Chemoprevention of Hepatocellular Carcinoma by Next-Generation Antipsychotic Aripiprazole

Abstract

Hepatocellular carcinoma (HCC) is the most common form of liver cancer and is a major global health burden, ranking sixth in incidence and third in cancer-related mortality. Despite therapeutic advances, treatment options for advanced liver disease and HCC are limited and strategies to prevent HCC development are lacking. To address the urgent need for preventive strategies, we identified aripiprazole, an oral atypical antipsychotic, as a candidate for HCC chemoprevention. Analyses of clinical liver tissues showed that aripiprazole targets are expressed in different liver cell compartments including fibroblasts, macrophages and epithelial cancer cells, and that target gene expression is associated with fibrotic liver diseases and HCC. In a rat model of MASH-induced HCC induced by choline-deficient L-amino acid-defined and high-fat diet, aripiprazole prevents liver disease progression and HCC development by modulating fibrogenesis related pathways and inflammation. Mechanistically, aripiprazole exerts antifibrogenic and anti-inflammatory effects by modulating the phenotype of liver fibroblasts and macrophages. Moreover, perturbation studies in cancer cell models showed that aripiprazole prevents tumor initiation and reduces cell proliferation via inhibition of the cMET and ERK pathways and perturbation of mitochondrial functions. Finally, treatment of patient-derived tumorspheroids demonstrated that aripiprazole modulates immune responses in the tumor microenvironment. Collectively, these findings suggest that treatment with aripiprazole is a clinically relevant approach for HCC chemoprevention.

Introduction

Liver cancer ranks as sixth in incidence with almost nine hundred thousand cases globally, while it’s at the third place regarding mortality(1). While liver cancer incidence rate is predicted to double by 2050, 60% of cases are preventable(2). Hepatocellular carcinoma (HCC) is the most common type of liver cancer, accounting for 80% of cases. A large majority of HCC cases arise on a fibrotic or cirrhotic liver, highlighting fibrosis as a critical predisposing factor regardless of the underlying etiology. Despite the diverse causes of chronic liver disease, the different etiologies share common patterns of disease progression. These include chronic inflammation, fibrosis progression, and ultimately malignant transformation. Metabolic dysfunction-associated steatohepatitis (MASH), alcoholic liver disease and chronic viral hepatitis are major causes of HCC(3). While lifestyle and dietary changes are primary prevention options for MASH-related HCC, in clinical practice these options are rarely successful due to difficulties in execution.

Chemoprevention aims to prevent cancer occurrence or recurrence in exposed patients through medical interventions or medication(4). However, the preclinical and clinical development of chemopreventive agents has been slow due to several challenges. These include limited understanding of hepatocarcinogenesis, limited translatability of animal models to patients, scarce liver biospecimens for translational research and validation, and the high cost and duration of clinical trials(5). While several candidate compounds for chemoprevention have been identified (e. g. statins, histamine receptor 2 antagonists, aspirin), their clinical validation in randomized clinical trials remains to be determined(4).

Serotonin signaling plays a functional role in metabolic dysfunction(6), steatosis(7), inflammation(8) and fibrosis(9). Aripiprazole is an FDA-approved third generation (or atypical) antipsychotic, which is used to treat mental health conditions including schizophrenia and bipolar disorders by modulating multiple serotonin receptor signaling pathways(10). Depending on the context, aripiprazole can act as an agonist, partial agonist and an antagonist at 5-hydroxytryptamine receptor 2A (HTR2A), 5-hydroxytryptamine receptor 2B (HTR2B), 5-hydroxytryptamine receptor 1A (HTR1A), 5-hydroxytryptamine receptor 7 (HTR7) and others(10). In this study, we investigated the potential of aripiprazole as an HCC chemoprevention in the setting of MASH by utilizing vivo and patient-derived cell-based models.

Experimental procedures

Human subjects

Human liver tissues were obtained from liver disease patients undergoing liver resection with informed consent from all patients for deidentified use at the Center for Digestive and Liver Disease of the Strasbourg University Hospitals University of Strasbourg, France (DC-2016-2616 and RIPH2 LivMod IDRCB 2019-A00738-49, ClinicalTrial NCT04690972). The protocols were approved by the local Ethics Committee of the University of Strasbourg Hospitals. All material was collected during a medical procedure strictly performed within the frame of the medical treatment of the patient. Informed consent is provided according to the Declaration of Helsinki. Detailed patient information and informed consent procedures are implemented by the Strasbourg University Hospital Biological Resources Center (HUS CRB). Patients were given an information sheet, which outlines that their leftover biological material (liver resection and blood samples) that was collected during their medical treatment is requested for research purposes. All patients received and signed an informed consent form to provide authorization or refuse the use of their biological samples (protocols DC-2016-2616 and RIPH2 LivMod IDRCB 2019-A00738-49 ClinicalTrial NCT04690972). The patients maintain the right to withdraw their consent at any time and to request the destruction of their biological material, which is strictly respected. While there was clinical descriptive data available, the identity of the patients was protected by internal coding. Summary of patient characteristics is provided in Supplementary Table 1.

