abstract

Background and aims:

Itaconate (ITA) is a metabolite produced from the tricarboxylic acid cycle (TCA) that has been shown to regulate atherosclerotic plaque growth and induce stability via immunomodulation. However, lipid metabolism regulation by ITA is currently underexplored in atherosclerosis. Here, we take advantage of plaque-targeting ITA-conjugated nanoparticles (ITA-LNPs) to investigate the effects of ITA on regulating lipid metabolism in foam cells/macrophages in atherosclerosis via ABCA1 stabilization and increased triglyceride metabolism.

Methods:

Apoe^−/− mice were fed a high-cholesterol/high-fat diet (HCHFD) for 12 weeks and injected once weekly with 50 mg/kg ITA-LNP or Ctrl-LNP. Aortas were tested for ITA-LNP biodistribution, followed by quantification of atherosclerotic plaque burden. Bone marrow-derived macrophages (BMDMs) or RAW 264.7 cells were treated with ITA-LNP or Ctrl-LNP in the presence of oxLDL, acLDL, or free cholesterol to investigate ITA’s actions on lipid metabolism, Abca1 expression, and ABCA1 stability under a variety of conditions, including stable gene knockdown.

Results:

ABCA1 was significantly upregulated with ITA-LNP treatment compared to Ctrl-LNP both in vivo and in vitro at the protein level, but not at the transcriptional level. ITA-LNPs were shown to prevent ABCA1 decay via the HO-1-calpain axis, resulting in significantly increased cholesterol efflux in macrophages. This was further confirmed in RAW 264.7 cells with a stable HO-1 knockdown. Additionally, ITA decreased lipid burden in conjunction with increased expression of Slc25a1 in ITA-LNP-treated BMDMs, suggesting enhanced fatty acid-derived citrate shuttling and increased fatty acid metabolism.

Conclusions:

ITA-LNPs regulate lipid metabolism in atherosclerosis by inducing triglyceride catabolism and cholesterol efflux.

1. Introduction

Itaconic acid (also known as itaconate, ITA) is a dicarboxylic acid metabolite, a product of the tricarboxylic acid cycle (TCA) enzyme aconitate decarboxylase 1 (ACOD1) [1,2]. ITA is a multifaceted metabolite and metabolic drug candidate because of its actions on inflammation, metabolism, and epigenetics through multiple mechanisms [1–5]. Most recently, this fascinating small molecule has gained a lot of momentum due to its pronounced effects on the cardiovascular system. Indeed, ITA and its derivatives, including lipid nanoparticle (LNP) formulations, alleviated the severity of myocardial infarction [4,6,7], diminished atherosclerotic plaque [3,8,9], and improved donor heart function in transplantation [10]. These effects are usually attributed to ITA targeting of multiple inflammation-related pathways [1,4,5,11–15], including inhibition succinate dehydrogenase (SDH) [4] and reduction of inflammasome activation leading to the subsequent reduction in IL-1β secretion [4,16–18]. ITA’s effects on atherosclerosis are of particular interest because of the magnitude of the atherosclerotic plaque reduction in different models of the disease [3,8,9], including those of vulnerable plaque atherosclerosis [3]. We have previously developed ITA-LNP technology to deliver free ITA intracellularly in atherosclerosis via ITA’s immunomodulatory effects and have demonstrated effects similar to that of endogenous unmodified ITA [3]. For example, in macrophages, ITA-LNPs inhibited succinate dehydrogenase (SDH) and induced heme oxygenase (HO-1), an antioxidant defense enzyme downstream of NRF2 [3].

Notably, in addition to these immunomodulatory effects, ITA and its derivatives have also been shown to regulate lipid metabolism and efflux in the liver [19,20]. This includes 4-octyl itaconate (4-OI) stabilizing CPT1a by inhibiting ubiquitination. Thus, it could very well be that the significant reduction of lipid burden in atherosclerotic plaque observed with ITA-LNP treatment is not only due to the anti-inflammatory effects of ITA [3]. This poses the question of whether the effects of ITA in atherosclerotic plaque extend beyond the regulation of inflammation into lipid regulation.

Lipid metabolism and transport are key targets for atherosclerotic therapy. The concept of “progressing plaques” [21,22] describes inflammation-amplified plaque progression. Progressing atherosclerotic plaques express retention molecules, such as netrin-1, semaphorin 3E, and cadherins, which inhibit myeloid cells from exiting the plaque [23,24]. Plaque myeloid cells, especially macrophages, accumulate cholesterol (i.e., give rise to foam cells) through continuous exposure to modified low-density lipoproteins. This ultimately precipitates atherothrombotic sequelae, resulting in major cardiovascular events [25,26]. Conversely, “regressing plaques” [27] show cellular egress from atherosclerotic plaque [28,29] and improved capacity to efflux cholesterol, known as reverse cholesterol transport (RCT) [30,31]. RCT removes the cholesterol from plaque, which is then transported and disposed of in the liver. Increased RCT is associated with lower risk of cardiovascular events [32–34]. Recent evidence supports the hypothesis that therapeutic modulation of the retention/egress balance in plaque may be an effective mechanism to target atherosclerosis [33,35,36]. Thus, if ITA can not only lower inflammation but also therapeutically target lipid metabolism and transport, it could potentially steer plaques toward this beneficial “regressive” phenotype. We postulated that a key RCT-related protein such as ABCA1 may be facilitating this phenomenon. ABCA1 regulates reverse cholesterol transport (RCT) via facilitating the efflux of cholesterol and phospholipids from cells to apolipoprotein A-I (apoA-I), which forms nascent high-density lipoprotein (HDL) particles [37,38]. Therefore, ABCA1 is critical for maintaining cellular cholesterol homeostasis and preventing the accumulation of cholesterol in peripheral tissues, which is a key factor in reducing the risk of atherosclerosis [39].

This work examines the relationship between ITA and major processes involved in RCT: the expression of ATP-binding cassette transporter ABCA1 in macrophages and cholesterol efflux. We hypothesized that ITA’s effects on plaque lipids are likely a direct consequence of ITA’s actions, rather than a byproduct of inflammation reduction. Using a model of intracellular ITA delivery to macrophages via ITA-LNPs, we demonstrate that ITA plays a previously unappreciated role in ABCA1 protein stabilization. This occurs through HO-1 induction, inhibition of lysosomal calpain activity, and reduced ABCA1 ubiquitination. As such, our work explores ITA in regulating macrophage lipid metabolism, which, considering the known effects of ITA on plaque inflammation, provides compelling evidence and reasoning to support the development of dual-action ITA therapies for atherosclerosis.

2. Materials and methods

Detailed information (including catalog numbers) on materials such as chemicals, biological agents, buffers and other supplies used in this work is presented in Supplemental Table S1.

