ACLY as a modulator of liver cell functions and its role in Metabolic Dysfunction-Associated Steatohepatitis

Primary Human Hepatocytes (HH, Lonza, Walkersville, MD, USA) were cultured in Hepatocyte Culture Medium (Lonza) according to the manufacturer’s protocol. HepG2 obtained from ATCC (Sigma-Aldrich, St Louis, MO, USA) were grown in CO₂ (5%) incubator at 37 °C in high glucose Dulbecco’s modified Eagle’s medium (DMEM) GlutaMAX Supplement (Thermo Fisher Scientific, San Jose, CA, USA) after addition of 10% (v/v) fetal bovine serum, 100 U penicillin, and 100 μg/mL streptomycin. Human cell lines were tested periodically for mycoplasma by using MycoAlert PLUS detection kit (Lonza). Cells were treated with 500 µM hydroxycitrate (HCA, Sigma-Aldrich) or 200 µg/mL red wine powder (RWP) obtained as previously described from the Italian red wine Aglianico del Vulture [21]. After 1 h of treatment with HCA or RWP, cells were stimulated with 5 ng/mL of TNFα (Sigma-Aldrich) for up to 24 h. Where indicated, cells were treated also with 20 μM IKK inhibitor VII (IKK 16, Sigma-Aldrich), 5 mM sodium acetate (Ac, Sigma-Aldrich) or 5 mM sodium malate (M, Sigma-Aldrich) alone or in combination with 500 μM NADPH (Sigma-Aldrich).

Oil Red O staining was used for the detection of neutral lipids. Stock solution was obtained by dissolving 0.06 g Oil Red O (Sigma-Aldrich) in 20 mL isopropanol 100% and left undisturbed at room temperature for 20 min. Oil Red O working solution was prepared by diluting 3 parts of stock solution with 2 parts of distilled water. Working solution, stable for 2 h, was filtered immediately before use. For qualitative analysis, HH were seeded (1 × 10⁵ cells/well) in a 24-well plate and, the day after, treated with TNFα 5 ng/mL in the presence or not of 500 µM HCA or 200 µg/mL RWP. Twenty-four hours later, HH were washed twice with phosphate buffered saline (PBS) and fixed with 4% paraformaldehyde in PBS for 30 min at room temperature. Following two washing with distilled water, isopropanol 60% was added for 5 min. The cells were incubated with Oil Red O working solution for 20 min at room temperature. Then, HH were thoroughly washed three times with distilled water to remove the unbound staining solution and observed under fluorescence microscopy FLoid Cell™ Imaging Station (Thermo Fisher Scientific).

For quantitative analysis, after the incubation with Oil Red O working solution, the staining was extracted in isopropanol 100%, and lipid accumulation was quantified at 490 nm by using a microplate reader (GloMax® Discover Microplate Reader (Promega). The results are shown as percentage of the control (untreated cells), defined as 100% of neutral fats accumulation.

Anonymized human whole venous blood samples from 8 patients with MASH were obtained from the Department of Infectious Disease of San Carlo Hospital in Potenza, Italy. The same number of healthy controls were enrolled among hospital staff and volunteers. Samples were collected from May 2022 to March 2023. The diagnosis of Hepatic steatosis was based on Hepatic Steatosis Index (HSI). This score is founded on the use of ultrasound in order to show hepatic steatosis and on the presence of any of the three conditions of overweight/obesity, diabetes, and increased transaminases [22]. Experienced ultrasound physician blinded to the study further classified it as [1] mild steatosis (presence of diffuse echogenic enhancement or hepatorenal contrast) or [2] moderate or severe steatosis (both bright echogenic and hepatorenal contrast enhancement visible or ultrasound beam attenuation observed). All subjects provided written, informed consent, approving, and authorizing the use of their material for research purposes. Research was carried out in accordance with the Declaration of Helsinki and in agreement with local Italian Committee on Human Research’s approved procedures (REF. TS/CEUR 20200034750—15 September 2020).