Cells

Huh7 cells were a gift from G. Christofori (University of Basel, Switzerland). Huh7.5.1 were a gift from F. Chisari (The Scripps Research Institute, La Jolla, San Diego, California, USA). SNU-423 LX2 were purchased from Merck. THP1 cells were purchased from ATCC. Huh7.5.1 and LX2 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Thermo Fisher Scientific) supplemented with 10% heat-decomplemented fetal bovine serum (FBS), gentamycin (0.05 mg/mL), and nonessential amino acids (complete DMEM) at 37 °C with 5% CO2. Cell lines were certified mycoplasma free. For proliferation arrest and differentiation (Huh7.5.1dif cells), Huh7.5.1 cells were cultured in complete DMEM containing 1% dimethylsulphoxide (DMSO)(11). THP1 cells were cultured and differentiated in RPMI 1640 medium with GlutaMAX-I supplement and HEPES, and they were supplemented with 10% FBS and gentamycin (0.05 mg/mL) (Thermo Fisher Scientific). To generate M1-like THP-1–derived macrophages, cells were treated with PMA (320 nM) (Promega) for 24h and then 100 ng/ml lipopolysaccharides (LPS) (Merck) and 20 ng/mL interferon gamma (IFNγ) (PeproTech) for 24h. Hepatic stellate cells (HHStec) were purchased from ScienCell. Human liver myofibroblasts (HLMF) were isolated from liver tissues of patients who had undergone liver resection as described(12). HHSteC and HLMF were cultured in Stellate Cell Medium SteCM (ScienCell), renewed every four days, on type I collagen pre-coated plates. To induce pro-fibrotic gene expression, HHStec and HLMF were stimulated by TGFβ 10 ng/mL (PeproTech) for 24h.

Animal model

In the MASH model of HCC, five-week-old male and female Wistar rats (Charles River Laboratories) were fed either standard chow (16% protein rodent diet, Teklad, 2916) or choline-deficient, L-amino acid-defined, high-fat diet (CDA/HFD; consisting of L-amino acid with 60 kcal% fat and 0.1% methionine and no added choline, Research diet, A06071302) for 18 weeks. After 10 weeks on CDA/HFD, rats (n = 9/group) were randomly assigned to daily oral gavage of vehicle control (DMSO) or drug treatment group (2 mg/kg aripiprazole) to coincide with the onset of fibrosis. At the time of sacrifice, animals were anesthetized and sedated. A terminal blood collection was performed by cardiac puncture, and livers were removed for measurement of weight, snap frozen for analysis, or fixed in formalin for histology. All animal experiments were performed after prior approval, and under the ethical guidelines of University of Texas Southwestern Medical Center, Institutional Animal Care and Use Committee (IACUC), APN-2019-102651 and received human care.

Statistics and reproducibility

Cell-based experiments were reproduced at least 3 times (as indicated in figure legends) in an independent manner, in triplicate (unless otherwise stated). The precise number (n) of biologically independent samples used to derive statistics is indicated in the figure legends. The data are presented as the mean ± sd and were analyzed by appropriate tests as indicated in figure legends, after determination of distribution using normality tests and sedasticity test. P < 0.05 was considered statistically significant. Significant P values are indicated by asterisks in the individual figures. Statistical analyses were performed with GraphPad Prism 10 software. No statistical analyses were performed if n < 4. For in vivo experiments, the sample size estimate was based on a P value of 0.05 at 0% power assuming a 50% difference in means in tumor burden with 33% SD between control and treated animals. For the prognostic liver signature (PLS) assay, variation of the high- and the low-risk genes was determined by a normalized enrichment score (NES) obtained using gene set enrichment analysis (GSEA). Significance of the data was determined by the false discovery rate (FDR) values. According to GSEA (https://www.gsea-msigdb.org/gsea/index.jsp), results are significant if FDR < 0.25 for discovery and FDR < 0.05 for validation. Results are expressed as a heatmap (log 10 of FDR).

Data Availability Statement

The authors declare that the data supporting the findings of this study are available within this Article and its Supplementary Information. RNA-Seq data were deposited in the NCBI Gene Expression Omnibus database with the accession number: GSE315143. Further information and requests for resources and reagents should be directed to Prof Thomas Baumert, Prof Yujin Hoshida and Dr Emilie Crouchet.

Results

Aripiprazole targets are expressed in the liver and are associated with liver fibrosis and cancer in patients

First, we sought to evaluate the clinical relevance of aripiprazole targets in liver disease progression and hepatocarcinogenesis in multiple clinical cohorts. Thereto, we analyzed the expression of serotonin receptors in non-diseased liver tissues and in tissues from patients with chronic liver disease (CLD) across several independent cohorts (Figure 1, Supplementary Figure 1). In non-diseased liver tissues ((GSE192742)(13) and (GSE124395)(14)), HTR7 and HTR2B were expressed across various liver cell types, mainly in endothelial cells (HTR2B), stromal cells and macrophages (HTR7) (Figure 1A, Supplementary Figure 1A). In patients with metabolic dysfunction-associated steatotic liver disease (MASLD)-related fibrosis (GSE162694)(15), we observed an upregulation of HTR2B and HTR7 in advanced fibrosis (Figure 1B). Across additional MASLD cohorts ((GSE193066)(16), (GSE193080)(17) and (GSE291313)) higher HTR7 was associated with higher body mass index (BMI), elevated aspartate aminotransferase (AST), alanine aminotransferase (ALT), advanced fibrosis, and severe inflammation (Supplementary Figure 1B). In addition, hierarchical clustering using HTR7 separated patients into two groups: one cluster (18% of patients) was enriched for high HTR7 expression and associated with advanced fibrosis and severe inflammation, potentially defining a population amenable to aripiprazole (Figure 1C). This association suggests a functional link between the expression of aripiprazole targets and the progression of liver fibrosis. Target expression analyses in HCC datasets ((GSE94660)(18) and (GEPIA)(19)) revealed that HTR7 expression is increased in HCC (Figure 1D, Supplementary Figure 1C). At the single cell level, HTR2B and HTR7 are predominantly expressed in tumor-associated macrophages (TAMs), tumor endothelial cells (TECs), cancer-associated fibroblasts (CAFs), and malignant cells (GSE151530)(20). Of note, target expression in isolated patient-derived cells and cell-based systems showed that HTR2B but not HTR7 expression is increased in TAM-like cells compared to macrophages, in association with an increase in CD163 expression, a marker of poor-prognosis and immunosuppressor macrophages in patients(21) (Supplementary Figure 1D). However, target expression remained unchanged in CAFs compared to stellate cells (Supplementary Figure 1D).