2.1. Animals

Male Apoe^−/− and Nrf2^−/− mice (5 weeks of age) were purchased from The Jackson Laboratory (Bar Harbor, ME) and kept in AAALAC-accredited facilities at Case Western Reserve University. The experimental procedures here described were approved by The Institutional Animal Care and Use Committee (IACUC). Animals were housed five per cage and allowed to acclimate in the facility for one week. Throughout all experiments, animals were kept on a 12:12 h light-dark cycle at 22 °C, and both diet and water were provided ad libitum.

2.2. Atherosclerosis targeting experiments

Male Apoe^−/− mice (n = 6–7) were randomized at 8 weeks of age to receive injections of Ctrl- or ITA-LNPs (50 mg/kg) once weekly through retro orbital plexus while providing high cholesterol high fat diet (HCHFD, 1.25 % cholesterol, 40 % calories from fat, Research Diets D12108C) ad libitum for 12 weeks. The mice were euthanized using CO₂ asphyxiation and aortas were isolated as described previously [3].

2.3. LNP synthesis and aortic biodistribution

LNPs were synthesized and characterized as previously described by us [3]. ITA-LNP concentrations reported in this manuscript are of total ITA-lipid conjugates (i.e., moles or grams of ITA-lipid conjugate per liter of buffer or kilogram of mouse). LNPs were labeled with Atto647 phospholipids and injected at 50 mg/kg in male Apoe^−/− mice (n = 3) that have been fed with HCHFD for 12 weeks. The whole aorta was extracted after extensive perfusion with 4 % paraformaldehyde (PFA) solution and imaged using Azure C400 imager utilizing Cy5 channel, as previously described [3].

2.4. Cell culture

RAW 264.7 (ATCC), iBMDMs (Kerafast) were cultured as per vendor recommendations. Bone marrow cells were isolated from mice by flushing the femurs and tibias with PBS as previously described [3]. The cells were cultured in Teflon flasks in 30 mL of Dulbecco’s Modified Eagle’s Medium (DMEM) containing 10 % fetal bovine serum, 1 % MEM Non-essential Amino Acid Solution, 1 % Anti-anti and 40 ng/ml mouse M-CSF until differentiation into macrophages (bone marrow derived macrophages, BMDM). The cells were treated with 100 μg/mL acetylated low-density lipoprotein (acLDL) or oxidized LDL (oxLDL) unless noted otherwise.

2.5. Preparation of acLDL

Acetylated low-density lipoprotein (acLDL) was prepared by mixing 1 mL of LDL solution (Athens Research and Technology) with 1 mL of saturated sodium acetate solution in an ice-cooled tube containing a magnetic stir bar. A final volume of 30 μL acetic anhydride per 5 mg of LDL was added incrementally: 2 μL aliquots were introduced every 5 min to ensure temperature stabilization. The reaction mixture was then dialyzed against 10 mM phosphate-buffered saline (PBS) using a dialysis cassette.

2.6. Preparation of BODIPY-acLDL

For BODIPY-labeled acLDL, 50 μL of 1 mg/mL BODIPY-cholesterol (Avanti Polar Lipids) in chloroform was dried under vacuum centrifugation. The resulting pellet was resuspended in 100 μL isopropanol and sonicated (30 s, 300 W). The suspension was transferred to a 37 °C water bath, and 3 mL of the prepared acLDL solution was added. The mixture was stirred magnetically at 37 °C in the dark for 1 h. Finally, BODIPY-acLDL was filtered through a 0.45 μm membrane prior to use.

2.7. Cholesterol efflux

RAW 264.7 cells were seeded in black clear-bottom 96-well plates at a density of 0.1 million cells per well. After cell attachment, the medium was replaced with a serum-free medium containing 0.2 % bovine serum albumin (BSA) and 75 μg/mL BODIPY-acLDL. Following a 24-h incubation, the cells were washed with pre-warmed PBS.

2.7.1. Experimental groups for efflux measurement

To isolate ABCA1-and ABCG1-mediated efflux, four experimental groups were established per treatment condition Ctrl-vs ITA-LNPs):

ABCA1/G1 experimental (exp[ABCA1/G1]): Cells treated with cAMP (0.3 mM) and ApoA1 or HDL (100 μg/mL).

ABCA1/G1 control (cnt[ABCA1/G1]): Cells treated with cAMP alone.

ABCG1 experimental (exp[ABCG1]): Cells treated with ApoA1/HDL alone.

ABCG1 control (cnt[ABCG1]): Untreated cells.

Each group included 8 biological replicates.

2.7.2. ABCA1 induction and LNP treatment

Under basal conditions, RAW 264.7 cells exhibit minimal Abca1 expression. To upregulate ABCA1, the cells in the ABCA1/G1 groups were pre-treated with 0.3 mM 8-Bromoadenosine 3′,5′-cyclic monophosphate (cAMP, Millipore-Sigma) and 1 μg/mL acyl-coenzyme A: cholesterol acyltransferase inhibitor (CP113818, Pfizer) for 24 h. Ctrl- or ITA-LNPs (50 μM) were added to respective wells, and cells were incubated overnight at 37 °C.

2.7.3. Efflux measurement

After washing with PBS, the cells were treated with 100 μg/mL ApoA1 (Millipore-Sigma) or HDL (Athens Research and Technology) for 24 h. Supernatants (efflux media) were transferred to a fresh 96-well plate, and cells were lysed using M-PER Mammalian Protein Extraction Reagent (Thermo Fisher) for 30 min at 37 °C. Fluorescence intensity was measured at excitation/emission wavelengths of 490/515 nm for both efflux media and lysates.

2.7.4. Efflux calculations

Cholesterol efflux was quantified using the following equations:

$$\%\text{efflux}=100\text{xEfflux media}/\text{efflux lysates}$$

$$\text{efflux}\left(\text{ABCA}1/\text{G}1\right)=\%\text{efflux}\left(\text{exp}\left[\text{ABCA}1/\text{G}1\right]\right)\text{-AVERAGE}\{\%\text{efflux}(\text{cnt}\left[\text{ABCA}1/\text{G}1\right])\}$$

$$\text{efflux}\left(\text{ABCG}1\right)=\%\text{efflux}\left(\text{exp}\left[\text{ABCG}1\right]\right)\text{-AVERAGE}\{\%\text{efflux}(\text{cnt}\left[\text{ABCG}1\right])\}$$

$$\text{efflux}\left(\text{ABCG}1\right)=\text{AVERAGE}\left\{\text{efflux}\left(\text{ABCA}1/\text{G}1\right)\right\}\text{-AVERAGE}\left\{\text{efflux}\left(\text{ABCG}1\right)\right\}$$

2.8. Gene expression analysis (qPCR)

For qRT-PCR analysis, cells were lysed in TRIzol reagent (Invitrogen), and total RNA was isolated using the Direct-zol RNA Miniprep Kit (Zymo Research) according to the manufacturer’s protocol. RNA concentration and purity were determined by spectrophotometry (NanoDrop 2000; Thermo Fisher Scientific). Complementary DNA (cDNA) was synthesized from 250 ng of total RNA using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems). Quantitative real-time PCR was performed in triplicate using TaqMan Fast Advanced Master Mix (Thermo Fisher Scientific) and gene-specific TaqMan probes on a LightCycler 480 system (Roche). Reaction conditions followed the manufacturer’s recommended thermal cycling parameters. Relative gene expression was calculated using the ΔΔC_T method and normalized to the geometric mean of the housekeeping genes Gapdh and Actb.

