UBE2Q2 promotes tumor progression and glycolysis of hepatocellular carcinoma through NF-κB/HIF1α signal pathway

Tumor and adjacent normal liver tissues of 80 HCC patients (68 male and 12 female patients) on whom the liver resections were performed during 2018 to 2020 in the Department of Hepatobiliary and Pancreatic Surgery, Zhongnan Hospital of Wuhan University were collected. None of the patients received any pre-operative therapy like radiofrequency ablation, anhydrous alcohol injection, TACE or immunotherapy. The histopathological examination of samples was confirmed by two experienced pathologists. Related clinical data of 80 patients were recorded for statistical analysis.

The cell lines Huh-7, HCC-LM3, Hep3B, HepG2, HepG2.2.15, and SK-Hep-1 were obtained from the Shanghai Cell Bank of the Chinese Academy of Medical Sciences. These were cultured under standard conditions: Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin, maintaining a temperature of 37 °C and a 5% CO₂ atmosphere. Cell line authenticity was verified through STR profiling and all were certified mycoplasma-free.

For specific experiments, Huh-7 and LM3 cell lines underwent treatment with distinct agents: 100 µM cobalt chloride for 12 h; 10 µM chlorhexidine, corynoxeine, or LY294002, each for a period of 24 h; 20 µM rapamycin for 24 h; 20 µM (-)-DHMEQ for 24 h and 2 mM 2-DG for 12 h. All aforementioned compounds were sourced from Sigma Aldrich (Germany) and MCE (USA).

RNA from both tissue specimens and cultured cells was isolated using TRIzol Reagent (Invitrogen, USA). The quality and concentration of RNA were assessed with a NanoDrop ND2000 spectrophotometer (Thermo Scientific, Massachusetts, USA). Reverse transcription was conducted using the HiScript III 1st Strand cDNA Synthesis Kit for mRNA and the Mir-X miRNA First-Strand Synthesis Kit (Takara, USA), following the provided protocols. Quantitative real-time PCR was performed with a SYBR Green PCR kit (Vazyme, China) on a CFX96TM Real-time PCR Detection System (Bio-Rad, California, USA). Relative RNA expression levels were quantified using the 2^(-ΔΔCt) method, with the threshold cycle (Ct) being determined. Reproducibility was ensured by conducting all procedures in triplicate. Primer sequences used are detailed in Table S1 of the supplementary materials.

UBE2Q2 expression was targeted for knockdown using small interfering RNAs (siRNAs). Three distinct siRNAs aimed at UBE2Q2 (referred to as siRNA #1, siRNA #2, and siRNA #3) were custom-synthesized by Genecreate (Wuhan, China). Among them, siRNA #2 was subsequently labeled as siUBE2Q2 in Figs. 4 and 5 and Fig. S4. The details of these siRNA and shRNA sequences are provided in Table S2 of the supplementary materials. For the siRNA transfection, the Genmute transfection reagent (Signagen, Maryland, USA) was used, following the manufacturer’s protocol.

The UBE2Q2 gene was synthesized and cloned into the pEZ-Lv105 vector to create an overexpression construct, hereafter referred to as UBE2Q2 plasmid. The empty pEZ-Lv105 vector was utilized as a control, designated as Vector. Both plasmids were products of GeneCopoeia (Rockville, USA). Additional plasmids, including pcDNA-HA-UBE2Q2, pcDNA-HIF1α, pcDNA-His-TRAF2, and pcDNA-Flag-cIAP1, were synthesized by Genecreate (Wuhan, China). Plasmids pCDEF-HA-Ub(K11), pCDEF-HA-Ub(K48), and pCDEF-HA-Ub(K63) were procured from Ke Lei Biological Technology (Shanghai, China). Transfections were carried out using Lipofectamine 3000 (Thermo Fisher Scientific, CA, USA) according to the manufacturer’s recommendations.

Stable UBE2Q2 knockdown and overexpression lines were established using lentiviral systems (Lv-shUBE2Q2 and UBE2Q2 respectively), also provided by Genecreate (Wuhan, China), with a virus titer of 6 × 10⁸ TU/mL. For infection, a multiplicity of infection (MOI) of 10 was applied to Huh-7 cells, with 100 µL of viral suspension coupled with 100 µL of Hitrans P virus infection reagent (GeneChem, Shanghai, China) per 1 × 10⁶ cells. After 24 h, the medium was replenished. Cells were collected 4 days post-infection and selected with puromycin. The knockdown efficiency was evaluated using qRT-PCR and Western blot analyses.

