A total of 31 pairs of tissue specimens, including tumor and adjacent tissues (distance from tumor edge >1 cm), from 31 patients with pathologically confirmed HCC were collected from the Department of Oncology and Vascular Intervention, First Hospital of Shanxi Medical University between January 2021 and December 2022. The patients (29 men and 2 women) ranged in age from 50-78 years. The clinicopathological features of these patients with HCC is shown in Table SI. This study was approved by the ethics committee of the First Hospital of Shanxi Medical University (approval number: 2021SLL086). All patients provided written informed consent. All research was conducted in strict accordance with the Helsinki Declaration and relevant institutional and national regulations.

The HCC-related GSE101728, GSE45050, GSE84598 and GSE121248 datasets from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/gds/?term=) were used for screening of shared differentially expressed genes (DEGs). The Cancer Genome Atlas (TCGA)-liver hepatocellular carcinoma (LIHC) dataset containing the clinical information and mRNA expression data of 371 HCC patients and 50 normal controls was downloaded from the TCGA database (https://portal.gdc.cancer.gov/). The somatic mutation profiles of HCC patients were obtained from the University of California, Santa Cruz (UCSC Xena Browser, https://xenabrowser.net/datapages/). HCC related single cell sequencing dataset GSE202642 came from the GEO database.

DEGs from the HCC-related GSE101728, GSE45050, GSE84598 and GSE121248 datasets were screened using the GEO 2R analysis tool (https://www.ncbi.nlm.nih.gov/geo/geo2r/). The 'limma' R package (version 3.52.0; https://bioinf.wehi.edu.au/limma/)was used to normalize the TCGA-LIHC dataset and identify DEGs. Thresholds were set as P<0.05 and |logFC|>1.

The list of mitochondrial-related genes was obtained from MitoCarta3.0 (https://www.broadinstitute.org/mitocarta/mitocarta30-inventory-mammalianmitochondrialproteins-and-pathways), which includes 1,136 human mitochondrial-related genes. Mitochondrial-related genes were intersected with the DEGs from the four HCC-related GEO datasets to obtain MRDEGs.

The GSE202642 dataset comprising seven HCC samples and 4 adjacent non-tumor samples was analyzed, it contained a total of 115,732 cells and 36,601 genes. The 'Seurat' R package (version 4.3.0; https://github.com/satijalab/seurat) was used to process single cell sequencing data as follows: i) 'DoubletFinder' was applied to remove doublets; ii) the 'PercentageFeatureSet' function calculated mitochondria and red blood cells content per cell. Quality control was performed using the criteria: i) Detected genes per cell >500; ii) detected genes per cell <6,000; iii) mitochondrial genes <20; iv) red blood cells <1. 'NormalizeData' and 'ScaleData' functions were then used to normalize and scale the data. 'RunPCA' function performed linear dimensionality reduction with dim=20. 'FindClusters' function clustered cells with resolution=0.9. 'RunTSNE' function performed tSNE dimensionality reduction. 'FindAllMarkers' function identified DEGs in each cell cluster. The 'SingleR' R package (version 2.0.0; https://github.com/SingleR-inc/SingleR) was used to identify and annotate cell types. KMO expression levels in normal and HCC samples were compared across different cell types.

Overall Survival analysis of genes in HCC patients was performed using the Kaplan-Meier Plotter online tool (http://kmplot.com/analysis/).

WGCNA was constructed using the DEGs matrix of 340 HCC samples via the 'WGCNA' R package (version 1.71; https://CRAN.R-project.org/package=WGCNA). Sample clustering was performed using the 'hcluster' function to identify and remove outliers. The optimal soft threshold power β for network construction was determined using the 'pickSoftThreshold' function. The adjacency degree matrix was transformed into a topological overlap matrix (TOM) using 'TOMsimilarity'. Gene expression similarity was then calculated and the minimum value of module genes was set to 30 genes. According to the similarity, genes were divided into different co-expression modules and the module allocation diagram under gene tree diagram was drawn by DynamicTreeCut. HCC samples were divided into high and low KMO expression groups based on the median value of KMO expression. Correlation analysis was performed between gene co-expression module and KMO expression status. Pearson's correlation coefficient and P-value were calculated using the 'cor' function. The co-expression module with the largest correlation coefficient P<0.05 was regarded as the key functional module and the genes within this module were considered KMO-co-expressed hub genes.