2.9. Immunoblot assays (Western blot)

Cells were lysed in 2 × Laemmli SDS loading buffer (4 % SDS, 0.1 M DTT, 6.25 % Ficoll, 1 × protease inhibitor cocktail, 0.001 % bromophenol blue) and denatured by boiling at 95 °C for 5 min. For ABCA1 detection, instead of standard denaturing protocol, the samples were heated at 37 °C for 2 min to minimize protein degradation. Protein concentrations were quantified using a Direct Detect Infrared Spectrometer (Millipore Sigma). Protein concentrations in lysates were normalized by mixing with the loading buffer, and resolved on 4–20 % Criterion TGX Precast Protein Gels (Bio-Rad) at 300 V in the Tris-Glycine running buffer. Proteins were transferred to PVDF membranes using the Trans-Blot Turbo Transfer System (Bio-Rad).

To enhance antigen detection, the membranes were fixed in pre-chilled acetone (−20 °C) for 30 min, followed by oven-drying at 50 °C for 30 min, as previously described [40]. All membranes were blocked with 5 % (w/v) nonfat dry milk in Tris-buffered saline containing 0.05 % Tween-20 (TBS-T) for 1 h at room temperature, then incubated overnight at 4 °C with primary antibodies (Supplementary Table S1). After washing with TBS-T, the membranes were incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies (1:50,000 dilution) for 1 h at room temperature. Signal detection was performed using enhanced chemiluminescence (ECL), and images were acquired with an Azure C400 Western Blot Imaging System (Azure Biosystems).

2.10. Gene silencing and overexpression

Immortalized bone marrow-derived macrophages (iBMDMs, Kerafast) or RAW 264.7 cells were seeded into 24-well plates at a density of 0.75 × 10⁶ cells per well. After 24 h, the culture medium was replaced with fresh medium containing 10 μg/mL polybrene (Merck Millipore). At the time of infection, cells had reached ~50 % confluency. Cells were transduced in technical duplicates with lentiviral particles encoding either IRG1-or HO-1-targeting shRNA or a copGFP control (Vector-Builder) all carrying puromycin resistance cassette. The transduction was performed using three viral titers (1, 2, and 4 × 10⁶ transduction units (TU) per milliliter based on preliminary titration). Following 24 h of incubation, the medium was replaced with polybrene-free medium to reduce cytotoxicity. To select stably transduced cells, puromycin dihydrochloride (Thermo Fisher Scientific) was added to the culture medium at 10 μg/mL for 7 d, with fresh puromycin-containing medium replenished every 48–72 h. Silencing efficiency was confirmed by Western blot analysis of IRG1 or HO-1 protein levels (antibody details in Supplementary Table S1), normalized to Ponceau S as a loading control.

To overexpress HA-tagged ubiquitin in RAW 264.7 cells, the cells were transduced with a lentivirus obtained from pLenti puro HA-Ubiquitin (Addgene, plasmid #74218) using HEK293 transfection and utilizing ABM 2nd generation lentivirus packaging system (Applied Biological Materials Inc. [abm], Cat. No.LV003) as per manufacturer’s instructions. The sable cell selection was performed with puromycin as described above.

2.11. Lipid accumulation and uptake assays in BMDMs

1. Oil Red O Staining for Lipid Accumulation

BMDMs were seeded into 96-well plates at a density of 0.05 × 10⁶ cells per well. To induce lipid accumulation, the cells were treated with 100 μg/mL acLDL (prepared as above) for 24 h. Subsequently, cells were incubated with either 50 μM Ctrl- or ITA-LNPs for an additional 24 h. The cells were fixed in 2 % (w/v) PFA for 10 min at room temperature. Lipid droplets were stained using Oil Red O (Sigma-Aldrich) as follows. The cells were rinsed with PBS, equilibrated in 60 % (v/v) isopropanol for 5 min, and stained with 0.5 % (w/v) Oil Red O in 60 % isopropanol/propylene glycol (1:1 v/v) for 10 min. Excess stain was removed by washing with 60 % isopropanol, followed by a distilled water rinse. Nuclei were counterstained with Richard-Allan Hematoxylin (Thermo Fisher Scientific) for 30 s, subjected to bluing in saturated sodium bicarbonate for 1 min, and rinsed with tap water. Bright-field images were acquired using a Keyence BZX700 microscope equipped with Nikon CFI 20x and 60x Plan Apo λ objectives.

2. BODIPY-acLDL Fluorescent Lipid Uptake Assay

For fluorescent lipid uptake analysis, BMDMs were treated with Ctrl- or ITA-LNPs (50 μM, 24 h) followed by 75 μg/mL BODIPY-labeled acLDL for 24 h. The cells were counterstained with 1 μM Hoechst 33342 for 10 min at room temperature. The cells were fixed in 2 % PFA for 10 min, and fluorescence images were captured using a Keyence BZX700 microscope.

2.12. Lipid quantification in BMDMs

BMDMs were seeded in 96-well plates and incubated for 24 h in BMDM medium supplemented with 75 μg/mL acLDL or phosphate-buffered saline (PBS; vehicle control). Cells were then treated for an additional 24 h with one of the following: PBS (vehicle), Ctrl-LNP, or ITA-LNPs (50 μM). After treatment, the medium was aspirated, and the cells were washed twice with PBS. Total cholesterol and triglyceride levels were quantified using the Cholesterol/Cholesterol Ester-Glo and Triglyceride-Glo luminescence assays (Promega) respectively, according to the manufacturer’s protocols.

2.13. Calpain activity experiments

WT or HO-1^KD RAW 264.7 cells were seeded in 12-well plates at a density of 1 × 10⁶ cells per well. They were incubated for 24 h in BMDM medium supplemented with 50 μM ITA- or Ctrl-LNP. After treatment, the medium was aspirated, and the cells were washed twice with PBS. The cells were lysed in a lysis buffer supplied with Calpain Activity Assay Kit (Abcam) and some lysates were treated with calpain inhibitor or active calpain I provided by the manufacturer, which served as plate negative and positive controls. Lysates were collected, and calpain activity was measured as per manufacturer instructions at excitation 400 nm, emission 505 nm on Molecular Devices Spectramax i3 multi-mode reader.