Figure 1. Aripiprazole targets expression in healthy and diseased human liver.

(A) Violin plots for HTR2B and HTR7 expression in the different liver cell compartments. Data extracted from GSE192742(13). (B) Analysis of HTR2B and HTR7 expression in fibrotic liver tissues of clinical cohorts at transcriptomic level (GSE162694(15)): 0, n = 35; 1, n= 30; 2, n = 27; 3, n = 8; 4, n = 12; normal, n = 31). * p < 0.05, ** p < 0.01, *** p < 0.001. Kruskal-Wallis test followed by Dunn multiple comparison test and BH correction. (C) Unsupervised clustering in a MASLD cohort (GSE193066)(16): 106 non-cirrhotic, HCC-naïve patients with MASLD) identified an HTR7-high patient subgroup (18%) enriched for advanced fibrosis/inflammation. (D) Analysis of HTR2B and HTR7 expression in paired HCC and non-tumoral liver tissues (GSE94660(18) : n = 21). * p < 0.05, Student’s T test. (E) Violin plots for HTR2B and HTR7 expression in the different cell compartments of HCC patient tissues. Data extracted from GSE151530(20). (F) HTR2B and HTR7 protein expression in HCC tumor tissues (Strasbourg cohort). Representative IHC pictures showing protein expression in adjacent non-tumoral tissue and HCC tumoral tissues (see Supplementary Figure 2 for other HCC cases and Supplementary Table 1 for clinical information). Scale bars = 200 µm. (G) HTR2B and HTR7 protein expression in normal liver tissues (Human Protein Atlas, images available from https://www.proteinatlas.org/ENSG00000135914-HTR2B/tissue/liver and https://www.proteinatlas.org/ENSG00000148680-HTR7/tissue/liver). Scale bars = 50 µm.

We then analyzed HTR2B and HTR7 protein expression in a Strasbourg patient cohort with CLD and HCC and in the Human Protein Atlas (proteinatlas.org)(22) (Figure 1F-G). While HTR7 showed a weak or absent expression in normal hepatocytes, HTR7 was detected in cancer cells in 3/7 patients, as well as in TAMs in the tumor stroma (Figure 1F-G, Supplementary Figure 2A-B). These results are in line with target expression analysis showing an increased expression of HTR7 in Huh7 cancer cells compared to normal primary human hepatocytes (PHH) (Supplementary Figure 2C). These results suggest a role of HTR7 in carcinogenesis. In contrast, HTR2B was detected in cancer cells in only 1/7 patients. However, a robust expression was detected in non-parenchymal cells including TAMs and CAFs in 7/7 patients (Supplementary Figure 2A-C), suggesting a role in carcinogenesis through the modulation of the tumor microenvironment (TME).

While HTR2B and HTR7 are not HCC prognosis markers (Supplementary Figure 3A-B, Supplementary Table 2), these data indicate that aripiprazole target expression in different liver cell compartments is associated with CLD and HCC, and suggest that treatment with aripiprazole could influence liver disease progression by modulating multiple pathways.

Aripiprazole inhibits fibrosis progression and prevents HCC development in vivo in a rodent model of MASH-driven HCC

To study the therapeutic effect of aripiprazole on fibrotic liver disease progressing to HCC in vivo, we used a rodent MASH-induced HCC model. This is an established model mimicking human MASH, where animals sequentially develop steatohepatitis, fibrosis and HCC, induced by CDA/HFD(23). At the onset of fibrosis at the end of 10 weeks, aripiprazole or vehicle control were administered via oral gavage in a therapeutic approach. Both sexes were used. All animals were sacrificed at the end of 18 weeks, after HCC development (Figure 2A). Treatment with aripiprazole significantly reduced liver fibrosis by ~ 30% (p = 0.0006) as measured by collagen proportionate area (CPA) (Figure 2B-C, Supplementary Figure 4). In addition, aripiprazole treatment had a significant effect on hepatocarcinogenesis reflected by a reduced number of liver tumor nodules (p = 0.0002) (Figure 2B-C, Supplementary Figure 4). A decreasing trend in the expression of the proliferating cell nuclear antigen (PCNA) was observed in treated animal livers, a marker associated with proliferation of mammalian cells (Figure 2B-C). Of note, HTR7 expression in animal livers was confirmed in macrophage subpopulations in CDA/HFD rat livers (Fig. 2D), as observed in diseased patient-derived tissues (Figure 1).

Figure 2. Aripiprazole slows fibrosis progression and prevents HCC development in a MASH/HCC rat model.

(A) Wistar rats were subjected to CDA/HFD for a total of 18 weeks. Oral gavage of aripiprazole or vehicle control was initiated after 10 weeks of CDA/HFD diet following the onset of fibrosis (n = 18 per group, 9 males and 9 females per group). The cartoon was generated using Biorender. (B) Representative morphometric analysis of animal livers at the time of sacrifice, Sirius red (collagen) and PCNA stainings are shown (Scale bar = 250 µm). (C) Body and liver weight, liver/body weight ratio, numbers of surface tumor nodules, quantification of PCNA and CPA are reported for each group. In box and whisker plots, boxes represent the 75th and 25th percentiles, the whiskers represent the most extreme data points within interquartile range × 1.5, and the horizontal bar represents the median. (D) Representative pictures of HTR7 (aripiprazole target), αSMA (myofibroblasts) and CD68 (macrophages) IHC staining. Scale bar = 250 µm for CD68 and αSMA, 100 µm for HTR7. (E) Quantification of αSMA and CD68 positive cells showed in (D) and serum CRP. In box and whisker plots, boxes represent the 75th and 25th percentiles, the whiskers represent the most extreme data points within interquartile range × 1.5, and the horizontal bar represents the median. The graphs show the number of positive cells per mm². (F) Analysis of liver function and inflammation by measurement of serum liver enzymes, albumin, total bilirubin, and CRP. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001, Kruskal Wallis followed by Dunn’s multiple comparison test, except for CPA, one way ANOVA followed by Tukey’s multiple comparisons test. For tumor nodules, “chow” group was excluded from statistical analysis because animals do not develop tumors (Mann Whitney test). Ns= non-significant.