Gene Ontology (GO) analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis were performed using clusterProfiler (version 4.8.0; https://yulab-smu.top/contribution-knowledge-mining/) and ggplot2 (version 3.4.1; https://CRAN.R-project.org/package=ggplot2) packages in R.

To investigate the relationship between KMO and Try metabolic pathway genes, Try metabolic pathway genes were downloaded from the GSEA official website (https://www.gsea-msigdb.org/gsea/msigdb/index.jsp) and analyzed using the STRING database (https://string-db.org/). The PPI network was visualized using Cytoscape 3.9.0 software (https://cytoscape.org/). The hub genes related to KMO were captured using CytoHubba of Cytoscape. Correlation heatmaps and lollipop plots were generated using the online SRplot tool (http://www.bioinformatics.com.cn/SRplot) (21).

The HCC cell lines MHCC-97H and HCCLM3 (Cell Resource Center, Institute of Basic Medicine, Chinese Academy of Medical Sciences) were cultured in DMEM supplemented with 10% FBS at 37°C with 5% CO₂.

KMO-overexpressing lentivirus were constructed by Shanghai GenePharma Co., Ltd. using the 3rd lentiviral, LV5 (EF-1a/GFP&Puro). 293T cells (Cell Bank of Chinese Academy of Sciences) was selected to package the lentivirus using shuttle plasmid H6370 and packaging plasmid (pGag/Pol, pRev, pVSV-G). The lentivirus titer of 293T cells after 72 h of infection with the collected lentivirus stock solution was 9×10⁸ TU/ml. Multiplicity of infection (MOI) was calculated according to virus titer and the optimal MOI (1 number/cell) was determined for HCC cell infection by pre-experiment. For lentiviral overexpression, cells were seeded in 24-well plates at 5×10⁴ cells/well. When confluence reached 40-60%, cells were infected with KMO-overexpressing lentivirus at 1×10⁵ TU/ml. After 24 h, the medium was replaced with fresh culture medium. Upon reaching 70-80% confluence, positive cells were screened using 1 µg/ml puromycin concentration (the optimal concentration determined by puromycin concentration gradient experiment), and 0.5 µg/ml was the maintenance concentration of cultured cells. For short interfering (si)RNA knockdown, siRNA targeting KMO (5′-GGA GCC TAT TCA ACT GTC A-3′, 5′-GCG AGC ACA TGT CAA CTC A-3′ and 5′-CCA AGG TAT TCC CAT GAG A-3′) and control siRNA was designed and synthesized by Guangzhou RiboBio Co., Ltd. Cells were seeded in 6-well plates at 3×10⁵ cells per well and transfected at 70-80% confluence using Lipo8000 reagent (Beyotime Institute of Biotechnology) with the screened siRNA sequence (5′-CCA AGG TAT TCC CAT GAG A-3′) having the highest knockdown efficiency and a control siRNA. Overexpression and knockdown efficiencies were assessed using qPCR and western blotting. KMO stably overexpressing cell lines and transiently transfected KMO-knockdown cells (harvested at 48 h) were used for subsequent studies.