Venous blood was collected into K2 EDTA-coated BD vacutainer tubes (Becton, Dickinson and Company, Franklin Lakes, NJ, USA). Peripheral blood mononuclear cells (PBMCs) were isolated using Histopaque-1077 (Sigma-Aldrich) density gradient centrifugation and human monocytes were obtained from PBMCs following incubation with CD14 antibody conjugated to magnetic beads (MACS®, Miltenyi Biotec GmbH, Bergisch Gladbach, Germany) as previously described [23]. The CD14⁺ monocytes were differentiated to macrophages by culturing in Roswell Park Memorial Institute (RPMI) 1640 medium (Thermo Fisher Scientific) supplemented with 10% fetal bovine serum, 2 mM l-glutamine, 100 U/mL penicillin, 100 μg/mL streptomycin, and 100 ng/mL recombinant human Macrophage Colony Stimulating Factor (M-CSF, Cell Guidance Systems, St. Louis, MO, USA) for 3 days at 37 °C in a humidified atmosphere of 5% CO₂.

Total RNA was extracted from 2 × 10⁶ cells by RNeasy Plus Mini Kit (Qiagen, Hilden, Germany) as per manufacturer’s instructions. Complementary DNA was synthesized from 1 μg of RNA by iScript™ cDNA Synthesis Kit (Bio-Rad Laboratories, Hercules, CA, USA) (5 min at 25 °C, 20 min at 46 °C and 1 min at 95 °C) as per manufacturer’s guidelines. Real-time PCR experiments were performed in triplicate on the 7500 Fast Real-Time PCR System (Thermo Fisher Scientific) with human ACLY (Hs00982738, RefSeq NM_001096.2), ME1 (Hs00159110, RefSeq NM_002395.5) and β-actin (Hs01060665, RefSeq NM_001101.3) TaqMan Gene Expression Assays (Thermo Fisher Scientific). Data were analyzed according to the ΔΔCt method, as previously reported [24]. In detail β-actin was used as endogenous reference gene to obtain ΔCt value by subtracting the β-actin Ct value from the target gene (ACLY or ME1) Ct value, where Ct refers to the threshold cycle. The fold changes in the ACLY or ME1 expression in treated cells or patients relative to unstimulated cells or healthy controls were estimated by ΔΔCt method [25] by using the ΔCt mean values for calculations, unless otherwise specified. Fold changes were determined as 2^−ΔΔCt.

The lysate obtained from 1 × 10⁶ cells as previously reported [26] was subjected to a Bradford assay (Pierce™ Coomassie (Bradford) Protein Assay Kit (Thermo Fisher Scientific) to determine protein concentration using Bovine Serum Albumin as standard. Thirty micrograms of proteins were resolved by 8–12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto nitrocellulose membranes. The membranes were blocked with a tris-buffered saline solution containing 5% non-fat dry milk and 0.5% Tween for 1 h at room temperature. Subsequently, membranes were probed overnight at 4 °C with anti-ACLY (ab157098, Abcam, Cambridge, MA), anti-ME1 (ab97445, Abcam), anti-NF-κB/p65 (ab16502, Abcam) or anti-β-actin (ab8227, Abcam) primary antibodies. The membranes were then incubated with horseradish peroxidase (HRP) conjugated goat anti-rabbit secondary antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA) for 1 h at room temperature. The immunoreactions were detected by WesternBright™ ECL (Advansta, Menlo Park, CA, USA) at Chemidoc™ XRS detection system (Bio-Rad Laboratories). Image Lab Software (Bio-Rad Laboratories) was used for image acquisition and densitometric analysis.