Interestingly, aripiprazole significantly decreased the number of CD68+ macrophages in the liver of treated males (p = 0.0110), in association with a reduction of serum C-reactive protein levels (CRP) (p = 0.0102) (Figure 2E), indicating that aripiprazole decreased inflammation in vivo. A modest reduction of inflammation was also observed in treated females (Figure 2E), however to a lesser extend which is likely due to less severe inflammation in female animals in the CDA/HFD (24). A decrease in the number of αSMA positive myofibroblasts was also observed after aripiprazole treatment, however not significant due to a high heterogeneity among animals (Figure 2D-E).

Importantly, aripiprazole did not show any major detectable liver toxicity in the liver as shown by the serum levels of ALT, AST, alkaline phosphatase (ALP), albumin (ALB), γ-glutamyl transferase (GGT) and total bilirubin level (TBIL) (Figure 2E, Supplementary Figure 4) and stable body weight (Figure 2C).

Together, these results show that aripiprazole inhibits fibrosis progression, reduces inflammation and prevents HCC development in vivo.

Aripiprazole reduces liver fibrosis and HCC by modulating immune and inflammatory responses and inhibiting tumor initiation

To elucidate the mechanism of action by which aripiprazole modulates liver disease progression, an RNA-Seq analysis on rat liver tissues was performed, followed by GSEA (Figure 3). Downregulation of serotonin signaling pathways suggested aripiprazole target engagement in vivo (Figure 3A). We observed that aripiprazole treatment modulated the expression of several gene sets associated with liver disease progression, inflammation, fibrosis and carcinogenesis (Figure 3A-B, Supplementary Table 3). In line with a decreased number of tumor nodules, tumor initiating and cell cycle processes were significantly downregulated by aripiprazole treatment (Figure 3B). Suppression of fibrosis-related genes, Col1a1, Col13a1 and Acta2 confirmed the antifibrotic efficacy (Figure 3C), consistent with the reduction of CPA levels (Figure 2B-C). Aripiprazole treatment also downregulated gene signatures related to immune cell activation, including helper T cells, and macrophages (Figure 3B). In addition, levels of Ccl21 and Cxcl13, genes encoding chemokines with a role in promoting fibrogenesis and chronic inflammation, decreased with aripiprazole treatment (Figure 3C). This finding is consistent with reduced CRP and CD68 marker levels (Figure 2D), likely reflecting an anti-inflammatory effect of aripiprazole. Collectively, these data confirm aripiprazole’s potential in modulating immune responses, mitigating liver fibrosis and carcinogenesis.

Figure 3. Aripiprazole modulates expression of cell circuits mediating fibrosis development and carcinogenesis in the MASH/HCC rat model.

RNA-Seq was performed on rat liver tissues (n = 3 per group) and analyzed by GSEA and gene set enrichment index (GSEI). (A) Heatmap shows induction (green) or suppression (orange) of the molecular pathways and fibrotic gene modules in animal livers (Left: vehicle CTRL vs Chow diet; right: Aripiprazole vs Vehicle CTRL). (B) The graph shows enrichment of gene sets relevant for liver disease and HCC expressed as NES from GSEA analysis. Negative NES indicates a suppression of the pathway in aripiprazole-treated animals compared to vehicle treated animals (n = 3). False discovery rate (FDR) < 0.05. (C) RNA-Seq-derived expression of pro-fibrotic genes in rat livers.

Aripiprazole reduces liver fibrosis by inhibiting collagen production in myofibroblasts and expression of pro-inflammatory cytokines in macrophages

Target expression analysis in patient-derived tissues had shown that aripiprazole targets are expressed in fibroblasts and macrophages (Figure 1). We therefore analyzed the therapeutic effect of aripiprazole in patient-derived human myofibroblasts and a model of monocyte-derived pro-inflammatory macrophages. We observed that aripiprazole treatment robustly decreased collagen 1 A1 (COL1A1) expression in HHSteC and HLMF stimulated with TGFβ (Figure 4A). Perturbation studies in HLMF showed that this effect is partially mediated by HTR2B (Figure 4B). Moreover, aripiprazole dose-dependently reduced expression of interleukin 6 (IL6) and tumor necrosis factor alpha (TNFα) in pro-inflammatory M1-like macrophages (Figure 4C), two cytokines associated with inflammation and carcinogenesis in the liver. This effect is also mediated through HTR2B (Figure 4D). Together, these results demonstrate a functional role of HTR2B in liver fibrosis progression by modulating collagen and pro-inflammatory cytokines production in myofibroblasts and macrophages.

Figure 4. Aripiprazole exhibits both anti-fibrotic and anti-inflammatory effects and reverses the poor prognosis status of the PLS across human cell-based liver disease models.