Cells were cultured in confocal dishes (1,000 cells/dish) for 24 h, then stained with Mito-tracker at 37°C for 25 min in the dark. After washing with ice-cold PBS, cells were fixed with 4% paraformaldehyde at 37°C for 1 h, permeabilized with 0.2% TritonX-100 for 10 min, after TBST wash buffer (0.05% Tween 20) washing 3 times (5 min/time), cells were blocked with 2% BSA (Beijing Solarbio Science & Technology Co., Ltd.) for 10 min. Cells were incubated with 1:100 diluted KMO or NR4A1 primary antibody (cat no. 10698-1-AP and 12235-1-AP, Proteintech Group, Inc.) overnight at 4°C, followed by fluorescent secondary antibody (1:200; cat no. SA00013-2; Proteintech Group, Inc.) at room temperature for 1 h. After DAPI staining (1:200) at room temperature for 15 min, cells were imaged at a magnification of ×100.

Paraffin-embedded sections (thickness: 5 µm) were dewaxed, cleaned with PBS and treated with 3% H₂O₂ to block endogenous peroxidase at room temperature for 10 min. Antigen retrieval was performed using sodium citrate buffer, followed by PBS washing. Sections were blocked with 5% BSA at 37°C for 30 min, then incubated with 1:100 diluted KMO primary antibody (cat no. 10698-1-AP, Proteintech Group, Inc.) overnight at 4°C. After washing, sections were incubated with 1:1,000 diluted HRP-conjugated secondary antibody (cat no. KIT-5010; Maxim Biomedical, Inc.) for 30 min at room temperature. DAB staining was performed for 1 min, followed by hematoxylin counterstaining for 5 min. Sections were dehydrated, mounted and imaged using a light microscope. The staining intensity (SI) and percentage of positive cells (PP) were used to calculate the IHC score (22). SI was graded as follows: 0 for negative, 1 for low positive, 2 for moderate positive, and 3 for high positive. PP was classified into four categories: 0 (0-9% positive cells), 1 (10-25%), 2 (26-50%), 3 (50-75%), and 4 (76-100%). The final immunohistochemical score (IS) was derived by multiplying SI and PP (IS=SI x PP). Quantitative analysis was performed using ImageJ software (version 1.53a, National Institutes of Health).

Total RNA was extracted from tissue samples or cultured cells using RNAiso Plus (Takara Bio, Inc.). RNA was reverse-transcribed into cDNA using the PrimeScript RT Master Mix Reverse Transcription Kit (Takara Bio, Inc.). The target gene expression was quantified by two-step qPCR using TB Green^® Premix Ex Taq II (Takara Bio, Inc.) with the procedure: pre-denaturation at 95°C for 30 sec, 1 cycle; denatured at 95°C for 5 sec, annealed at 60°C and extended for 30 sec, 40 cycles. All operations of RNA extraction, cDNA synthesis, and qPCR were carried out according to the manufacturer's protocols. Relative expression levels were calculated using the 2^−ΔΔCq method according to previous study (23). Primer sequences are listed in Table I. Each sample was analyzed in triplicate in three independent experiments.

Cells were lysed in RIPA buffer (Beyotime Institute of Biotechnology), incubated on ice for 30 min and centrifuged at 12,000 × g for 20 min at 4°C. The supernatant was collected and protein concentration was measured using a BCA kit (Beyotime Institute of Biotechnology) at 562 nm.

Equal amounts of protein (≥20 µg) were separated by 10% SDS-PAGE and subsequently transferred to the PVDF membrane at 100 V for 1 h. Membranes were blocked with 5% non-fat milk for 1 h at room temperature. The membrane was incubated with 1:500 diluted KMO (cat no. 10698-1-AP; Proteintech Group, Inc.) or 1:10,000 diluted β-Actin (cat no. 66009-1-Ig; Proteintech Group, Inc.) primary antibodies overnight at 4°C. After washing, membranes were incubated with 1:10,000 diluted horseradish peroxidase-conjugated secondary antibody (cat no. SA00001-2; Proteintech Group, Inc.) for 1 h at room temperature. Finally, the chemiluminescent substrate (1:1 ratio of solutions A and B) was added to the PVDF membrane for imaging after washing. Relative protein expression was calculated using ImageJ software (version 1.53a, National Institutes of Health).