HH (1 × 10⁷) were treated for 3 h with 5 ng/mL TNFα, collected, and washed twice in ice-cold PBS. ACLY activity was assessed as previously reported [23] on the cell lysate obtained after three freeze-melt cycles (− 80 °C for 8 min/40 °C for 4 min) in ice-cold 0.1% Nonidet P40 (NP40)/PBS solution. The specific ACLY activity was normalized to the protein concentration and expressed as a percentage of the control (set at 100%).

To detect reactive oxygen species (ROS) levels, HH (5 × 10⁵) were triggered by 5 ng/mL TNFα for 24 h in the presence or absence of 500 µM HCA or 200 µg/mL RWP. Where indicated, cells were co-treated with 5 mM Ac or 5 mM M alone or in combination with 500 μM NADPH. After 24 h, ROS concentrations were measured by using Di(Acetoxymethyl Ester) (6-Carboxy-2′,7′-Dichlorodihydrofluorescein Diacetate) (Thermo Fisher Scientific) as described in [27].

Lipid peroxidation was monitored by the Lipid Peroxidation (MDA) Colorimetric/Fluorometric Assay Kit (BioVision, Milpitas, CA, USA) as previously reported [28]. HH (1 × 10⁷) were triggered by TNFα in the presence or not of HCA or RWP. In some experiments, cells were also treated with Ac or M alone or in combination with NADPH. After 24 h treatment, cells were collected and homogenized on ice in 300 μL of Malondialdehyde (MDA) Lysis Buffer. The supernatants, collected by centrifugation, were placed in microcentrifuge tubes to which Thiobarbituric Acid (TBA) reagent was added. After the incubation at 95 °C for 1 h and cooling at room temperature in an ice bath for 10 min, 200 μL of each sample, run in triplicate, was added to a 96-well plate. The MDA-TBA adduct was quantified fluorometrically (Ex/Em = 532/553 nm) by using GloMax® Discover Microplate Reader (Promega). The MDA amount was further calculated according to the kit manufacturer’s protocol.

The levels of oxidized glutathione (GSSG) in human hepatocytes were evaluated by Glutathione Fluorometric Assay Kit (BioVision) at the end of 24 h treatment with TNFα in the presence or not of HCA or RWP. Cells (4 × 10⁶) were collected and homogenized on ice with 100 μL of ice-cold Glutathione Assay Buffer. Sixty μL of each homogenate was added in a prechilled tube containing perchloric acid (PCA) and vortexed for a few seconds to obtain a uniform emulsion. Supernatants were collected by centrifugation and neutralized by adding ice-cold 6N Potassium hydroxide (KOH) to precipitate PCA. To 10 µL of neutralized samples transferred into a 96-well plate, Assay Buffer and GSH Quencher were added. Following 10 min incubation at room temperature, Reducing Agent Mix was added to destroy the excess GSH Quencher and convert GSSG to GSH (reduced glutathione). After 40 min of incubation with OPA Probe, the fluorescence of each well was measured with GloMax® Discover Microplate Reader (Promega) by the mean of Fluorescence Excitation Modules and Emission Filters for UV (365nm_Ex/415–445nm_Em). The GSSG levels were further calculated according to the kit manufacturer's protocol and expressed as percentage of control.

For monitoring the promoter activity of ACLY gene, transient transfection was achieved using pGL3 basic-LUC vector (Promega) containing the − 3116/ − 20 bp region of the ACLY gene promoter (called “3000” and including the binding site for NF-κB) or a deletion fragment of this region (without NF-κB binding site, indicated as “1000”), and with 10 ng of pRL-CMV (Promega) to normalize the extent of transfection. The transfection was performed when the cells became 50% to 70% confluent, using Lipofectamine 3000 (Thermo Fisher Scientific) at a ratio of ~ 3.5 μg:5 μL/well (DNA: Lipofectamine). Briefly, HepG2 cells were seeded into a 24-well plate (5 × 10⁴ cells/well). The day after, the DNA-lipofectamine complex was made by mixing 25 μL of Opti-MEM Reduced Serum Medium containing 0.75 μL of lipofectamine 3000 and 25 μL of the medium containing 0.5 μg of DNA. The DNA-lipofectamine mixture was incubated for 15 min at room temperature. Finally, 50 μL of the DNA-lipofectamine mixture was added to the well containing the cells. Twenty-four hours after transfection, HepG2 cells were triggered by TNFα in the presence or absence of HCA. The next day, cells were lysed and assayed for LUC activity by using the Dual-Luciferase® Reporter Assay System (Promega), according to the manufacturer’s protocol. Luminescence was measured on GloMax® Discover Microplate Reader from Promega.