(A) Aripiprazole decreases expression of collagen 1 in patient-derived stellate and myofibroblasts. HHStec and HLMF were stimulated with TGFβ to induce fibrotic-gene expression and then treated with DMSO or aripiprazole (10 µM). The graphs show means +/- sd of % of normalized mRNA to GAPDH for ACTA2 and COL1A1, from 3 independent experiments performed in triplicate (n = 9). * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001, Kruskal Wallis followed by Dunn’s multiple comparison test. (B) Effect of aripiprazole on collagen expression in HLMF is partially mediated though HTR2B. Left panel: HTR2B and HTR7 expression was knocked down (KD) in HLMF. Gene expression was assessed by qRT-PCR. Center panel: effect of HTR2B and HTR7 KD on collagen expression in aripiprazole-treated cells. The graphs show means +/- sd of % of normalized mRNA to GAPDH for ACTA2 and COL1A1, from 2 independent experiments performed in triplicate (n = 6). * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001, Unpaired Student’s T test. Right panel: the graph shows the fold change of COL1A1 gene expression between aripiprazole-treated and DMSO-treated cells for each condition (siCTRL, siHTR2B, siHTR7). (C) Aripiprazole decreases expression of pro-inflammatory cytokines in M1-like THP1-derived macrophages. THP1-derived macrophages were differentiated in M1-like macrophages using LPS and IFNγ before aripiprazole or DMSO treatment. Expression of CD204, CD206 and IL23 was assessed by qRT-PCR to validate differentiation of M1-like macrophages. The graphs show mean percentages +/- sd of normalized mRNA to GAPDH, from 3 independent experiments performed in triplicate (n = 9). * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001, Kruskal Wallis followed by Dunn’s multiple comparison test, except for IL6 and TNFA, one way ANOVA followed by Tukey’s multiple comparisons test. Ns = non-significant. (D) Effect of aripiprazole on IL6 and TNFA expression in M1-like macrophages is mediated though HTR2B. The graphs show mean percentages +/- sd of normalized mRNA to GAPDH. One representative experiment out of 3, performed in triplicate is shown (variability between the assays due to the presence of siRNAs). (E). Aripiprazole reverses the poor-prognosis PLS in different liver fibrosis models. The poor-prognosis PLS was induced by persistent HCV infection in the 2D cPLS model, by FFA treatment in the 2D co-culture cPLS model and by using a mixture of TGFβ/FFA/LPS in the 3D cPLS model containing stellate cells and macrophages (see method). Heatmaps show the PLS global status (bottom) and the global variation of the high- and low-risks genes (top), calculated using GSEA. One representative experiment performed in triplicate out of 3 is shown. The cartoons were generated using Biorender.

Next, we evaluated the antifibrotic effect of aripiprazole in human liver fibrosis models recapitulating the clinical PLS. PLS is a transcriptomic signature identified in fibrotic/cirrhotic patients and predictive of disease progression, HCC risk, and outcome for all major etiologies(11,23,25–27). We applied complementary cell culture-derived PLS (cPLS) models, based on the use of hepatocyte-like cells in monoculture, or in co-culture with LX2 stellate cells, and treated with inducers of chronic liver disease (hepatitis virus C (HCV) infection, or metabolic injury induced by free fatty acids (FFAs), respectively) (Figure 4E). Expression of HTR2B and HTR7 was confirmed in the different cell types, indicating that the cPLS model systems are applicable to investigate the therapeutic and mechanistic effects of aripiprazole (Supplementary Figure 5A). In the cPLS models, aripiprazole treatment reverted the poor-prognosis PLS to good prognosis PLS, by significantly suppressing the expression of high-risk genes and increasing low-risk gene expression (Figure 4E, Supplementary Figure 5B). In addition, the PLS status was assessed in a next-generation triple culture model cPLS system with LX2 stellate cells and macrophages to mimic the liver immune environment in 3D. Fibrosis and poor-prognosis PLS were induced by a combination of transforming growth factor beta (TGF-β), FFA, and LPS (Figure 4E). Similarly, aripiprazole treatment reversed the poor PLS prognosis to good prognosis status in the 3D cPLS model (Figure 4E). Interestingly, PLS reversal was more robust in the triple culture cPLS 3D model (consisting of epithelial cells, macrophages and stellate cells) compared to the monoculture cPLS using epithelial cells only. These results confirm the functional role of non-parenchymal cells expressing aripiprazole targets for the therapeutic efficacy of the compound (Figure 4E, Supplementary Figure 5B).

Collectively, these data demonstrate that aripiprazole shows robust antifibrotic efficacy and reverses the expression of a clinically relevant signature predicting disease progression and cancer risk across different human cell-based liver disease models and etiologies.

Aripiprazole exhibits a direct anti-cancer effect by inhibiting ERK signaling and mitochondrial functions in patient-derived HCC cell lines

Next, to assess whether aripiprazole also contributes to a direct effect on arising HCC in advanced liver disease, we studied the effects on cancer hallmarks in HCC cancer cells. To assess the effect of aripiprazole on cancer cell viability and proliferation, an MTT and Edu assay were performed respectively. MTT assay showed a time and dose dependent reduction in Huh7 cancer cells viability, with a 50% reduction in metabolic activity with the highest doses (Figure 5A). Similarly, EdU proliferation assay showed a dose-dependent reduction in cell proliferation, with a 50% reduction in the percentage of EdU-positive cells after 24 hours of aripiprazole treatment (Figure 5B), indicating that aripiprazole inhibits cell proliferation in a concentration-dependent manner. Of note, the caspase 3/7 assay also showed an induction of apoptosis in Huh7 treated with aripiprazole, in line with MTT assay (Supplementary Figure 6A). Then, the multipotent nature of the cancer stem cell subpopulations was assessed using the tumorspheroid formation assay (Figure 5C). Aripiprazole significantly reduced tumorspheroid formation in Huh7. The area of the formed tumorspheroids decreased by 40-50% at the highest aripiprazole concentration (10 μM) compared to the non-treated condition (Figure 5C). These findings suggest that aripiprazole inhibits the stem cell-like properties of HCC cells.

Figure 5. Aripiprazole treatment shows a direct anti-cancer effect in Huh7 cancer cells.