Cell viability was assessed using the Cell Counting Kit-8 (CCK-8, Biosharp, China) assay. Briefly, cells were seeded at a density of 5,000 per well in 100 µL of culture medium in a 96-well plate. Afterward, 10 µL of CCK-8 solution was added to each well. The plate was incubated at 37 °C for 1 h before measuring the absorbance at 450 nm using a microplate reader.

A colony formation assay was performed by plating 500 to 1,000 cells in a six-well plate and culturing for 14 days. The colonies were then fixed with 4% paraformaldehyde and stained for 1 h with 0.1% crystal violet (Sigma Aldrich, Germany). Post-staining, colonies were counted and imaged using a stereomicroscope.

Cell transfection and culture were performed in a six-well plate. Upon reaching confluence, a 100 µL pipette tip was used to create a straight scratch through the cell monolayer. Subsequently, cells were incubated in serum-free DMEM. Wound closure was documented at 0 h, 24 h and 48 h post-scratch, and analyzed using Image J software.

Migration and invasion assays were conducted using a 24-well plate with a Transwell insert (8 μm pore size, BD Biosciences, USA). Medium with 10% FBS (600 µL) was added to the lower chamber. For invasion assays, the upper insert was precoated evenly with Matrigel (Sigma Aldrich, Germany). Subsequently, 1 × 10⁴ cells in 200 µL of serum-free medium were seeded into the upper chamber. Following a 48 h incubation period, non-invading or non-migratory cells were removed. Cells were fixed with 4% paraformaldehyde for 30 min and then stained with 0.1% crystal violet (Sigma Aldrich, Germany) for 1 h. After washed by sterile water, cells were observed using an inverted fluorescence microscope (Olympus, Japan). Results were examined and documented in three random fields of view.

Cell cycle and apoptosis analyses were performed utilizing the Annexin V FITC/PI Apoptosis and Cell Cycle Kit (MultiSciences, China) following the manufacturer’s protocol. Flow cytometric analysis was performed using a Beckman instrument (USA).

Four-week-old male athymic BALB/C-nude mice were acquired from SPF Biotechnology (Beijing, China). All experimental procedures and husbandry were conducted at the Center for Animal Experiment at Wuhan University. The mice were randomly assigned and identified in two groups, with five mice per group, for the establishment of a subcutaneous tumor model. A stable UBE2Q2 knockdown cell line (Lv-shUBE2Q2) was generated using Huh-7 cells. For tumor induction, 5 × 10⁶ cells were injected subcutaneously into the right axillary region of each mouse. Commencing on day 7 post-inoculation, mice received 100 mg/kg of PX-478 in PBS by intraperitoneal injection bi-daily. Tumor sizes were measured tri-weekly. After 3–4 weeks, the mice were euthanized, tumors were excised, and preserved in formalin for further immunohistochemical analysis.

The HCC-LM3 cells transfected with UBE2Q2 plasmids constituted the experimental group, while those transfected with empty vector plasmids served as the control. Cells were harvested 48 h post-transfection. Cell pellets were lysed using a protein lysis buffer containing 7 M urea, 2 M thiourea, 4% SDS, 40 mM Tris-HCl (pH 8.5), 1 mM PMSF, and 2 mM EDTA. The lysate was sonicated for 15 min, then centrifuged at 4 °C for 20 min to collect the protein precipitate. The resulting samples were submitted in triplicate for proteomic analysis at Singleronbio (Nanjing, China). Peptides were tagged using the iTRAQ-8 kit (SCIEX), and peptide detection was conducted with TripleTOF 5600plus mass spectrometry. ProteinPilot V4.5 software was utilized for protein identification and quantification. Differential protein expression was determined based on fold change and P-value; proteins exhibiting a fold change of ≥ 1.5 (upregulated) or ≤ 0.67 (downregulated) with a P-value of ≤ 0.05 were considered significant.