2.14. ABCA1 stability experiments

RAW 264.7 cells were seeded into 12-well plates at a density of 1 × 10⁶ cells per well. Following cell attachment for 2 h, the medium was replaced with DMEM supplemented with 1 % FBS, 1 % penicillin-streptomycin, and 0.3 mM 8-Bromoadenosine 3′,5′-cyclic monophosphate (cAMP; Sigma-Aldrich). After 24 h of incubation, the cells were washed with PBS, and the medium was replaced with fresh DMEM (1 % FBS, 1 % penicillin-streptomycin) without cAMP. The cells were then treated with either Ctrl- or ITA-LNPs (50 μM). At 0, 4, 6, 18, and 24 h post-treatment, the cells were lysed in 2 × Laemmli SDS loading buffer containing containing 100 mM dithiothreitol (DTT) and 10 μM MG132 (Adipogen) to prevent protein degradation. Lysates were sonicated (3 × 5 s pulses), centrifuged (12,000×g, 10 min, 4 °C), and quantified with a Direct Detect Infrared Spectrometer. Equal amounts of protein (30 μg per lane) were resolved on 4–20 % Criterion TGX Precast Protein Gels (Bio-Rad) in Tris-Glycine-SDS running buffer at 300 V for 25 min. Proteins were transferred to PVDF membranes using a Trans-Blot Turbo Transfer System (Bio-Rad). Immunoblotting was performed as described above.

2.15. Lipid metabolism and ABCA1 degradation analysis in BMDMs

Bone marrow-derived macrophages (BMDMs) were seeded into 12-well plates at a density of.

0.5 × 10⁶ cells per well. Following attachment, the cells were treated with 100 μg/mL acLDL for 6 h to induce lipid loading. The cells were then incubated with either Ctrl- or ITA-LNPs (50 μM) for 18 h. After treatment (18 h), the medium was replaced with DMEM supplemented with: 1 % fatty acid-free bovine serum albumin (BSA; Sigma-Aldrich), 1 % MEM Non-Essential Amino Acids (Thermo Fisher Scientific), 1 % Antibiotic-Antimycotic (Anti-Anti; Gibco), 10 μg/mL cycloheximide (CHX; protein synthesis inhibitor; MedChemExpress). To inhibit specific degradation pathways, the cells were co-treated with: 10 μM MG132 (proteasome inhibitor; Adipogen), 10 μM Pepstatin A (lysosomal protease inhibitor; Cayman Chemical), 50 μM chloroquine phosphate (autophagy inhibitor; Sigma-Aldrich). After 4 h of incubation, the cells were lysed in a 2 × Laemmli SDS loading buffer containing 100 mM dithiothreitol (DTT). Lysates were sonicated (3 × 5 s pulses), centrifuged (12,000×g, 10 min, 4 °C), and quantified using Direct Detect Infrared Spectrometer. Equal protein amounts (20–30 μg per lane) were resolved by SDS-PAGE, and immunoblotting was performed as described above.

2.16. Ubiquitin immunoprecipitation

RAW 264.7 cells were cultured and treated with cAMP and Ctrl- and ITA-LNPs as described above. The cells were lysed in a buffer containing 50 mM Tris-HCl, pH 7.5, 0.15 M NaCl, 1 mM EDTA, 1 % IGEPAL CA-630, 10 % glycerol, 1x Roche cOmplete protease inhibitors (Millipore Sigma), 50 μM PR-619 (LifeSensors, SI9619), and 1 μM 1,10-phenanthroline. The lysates containing at least 20 mg/mL total protein were treated with Tandem Ubiquitin Binding Entities (TUBEs, LifeSensors) as per manufacturer’s protocol. Following pulldown of polyubiquitinated proteins as per manufacturer’s protocol and elution using 2 × Laemmli SDS buffer containing 100 mM DTT, 1 M urea and 20 μM MG132, the immunoprecipitates were resolved on 4–20 % Criterion TGX Precast Protein Gels (Bio-Rad) and subjected to immunoblotting as outlined above. The membranes were probed with monoclonal anti-ABCA1 antibodies.

2.17. Sequencing and transcriptomic analysis

Whole abdominal aortas were excised, immediately placed in ice-cold TRIzol Reagent (Thermo Fisher), and homogenized using a sterile pestle. Total RNA was extracted following Direct-zol RNA Miniprep (Zymo Research) and quantified via a NanoDrop 2000 (Thermo Fisher Scientific), and assessed for integrity using 1 % agarose gel electrophoresis and an Advanced Analytical Fragment Analyzer (Agilent Technologies). The samples that exhibited RNA Integrity Number (RIN) > 7.0 were further subjected to bulk RNA sequencing.

RNA samples were normalized to 50 ng/μL and submitted for single-end sequencing at the Beijing Genomics Institute (BGI). Libraries were prepared using standard protocols and QuantSeq 3’ mRNA-Seq V2 Library Prep Kit and sequenced using Illumina HiSeq 3000. Bioinformatics analysis was performed using a NextFlow nf-core computational pipeline, as previously described [41]. The RNA-seq data was uploaded to the GEO database (GEO: GSE288359).

2.18. Statistical analysis

Data was first tested for normality using the Shapiro-Wilk test and for equality of variances using Bartlett’s test. If the normality and equality of variances are satisfied (p ≤ 0.05), the Student’s t-test (for 2 groups) or ANOVA with Tukey’s post-hoc test (>2 groups) were used to compare the means. For experiments with >2 groups, pairwise t-test was applied with Holm post-hoc test. The specific statistical analysis employed is indicated in the figure legends. If the data displayed non-normal distribution or data with unequal variances, the nonparametric Mann-Whitney U test (2 groups) was used, and statistical analysis was performed using R version 4.3.0 (“Already Tomorrow”) or later. Graphs and plots were created using Plotly Chart Studio or with R packages tidyverse/ggplot2. The results are presented as mean with standard error of mean (SEM). The SEM values are displayed as error bars in the figures or values following plus-minus sign in the main text.