For chromatin immunoprecipitation (ChIP) experiments, HH cells (5 × 10⁶) were treated with TNFα in the presence or absence of HCA for 3 h, then fixed by 1% formaldehyde at 37 °C for 10 min. Subsequently, HH cells were lysed in a 0.1% NP40/PBS buffer supplemented with protease inhibitors, put on ice for 10 min, and centrifuged at 1000×g for 2 min at 4 °C. Pellet resuspension in a 0.1% NP40/PBS buffer was followed by sonication at 70% power at cycle 9 for 10 min in a 0.1% SDS lysis buffer to generate cellular chromatin fragments of about 400 bp. Chromatin immunoprecipitation was performed overnight at 4 °C on a rocking platform using A/G PLUS agarose beads (Santa Cruz) with anti-NF-κB/p65 antibody (ab16502, Abcam). The following day, all samples, including the input (total chromatin extract) and mock (immunoprecipitation without the antibody), were recovered. The protein/DNA complexes were washed with PBS, then treated with RNase and Protease K. DNA was purified by means of PureLink™ PCR Purification Kit (Thermo Fisher Scientific) and analyzed by qPCR using the following primers: For-ACLY 5′-CTTTCCAAAGTTGGGTCTTGTG-3′ and Rev-ACLY 5′-CCTCAGCAATTCAGACTCCTT-3′. Each qPCR reaction mixture contained SYBR™ Green PCR Master Mix (Thermo Fisher Scientific), 20-pmol forward and reverse oligonucleotides, 200 ng DNA, and nuclease-free water was run on a 7500 Fast Real-Time PCR System (Thermo Fisher Scientific). Melting curve analysis was performed to verify the specificity of designed primers.

HH (5 × 10⁵), following 3 h treatment with 5 ng/mL TNFα in the presence or absence of 200 µg/mL RWP, were subjected to immunocytochemistry experiments by using anti-NF-κB/p65 (ab7970, Abcam) as primary antibody and Alexa Fluor 488 (Thermo Fisher Scientific) as a secondary antibody according to the protocol described in [21]. Nuclear localization was calculated as % Nuclear according to the equation reported in [29] : % Nuclear =  [Total Nuclear Intensity/(Total Cytoplasmic Intensity + Total Nuclear Intensity)] × 100. Total fluorescence intensities were determined by using ImageJ software (https://​imagej.​nih.​gov/​ij/​index.​html, NIH, accessed on 12 July 2023).

The effects of TNFα alone or in combination with HCA on H3 and H4 histone acetylation were evaluated with the EpiQuik™ Global Histone H3 Acetylation Assay Kit and EpiQuik™ Global Histone H4 Acetylation Assay Kit (Epigentek, Farmingdale, NY, USA). Following the manufacturer’s instructions, histones were extracted from HH that had been treated for 24 h with 5 ng/mL TNFα alone or combined with 500 μM HCA. The histones were spotted onto the wells, and then an antibody specific for acetylated histone H3 or H4 was added. After washing the wells, HRP-conjugated secondary antibody was added, followed by the detection reagent, and absorbance was measured with GloMax® Discover Microplate Reader at 450 nm. Percent acetylation was calculated as OD (treated sample − blank)/OD (untreated control − blank) × 100%.