(A) MTT assay shows a time and dose dependent reduction of Huh7 cell viability. The graph shows mean percentages +/-sd of viability compared to the mock non-treated cells. One representative experiment performed in triplicate out of 3 is shown. (B) Aripiprazole dose-dependently reduces Huh7 cell proliferation. Cell proliferation was assessed by EdU assays (flow cytometry). Left panel: the graph shows mean percentages +/- sd of Edu positive cells compared to the DMSO-treated cells from 3 independent experiments performed in duplicate. Kruskal-Wallis test followed by Dunn’s multiple comparisons test. Right panel: representative histogram of the flow cytometry experiment. (C) Aripiprazole dose-dependently suppresses tumorspheroid formation. Representative pictures are shown for each condition. Scale = 300 µm. The graph shows means +/- sd of sphere size from 3 independent experiments performed in duplicate. Each dot represents one spheroid. Kruskal-Wallis test followed by Dunn’s multiple comparisons test. (D) Aripiprazole suppresses ERK and AKT phosphorylation in Huh7 upon EGF and HGF induction. Activation of ERK, AKT, cMet and EGFR by phosphorylation was assessed by Western blot analysis. The graph shows protein quantification as means +/- sd of normalized protein intensity (normalization to total proteins), from 4 independent experiments (n = 4). * p < 0.05, ** p < 0.01. (E) Aripiprazole impairs mitochondrial functions. Mitochondrial respiration in Huh7 cells was assessed in a two-chamber respirometer Oroboros Oxygraph-2k (O2k; Oroboros Instruments, Innsbruck, Austria) at 37 °C. Left panel: the graph shows time course of oxygen consumption upon successive activation/inhibition of the different mitochondria complexes (see method). One representative experiment out of 4 is shown. The respiration of each pathway control state was corrected for residual endogenous substrates (REN) and reported on the right panel. The graph shows oxygen consumption as means +/- sd, from 3 independent experiments performed in duplicate (n = 2). * p < 0.05, Student T test. Ns = non-significant.

Given the observed regulation of proliferation and tumorspheroid formation, we investigated the effect of aripiprazole on key signaling pathways known to regulate cell proliferation, survival and apoptosis in HCC, namely the extracellular signal-regulated kinases 1 and 2 (ERK1/2) and the protein kinase B (AKT) pathways(28). ERK was activated by using epidermal growth factor (EGF) and AKT was induced by using hepatocyte growth factor (HGF), as shown by increased phosphorylation on the western blot analysis (Figure 5D). Interestingly, aripiprazole suppressed ERK and AKT phosphorylation in stimulated cancer cells (Figure 5D). To determine whether this effect on ERK and AKT phosphorylation was due to upstream regulation, the activation status of two key tyrosine kinase receptors was assessed: epidermal growth factor receptor (EGFR) and hepatocyte growth factor receptor (cMet). A decrease in phosphorylation of the cMet receptor was observed in Huh7, but no significant change in EGFR phosphorylation, indicating that aripiprazole can inhibit cMet and ERK signaling independently from EGFR (Figure 5D).

Of note, aripiprazole is known to regulate calcium levels and ion channels (such as NMDARs)(29). As calcium signaling is an emerging pathway linked to cancer development and CLD(30,31), calcium influx in Huh7 cells was measured and no difference in calcium levels was noted between aripiprazole-treated cells and control in the first 3.3 minutes of live-cell stimulation, indicating that the antiproliferative effect of aripiprazole is not mediated through calcium signaling (Supplementary Figure 6B).

Recently, it was suggested that atypical antipsychotic compounds such as aripiprazole may exert an anticancer effect by affecting mitochondrial function and inducing apoptosis(32,33). Given the observed inhibition of ERK and AKT pathways and leading to decrease in cancer cell proliferation, we investigated the effect of aripiprazole on mitochondrial function (Figure 5E). We observed that aripiprazole treatment significantly reduced F-, FN- and S-mitochondrial pathways, indicating an impaired activity of oxidative complex I and complex II activity (Figure 5E). Importantly, the integrity of the mitochondrial membranes was not impacted by the treatment (Supplementary Figure 6C). These data demonstrate that aripiprazole impairs mitochondrial respiration by reducing electron flux through complexes I and II, leading to decreased cellular energy metabolism and cell proliferation.

To validate our findings in another cancer cell line, we performed tumorspheroid- and signaling-assays in Hep3B cancer cells. As observed in Huh7, aripiprazole dose-dependently decreases tumorspheroids size (Supplementary Figure 7A) and inhibits ERK phosphorylation in Hep3B (Supplementary Figure 7B). Altogether, these results indicate that aripiprazole may have a direct effect on HCC arising on fibrotic liver.

Finally, to investigate whether aripiprazole may impact HCC progression, we studied the effect of aripiprazole on the metastatic potential of HCC cells (Figure 6A). The 3D-invasion assay showed a dose-dependent reduction in cellular protrusions within the Matrigel in spheres treated with aripiprazole, indicating that the compound may inhibit the invasive property of cells—a key hallmark of cancer. Interestingly, the effect of aripiprazole was more pronounced in the co-culture model (Huh7 cancer cells and LX2 stellate cells) compared to the mono-culture model (Figure 6B). These data again confirm the functional role of non-parenchymal cells for the therapeutic efficacy of aripiprazole as observed in the cPLS model (Figure 4E). Overall, our results demonstrate that aripiprazole exhibits a direct anti-cancer effect in HCC cells.

Figure 6. Aripiprazole dose-dependently suppresses cancer cell migration in 3D HCC co-culture models.

Invasion in a Matrigel was induced by a mix of EGF, HGF and TGFβ in a monoculture system (Huh7 cancer cells) (A) or a co-culture system (Huh7 cancer cells + LX2 stellate cells) (B). Representative pictures are shown for each condition. Scale = 500 µm. The graphs show mean percentages +/- sd of cell invasion compared to the mock non-treated spheres. One representative experiment out of 3 performed in triplicate is shown.