Huh-7 cells were grown in 10 cm diameter culture dishes and co-transfected with various plasmids for subsequent co-immunoprecipitation experiments. The cells were harvested and placed into 2 mL centrifuge tubes. Each tube was treated with 1 mL of IP buffer (comprising 20 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1 mM EDTA (pH 8.0), 1% NP-40, and 1×Protease and Phosphatase Inhibitor Cocktail) to lyse the cells for 30 min. Pre-washed Protein A/G Magnetic beads from MCE (USA) were introduced into the tubes and the mixture was agitated on a disc rotator mixer for another 30 min. Target-specific antibodies were then added to the respective tubes and incubated for over 3 h at 4 °C. After the incubation, beads were collected by centrifugation at 5,000 rpm for 2 min at 4 °C and washed 4 times with ice-cold wash buffer (10 mM Tris/Cl pH 7.5, 150 mM NaCl, 0.5 mM EDTA). The beads were resuspended with 40 µl of 2× sample buffer and boiled for 3 min. Immunoprecipitates were separated by SDS-PAGE and sent for mass spectrometry analysis (Hangzhou Cosmos Wisdom Biotechnology).

Designated cells were transfected with UBE2Q2 plasmids, HIF1α plasmids, empty vector plasmids, siUBE2Q2, and siRNA NC in line with the experimental protocol, and then seeded into 96-well plates. At the 24 h mark, glucose uptake and lactate secretion were quantified using the Glucose Uptake Colorimetric Assay Kit and the L-Lactate Assay Kit, respectively, both from Sigma Aldrich (Germany) and in strict accordance with the provided manufacturer’s protocols. Absorbance readings at 412 nm were captured using an enzyme-linked immunosorbent assay reader.

Huh-7 and HCC-LM3 cells fixed in 4% paraformaldehyde were permeabilized using 0.5% Triton X-100 for 15 min. Subsequently, cells were blocked with 5% bovine serum albumin (BSA) for 1 h at room temperature before overnight primary antibody incubation at 4 °C. After thorough washing with PBS, the cells were incubated with fluorescently labeled secondary antibodies for 1 h at room temperature within a humidified chamber. Following this, cells were mounted with DAPI to stain nuclei and visualized with an Olympus FluoView™ FV1000 confocal microscope.

Cells transfected with the appropriate plasmids or siRNA were plated at a density of 4 × 10⁴ cells per well in XF24 microplates (Seahorse Agilent, USA). The assay cartridge was hydrated in a CO₂-free incubator at 37 °C overnight. The next day, cells were switched to a specialized assay medium before the extracellular acidification rate (ECAR) or oxygen consumption rate (OCR) was measured using a Seahorse XF24 Analyzer’s sensor probe, adhering to the manufacturer’s guidelines. For ECAR, glycolysis stress tests were conducted with Seahorse XF cell glycolysis stress test kits, and the data were processed using Wave software. For OCR, oxygen consumption tests were conducted with Seahorse XF cell mito stress test kits, and the data were processed using Wave software.

Following 48 h post-transfection, Huh-7 cells transfected with pEZ-UBE2Q2 or vector plasmids and treated with cobalt chloride were harvested. Additionally, Huh-7 cells co-transfected with pEZ-UBE2Q2 and mutant ubiquitin plasmids—pCDEF-HA-Ub(K11) (all lysine was mutant to arginine except from wild-type lysine 11), pCDEF-HA-Ub(K48) (all lysine was mutant to arginine except from wild-type lysine 48), or pCDEF-HA-Ub(K63) (all lysine was mutant to arginine except from wild-type lysine 63)—were collected after 48 h. Cell pellets from each group were lysed in IP lysis buffer. Subsequent procedures were coherent with those previously described for co-immunoprecipitation (Co-IP).

UBE2Q2 expression in paired tissue samples and associated Kaplan-Meier survival curves for hepatocellular carcinoma (HCC) patients were extracted from The Cancer Genome Atlas (TCGA) database. Additional analyses of overall survival curves and correlations with UBE2Q2 in HCC were conducted using the GEPIA platform (http://​gepia.​cancer-pku.​cn). The data for single-gene enrichment analysis were sourced from the Gene Set Enrichment Analysis (GSEA) website (http://​www.​gsea-msigdb.​org/​gsea/​index.​jsp), with the analysis performed using GSEA-3.0 software.