3. Results

3.1. Itaconate reduces atherosclerotic burden and increases ABCA1

To gain insight into ITA-LNP’s effects on plaque lipid metabolism in atherosclerosis, we intravenously administered ITA-LNPs (50 mg/kg) into Apoe^−/− mice while providing HCHFD (1.25 % cholesterol, 40 % calories from fat) ad libitum for 12 weeks (Fig. 1a). As a control, Ctrl-LNPs were injected at the same dose, similar to the experiments previously described by us [3]. ITA-LNP accumulation in atherosclerotic lesions was confirmed by a single dose injection of Atto647 fluorochrome-labeled ITA-LNPs (50 mg/kg) followed by excision of descending thoracic aorta with aortic arch and fluorescence imaging (Fig. 1b). As expected, ITA- but not Ctrl-LNPs significantly reduced atherosclerosis burden as seen by whole aorta en face ORO staining analysis (Fig. 1c) and in aortic root sections (Supplementary Fig. S1a and b). Strikingly, ABCA1 immunostaining in brachiocephalic arteries (BCAs) from the same animals demonstrated significant upregulation of the ABCA1 in the ITA-LNP group compared to Ctrl-LNP animals (Fig. 1d and e). To gain insight into changes at the transcriptional level in these aortas, we performed bulk RNA sequencing (RNAseq) of the RNA isolated from whole abdominal aortas (GEO: GSE288359). After differential expression gene analysis (Fig. 1f), we noted a significant (false discovery rate, FDR <0.05) yet small increase in the expression of Lpl, encoding for lipoprotein lipase, a key enzyme responsible for triglyceride metabolism, and Cd36, scavenger receptor Class B, known for its role in the uptake of oxidized low-density lipoprotein (oxLDL) [42,43]. The analysis further revealed downregulation of inflammation mediators such as Il1b, encoding for cytokine interleukin-1β, and chemokine Cxcl3, which are consistent with the anti-inflammatory actions of ITA [4,5,44]. However, we did not observe any other notable differences in the expression level of mRNAs frequently associated with lipid metabolism (Abca1, Abcg1, Fabp1, Lipc, Fasn, and others). Gene ontology (GO) analysis (Supplementary Fig. S1c) was in line with these observations, demonstrating non-significant upregulation of Fatty acid omega oxidation as the only lipid metabolism-related GO term. In light of the upregulated ABCA1 expression on the protein level in BCAs from animals injected with ITA-LNPs (Fig. 1d), we sought to understand if this effect could be mediated by endogenous, non-LNP-derived itaconate. To demonstrate this, we knocked down the expression of ACOD1, an ITA-producing enzyme, in immortalized bone marrow-derived macrophages (iBMDMs). The treatment of wild type (WT) iBMDMs with free cholesterol produced significant upregulation of ABCA1 protein, in contrast to ACOD1-deficient iBMDMs (ACOD1 KD) where ABCA1 expression was blunted (Fig. 1g). We chose to use free cholesterol in these experiments and prevent its intracellular esterification with acyl-coenzyme A: cholesterol acyltransferase inhibitor [45–47] because our prior work demonstrated that ACOD1 cannot be induced by other types of cholesterol loading [3]. We also examined ABCG1 expression, which remained unchanged across different treatments and was insensitive to ACOD1 expression (Fig. 1g). Collectively, this data demonstrates that ITA selectively upregulates ABCA1 on the protein level in atherosclerosis but not at the transcription level, and this change is not accompanied by notable differences in the expression of other lipid metabolism genes.

Fig. 1. Itaconate alleviates atherosclerosis and upregulates plaque ABCA1.

(a) Schematic depiction of the injection schedule of ITA-LNPs in Apoe^−/− mice under continuous HCHFD feeding. (b) ITA-LNPs accumulation in atherosclerosis as detected by ITA-LNP Atto 647 fluorescence. (c) En face Oil Red O staining of atherosclerotic lesions in the aorta at the end of the experiments depicted in (a). Quantification is presented on the right. (d) ABCA1 immunohistochemical staining in brachiocephalic arteries (BCA) at the end of the experiments depicted in (a). (e) Quantification of ABCA1 expression from (d). (f) Volcano plot after differential gene expression analysis from bulk RNAseq in whole abdominal aortas at the end of the experiments depicted in (a). (g) Immunoblotting experiments probing for ABCA1 and ABCG1 in iBMDMs, either wild type (WT) or with an ACOD1 knockdown (KD). The cells were loaded with cyclodextrin-stabilized cholesterol (CD-Chol, 10 μg/mL) in the presence or absence of acyl-coenzyme A:cholesterol acyltransferase inhibitor (CP113818, 1 μM). Statistical analysis was performed using pairwise t-test with Holm post hoc correction. n = 7 for ITA-LNP groups and n = 6 for Ctrl-LNP groups in (a–e). n = 3 for ITA-LNP and n = 4 for Ctrl-LNP groups in (f). n = 2 in (g).

3.2. Itaconate regulates lipids via citrate metabolism

RNAseq data (Fig. 1f) suggests ITA has a role in regulating lipid metabolism. One possible scenario by which ITA could mediate lipid metabolism is through disruption of the tricarboxylic acid cycle (TCA). Fig. 2a depicts the hypothetical fate of citrate, a key metabolite in fatty acid and cholesterol synthesis and metabolism. We previously demonstrated that ITA inhibits succinate dehydrogenase A (SDHA) while increasing levels of pyruvate dehydrogenase (PDH), which catalyzes the conversion of pyruvate to acetyl-coenzyme A, further feeding into the TCA cycle [3]. The TCA citrate could then either accumulate or be transported into the cytosol for fatty acid and cholesterol synthesis. We investigated the fate of citrate using datasets from our previous study [3] utilizing stable isotope metabolomics in BMDMs treated with ITA- or Ctrl-LNPs. In these experiments, BMDMs were cultured in a medium supplemented with uniformly labeled with [U–¹³C]-glucose as the only carbohydrate source. The results showed that the amount of total citrate was slightly, but not significantly upregulated in ITA-LNP-treated cells. Notably, the citrate in the ITA-LNP group was mostly unlabeled (derived from ¹²C source) in contrast to that in the Ctrl-LNP group where the citrate was largely ¹³C-labeled (Fig. 2b). ITA inhibits SDH, which would lead to accumulation of TCA components such as citrate. However, this observed increase in unlabeled citrate without a significant increase in total citrate indicates that ITA-LNPs may also induce significant citrate utilization as well as generation from a non-glucose source, such as fatty acids.

Fig. 2. Itaconate alleviates lipid burden in acetyl-LDL-loaded macrophages.

(a) Schematic representation of ITA actions on tricarboxylic acid cycle (TCA) enzymes and the role of citrate in fatty acid synthesis. (b) The lysates from BMDMs cultured with [U–¹³C]-glucose and treated with Ctrl- or ITA-LNPs (50 μM, 24 h) were subjected stable isotope resolved metabolomics (SIRM) to analyze the levels of intracellular citrate and its isotopologues. (c) Schematic of experiments in acetyl-LDL (acLDL)-loaded BMDMs subjected to PBS, Ctrl- or ITA-LNP treatment. (d) Total triglyceride levels at day 8 in BMDM lysates from (c). (e) Oil Red O (ORO) staining and quantification in BMDMs from (c). Nuclei were counterstained with hematoxylin. (f) Intracellular lipid droplet staining and quantification with 4,4-difluoro-4-bora-3a,4a-diaza-s-indacene (BODIPY, green). Nuclei were counterstained with Hoescht (blue). Low (20x) and high (60x) power images are shown. Data were analyzed and p values were obtained by one-way ANOVA with Tukey’s multiple comparisons test or pairwise t-test with Holm post hoc correction. n = 3 for ITA- and Ctrl-LNP groups in (b). n = 8 for all groups in (d). n = 16, 17, and 19 for PBS, Ctrl-LNP, and ITA-LNP respectively in (e–f).