Cell viability was assessed using the CCK-8 assay kit (MeilunBio Co., Ltd.). Cells were seeded in 96-well plates at 3,000 cells per well. Upon cells reaching the logarithmic growth phase, siRNA transfected experiment was performed. For the 3-HAA group, the medium was replaced with one containing 100 µM 3-HAA (24). CCK-8 reagent was added at 0, 24, 48, 72 and 96 h post-transfection and incubated at 37°C for 2 h. Absorbance at 450 nm was measured using a multifunctional enzyme-linked immunosorbent assay instrument.

Cell proliferation was assessed using the EdU assay kit (Guangzhou RiboBio Co., Ltd.). Cells were seeded in a 96-well plate at 3,000 cells per well and treated as described. After 24 h, cells were transferred to confocal dishes. After 48 h, 200 µl of 50 µM EdU was added and incubated at 37°C for 2 h in the dark. Cells were then treated sequentially with glycine (2 mg/ml, room temperature, 5 min), Triton X-100 (0.5%, room temperature, 10 min), Apollo (30 min, room temperature, in the dark) and Hoechst 33342 staining for 30 min (room temperature, in the dark).

Cells were seeded in a 6-well plate at 500 cells per well. When visible colonies appeared after 10-14 days, cells were fixed with 4% paraformaldehyde at room temperature for 15 min, then stained with 0.1% crystal violet at room temperature for 30 min. Colonies were defined as cell clusters containing ≥50 cells. Plates were rinsed, inverted and colonies were photographed and counted.

The HCC cell lines MHCC-97H and HCCLM3 were cultured for 24 h to 80% confluence, then scratched with a 200 µl pipette tip. For siKMO groups, cells were transfected with siKMO RNA or siControl RNA (20 nM) and cultured at 37°C. After 6 h of transfection, the complete medium was replaced and the culture medium was replaced with fresh medium containing 1% FBS after 24 h of transfection. For KMO overexpression groups, the cells with stable overexpression of KMO and the EV control group were directly scratched after 24 h of culture and then cultured in 1% FBS medium. The wound closure was imaged at 0 and 48 h after scratch and ImageJ software (version 1.53a, National Institutes of Health) was used to quantify the migration distance. For other cells to be tested by Transwell, transfection was performed followed by 24 h of culture. A 24-well plate was prepared with 700 µl of complete medium per well. 200 µl of serum-starved cell suspension (KMO knockdown, KMO overexpression, or corresponding control cells) was added to each well and cultured for 48 h. Non-migrated cells on the upper layer of the chamber membrane were removed with a cotton swab and cells on the lower surface were fixed with methanol at room temperature for 10 min, then stained with crystal violet at room temperature for 20 min and counted under a light microscope.

Following the instructions for the Annexin V-FITC apoptosis detection kit (Bestbio Technology Co., Ltd.), cells were washed twice with pre-cooled PBS to obtain cell pellets, then re-suspended in binding buffer. Cells were stained with 5 µl Annexin V-FITC at 4°C for 15 min, followed by 10 µl propidium iodide dye for 10 min at 4°C in the dark. Samples were immediately detected using flow cytometry (FACSCelesta; BD Biosciences). And the apoptotic rate of HCC cells (the sum of the percentages of early + late apoptotic cells) was analyzed using FlowJo 7.6 software (BD Biosciences).

Mitochondrial biogenesis was detected by using MitoTracker staining method. Cells were seeded in confocal dishes at 1,000 cells/dish. At 48 h after siKMO transfection or 100 µM 3-HAA treatment, cells were stained with MitoTracker (200 nM) for 25 min at 37°C in the dark. After washing twice with DMEM, cells were imaged by confocal microscopy at a magnification of ×100. The total mitochondrial fluorescence intensity was quantified using ImageJ (version 1.53a, National Institutes of Health).

Targeted Try metabolism analysis was performed by Shanghai Zhongke New Life Biotechnology Co. Ltd.