Statistical analysis was conducted using GraphPad Prism version 8.0 (GraphPad Software) and SPSS version 22.0 (SPSS, USA). Data are expressed as mean ± standard deviation (SD). The Student’s t-test and non-parametric test were utilized for comparing differences between two groups, while two-way ANOVA was employed for analyses involving more than two groups. Correlations between UBE2Q2 expression and clinical parameters were determined using the chi-squared (χ2) test, with two-tailed P values. Significance levels for all tests were set as follows: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, with ns denoting ‘not significant’.

The expression levels of UBE2Q2 were evaluated in 80 paired tissue specimens utilizing quantitative real-time PCR (qRT-PCR). The analyses revealed a notable overexpression of UBE2Q2 in hepatocellular carcinoma (HCC) tissues compared with adjacent non-cancerous tissues (Fig. 1a), a finding that was further supported by data from The Cancer Genome Atlas (TCGA) (Fig. 1b). Furthermore, elevated UBE2Q2 expression was found to be significantly associated with decreased overall survival (OS) and disease-free survival (DFS) rates in HCC patients (Fig. 1c and d). Protein expression analyses through Western blotting (WB) and immunohistochemistry (IHC) consistently demonstrated a substantial increase in UBE2Q2 levels in HCC samples (Fig. 1e and g). The overexpression of UBE2Q2 was also prominently observed in HepG2, Hep3B, and Huh-7 cell lines when contrasted with LM3 and SK-Hep-1 cell lines, as evidenced by qRT-PCR and WB data (Fig. 1f). Immunofluorescence staining showed that UBE2Q2 primarily localizes to the cytoplasm (Fig. 1h). Additionally, high UBE2Q2 expression was significantly correlated with key clinical parameters, including increased tumor size (p = 0.010), higher Barcelona Clinic Liver Cancer (BCLC) stage (p = 0.003), and the presence of hepatocirrhosis (p = 0.005) (Table 1).

To investigate the biological role of UBE2Q2 in hepatocellular carcinoma (HCC) cells, siRNA was employed to effectively silence the gene in Huh-7 and Hep3B cell lines. We selected siRNA #1 and siRNA #2 for subsequent experiments after the detection of gene knockdown efficacy verified through quantitative real-time PCR (qRT-PCR) and WB analysis (Fig. 2a and b). Wound healing assays indicated that UBE2Q2 suppression hindered the repair and migratory abilities of these cells (Fig. 2c). Proliferative capacities were significantly reduced as evidenced by CCK-8 and colony formation assays (Fig. 2d and e).

Conversely, UBE2Q2 expression was enhanced in HCC-LM3 and Huh-7 cells through transfection with a UBE2Q2-expressing plasmid, and the success of this overexpression was confirmed (Fig. S1a, Fig. S1c). This overexpression of UBE2Q2 was associated with increased cell proliferation (Fig. S1d, Fig. S1f). Furthermore, migration and invasion assays showed that knockdown of UBE2Q2 led to the decreased migratory and invasive potential of Huh-7 and Hep3B cells (Fig. S2a) while overexpression of UBE2Q2 enhanced migratory and invasive capacity of Huh-7 and LM3 cells (Fig. S2b).

Furthermore, UBE2Q2 knockdown led to cell cycle arrest, particularly at the G2/M transition, and increased apoptotic rates in HCC cells (Fig. 3a, c). On the other hand, overexpression of UBE2Q2 produced antithetical outcomes (Fig. 3b, d).

For in vivo analysis, a subcutaneous tumor model was generated using Huh-7 cells transduced with lenti-virus carrying shUBE2Q2 or shControl vectors (Fig. 4a), where the silencing efficiency was quantified employing qRT-PCR and WB (Fig. S1b, Fig. S1c). The volume and weight of the tumors from animals inoculated with UBE2Q2-knockdown cells were significantly lower, demonstrating a substantial inhibition of HCC cell proliferation in vivo (Fig. 4b and c), and IHC analyses showed decreased expression of the proliferation marker Ki-67 in the tumors with UBE2Q2 knockdown (Fig. 4d).

