PIK3CB depletion restores immunosurveillance and enhances radioactive iodine response in hepatocellular carcinoma

Abstract

Hepatocellular carcinoma (HCC) has a high incidence and mortality rate despite the existence of various treatments, including iodine-125 (¹²⁵I) seed implantation. Nevertheless, resistance to these treatments persists as a challenge. Here, we identified PIK3CB as a key molecule involved in the nonresponse to ¹²⁵I and immune evasion in HCC and expounded its upstream modifiers. Differentially expressed genes in HCC tissues of ¹²⁵I non-responders or responders were screened using transcriptomics sequencing. PIK3CB expression was increased in the tumors of HCC ¹²⁵I non-responders. PIK3CB knockout increased HCC cell response to ¹²⁵I and inhibited tumor growth. PIK3CB induced PD-L1 expression to support the evasion of HCC cells from CD8⁺ T cells and NK cells-mediated immunosurveillance. The binding of ZBED4 at the PIK3CB promoter induced transcription of PIK3CB, and the knockout of ZBED4 enhanced the ¹²⁵I response and inhibited the immune evasion of HCC cells. Overexpression of ASB2 enabled ubiquitination of ZBED4. ASB2 overexpression increased ¹²⁵I sensitivity and inhibited immune evasion in HCC, which was reversed by overexpression of ZBED4 or PIK3CB. Consequently, these findings offer significant insights into the ASB2/ZBED4/PIK3CB regulatory axis, illuminating the potential for enhancing the efficacy of immune checkpoint blockade and ¹²⁵I seed brachytherapy in HCC treatment.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s10565-026-10229-5.

Introduction

Liver cancer has been identified as the fourth leading cause of cancer-related mortality on a global scale, and hepatocellular carcinoma (HCC) represents the most prevalent form of primary liver cancer (Sankar et al. 2024). Hepatectomy, as a major curative treatment, is appropriate for only a small population since preexisting liver dysfunction and the tumor location in the major vessels of the liver may limit the application of this option (Xie et al. 2023). Due to an insufficient number of donors and the unsatisfactory long-term efficacy of radiofrequency ablation (RFA) with or without transhepatic arterial chemotherapy and embolization (TACE), clinicians have explored more effective strategies (Wei et al. 2021). For instance, the combination of iodine-125 (¹²⁵I) seed brachytherapy and TACE has been revealed to be a more effective option for patients with HCC and portal vein-branch tumor thrombus than TACE alone (Hong et al. 2021). However, despite the advancements in therapies for advanced diseases, resistance remains a common challenge, and every available single-line treatment encounters some degree of these issues, highlighting the need for continued research and innovation (Winter et al. 2024).

Immunotherapy combination strategies, including anti-PD-1 and anti-PD-L1 monoclonal antibodies in conjunction with anti-angiogenic agents, are undergoing rigorous testing with the objective of expanding the population that exhibits responsiveness, countering resistance, and enhancing efficacy (Rimassa et al. 2023). Here, we identified the enrichment of the PD-L1 expression and PD-1 checkpoint pathway in cancer by the differentially expressed genes (DEGs) screened using transcriptome sequencing from patients with complete response/partial response (CR/PR) or stable disease (SD) after ¹²⁵I seed brachytherapy, and phosphoinositide 3-kinase subunit beta (PIK3CB) as a DEG enriched to this pathway that relates to prognosis of HCC patients. PI3Ks are lipid kinases and can be divided into three main classes: class I, II, and III, depending on their structure and substrate preference (Mazloumi Gavgani et al. 2018). Class I PI3Ks are further subdivided into two subgroups, IA and IB, and Class IA is comprised of a catalytic subunit, p110 (p110α, p110β, or p110δ, encoded by genes PIK3CA, PIK3CB, and PIK3CD, respectively) and a p85 regulatory subunit (Sun and Meng 2020). PIK3CB has been documented as a regulator of metastasis, specifically by modulating cell adhesion to collagen I in the context of pancreatic cancer (Qu et al. 2021). More importantly, cancer intrinsic and exosome-derived PIK3CB promoted the transcription of the PD-L1 promoter in esophageal squamous cell carcinoma cells (Xu et al. 2023), indicating its possible association with immune evasion. PIK3CB levels were found to be elevated in rectal adenocarcinoma tissues exhibiting a suboptimal response to preoperative radiotherapy, and enhanced PIK3CB expression was identified as a prognostic factor for both local recurrence-free survival and metastasis-free survival (Yu et al. 2014). Nevertheless, the functional role and mechanism of PIK3CB in immune evasion and ¹²⁵I irradiation response in HCC remain to be established.