To understand if these changes in citrate metabolism have an impact on lipid handling in ITA-vs. Ctrl-LNP-treated cells, we preloaded BMDMs with acetylated low-density lipoprotein (acLDL) to induce the formation of foam cells and triglyceride overload (Fig. 2c). Next, the BMDMs were incubated with LNPs followed by the analysis of total triglycerides in the cell lysates. Interestingly, ITA-LNPs, but not Ctrl-LNPs, significantly reduced triglyceride load in these BMDMs (Fig. 2d). This occurred without changes in total cholesterol as analyzed in the same cell lysates (Supplementary Fig. S1d). These results were further supported by ORO staining in BMDMs subjected to the same treatment schedule (Fig. 2e), which demonstrated a significant reduction in the formation of lipid droplets in cells treated with ITA-LNPs. Additionally, we performed an experiment where LNPs were first added to the BMDMs, followed by the acLDL loading, which also demonstrated similar results (Fig. 2f). These results could be explained by more facile lipid utilization in ITA-LNP-treated BMDMs as compared to Ctrl-LNP-treated cells. To gain insight into this possibility, we analyzed the expression of Slc25a1, a bidirectional transporter of the citrate between the mitochondria and cytosol (Supplementary Fig. S1e). The levels of Slc25a1 mRNA expression were significantly increased in ITA-LNP-treated BMDMs, suggesting enhanced fatty acid-derived citrate shuttling. These data are in line with metabolomics experiments (Fig. 2b) that suggested a non-carbohydrate source of citrate in ITA-LNP-treated cells. Overall, the in vitro data above are supportive of ITA’s role in lipid metabolism in macrophages and could explain, at least in part, the enhanced Lpl and CD36 expression seen in atherosclerotic plaque from animals subjected to ITA-LNP injections.

3.3. Itaconate regulates cholesterol via ABCA1 upregulation

Given that ABCA1 was also increased in the plaques of mice treated with ITA-LNPs (Fig. 1d and e), we investigated ITA’s role in regulating cholesterol. ABCA1 is a key shuttle protein responsible for cholesterol removal from the cell and an important player in reverse cholesterol transport (RCT) [48,49]. The protein levels of ABCA1 were highly upregulated in unstimulated BMDMs that were treated with ITA-LNP for 48 h compared to Ctrl-LNP-treated cells (Fig. 3a). Stimulation of BMDMs with oxLDL further amplified ITA-LNP-induced expression of ABCA1. ITA-LNP treatment also increased the levels of ABCG1 in unstimulated BMDMs. However, no discernible differences in ABCG1 expression between LNP groups were noted upon oxLDL stimulation. To evaluate whether the increase in ABCA1 levels is reflective of the functional capacity to efflux cholesterol in macrophages, we performed сhоlеsterol efflux assays [32,50–52] using a well-established model of RAW 264.7 cells pretreated with cyclic-AMP (cAMP) [51,53]. We observed significantly increased cholesterol efflux in ITA-vs. Ctrl-LNPs treated RAW 264.7 macrophages towards two classic cholesterol acceptors, high-density lipoprotein (HDL) and ApoA-I (Fig. 3b).

Fig. 3. Itaconate stimulates cholesterol efflux from macrophages through ABCA1 protein stabilization.

(a) Immunoblot analysis probing for ABCA1 and ABCG1 in BMDM lysates following cell treatment with Ctrl- and ITA-LNPs (50 μM) in the presence or absence of oxLDL (50 μg/mL) for 48 h. (b) Cholesterol efflux experiments in RAW 264.7 cells preloaded with 100 μg/mL acLDL containing BODIPY-cholesterol for 24 h followed by Ctrl- or ITA-LNPs (50 μM) treatment for additional 24 h. The efflux was performed in the presence of HDL (100 μg/mL) or recombinant ApoA-I (100 μg/mL). (c) BMDMs were treated for 24 h as indicated, including with Ctrl- and ITA-LNPs (50 μM), and were subjected to analysis of Abca1 mRNA expression by qPCR. LXR agonist GW3965 (1 μM) and oxidized LDL (oxLDL, 100 μg/mL) served as positive controls, and LXR antagonist SR9243 (1 μM) served as a negative control. (d) RAW 264.7 cells were pretreated with cAMP to induce Abca1 expression followed by ITA- or Ctrl-LNP treatment and cAMP withdrawal as indicated. Left: ABCA1 levels were then determined by immunoblotting. Right: Quantification of ABCA1 protein levels over time in experiments from (d), reported as % ABCA1 protein. Statistical analysis was performed using pairwise t-test with Holm post-hoc analysis (in a), two-tailed t-test (in c), and ANOVA with Tukey HSD post-hoc (in d) analysis. *p < 0.05, **p < 0.01. n = 16 for ITA- and Ctrl-LNP groups for HDL efflux in (b). n = 10 for Ctrl-LNP and n = 11 for ITA-LNP for ApoA-1 efflux in (b). n = 4 for all groups in (c). n = 3 for all groups in (d).

Intriguingly, the levels of Abca1 mRNA transcripts in unstimulated BMDMs remained unchanged across LNP treatment groups and PBS-treated cells, as compared to relevant positive and negative controls (Fig. 3c), indicating that the observed increase in ABCA1 was occurring on the protein level. We recently demonstrated that the persistent expression of the ABCA1 protein without significant changes in levels of Abca1 mRNA transcripts is suggestive of enhanced ABCA1 stability in the cell and its resistance to proteasomal degradation [54]. To investigate this in the context of ITA’s effect on ABCA1 in macrophages, we performed a kinetic study examining ABCA1 expression levels over time following treatment with ITA- or Ctrl-LNPs. We took advantage of the fact that RAW 264.7 cells express very low basal levels of ABCA1 and that ABCA1 induction by cAMP is transient upon cAMP removal [55]. We incubated RAW 264.7 cells with cAMP for 24 h followed by the replacement of the cell culture medium with ITA- or Ctrl-LNP-containing medium, however without cAMP, for an additional 24 h (Fig. 3d). The levels of ABCA1 protein were then determined by immunoblotting at different time points. The analysis (Fig. 3d) showed enhanced stability of ABCA1 in ITA-vs. Ctrl-LNP-treated cells at 18 and 24 h. Collectively, these data suggest that ITA plays an important role in ABCA1 protein stabilization in macrophages, which manifests in improved cholesterol efflux to common cholesterol acceptors.

3.4. Itaconate inhibits ABCA1 degradation via regulation of the HO-1-calpain axis

One possible explanation of ITA stabilizing ABCA1 in macrophages is through heme oxygenase (HO-1)-calpain axis (Fig. 4a). This scenario assumes the induction of HO-1, known to inhibit calpains [56–58], a family of enzymes responsible for proteolysis in the cell. Calpain-driven protein degradation provides cleavage products serving as substrates for the ubiquitin-proteasome pathway (UPP) [59]. Thus, calpain inhibition may also result in attenuation of ubiquitin-dependent ABCA1 degradation [60]. We and other investigators have previously demonstrated that ITA and its derivatives robustly upregulate nuclear factor erythroid 2-related factor 2 (NRF2) levels in macrophages [1,3,8,61]. NRF2 is a transcription factor that coordinates the expression of a variety of genes involved in the cellular response to oxidative stress, most notably HO-1 [62,63]. To confirm HO-1 induction by ITA-LNPs and its dependence on NRF2, we conducted immunoblotting experiments in lysates from wild-type (WT) and Nrf2^−/− BMDMs treated with ITA-LNPs (Supplementary Fig. S2a). Indeed, ITA-LNPs significantly induced HO-1 expression in WT BMDMs, but not in Nrf2^−/− cells. Logically, then, HO-1 induction should also attenuate ABCA1 ubiquitylation (Fig. 4a).