To determine the impact of UBE2Q2 on HCC, a proteomics analysis was performed, revealing 209 upregulated proteins and 153 downregulated proteins in the UBE2Q2 overexpression group (Fig. 4e). Notably, several upregulated proteins were implicated in glycolysis, including PKM2, PFKL, LDHA, and GLUT1 (Fig. 4e). Enrichment analysis utilizing the Kyoto Encyclopedia of Genes and Genomes (KEGG) dataset corroborated that the Glycolysis/Gluconeogenesis pathway was among those enriched (Fig. S3a). This finding aligns with results from the Gene Set Enrichment Analysis (GSEA) of single gene expressions from the TCGA Liver Cancer RNA-seq data, where gene sets associated with glycolysis were predominantly upregulated in tumor tissues with high UBE2Q2 expression (Fig. S3b). These results led to the hypothesis that UBE2Q2 could play a role in modulating glycolytic processes in HCC. Assays for glucose uptake and lactate secretion confirmed that UBE2Q2 knockdown attenuated glycolysis in HCC cells, while its overexpression had the opposite effect (Fig. 4f and g, Fig. S3c, Fig. S3d). Moreover, energetic metabolism assessment through extracellular acidification rate (ECAR) and oxygen consumption rate (OCR) measurements showed that UBE2Q2 elevated glycolytic function and mitochondrial respiration capacity in Huh-7 cells (Fig. 4h, Fig. S3e, Fig. S3f). CCK-8 suggested that inhibition of glycolysis using 2-DG could block the proliferative effect of UBE2Q2 overexpression on HCC cells (Fig. S3g). The use of 2-DG effectively blocked the increase of lactate secretion and migration capacity in HCC cells caused by UBE2Q2 overexpression (Fig. S3h, Fig. S3i).

Delving deeper into the mechanism by which UBE2Q2 enhances glycolysis in HCC, we observed an upregulation of HIF1α in the UBE2Q2 overexpression cohort. The KEGG pathway enrichment analysis of our proteomic data also implicated the HIF1 signaling pathway (Fig. S3a). Building upon existing literature where HIF1α elevation was shown to activate glycolysis via induction of glycolytic enzymes and GLUTs [22, 23], and given our prior findings of similar protein upregulations, we surmised that UBE2Q2 overexpression would heighten HIF1α levels. IHC staining in subcutaneous tumors also demonstrated that the expression levels of glycolytic proteins such as PFKL, PKM2, GLUT1, LDHA and HIF1α were reduced upon UBE2Q2 knockdown (Fig. S4a). WB revealed HIF1α presence in Huh-7 and LM3 cells under normoxic conditions, further increased with UBE2Q2 overexpression. Correlated with HIF1α, proteins such as VEGFA, LDHA, and GLUT1 were upregulated in the overexpression context (Fig. 5a). Even under simulated hypoxia using cobalt chloride, UBE2Q2 overexpression continued to promote HIF1α expression (Fig. 5a). Besides, the expression level of UBE2Q2 did not affect the protein levels of VHL and HIF2α (Fig. 5a). Further investigation revealed increased transcription levels of HIF1α and associated genes in UBE2Q2-augmented cells, with unchanged HIF1β, HIF2α and MYC levels (Fig. S4b). Concurrent silencing of UBE2Q2 and forced overexpression of HIF1α, confirmed by WB (Fig. 5b), suggested that HIF1α overexpression could mitigate the reduced cellular proliferation caused by UBE2Q2 knockdown, as evidenced by CCK-8 and colony assays (Fig. 5c, Fig. S4c). Glucose uptake, lactate secretion, ECAR and OCR assays indicated HIF1α could recover glycolytic and mitochondrial respiration capabilities impaired by UBE2Q2 knockdown (Fig. S4d-Fig. S4h). Subcutaneous tumors demonstrated UBE2Q2 overexpression fostered tumor growth, an effect that was abrogated by the HIF1α inhibitor PX-478 (Fig. 5d and f). IHC confirmed the increased expression of HIF1α and its target proteins with UBE2Q2 overexpression, which was inhibited by PX-478, indicating UBE2Q2 may promote HCC glycolysis and tumorigenesis via HIF1α regulation. The IHC staining of CD31 indicated that up-regulated HIF1α promoted the infiltration of endothelial cells in tumor tissue (Fig. 5g, Fig. S4i).