Twenty-four hours after stimulation with TNFα in the presence or not of HCA and/or RWP, cell-free supernatants were collected and assayed for the concentration of interleukins 1β and 6 by the mean of Lumit™ IL-6 (Human) Immunoassay and Lumit™ IL-1β (Human) Immunoassay (Promega) following the manufacturer’s recommendations.

The statistical analysis of data was carried out employing the statistic tools implemented in the GraphPad Prism software (La Jolla, CA, USA). All results are presented as mean ± standard deviation (SD) from at least three independent experiments, each run in triplicate. For pairwise comparisons, Student’s t-test test was performed. Comparisons of more than two groups were made using one-way ANOVA followed by Tukey’s post hoc test. The statistical method used for each experiment is detailed in the figure legends. The asterisks in the figures denote statistical significance (*p < 0.05; **p < 0.01; and ***p < 0.001). When Tukey’s test was performed, different letters indicated significant differences between treatments at p < 0.05.

Circulating TNFα shows increased levels in animal model with MASH [30] as well as in patients with MAFLD compared to control subjects, along with different disease markers. Moreover, a high amount of circulating TNFα has been linked to MAFLD severity [31] while TNFα signaling inhibition lowers liver steatosis and hepatocellular injury in MAFLD mouse model. It is well known that treatment with TNFα promotes inflammation, lipid accumulation (a hallmark of MASH [32])—including fatty acid and cholesterol—and injury in hepatic cells [33‐35]. Therefore, we used TNFα to trigger liver damage pathways and in turn to explore ACLY function.

First of all, we tested the effect of TNFα on the expression levels of ACLY, the key metabolic enzyme in fatty acid metabolism as well as in oxidative stress and inflammation, both peculiar features of MASH, in primary human hepatocytes. Following 3 h stimulation with TNFα, we observed increases in ACLY mRNA (1.45 ± 0.01-fold change, Fig. 1A), protein levels (1.87, Fig. 1B) and enzymatic activity (128.35 ± 21.06, Fig. 1C) compared to untreated cells.

In the same condition, we recorded a double amount of cytosolic ME1 mRNA (Fig. 1D) and a 4.4-increment of protein (Fig. 1E) in TNFα-triggered HH cells than control. ME1 participates in the recycle of ACLY-produced OAA from cytosol to mitochondria with concomitant formation of NADPH reducing equivalents, essential to both fatty acid and inflammatory mediator biosynthesis.

In light of these results, we investigated the effect of HCA, a potent competitive inhibitor of ACLY, on lipid accumulation. Lipid content was assessed by using Oil Red O staining in HH treated with TNFα, individually or in combination with HCA. The concentration of HCA used was the same used in [36], that did not affect HH cell viability measured as described in Additional file 1  (Additional file 2: Fig. S1A). In comparison to the control, treatment with TNFα increased 2.2-fold the lipid accumulation (p < 0.001) (Fig. 1F, G). Instead, the fat droplet content was radically reduced to levels lower than the control ones by the combination treatment of TNFα and HCA (Fig. 1F, G). These findings suggest that ACLY and ME1 are overexpressed and the specific ACLY inhibition by HCA reduces fat droplet accumulation in TNFα-triggered HH.

Increased oxidative stress plays a central role in MAFLD pathogenesis and in its progression to MASH [17]. Indeed, MASH development is characterized by an increase in lipid peroxidation and ROS levels, together with a decrease in antioxidant defenses, GSH amount, and an increment in GSSG levels [37].