CD8⁺ cytotoxic T cells and natural killer (NK) cells comprise the predominant components of the infiltrating T cells within the tumor microenvironment (TME) of HCC (Du et al. 2024). Given our preliminary prediction results that PIK3CB expression was conversely correlated with CD8⁺ and NK cells in HCC, we elucidated the function of PIK3CB in the evasion from CD8⁺ and NK cells-mediated immune surveillance and ¹²⁵I irradiation response in HCC and identified upstream regulators.

Materials and methods

Clinical samples

From February 2024 to March 2025, a total of 45 samples of HCC patients who underwent surgery combined with ¹²⁵I seed brachytherapy were included (mean age: 61.3, median age: 62, age range: 37–85), including 28 men and 17 women. According to the BCLC staging system, there were 14 cases of BCLC stage 0/A, 25 cases of BCLC stage B, and 6 cases of BCLC stage C. All patients underwent TACE and RFA, followed by ¹²⁵I radioactive seed implantation after TACE-RFA treatment. Tumor tissues via image-guided needle biopsy before ¹²⁵I seed implantation were collected, and patients were classified into CR/PR (n = 21) and SD (n = 24) based on their response after ¹²⁵I implantation. CR was defined as a complete disappearance of all visible lesions, which was maintained for a period of at least four weeks. PR was defined as the sum of the product of the two largest vertical diameters of the tumors, which was reduced by more than 50%. This reduction was sustained for a period exceeding four weeks, during which no further progression of the lesions was observed, and no new lesions emerged. SD was defined as either tumor regression or growth of less than 25%, with the absence of new lesions.

Transcriptome sequencing

RNA was extracted from fresh tumor tissue from patients with CR/PR (n = 5) and SD (n = 5) (all with BCLC Stage B) using the QIAwave RNA Plus Mini Kit (74636, Qiagen, Valencia, CA, USA). The MGIEasy Fast RNA Library Preparation Kit (940–002921-00, BGI, Beijing, China) and the MGIEasy Dual Barcode Cylization Kit (1000020570, BGI) were used to prepare sequencing libraries, which were sequenced on the MGISEQ-2000 sequencing platform. Fastq files were aligned using STAR (v 2.5.1b) and the human reference genome (GRCh38/hg38) and were quantified with featureCounts (v1.5.2). The R software package "DESeq2" (v. 1.40.0) was used for differential expression analysis, with FDR employed to recalibrate p-values. DEGs were identified using an adjusted p-value < 0.05 and |Log₂FoldChange|> 1 as the significance thresholds.

The resources of public data

The GEPIA database (Tang et al. 2017), including data from 364 HCC patients, was used to analyze survival rates and gene expression correlations in HCC patients. The CIBERSORT analysis tool on SangerBox (Shen et al. 2022), encompassing 363 HCC patient datasets from the TCGA project, was utilized to evaluate the correlation between gene expression and immune cell infiltration. The “JASPAR CORE 2024-Predicted Transcription Factor Binding Sites” track in the UCSC Genome Browser (Casper et al. 2026) was applied to analyze transcription factor binding sites within the PIK3CB promoter region (chr3: 138834749–138835095). UbiBrowser (Wang et al. 2022) was used to predict E3 ubiquitin ligases that target ZBED4.

Cell lines, transfection, and ¹²⁵I irradiation model

Human HCC cells Huh-7 (SNL-085) and PLC/PRF/5 (SNL-086) were purchased from Sunncell (Wuhan, Hubei, China), and HEK293T cells (CL-0005) and mouse HCC cells Hepa 1–6 (CL-0105) were procured from Procell (Wuhan, Hubei, China). They were all cultured in DMEM containing 10% FBS +1% P/S. The human NK cell line NK-92MI (CL-0533, Procell) was maintained in α-MEM (CM-0533, Procell) containing 0.2 Inositol, 0.1 β-mercaptoethanol, 0.02 mM Folic Acid, 100–200 U/mL recombinant IL-2, 12.5% HS, 12.5% FBS, and 1% P/S. All cells have been successfully identified via STR analysis. The culture conditions were 37 °C and 5% CO₂.

Terazosin hydrochloride (HY-B0371F, MCE, Monmouth Junction, NJ, USA) was used in subsequent experiments after treating HCC cells at 1, 5, 10, and 20 μg/mL for 24 h, and cells treated with the solvent DMSO served as controls.