Considering UBE2Q2’s role as a potential ubiquitin-conjugating enzyme, we preliminarily suspected an influence on HIF1α degradation. Nonetheless, UBE2Q2’s overexpression still elevated HIF1α under cobalt chloride culture, excluding an effect on HIF1α proline hydroxylation and VHL-mediated degradation (Fig. 5a line1), and upon treatment with the protein synthesis inhibitor CHX, HIF1α’s stability was unaffected (Fig. 6a). This led us to rule out UBE2Q2’s impact on HIF1α protein degradation. TCGA RNA-seq data supported a strong correlation between UBE2Q2 and HIF1α-associated genes (Fig. S5a). Therefore, it’s posited that UBE2Q2 bolsters HCC glycolysis predominantly by amplifying HIF1α transcription.

Current literature indicates that multiple signaling cascades, including WNT, MAPK, PI3K/AKT/mTOR, and NF-κB, can modulate the transcriptional activity of HIF1α [12‐16]. WB analysis revealed that UBE2Q2 overexpression increased the levels of phosphorylated p65 (p-p65), whereas β-catenin, ERK1/2, and Ras remained unchanged (Fig. 6b). Pharmacological blockade of specific pathways using corynoxeine, rapamycin, LY294002, and (-)-DHMEQ to inhibit ERK1/2, mTOR, PI3K, and NF-κB respectively, demonstrated that NF-κB inhibition markedly prevented the UBE2Q2-mediated upregulation of HIF1α (Fig. 6c). This led to an investigation into the NF-κB pathway activation correlated with UBE2Q2 overexpression and subsequent HIF1α transcription elevation. Further confirmation came from observing increased levels of p-p65, IκBα, and p-IκBα, indicating UBE2Q2-mediated canonical NF-κB pathway activation (Fig. 6d). Nuclear-cytoplasmic fractionation followed by Western blot suggested that UBE2Q2 overexpression fostered the nuclear localization of p-p65 and HIF1α, effects attenuated by (-)-DHMEQ (Fig. 6e). Additionally, knockdown p65 by shRNA in Huh-7 cells overexpressed UBE2Q2 significantly inhibited the transcription and protein expression of HIF1α (Fig. 6f and g). RNA sequencing analyses reinforced a strong correlation between UBE2Q2 and downstream NF-κB pathway-related genes (Fig. S5b). Consequently, UBE2Q2 appears to activate the NF-κB pathway, enhancing HIF1α transcription.

To elucidate the mechanism through which UBE2Q2 activates the NF-κB pathway in HCC cells, endogenous co-immunoprecipitation (Co-IP) using an anti-UBE2Q2 antibody and mass spectrometry were employed to pinpoint interacting proteins of UBE2Q2, complemented by searches in the National Center for Biotechnology Information (NCBI) database as a cross-reference [24]. Our focus narrowed to TNF receptor-associated factor 2 (TRAF2) and baculoviral IAP repeat-containing 2 (cIAP1), both ubiquitin ligases known for forming a complex where cIAP1/2-mediated ubiquitin chains are transferred to receptor-interacting serine/threonine kinase 1 (RIP1), subsequently facilitating the activation of the NF-κB pathway (Fig. 7a) [25]. UBE2Q2 was found to bind with cIAP1 in HCC cells (Fig. 7b). Building on prior research indicating cIAP1’s role in catalyzing K11, K48, and K63 polyubiquitination of RIP1 [26], and considering UBE2Q2 as a potential ubiquitin-conjugating enzyme, we posited that UBE2Q2 might alter RIP1’s ubiquitination. Ubiquitination assays confirmed UBE2Q2 overexpression promoted RIP1 ubiquitination (Fig. 7c). Enhanced RIP1 ubiquitination was observed specifically with transfection of ubiquitin plasmid containing all mutant lysine sites (lysine mutates to arginine) except from wild-type lysine 63 (HA-Ub(K63)), while transfection of ubiquitin plasmid containing all mutant lysine sites except from wild-type lysine 11 (HA-Ub(K11)) or transfection of ubiquitin plasmid containing all mutant lysine sites except from wild-type lysine 48 (HA-Ub(K48)) had no such effect (Fig. 7d). Immunoblotting with a K63-linkage-specific polyubiquitin antibody corroborated increased K63-linked ubiquitination of RIP1 upon UBE2Q2 overexpression (Fig. 7e). Analysis of Kaplan-Meier curves from the TCGA database linked the UBE2Q2/NF-κB/HIF1α axis to poor HCC patient prognosis (Fig. 7f), suggesting that targeting this axis could be pivotal in improving survival outcomes.