Cells were seeded in a 6-well plate at a density of approximately 3×10⁵ cells per well. For KMO knockdown groups, after 24 h of culture, the cells were transfected with siKMO small interfering RNA or siControl RNA using Lipo8000^® transfection reagent (Beyotime Biotechnology). For the siKMO + 3-HAA treated groups, upon reaching 70-80% confluence, the medium was changed to 2 ml culture medium containing 100 µM 3-HAA or an equivalent volume of DMSO (control). The cells were then transfected with siKMO small interfering RNA. At 48 h after transfection, cell DNA was extracted using a DNA extraction kit (Tiangen Biotech Co., Ltd.) according to the manufacturer's protocol. For KMO overexpression groups, DNA was extracted after 48 h of culture. mtDNA copy number was detected by qPCR using primers specific for the mitochondrial gene ND1 and HGB was used as reference gene.

Cells were lysed using ATP detection lysate, then centrifuged at 12,000 × g, for 5 min at 4°C. The supernatant was collected as the test sample. An ATP standard curve was generated by diluting the ATP standard solution (10 µM) to concentrations ranging from 0.01-10 µM. The ATP detection working solution was prepared at a 1:9 ratio of reagent to diluent. Working solution (100 µl) was added to 96-well black plate and incubated at room temperature for 5 min. Subsequently, 20 µl of sample or standard was added to each well, mixed and the relative light unit value was measured after 2 sec. ATP concentration was calculated using the standard curve.

Statistical analyses were performed using R (version 4.2.2, http://www.R-project.org/), GraphPad Prism (version 8.0; Dotmatics) and SPSS (version 25; IBM Corp.). All experiments were performed three times independently. The data were presented as the mean ± SD. The unpaired Student's t-test was employed for comparisons between two groups with normally distributed continuous variables, while Mann-Whitney U test was used for non-normally distributed data. One-way ANOVA with Tukey's post hoc test was used for comparisons among three groups. Correlations between continuous variables were assessed using Spearman's rank correlation analysis. Categorical variables were compared between groups using the χ² test. P<0.05 was considered to indicate a statistically significant difference.

Analysis of GSE101728, GSE45050, GSE84598 and GSE121248 datasets identified 31 upregulated and 109 downregulated genes commonly associated with HCC (Fig. S1A-C). Intersection with mitochondrial-related genes yielded one upregulated and 14 downregulated MRDEGs (Fig. S1D). Notably, KMO expression was consistently downregulated across all four datasets (LogFC ranging from -2.094 to -3.453; Fig. S1E). The finding was validated in TCGA data which demonstrated significantly lower KMO expression in HCC patients compared to normal individuals (Fig. 1A). qPCR and immunohistochemical results both showed that compared with matched adjacent normal tissues, the expression of KMO in HCC tumor tissues decreased (Fig. 1B and C). The present study grouped HCC tumor samples from TCGA database according to KMO median expression value, including 183 samples in KMO high expression group and 182 samples in KMO low expression group. Association analysis between KMO and clinicopathological features of HCC patients in TCGA database are shown in Table II. The results showed that KMO expression was related to Grade, T stage and tumor stage. Kaplan-Meier survival analysis demonstrated that the overall survival time of HCC patients with low KMO expression was significantly shorter than those with high expression (Fig. 1D).

The present study also analyzed the KMO expression within the HCC tumor microenvironment using single cell sequencing data. Analysis of 85,168 cells (28,998 genes) identified 10 cell types. Compared with normal tissues, HCC samples exhibited increased proportions of T cells, fibroblasts, epithelial and tumor cells, with decreases in endothelial cells, dendritic cell (DC), myeloid suppressor cell (MDSC) and natural killer cells (Fig. 1E). The expression of KMO in HCC was lower in HCC compared with that in normal (Fig. 1F) and was specifically reduced in DC, epithelial and tumor cells and MDSC (Fig. 1G), suggesting KMO may play important roles in these cells.