Fig. 4. Itaconate limits ABCA1 ubiquitination through induction of heme oxygenase (HO-1) and inhibition of lysosomal calpain activity.

(a) Schematic representation of a relationship between ITA, HO-1, calpains, and proteasomal degradation of ABCA1 protein. (b) BMDMs were treated with ITA- or Ctrl-LNPs (50 μM, 24 h) followed by immunoblotting (IB) for ubiquitin. (c) Polyubiquitylated proteins were captured by tandem ubiquitin-binding entities (TUBEs) followed by immunoblotting for ABCA1 in lysates from RAW 264.7 cells. (d) Left: HO-1 expression was knocked down (HO-1^KD) in RAW 264.7 cells and the cells were subjected to ABCA1 stability testing. Right: Quantification of ABCA1 protein levels over time, reported as % ABCA1 protein. (e) Calpain activity as measured in WT or HO-1^KD RAW 264.7 cells treated with ITA- or Ctrl-LNPs. (f) BMDMs were treated with 50 μM ITA-LNPs for 24 h or left untreated followed by the inhibition of HO-1 activity with zinc protoporphyrin (ZnPPIX, 10 μM), inhibition of calpain activity with ac-calpastatin (acCal, 50 μM), or the use of both inhibitors simultaneously for an additional 24 h. Dimethylsulfoxide (DMSO) served as a vehicle control. The cell lysates were then immunoprobed for ABCA1. (g) BMDMs were loaded with acLDL and treated with Ctrl- and ITA-LNPs (50 μM) for 24 h in the presence or absence of various inhibitors as indicated. The cell lysates were then immunoprobed for ABCA1. The treatments included protein synthesis inhibitor cycloheximide (CHX, 10 μg/mL), proteasome inhibitor MG132 (10 μM), inhibitor of aspartyl proteases Pepstatin A (PepA, 50 μM) and chloroquine (CHQ, 100 μM), the inhibitor of maturation and fusion of endosomes and lysosomes. Statistical analysis was performed using ANOVA with Tukey HSD post-hoc (in d) or pairwise t-test with Holm post-hoc analysis (in e), **p < 0.01 vs. Ctrl-LNP, ns = not significant. n = 3 for all ITA- and Ctrl-LNP groups in (b–d). n = 8 for all groups in (e). n = 2 in all groups in (f–g).

To test this, we first performed a series of experiments using three independent methods of ubiquitylation detection and two types of macrophages. First, pan-ubiquitylation was determined by densitometry analysis of ~250 ± 50 kDa ubiquitin-immunoreactive bands using FK2 anti-ubiquitin antibodies (Fig. 4b). This indirect analysis was deemed possible due to low ubiquitin-immunoreactivity at <250 kDa and because ABCA1’s molecular weight is 250 kDa. Indeed, ITA-LNPs significantly attenuated the ubiquitylation of ~250 kDa proteins in BMDMs as compared to Ctrl-LNPs. Next, we subjected RAW 264.7 cells to cAMP-driven induction of ABCA1 and LNP treatment as above (Fig. 3d), followed by the capture of polyubiquitylated proteins with agarose-bound tandem ubiquitin-binding entities (TUBEs) [64]. The TUBE pulldowns were then eluted, followed by immunoblotting for ABCA1 (Fig. 4c). The results show drastic attenuation of ubiquitylation of ABCA1 upon treatment with ITA-LNPs, while total levels of ABCA1 were high in this treatment group, as detected in inputs from these immunoprecipitation experiments. Finally, we overexpressed epitope-tagged ubiquitin (hemagglutinin, HA) in RAW 264.7 cells using a lentiviral vector, followed by the same treatment schedule with cAMP and LNPs as above. Immunoprecipitation of the lysates with anti-ABCA1 and probing with anti-HA antibodies confirmed the findings above (Supplementary Fig. S2b).

After confirming that ubiquitylation was decreased by ITA-LNP, we investigated HO-1’s hypothetical role in preventing ABCA1 degradation via calpain inhibition (Fig. 4a) [65,66]. We knocked down the expression of HO-1 in RAW 264.7 cells (HO-1^KD) using an shRNA-expressing lentiviral vector (Supplementary Fig. S2c). Control cells transduced in the same manner but with a vector encoding for green fluorescent protein (RAW GFP) were also generated. Next, we tested ABCA1 stability in the presence of ITA-LNPs using the same experimental conditions as above (Fig. 4d). Notably, the enhanced persistence of ABCA1 expression over time upon ITA-LNP treatment was completely abolished in HO-1^KD cells (Fig. 4d). However, ABCA1 expression was still prominent and statistically significant in RAW GFP cells (Supplementary Fig. S2d and e). Further, we measured calpain activity in HO-1^KD cells treated with ITA- and Ctrl-LNPs compared to that in WT cells (Fig. 4e). The calpain activity was significantly downregulated in WT cells in response to ITA-LNPs. The same was observed in HO-1^KD cells; notably, however, calpain activity was somewhat elevated in HO-1^KD cells in HO-1^KD cells compared to WT (Fig. 4e), suggesting that the effects of ITA-LNPs on HO-1 mediate calpain activity, at least in part. In line with these findings, we further investigated the effects of inhibiting HO-1 and calpain activity separately or in combination in BMDMs (Fig. 4f). ITA-LNPs were able to partially restore ABCA1 in the setting of HO-1 inhibition (zinc protoporphyrin, ZnPPIX). Additionally, ITA-LNPs further increased ABCA1 levels in cells treated with a calpain inhibitor (ac-calpastatin, acCal). These results in primary macrophages additionally corroborated the aforementioned findings in genetically manipulated RAW 264.7 cells, indicating that ITA-LNPs stabilize ABCA1 in part via the HO-1-calpain axis.

Finally, we sought to gain insight into whether ITA impacts other selected proteases in primary cells by utilizing an experimental setup previously described by us [54]. We preloaded BMDMs with acLDL, then treated the BMDMs with cycloheximide (CHX) simultaneously with chemical inhibitors of lysosomes and lysosomal proteases (Fig. 4g). Notably, only proteasome inhibitor MG132 was able to equalize the levels of ABCA1 between ITA- and Ctrl-LNP-treated BMDMs. Collectively, these data show that ITA-LNPs prevent ABCA1 decay, in part, by decreasing its proteasomal and calpain-mediated degradation via the HO-1-calpain axis, thus increasing cholesterol efflux in macrophages.