Lentiviral vectors harboring overexpression (OE)-ASB2, OE-ZBED4, OE-PIK3CB, and their respective controls were acquired from Genechem (Shanghai, China). Additionally, to achieve specific knockout of ZBED4 and PIK3CB in cells, we obtained two distinct gRNA-based human PIK3CB CRISPR and human ZBED4 CRISPR lentiviral vectors. After infection of Huh-7, PLC/PRF/5, or Hepa 1–6 cells for 48 h, the medium was refreshed, followed by stable infection screening using puromycin or neomycin. The effect of the intervention was verified by RT-qPCR or Western blot detection of target gene expression. The gRNA sequences of PIK3CB were: 1#, ATCGTATTTACCCACGCTAC; 2#, TGTAGCGTGGGTAAATACGA. The gRNA sequences of ZBED4 were: 1#, GTCCCCCGATCGAATAACAG; 2#, CAGCAGCACCCTATACGACG.

Briefly, ¹²⁵I radioactive seeds were evenly distributed on a 3 cm diameter circular plate with 1 ¹²⁵I seed placed in the center of the plate. The HCC cell culture dish was placed on the circular plate, leaving a 5 mm gap, and the gap was filled with water. During irradiation, the cell culture dish was rotated by 22.5° for a more even dose distribution (Lu et al. 2016).

Colony formation assay

Huh-7 and PLC/PRF/5 cells in 6-well plates were irradiated at 0, 2, 4, 6, and 8 Gy for colony formation assay as previously described (Lin et al. 2012). Plating efficiency was defined as the number of colonies formed at 0 Gy/the number of seeded cells ×100% (Zhao et al. 2021). Single-hit multitarget modeling (SF = 1-(1-EXP(-D/D₀))^N) was used to analyze various parameters: Dq, D₀, N values, and changes in SF2 (Sun et al. 2020).

Co-culture of HCC and CD8⁺ T cells

CD8⁺ T cells, isolated from PBMCs of healthy donors using CD8 MicroBeads, human (130–045–201, Miltenyi Biotec Inc., Auburn, CA, USA), were cultured in RPMI-1640 medium and activated for 7 days using Dynabeads Human T-Activator CD3/CD28 (11161D, Thermo Fisher Scientific Inc., Waltham, MA, USA).

Huh-7 and PLC/PRF/5 cultured overnight in the culture dishes were co-cultured with activated CD8⁺ T cells at E: T = 8:1 (Zhang et al. 2024b) for 48 h at 37 °C under 5% CO₂. T cells were separated from cell debris, and the remaining surviving HCC cells were stained using crystal violet. The staining intensity was analyzed quantitatively using a spectrophotometer at an optical density (OD) of 570 nm. To inhibit the binding of PD-L1 to PD-1 in the co-culture system, we added atezolizumab (10 ng/mL, HY-P9904, MCE) to the co-culture system, and an equal concentration of isotype control human IgG (HY-P99001, MCE) served as the negative control.

Co-culture of HCC and NK cells

Huh-7 and PLC/PRF/5 cultured overnight in the culture dishes were co-cultured with NK-92 cells at the ratio of E: T = 5:1. After 24 h, NK cells were separated from cellular debris by PBS washing, and the remaining surviving HCC cells were stained using crystal violet, and the intensity of staining was analyzed quantitatively using a spectrophotometer at OD of 570 nm. Similarly, 10 ng/mL of atezolizumab or isotype control human IgG was added to the co-culture system.

Cellular cytotoxicity assay

The lactate dehydrogenase (LDH) content in the culture supernatant from the co-culture system was assayed using the L-LDH Activity Assay Kit (D799207, Sangon, Shanghai, China) to assess the cytotoxicity (Shi et al. 2016).

Elisa

The kits for Perforin (PFN) (A270341, Antibodies-online GmbH, Aachen, Germany), GZMB ELISA (A78219, Antibodies-online GmbH), interferon gamma (INF-γ) (A270350, Antibodies-online GmbH), and TNF-α (A270361, Antibodies-online GmbH) were used to assess cytokine levels in supernatant from the co-culture system.