Although some of the mechanisms responsible for altered redox status are known, it cannot be ruled out that other processes, currently unknown, may contribute to it. In keeping the activation of ACLY (Fig. 1A–C) and its involvement, as part of the citrate pathway, in oxygen radical production [14] we assessed ROS levels in TNFα-triggered HH in the presence of HCA. As expected, TNFα caused an enhanced and significant release of ROS (mean ± SD = 132.5 ± 4.8, Fig. 2A) in comparison to control (mean ± SD = 100 ± 5.9). HCA, on the other hand, reduced by 35% the levels of ROS compared to cells treated only with TNFα (Fig. 2A). The addiction of exogenous acetate, which can provide acetyl-CoA (a product of ACLY activity) did not revert this decrease of ROS levels (Fig. 2A). On the contrary, the addiction of exogenous sodium malate alone or combined with NADPH, two players of the OAA recycle and produced downstream of ACLY, reverted the effect of ACLY inhibition. In fact, the malate alone led to a significant increase (p < 0.001) of about 30% in ROS levels compared to cells treated with TNFα and HCA and even a more marked increase, greater than 40%, when NADPH was also present (TNFα + HCA = 87.7 ± 6.9; TNFα + HCA + M + NADPH = 123.5 ± 6.0, Fig. 2A). Among the different consequences of oxidative stress with increase of ROS levels is the lipid peroxidation of membrane phospholipids. The peroxidation of arachidonic acid produces different peroxides and finally toxic aldehydes as well as MDA. The latter reacts with amino groups of proteins and DNA by forming adducts that can alter the function of these macromolecules in cells [38]. Therefore, we evaluated the outcome of ACLY inhibition by HCA on MDA levels. A significant increase of about 80% in lipid peroxidation was observed in TNFα-triggered HH compared to untreated cells while the addition of HCA drastically lowered the levels bringing them back to values similar to the control ones (Fig. 2B). As observed for ROS, the addition of sodium acetate had no effect; on the contrary, sodium malate alone or plus NADPH led to strong increase (of about 45% and 52%, respectively) in MDA levels compared to cells treated with TNFα and HCA (Fig. 2B). As expected, similar results of ROS and MDA reductions were obtained when ACLY was silenced - as described in Additional file 1 - by a specific siRNA targeting human ACLY (siACLY) (Additional file 2: Fig. S2). In detail, we recorded significant decreases (p < 0.001) of 30 and 25% in the content of ROS and MDA, respectively, in the presence of siACLY compared to cells treated only with TNFα (Additional file 2: Fig. S2).

GSSG levels increased by 65% in HH cells treated with the pro-inflammatory cytokine TNFα compared to those of the untreated cells, highlighting the oxidative stress condition. It is interesting to note that GSSG levels were significantly reduced by about 30% in HH cells pretreated with HCA and exposed to TNFα for 24 h (Fig. 2C). These findings confirm the alteration in glutathione homeostasis, observed in MASH where the hepatic accumulation of GSSG has been related to the regulation of oxidative stress effects, suggesting a synergistic effect of TNFα and GSSG levels in the induction of hepatocyte damage [37].

These outcomes, running parallel to ACLY and ME1 activation, indicate that the blocking of ACLY, through inhibition or gene silencing, may improve the redox status of hepatocytes.

To further confirm the central role of ACLY in liver function, we tested the effect of ACLY inhibition by RWP, deriving from Aglianico del Vulture red wine, in our in vitro model. Notably, RWP restores LPS-triggered macrophage homeostasis via inhibition of both NF-kB and the citrate pathway [21]. After verifying that RWP did not affect HH viability (Additional file 2: Fig. S1B), we treated TNFα-activated HH with 200 μg/mL RWP, and RT-qPCR experiments evidenced a 35% reduction in ACLY mRNA (Fig. 3A) and a halving of ME1 gene expression levels (Fig. 3B). A similar trend of more marked reduction effect of ME1 than of ACLY by RWP was observed on protein expression: ACLY was reduced by about 20% (Fig. 3C) while ME1 by ~ 75% (Fig. 3D). Similarly to HCA, RWP drastically reduced lipid content in TNFα-activated HH restoring levels to values comparable to untreated cells (Additional file 2: Fig. S3A, B). Our analysis showed that RWP was able, like HCA, to improve the cellular redox status leading to a decrease in the levels of ROS (Fig. 3E), MDA (Fig. 3F) and GSSG (Fig. 3G) with respect to cells treated with TNFα alone. Once again, the modulation of ROS and lipid peroxidation occurred through ACLY since the addition of M, alone or plus NADPH, led to increased levels of both ROS and MDA (Additional file 2: Fig. S3C, D). Conversely, sodium acetate, which supplies acetyl units, i.e. the other metabolite produced by ACLY, had no effect (Additional file 2: Fig. S3C, D). Thus, the significant role played by RWP in restoring liver homeostasis is mainly dependent on ACLY.