Using KMO knockdown and overexpression HCC cell models (Fig. S2), the present study evaluated the role of KMO in tumor progression. The CCK-8 test, EdU assay and colony formation experiment revealed that, compared with siControl group, the cell viability, cell proliferation and cell clone formation in siKMO group were increased significantly. Conversely, compared with empty vector (EV) control group, the cell viability, cell proliferation and cell clone formation in KMO-OE group were significantly decreased (Fig. 2). The scratch and Transwell assays demonstrated that cell migration ability in siKMO group were increased significantly, whereas the opposite effect was observed in KMO-OE group (Fig. 3). Flow cytometry analysis of cell apoptosis indicated that the apoptosis rate was increased in KMO-OE group compared with that in EV control group (Fig. S3).

Collectively, these results suggested that downregulation of KMO in HCC cells is associated with enhanced malignant progression.

WGCNA showed MEturquoise module was correlated with KMO expression with a biggest correlation coefficient (0.45) and designating it as a key gene module that may be co-expressed with KMO (Fig. 4A). A total of 1,433 genes in MEturquoise module were extracted for gene function enrichment analysis. GO enrichment analysis showed in biological processing, KMO co-expressed related genes participated in metabolic related pathways; in terms of cell components, KMO co-expression related genes are closely related to mitochondria; In terms of molecular function, KMO co-expression related genes may affect oxidoreductase activity, cofactor and coenzyme binding (Fig. 4B). These findings suggest KMO's potential role in many metabolic processes and its close association with mitochondrial structure and function, which was further supported by immunofluorescence experiments confirming the co-localization of KMO with mitochondria in HCC cells (Fig. 4C). KEGG enrichment results indicated that KMO co-expression related genes were related to Try metabolism (Fig. 4D). A PPI network constructed by six key genes of Try metabolic pathway from GSEA official website showed that KMO had direct interaction with IDO1, TDO2, IDO2, arylformamidase (AFMID), TDO2 and kynureninase (KYNU) with the correlation coefficients were 0, 0.47, 0.15, 0.37 and 0.47 respectively (Fig. 4E). These results suggest that KMO may reside in mitochondria and functionally interacts with other genes to influence the Try metabolic pathway.

Mitochondrial dysfunction is closely related to mitochondrial metabolic disorder and the occurrence and progression of HCC. Mito-Tracker staining was used to detect mitochondrial biogenesis and the 3D image statistical results showed that in MHCC-97H cells and LM3 cells, the mitochondrial mass in siKMO group was increased compared with that in siControl group. Conversely, KMO overexpression reduced mitochondrial mass relative to the EV control group (Fig. 5A and B). The present study detected the copy number of mtDNA coding gene ND1 to further verify whether KMO is related to mitochondrial biogenesis. The results showed that compared with siControl group, the ND1 copy number in siKMO group was increased significantly. Compared with EV group, ND1 copy number in KMO-OE group was decreased (Fig. 5C).

In addition, analysis of transcriptomic data revealed that all 13 genes encoded by the mitochondrial genome were upregulated in high-KMO expression patient samples and were positively correlated with KMO expression (Fig. 5D). Then the expression of two genes with the strongest correlation with KMO, MT-ATP6 and MT-CO1 was validated by using qPCR in KMO knockdown and overexpression HCC cells (Fig. 5E).

These results indicated that the downregulation of KMO in HCC cells promoted mitochondrial biogenesis. The uncoordinated changes in mitochondrial DNA copy number and mitochondrial gene expression suggested that KMO dysregulation can lead to imbalances in mitochondrial mass and function.

In order to investigate how KMO influences Try metabolism, HPLC-MS methods were used to test the content of Try metabolites in HCC LM3 cells overexpressing KMO. Fig. S4 showed the total ion flow mass spectrometry of HCC cells. Among the total 16 metabolites detected, 3-HAA was the only one significantly elevated in HCC cells overexpressing KMO (Figs. S5 and 6A and B). Next, the expression of kynureninase (KYNU), the enzyme immediately downstream of KMO that catalyzes the conversion of 3-hydroxykynurenine (3-HK) to 3-HAA was detected. No statistical significance was found in siKMO or KMO-OE group compared with their control groups (Fig. 6C), suggesting the changes in 3-HAA levels was directly mediated by KMO and independent of KYNU.