Flow cytometry

At the end of the co-culture, the medium of the co-culture system was centrifuged at 400 g to collect the cell suspension. Cells were dispensed into tubes, and the Fc receptor was sealed using Human TruStain FcX (422,301, BioLegend, San Diego, CA, USA). For cell surface marker labeling, the cells were stained with antibodies to APC-CD8 (17–0088-42), APC-CD56 (17–0567-42), and FITC-NKG2D (11–5878-42) for 30 min at 4 °C in darkness. Subsequently, 0.05% Triton X-100 was utilized for permeabilization, followed by intracellular marker staining using FITC-IFN-γ (11–7319-82) for 30 min at 4 °C in darkness. All antibodies were obtained from Thermo Fisher.

Transwell assay

The cell supernatant from HCC cells after 24-h culture was used as a chemotactic migration inducer and added to the lower chamber of a 24-well Transwell insert (Corning Incorporated, Corning, ME, USA). CD8⁺ T cells or NK cells (1 × 10⁶ cells/well) were supplemented to the upper chamber of the Transwell insert. After 12 h, a flow cytometer was used to count the number of cells that had migrated to the lower chamber.

Chromatin immunoprecipitation (ChIP)

Cultured Huh-7 and PLC/PRF/5 cells were treated with 1% formaldehyde for 10 min at room temperature (RT), and glycine was supplemented to quench cross-linking. The cells were lysed in a lysis buffer. After sonication to break chromatin and centrifugation to obtain chromatin fragments (2% was left as input), the cell lysates were incubated overnight with an anti-ZBED4 (1:100, bs-13554R, Bioss Inc., Woburn, MA, USA) or IgG (1:100, ab172730, Abcam, Cambridge, MA, USA). After a 2-h incubation with protein A/G beads, the eluted immune complexes were de-crosslinked. After purification of the DNA fragments, the abundance of ZBED4 in the PIK3CB promoter was assessed by qPCR using promoter-specific primers against PIK3CB (LEFT PRIMER: CTGCCTCTCTCGATCACCTC; RIGHT PRIMER: GGGATGCAGGCCTGAGAG).

Dual-luciferase assay

The wild-type (WT) promoter sequence of PIK3CB and its corresponding mutant sequence (Mut) were inserted into the pGL4 luciferase vector (Promega Corporation, Madison, WI, USA). Huh7 and PLC/PRF/5 cells with stable knockout of ZBED4 were plated into 24-well plates and grown to 70% confluence. The cells were co-transfected with a luciferase reporter plasmid and the Renilla luciferase pRL-TK plasmid (which served as an endogenous standardized control) using Lipofectamine 3000 (L3000075, Thermo Fisher). After 24 h, a dual-luciferase reporter gene assay kit (E1910, Promega) was used.

Protein half-life measurement

HCC cells overexpressing ASB2 were treated with CHX at 100 µg/mL. Cells were collected at 0, 3, 6, and 9 h after CHX treatment, respectively, and the protein half-life was assessed by evaluation of ZBED4 protein expression in cells.

Co-immunoprecipitation (Co-IP)

HEK293T cells were transfected with the constructed HA-Ub, Flag-ZBED4, and Myc-ASB2 plasmids using Lipofectamine 3000 and cultured for 48 h. HCC cells overexpressing ASB2 and HEK293T cells transfected with the relevant plasmids were treated with 20 µM MG132 (S2619, Selleck, Houston, TX, USA) for 8 h and lysed. The lysates were centrifuged to obtain cytoplasmic proteins, and an immunoprecipitation reaction was carried out using anti-ZBED4 (1:100, bs-13554R, Bioss) or Flag antibody (1:50, 20,543–1-AP, Proteintech, Wuhan, Hubei, China) at 4 °C overnight. The immunoprecipitation complexes were enriched using Protein A/G magnetic beads, resuspended in SDS-PAGE buffer, and subjected to Western blot analysis.

RT-qPCR

Total RNA was extracted from tumor tissues from clinical HCC patients or HCC cells. cDNA was synthesized from purified RNA using the ZymoScript RT PreMix Kit (R3012, Zymo Research, Irvine, CA, USA). RT-qPCR was performed using ZymoTaq qPCR Premix (E2054, Zymo Research). To calculate relative expression changes, the 2^−ΔΔCt method was employed. The primer details can be found in Table 1.

Table 1.

Sequences of primer pairs

Gene Name	Forward Sequence (5’−3’)	Reverse Sequence (5’−3’)
PIK3CB	GGTAATCGGAGGATAGGGCAGT	CGGCAGTATGCTTCAAGGATGAC
ZBED4	AGCGGTTCCATAGCAACGTGCT	GGTGTCATCAAAAGAGGAAGCGC
ASB2	GGACATCTCCAACAAATCCCGAG	TTGGTGTCTGCGTTGTGCTGCA
GAPDH	GTCTCCTCTGACTTCAACAGCG	ACCACCCTGTTGCTGTAGCCAA

Animal assays

All animal studies and experiments were approved by the Animal Research Committee of Harbin Medical University (approval no. SYDW 2023–080). BALB/c nude and C57BL/6 J mice (both 4–6 weeks old, female) were procured from Shulaibao (Wuhan, Hubei, China).