It is well known that TNFα modulates various inflammatory and immune responses through the activation of NF-kB [39]. As previously reported, TNFα treatment upregulates ACLY through NF-kB in macrophages [36]. More recently, we demonstrated that ACLY-mediated p65 acetylation is required for NF-kB full activation and in turn for ACLY transcriptional upregulation in LPS-triggered macrophages [16]. Therefore, we tested if ACLY overexpression was linked to its own activity and depended on NF-kB in TNFα induced hepatocytes. To this aim, liver cells were transiently transfected with 3000 (a pGL3 basic-LUC vector containing the − 3116/− 20 bp region of the ACLY gene promoter including the NF-κB response element localized at − 2048/− 2038 bp) or 1000 (a truncated version without the NF-κB response element) (Fig. 4A). After 48 h, cells were triggered by TNFα in the presence or absence of HCA. Clearly, ACLY gene promoter activity was drastically smaller in cells transfected with 1000 than 3000 due to the lack of NF-κB response element (Fig. 4A). As expected, TNFα increased 2.8-fold luciferase activity in liver cells transfected with 3000 (Fig. 4A). HCA significantly reduced by 40% the ACLY promoter activity in comparison to TNFα activated cells (3000 + TNFα = 278.6 ± 26.7; 3000 + TNFα + HCA = 165.3 ± 14.5, Fig. 4A). IKK 16, an inhibitor of NF-kB signaling, specifically blocks NF-kB activation [40]. In liver cells transfected by 3000 and triggered by TNFα, IKK 16 halved the luciferase activity (Additional file 2: Fig. S4), confirming that ACLY gene upregulation is under NF-kB control. Moreover, ChIP with a specific antibody against p65 revealed a strong binding of NF-κB to the human ACLY gene promoter upon TNFα treatment of HH which was significantly reduced in the presence of HCA (Fig. 4B).

Next, we analyzed protein expression and cellular localization of NF-kB in TNFα-triggered HH in the presence or in the absence of RWP. We observed a reduction of NF-kB p65 subunit protein levels by slightly more than twofold when RWP was used compared to TNFα-triggered HH (Fig. 5A). After 3 h of treatment with TNFα, the subunit p65 of NF-κB translocated from the cytosol to the nucleus, indeed % Nuclear rose by 26% with respect to unstimulated HH (C, Fig. 5B). Instead, RWP restored % Nuclear to values comparable to C (Fig. 5B).

Our data indicate that ACLY transcription is mediated by NF-κB which in turn requires ACLY activity for its full activation.