Aripiprazole treatment modulates the TME in patient-derived HCC tissue spheroids

Finally, the effect of aripiprazole on TME was investigated. To address this question, we applied a recently established patient-derived liver tumorspheroid system, modeling both cancer cells and cells from the TME (CAFs, TAMs, endothelial cells, T cells) (Figure 7A)(34,35). These multicellular tumorspheroids generated from authentic patient tissues maintain the sample-specific features, reflect HCC heterogeneity and, coupled to single cell RNA-Seq (scRNA-Seq), allows for more clinically relevant testing of compounds, offering a more individualized approach to treatment(34,35).

Figure 7. ScRNA-Seq of HCC patient tumorspheroids treated with aripiprazole reveals reprogramming of the TME.

(A) Experimental approach. Patient-derived HCC tissue was processed to generate cell suspension including cancer cells and cells from the tumor microenvironment. Multicellular tumorspheroids were generated by combining cell suspension and patient autologous serum in low attachment plates (see method) and subjected to aripiprazole or DMS0 control for 48 hours. The cartoon was generated using Biorender. (B) 2D-visualization of single-cell transcriptomics of cells from aripiprazole and DMSO-treated tumorspheroids and cell clustering using umaps. Each dot represents one cell. Of note, cluster 7 corresponds to a mix of tumor epithelial cells with mesenchymal-like profile, and of CAFs, which cluster together. (C-D-E) GSEA analysis of the CAFs/cancer cell (cluster 7), pro-inflammatory macrophages (cluster 9), and CD4+ T cells (cluster 0, cluster3). The graphs show enrichment of gene sets relevant for carcinogenesis expressed as normalized enrichment score (NES) from GSEA analysis. Negative NES indicates a suppression of the pathway whereas positive NES indicates activation of the pathway in aripiprazole treated tumorspheroids compared to DMSO. False discovery rate (FDR) < 0.05.

A total of 3,696 and 2,852 cells were analyzed in the aripiprazole- and DMSO-treated condition, respectively (Figure 7B). Clustering and annotation revealed the presence of main immune cell types, including macrophages, several subtypes of T cells and B cells, as well as a cluster of cancer cells based on marker expression (Figure 7B, Supplementary Figures 8-9). Expression of aripiprazole targets HTR2B and HTR7 was observed, mainly enriched in CD163+ TAMs (cluster 8) (Supplementary Figure 8). While treatment with aripiprazole did not result in a marked difference in cell type populations (Figure 7B), several inflammatory pathways were downregulated across clusters, including CAFs and cancer cells (cluster 7), pro-inflammatory TAMs (cluster 9) and CD4+ T cells (clusters 0 and 3), indicating an overall effect on the TME (Figure 7C-D). These observations are in line with the anti-inflammatory effect of aripiprazole observed in CDA/HFD rats (Figure 2). Importantly, a suppression of the cMet pathway was observed in the CAFs/cancer cells (cluster 7) (Figure 7C), as observed in our cell-based models (Figure 5D), as well as suppression of MYC target gene expression (Figure 7C), as observed in the CDA/HFD rat model (Figure 3B). Collectively these data confirm key mechanistic findings in an authentic patient-derived model system comprising the liver microenvironment.

Discussion

Effective HCC chemopreventive strategies are urgently needed. Here, we discovered aripiprazole, an atypical antipsychotic, as a novel candidate for preventive treatments of HCC. Our key findings are: (i) Aripiprazole target expression is associated with CLD and HCC. (ii) Aripiprazole suppresses liver fibrosis and inflammation and reduces HCC development in a MASH-induced HCC state-of-the-art rat model. (iii) Mechanistically, aripiprazole inhibits collagen expression in HLMF and decreases the expression of pro-inflammatory cytokines in macrophages. (iv) Aripiprazole exerts a direct anti-cancer effect by inhibiting oncogenic signaling and mitochondrial functions in HCC cells and modulating the TME.

Previous studies have suggested a functional role of the serotonin pathway in inflammation, which is subtype- and receptor-dependent in immune cells, including macrophages(8). In this study, aripiprazole treatment reduced liver inflammation in vivo reflected by a decrease in CD68+ cell infiltration and CRP levels, MASH-related CD68-CCR2 signaling and other immune datasets and reduced pro-inflammatory datasets in patient-derived tumorspheroids. This anti-inflammatory effect likely contributes to the attenuation of disease progression and supports the restoration of hepatic function. Aripiprazole treatment also slowed fibrosis progression in rats (Figure 2) and decreased the expression of fibrosis-related markers in cell-based models (Figure 4). The antifibrotic effect of aripiprazole may be attributed to its activity as an inverse agonist at HTR2B, a serotonin receptor implicated in fibrogenesis(9). Finally, aripiprazole also exerts an anti-carcinogenic effect in different models through the inhibition of ERK-, AKT-, and cMET-signaling pathways. In conclusion, our data indicate that aripiprazole mitigates liver fibrosis and inflammation resulting in the prevention of early hepatocarcinogenesis, with preserved viability of healthy hepatocytes.

Aripiprazole has been studied in preclinical models of cancers outside HCC, including glioblastoma(36) and colorectal cancer(37). In the case of colorectal cancer (CRC), aripiprazole was found to inhibit cancer cell proliferation by targeting HTR2B(37). In our study, HTR2B seems to play a crucial role in the regulation of fibroblasts and macrophage phenotypes in the liver microenvironment (Figure 4), whereas HTR7 may be associated with carcinogenesis (Figure 1, Supplementary Figure 2). Additional studies have shown functional roles of aripiprazole other targets in MASH and fibrosis. For example, DRD2 has been implicated in mitigating the profibrotic macrophage-endothelial crosstalk(38), or HTR2A in hepatic stellate cell activation and cirrhosis(39). In addition, the presence of aripiprazole targets in various liver cell types suggest that aripiprazole acts on liver disease progression and hepatocarcinogenesis via several and complementary mechanisms of action. This is supported by a most pronounced therapeutic effect of aripiprazole in liver disease and cancer models consisting of both epithelial and non-parenchymal cells. Moreover, transcriptome modulations of cells that do not express aripiprazole targets, such as CD4+ T cells (Figure 7), also suggest an indirect and paracrine effect of the treatment, contributing to its effects on CLD. Further studies are needed to investigate the role of the different aripiprazole targets in cellular response and whether associated pathways are relevant for aripiprazole’s therapeutic effects.