For subcutaneous xenografts, 84 BALB/c nude mice were randomized into 7 groups of 12 mice each. Stably transfected Huh-7 cells (1 × 10⁷ cell) were injected into mice. When the tumors grew to approximately 60 mm³, three 0.6 mCi radioactive ¹²⁵I particles were implanted into the xenografts of six mice in each group, while the remaining six mice in each group served as non-irradiated controls. Tumor length (a) and width (b) were measured every 7 d after implantation, and the tumor volume (mm³) was calculated as (a × b²)/2. All mice were euthanized after 4 weeks of treatment, and the resected tumors were weighed, photographed, and made into paraffin sections. ¹²⁵I-mediated tumor inhibition rate (%) = (non-irradiated group total value—¹²⁵I-irradiated group total value)/non-irradiated group total value ×100.

For orthotopic tumors, 60 C57BL/6 mice were categorized into 5 groups of 12 mice each. Hepa1-6 cells with stable OE-NC, OE-ASB2, OE-ASB2 + OE-NC, OE-ASB2 + OE-ZBED4, and OE-ASB2 + OE-PIK3CB (3 × 10⁵ cells suspended in 0.025 mL of Matrigel) were administered under the left liver capsule. Six mice were implanted with three 0.6 mCi radioactive ¹²⁵I particles simultaneously, while the other six mice in each group served as non-irradiated controls. Three weeks after the injection, all mice were euthanized, and the liver tissues were harvested and made into paraffin sections.

Western blot, immunofluorescence staining, and immunohistochemistry (IHC)

Proteins were isolated from HCC tissues and cells as described following the indicated treatment (Lin et al. 2012). Protein expression of PIK3CB (1:1000, ab151549, Abcam), PD-L1 (1:1000, ab237726, Abcam), ZBED4 (1:1000, bs-13554R, Bioss), ASB2 (1:1000, A17923, ABclonal, Wuhan, Hubei, China), AKT1(1:1000, A17909, ABclonal), Phospho-AKT1 (S473) (1:1000, AP0140, ABclonal), NF-kB p65 (1:10,000, A19653, ABclonal), Phospho-NF-kB p65 (S536) (1:10,000, AP0124, ABclonal), Ubiquitin (1:4000, A19686, ABclonal), HA (1:4000, ab9110, Abcam), Flag (1:20,000, 20,543–1-AP, Proteintech), Myc (1:2000, 16,286–1-AP, Proteintech), and GAPDH (1:5000, 10,494–1-AP, Proteintech) was detected by Western blot.

HCC cells were harvested at 0.5, 6, 12, and 24 h following irradiation for analysis and subjected to immunofluorescent staining with anti-γ-H2AX (1:500, MA5-33,062, Thermo Fisher). The percentage (%) of cells with γ-H2AX foci > 10 in the nucleus was calculated. For mouse tumor tissues, paraffin-embedded sections of liver tumor tissues were stained with the primary antibodies to CD8 (1:200, GTX00493, GeneTex, Inc., Alton Pkwy, Irvine, CA, USA), GZMB (1:250, 701,395, Thermo Fisher), CD56 (1:200, 60,238–1-Ig, Proteintech), NKG2D (1:100, PA5-102,038, Thermo Fisher), and anti-γ-H2AX (1:500, MA5-33,062, Thermo Fisher).

Tumor tissue sections from clinical HCC patients or mice were subjected to IHC with antibodies against PIK3CB (1:500, ab232997, Abcam), ZBED4 (1:200, bs-13554R, Bioss), ASB2 (1:250, ab200370, Abcam), KI67 (1:50, GTX20833, GeneTex), CD8 (1:50, GTX16696, GeneTex), CD56 (1:100, GTX133217, GeneTex), and PD-L1 (1:300, 17,952–1-AP, Proteintech). The IHC score of proteins in tissue sections from clinical patients was scored as previously reported (Lin et al. 2012). The number of positive cells was counted for the tumor tissues of mice.

TUNEL assay

TUNEL Apoptosis Detection Kit (PF00009, Proteintech) was used. The apoptosis rate (%) in HCC cells or subcutaneous tumors was viewed and calculated under fluorescence microscopy.