The present study further explored whether KMO downregulation induced cell proliferation is caused by 3-HAA. In siKMO HCC cells, 3-HAA can reverse the increased cell proliferation and vitality caused by KMO knockdown (Fig. 6D and E).

Similarly, it was investigated whether the mitochondrial changes induced by KMO are mediated by 3-HAA. As shown in Fig. 7A and B, in both MHCC97H and LM3 cells, after treatment with 3-HAA in siKMO cells for 48 h, the mitochondrial mass was decreased compared with the control group. The mtDNA copy number was significantly reduced in siKMO cells treated with 3-HAA compared to those treated with DMSO (Fig. 7C). ATP content in siKMO HCC cells was lower than that in control cells, while in KMO-OE HCC cells was higher. And 3-HAA can up regulate the ATP content decreased by KMO knockdown (Fig. 7D). ROS generation showed an opposite trend to ATP production and 3-HAA inhibited the ROS generation upregulated by KMO knockdown (Fig. 7E)

These results indicate that downregulation of KMO in HCC reduces the production of its downstream metabolite 3-HAA, thereby promoting HCC progression and inducing alterations in mitochondrial mass and function.

To investigate the mechanisms underlying the mitochondrial alterations induced by KMO downregulation, the present study screened mitochondrial transcription factors associated with KMO expression. A total of 1,670 transcription factor genes downloaded from GTRD website (http://gtrd.biouml.org), were intersected with 2,030 mitochondrial-related genes and 1,721 DEGs between KMO high and low groups (26). A total of four KMO related mitochondrial transcription factors were obtained (Fig. 8A). In these four transcription factors, the expression of KMO was positively correlated with the expression of FOXO1 and NR4A1, while negatively correlated with the expression of HMGA1 and SOX4 (Fig. 8B). After qPCR verification, only NR4A1 expression was consistent with the data analysis: It positively correlated with KMO expression in HCC cells and 3-HAA treatment rescued in the downregulation of NR4A1 expression caused by KMO knockdown (Fig. 8C). Low NR4A1 expression was associated with poor prognosis in HCC patients (Fig. 8D). Similar to KMO, NR4A1 expression was positively correlated with the expression of ten mitochondrial genome-encoded genes (Fig. 8E).

Since NR4A1 is a mitochondria-associated transcription factor, the present study was interested in finding out the subcellular localization of NR4A1. After immunofluorescence staining, it was observed that KMO expression in HCC cells was distributed in both nucleus and cytoplasm. Combined with fluorescent staining of mitochondrial Mito-Tracker, NR4A1 showed increased expression in mitochondria in KMO knockdown HCC cells, whereas NR4A1 expression in mitochondria was reduced after overexpression of KMO. 3-HAA reversed the mitochondrial translocation of NR4A1 induced by KMO knockdown (Fig. 9).

These results illustrated that downregulation of KMO in HCC cells can lead to reduced NR4A1 expression and promotes its mitochondrial translocation, both of which are mediated by 3-HAA.

In KMO overexpressed HCC cells, the expression of NR4A1 was knocked down to investigate the effect of KMO-mediated NR4A1 on HCC cell growth and mitochondrial mass and function. CCK8 and EdU results showed that silencing of NR4A1 enhanced cell viability and proliferation in HCC cells overexpressing KMO (Fig. 10A and B). NR4A1 knockdown increased mitochondrial mass in HCC cells overexpressing KMO compared with siControl cells (Fig. 10C). Additionally, ATP production also decreased by NR4A1 knockdown in KMO overexpressed HCC cells (Fig. 10D). These results indicated that NR4A1 inhibition promotes mitochondrial biogenesis and the malignant growth of HCC cells.