This study has the following strengths: first, we identified an urgently needed novel approach for chemoprevention using a generic drug. Second, our results are robust by showing a consistent effect of aripiprazole treatment across different models, including in vivo studies and patient-derived models. The study has the following limitations: first, we cannot exclude that additional mechanistic events may apply. Second, while there has been scarce evidence showing direct and permanent consequences of aripiprazole on the brain-gut-liver axis, we cannot exclude the potential role of brain-liver communication in the response to treatment we show in this study.

Translating these findings to chemoprevention in clinical practice requires consideration of aripiprazole’s pharmacokinetics and safety profile in patients with liver disease. Aripiprazole is primarily metabolized in the liver by the cytochrome P450 enzymes CYP3A4 and CYP2D6 and is excreted through both the liver and kidneys(40). It is generally well tolerated and presents a safer profile than other antipsychotics due to low risk of hepatotoxicity, which was also confirmed in our animal model. Since no significant differences have been found in its pharmacokinetics between individuals with normal liver and kidney function and those with impaired liver or kidney function(41,42), it could potentially be used in patients with compensated cirrhosis. Additionally, given its effect on arising HCC, and the modulation of the surrounding TME, aripiprazole could potentially be used in combination with existing HCC treatment modalities, notably immunotherapies. Future studies should investigate its potential as an adjuvant therapy in combination with standard-of-care treatments, including immune checkpoint inhibitors.

In our cPLS systems, aripiprazole reversed the poor-prognosis status of the PLS, induced by persistent HCV infection, metabolic injury, and pro-fibrotic molecules (Figure 4). These results indicate that aripiprazole may prevent hepatocarcinogenesis in different etiologies, including viral infection. This hypothesis is supported by our proof-of-concept study and mechanistic data, which have demonstrated that aripiprazole reduces inflammation and fibrosis independently from lipid metabolism. This supports broader applicability of aripiprazole and its clinical translatability potentially slowing disease progression, decreasing HCC risk and improving patient survival across all etiologies.

The presence of aripiprazole target receptors in various liver cell types and its effects on fibrosis and liver disease pathways suggest aripiprazole is a viable and differentiated candidate for chemoprevention. Moreover, expression analysis in patient tissues showed heterogeneity in aripiprazole target expression and identified a HTR7^high subgroup of patients with more advanced fibrosis/inflammation, suggesting a population amenable to aripiprazole treatment. Future clinical studies should focus on assessing tolerability, optimal dosing, and identifying biomarkers for patient stratification to enhance treatment efficacy.

A blood-based derivative of the PLS, termed PLSec, was identified to predict HCC risk in patients in a non-invasive manner. In association with α-fetoprotein and clinical variable-based, the PLSec has been validated recently in a randomized phase 3 clinical trial enabling accurate HCC risk stratification in patients with cirrhosis from contemporary etiologies(43,44). Given that aripiprazole has a significant impact on the PLS status, the PLSec may be a biomarker candidate for patient stratification and a non-invasive biomarker for future aripiprazole treatment response.

Overall, our findings highlight its potential for repurposing in HCC chemoprevention – a major unmet medical need world-wide.

List of abbreviations

HCC

Hepatocellular carcinoma

MASH

Metabolic dysfunction-associated steatohepatitis

HTR2A

5-hydroxytryptamine receptor 2A

HTR2B

5-hydroxytryptamine receptor 2B

HTR1A

5-hydroxytryptamine receptor 1A

HTR7

5-hydroxytryptamine receptor 7

FBS

Fetal bovine serum

DMEM

Dulbecco’s modified Eagle’s medium

LPS

Lipopolysaccharides

IFNγ

Interferon gamma

HHStec

Hepatic stellate cells

HLMF

Human liver myofibroblasts

CDA/HFD

Choline-deficient, L-amino acid-defined, high-fat diet

DMSO

Dimethylsulphoxide

PLS

Prognostic liver signature

NES

Normalized enrichment score

GSEA

Gene set enrichment analysis

FDR

False discovery rate

CLD

Chronic liver disease

MASLD

Metabolic dysfunction-associated steatotic liver disease

BMI

Body mass index

AST

Aspartate aminotransferase

ALT

Alanine aminotransferase

TAMs

Tumor-associated macrophages

TECs

Tumor endothelial cells

CAFs

Cancer-associated fibroblasts

PHH

Primary human hepatocytes

TME

Tumor microenvironment

CPA

Collagen proportionate area

PCNA

Proliferating cell nuclear antigen

CRP

C-reactive protein levels

ALP

Alkaline phosphatase

ALB

Albumin

GGT

γ-glutamyl transferase

TBIL

Total bilirubin level

COL1A1

Collagen 1 A1

IL6

Interleukin 6

TNFα

Tumor necrosis factor alpha

cPLS

Cell culture-derived PLS

HCV

Hepatitis virus C

FFAs

Free fatty acids

TGF-β

Transforming growth factor beta

ERK1/2

Extracellular signal-regulated kinases 1 and 2

AKT

Protein kinase B

EGF

Epidermal growth factor

HGF

Hepatocyte growth factor

EGFR

Epidermal growth factor receptor

cMet

Hepatocyte growth factor receptor

scRNA-Seq

Single cell RNA-Seq

Ns=

non-significant

GSEI

Gene set enrichment index

CTRL

Control

KD

Knocked down