ACLY represents the major source of acetyl-CoA in mammals and is essential for increasing histone acetylation [41] which controls the accessibility to chromatin and gene transcription. Upon TNFα stimulation the levels of global acetylation of both histone H3 and histone H4 significantly increased compared to control HH cells (Fig. 6A, B). Treatment with HCA resulted in significant reductions by 25% in acetylated H3 (Fig. 6A) and by 35% in acetylated H4 (Fig. 6B). Taking into consideration that: (a) ACLY mediates NF-kB full activation and is under the transcriptional control of NF-kB; (b) the activity of transcription factors depends on the state of the chromatin and therefore on its accessibility; (c) NF-kB fosters the transcription of several proinflammatory genes; (d) HCA determines the inhibition of NF-kB and the reduction of histone acetylation, we decided to analyze the levels of proinflammatory cytokines IL-6 and IL-1β, which are both NF-kB target genes. The treatment with TNFα alone resulted in an approximately sixfold increase in both IL-6 (Fig. 6C) and IL-1β (Fig. 6D) secretion. In the presence of HCA, IL-6 levels were decreased by about 30% (Fig. 6C), those of IL-1β by 37% (Fig. 6D) compared to TNFα-activated hepatocytes. Interestingly, RWP, for which we have previously demonstrated a significant effect on gene expression reprogramming through the modulation of the histone acetylation [21], also affected the secretion of both IL-6 and IL-1β. As shown in Additional file 2: Fig. S5, IL-6 (A) as well as IL-1β (B) levels were lowered by about 50% after addition of RWP when compared to HH treated with TNFα alone. These results confirm the central role of ACLY in modulation of proinflammatory genes encoding IL-6 and IL-1β by promoting histone acetylation also in human hepatocytes.

In light of the results described above and previously reported about increased levels of ACLY in liver of patients with MASH [20], we investigated whether these dysregulations were also present at a systemic level. Therefore, we evaluated ACLY and ME1 gene expressions in macrophages from MASH patients and healthy controls. ACLY resulted overexpressed in MASH patients: the mean values of ACLY mRNA were ~ 1.4-fold (mean ± SD: 1.36 ± 0.35) greater than in control subjects (mean ± SD: 0.98 ± 0.06) (Fig. 7A). The interquartile range (IQR) in MASH patients, equal to 0.55, suggested that 50% of NASH patients enrolled had ACLY mRNA from 1.14 to 1.69-fold changes of healthy subjects (Fig. 7A). We observed a 4.5-fold increase in ME1 mRNA in MASH patients (mean ± SD: 5.68 ± 3.74, median 4.47) compared to controls (mean ± SD: 1.16 ± 0.45, median: 1.03) (Fig. 7B). Noteworthy, parallel to what happens in hepatocytes, both ACLY and ME1 are overexpressed and ME1 is more markedly expressed than ACLY.

Once we confirmed the upregulation of both ME1 and ACLY, we decided to treat macrophages differentiated PBMCs from MASH patients with HCA + RWP to evaluate the effects on both the expression of these genes and specific markers of MASH. In RT-qPCR experiments, for the calculation of ΔΔCt, we subtracted the Ct of the untreated cells (UNTD) from the Ct of UNTD and HCA + RWP treated PBMCs. Consequently, the 2^ΔΔCt of UNTD was always equal to 1. As expected, the co-treatment with HCA and RWP significantly (p < 0.05) reduced by ~ 30% ACLY mRNA (mean ± SD: 0.72 ± 0.33) with a range spreading from 0.35 to 1.25-fold changes of control (Fig. 8A). ME1 mRNA content was reduced by even more than half (mean ± SD: 0.43 ± 0.29) when cells were treated with HCA + RWP (Fig. 8B). Except for one patient in whom ME1 was unaffected by HCA + RWP treatment, all the others had ME1 mRNA lower than UNTD (Fig. 8B). Finally, we focused on pro-inflammatory cytokines IL-6 and IL-1β, which represent MASH biomarkers [42, 43]. At the end of 24 h treatment, the cell culture supernatants were collected from UNTD and HCA + RWP cells. We observed significant reductions (p < 0.05) in the secretion of both cytokines following treatment: IL-6 concentration decreased from 302.8 (± 69.9) in UNTD to 221.2 (± 63.4) pg/mL in HCA + RWP treated cells (Fig. 8C); IL-1β levels reduced nearly by the half from 36.9 (± 17.8) in UNTD to 20.4 (± 8.7) pg/mL in HCA + RWP treated cells (Fig. 8D). These results strengthen the role of both ACLY and ME1 in MASH and open glimmers on the possibility of using them as biomarkers as well as drug targets for MASH.