4. Discussion

In this study, we utilized ITA-LNPs to broaden ITA’s mechanisms of action in atherosclerosis: not only does ITA induce immunomodulation, but it also regulates lipid metabolism and cholesterol efflux in macrophages. This highlights ITA as a multifaceted therapeutic that simultaneously targets inflammation and lipid dysregulation to promote a beneficial phenotype in atherosclerotic macrophages (Fig. 5).

Fig. 5.

ITA-LNPs inhibit atherosclerotic progression via ABCA1 stabilization and immunomodulation.

Itaconate is a molecule with a number of molecular derivatives, two of which (e.g., ITA-LNPs, 4-OI) have been shown to modulate atherosclerosis [3,8,67]. ITA-LNPs were selected for this study for several reasons. First, they increase the delivery of unconjugated ITA intracellularly, which is unique amongst the ITA derivatives. Second, they exhibit specific targeting to plaque in atherosclerotic mice. Notably, we have previously shown that another derivative of ITA, dimethyl itaconate (DMI), does not accumulate in cardiac tissue [68]. Third, they recapitulate the effects of ITA, allowing us to study the effects of endemic ITA [3]. Finally, the study of endogenous ITA can be made possible by upregulating its producing enzyme, ACOD1, to increase intracellular ITA. However, we previously demonstrated that several lipid-loading models, such as acLDL, do not increase Acod1 (Irg1) expression [3]. Thus, to observe the effects of ITA in more foam cell-like models (e.g., triglyceride-laden macrophages using acLDL), we used ITA-LNPs to increase intracellular ITA.

This study expands upon the ability of an ITA derivative, 4-octyl itaconate (4-OI), to regulate lipids in non-alcoholic steatohepatitis (NASH) in hepatocytes [19] and in endotoxemia [20] to encompass macrophages in atherosclerosis. Although hepatocytes are not known to produce Acod1, they and other cell types are responsive to ITA and its derivatives. Notably, itaconate-treated hepatocytes had significantly increased citrate levels, which was attributed to increased fatty acid oxidation [19]. Additionally, 4-OI was found to stabilize CPT1a, a mitochondrial fatty acid transporter, to increase lipid metabolism, by decreasing ubiquitination [20]. Combined with the results of our study, ITA appears to regulate fatty acid transport in both the cytosolic (ABCA1) and mitochondrial (CPT1a, Slc25a1) compartments.

The findings of this study echo the effects observed with mir-223, which has also been shown to ameliorate atherosclerosis and regulate cholesterol via upregulation of Abca1 [35]. Additionally, both mir-223 and ITA-LNPs increase ABCA1 via protein stabilization [54]. Notably, ITA-LNPs have several significant benefits over mir-223 as a therapeutic, including ITA’s multi-pronged mechanisms of action that act on inflammation and lipid regulation, both of which are therapeutic targets of atherosclerosis. Additionally, ITA-LNPs have the additional benefits of not being a biologic, including no additional challenge of antibody humanization, lower expenses to produce, and greater ease of manufacturing [69].

In analyzing ABCA1 stabilization, we utilized three different methods of ubiquitination testing (Fig. 4b and c), as there are significant challenges and limitations in the analysis of this post-translational modification that are widely acknowledged in the literature [64,70]. Ubiquitination is often a transient modification, leading to low stoichiometry, which makes it challenging to detect using conventional methods such as Western blotting or mass spectrometry (MS).

Itaconate is a well-known NRF2 activator, which leads to the upregulation of enzymes such as HO-1. In studying the effects of ITA-NPs on ABCA1, we focused on the HO-1-calpain axis and demonstrated that this pathway is indeed a part of the ABCA1 stabilization effect. There may be other pathways affected by NRF2 that may also contribute to ABCA1 stability that were not discussed here. For example, 4-OI was shown via iTalk probe to interact with the protein ubiquitination pathway [20].

In addition to examining the ABCA1 ubiquitination and actions of ITA on HO-1 and calpain, we expanded our study to include other selected proteases that have also been implicated in ABCA1 degradation (Fig. 4g). The use of cycloheximide (CHX), an inhibitor of protein synthesis [71,72], in the presence of acLDL allowed us to focus on ABCA1 degradation rather than synthesis. This is an important consideration because acLDL upregulates ABCA1 synthesis and ITA-LNPs were shown to not affect the levels of ABCA1 mRNA transcripts (Fig. 3c) [3,73,74]. The use of chloroquine (CHQ), a lysosome inhibitor [75,76], demonstrated that the ITA must be released from particles via lysosomal degradation to cause ABCA1 stabilization. This is in line with previous studies showing that ITA-LNPs, unlike other ITA derivatives, can induce ITA-specific effects such as SDH inhibition due to its release of ITA intracellularly [3,61]. Pepstatin A (PepA) achieved similar levels of ABCA1 degradation as CHQ, likely because PepA inhibits acid enzymes such as those found in lysosomes [77,78]. Finally, MG132, which inhibits ubiquitin-proteasome pathways [79], stabilized ABCA1 on its own, suggesting that ABCA1 levels were even further stabilized in the presence of ITA. Notably, ABCA1 was degraded in Ctrl-LNP-treated groups in the presence of MG132 and CHQ but not solely MG132.

5. Limitations

Calpains have more than 10 isoforms; we did not investigate which were specifically affected by ITA-LNPs. This is a subject of future investigation.

In addition to direct interactions, the regulation of calpain by HO-1 may also involve downstream signaling pathways. The protective effects of HO-1 against oxidative stress could, therefore, be partly attributed to its ability to inhibit calpain activation, thereby preventing the proteolytic cleavage of critical cellular substrates that would otherwise lead to cell death.

We did not study other ITA derivatives, such as 4-OI. However, ITA-LNPs have been shown to have more similar biological effects akin to ITA itself. Additionally, ITA-LNPs specifically target plaque and plaque macrophages.

While we have previously shown that ITA-LNPs release itaconate in BMDMs and that these particles accumulate in plaque myeloid cells in vivo, it is unclear whether all of the observed effects are from unconjugated ITA released from ITA-LNPs or with additional effects from the ITA-LNP lipid conjugate itself. This is a subject of future investigation.

We did not evaluate the mechanisms discovered in vitro in in vivo settings. Although in vitro models may not fully replicate the complex interactions and microenvironment found in vivo, cell culture models using BMDMs are well-established and reproducible, and allow for precise control over experiments, which is critical for dissecting specific molecular mechanisms.

6. Conclusions

In summary, ITA-LNPs have demonstrated that ITA not only has an immunomodulatory role in atherosclerosis, but also a lipid regulation role. Because ITA-LNPs shifted macrophages to anti-inflammatory and lipid effluxing and metabolic pathways, we posited that ITA-LNPs are likely to have significant effects on atherosclerosis through multiple mechanisms.

HIGHLIGHTS.

• ITA-LNPs significantly inhibit plaque growth.

• ITA-LNPs stabilize ABCA1 in atherosclerosis and promote a cholesterol efflux phenotype.

• ITA-LNPs promote lipid metabolism via citrate shuttling and fatty acid metabolism.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.atherosclerosis.2025.120445.