Statistics and reproducibility

GraphPad Prism 10.4.2 software (GraphPad, San Diego, CA, USA) was used. Data are reported as the mean ± SEM. Based on a post hoc analysis using the GPower software (v 3.1.9.6), under a significance level of α = 0.05, the statistical power (1-βerro prob) of the experimental data obtained with the current sample size met the experimental requirements (> 0.8). Unpaired t-tests were used for analyses of two populations, and ANOVA followed by Tukey's or Sidak's test was used for multiple-comparison testing. The significance of the observed results was determined by a p-value less than 0.05.

Results

PIK3CB is associated with ¹²⁵I nonresponse in patients with HCC

Tumor tissues via image-guided needle biopsy before treatment with ¹²⁵I radioactive seed implantation were collected, and five samples from patients with CR/PR or SD were collected for transcriptome sequencing. DEGs were screened out (Fig. 1A) and subjected to KEGG pathway enrichment analysis (Fig. 1B). Among the top ten enriched pathways, there was only one pathway that exhibited a strong correlation with cancer, KO05235: PD-L1 expression and PD-1 checkpoint pathway in cancer. DEGs (PIK3CB/IFNG/NFATC2/CD28/LAT/PRKCQ/TLR2/CD3E/MAPK13/MAP2K3/BATF/MYD88/TLR4) enriched into this pathway were selected for the following analysis.

Fig. 1.

PIK3CB expression is associated with ¹²⁵I nonresponse in HCC patients. A) The volcano diagram of DEGs in tumor tissues of patients with CR/PR (n = 5) versus patients with SD (n = 5) after ¹²⁵I implantation. B) The bubble diagram for KEGG pathway enrichment analysis of DEGs. C-D) GEPIA database analysis of PIK3CB and MAPK13 on OS C) and DFS D) in HCC patients (n = 364). E) The expression changes of DEGs enriched in the KO05235 pathway were visualized using GeneRanking dot plot. F) IHC score of PIK3CB in tumor tissues of patients with CR/PR (n = 21) versus patients with SD (n = 24) after ¹²⁵I implantation. G-H) The mRNA G) and protein H) expression of PIK3CB in tumor tissues of patients with CR/PR (n = 21) versus patients with SD (n = 24) after ¹²⁵I implantation was examined using RT-qPCR and Western blot analysis. The data are presented as the mean ± SEM, and p values were calculated using an unpaired t-test. **** p < 0.0001

The prognostic significance of the expression of DEGs enriched in KO05235 for HCC patients was analyzed in GEPIA (Group Cutoff: median). A remarkable (p < 0.05) prognostic value for overall survival (OS) in patients with HCC was exhibited only by PIK3CB and MAPK13 (Fig. 1C). Also, PIK3CB had prognostic significance for disease-free survival (DFS) in patients, while MAPK13 showed insignificant value for DFS (Fig. 1D). In addition, a gene ranking dot plot of DEGs enriched in the KO05235 pathway was generated using TOmicsVis Cloud (https://shiny.hiplot.cn/tomicsvis-shiny/#). It was observed that PIK3CB exhibited the most significant change in expression (Log₂FoldChange = 3.5461590137285), indicating that it is a key driver of this pathway (Fig. 1E). Therefore, PIK3CB was selected for further study.

IHC for a larger sample size showed consistent results that the IHC scores of SD patients (n = 24) were higher than those of patients with CR/PR (n = 21) (Fig. 1F). Significantly increased expression of PIK3CB was also found in the tumors of SD patients (Fig. 1G, H).

PIK3CB loss enhances HCC cell sensitivity to ¹²⁵I irradiation

We performed PIK3CB depletion in HCC cells, and both gRNAs successfully targeted PIK3CB, mediating its expression loss in the cells (Fig. 2A). Next, we assessed cell sensitivity after treatment with ¹²⁵I by using a colony formation assay (Fig. S1A). Loss of PIK3CB reduced cell survival in the presence of irradiation, with parameters such as D₀, Dq, N, and SF2 all decreasing (Fig. 2B). γ-H2AX staining following ¹²⁵I irradiation revealed that PIK3CB knockout resulted in stronger γ-H2AX nuclear foci, increased DNA damage, and delayed clearance of γ-H2AX nuclear foci, indicating the presence of persistent DNA damage (Fig. 2C). TUNEL assay confirmed that PIK3CB loss enhanced apoptosis induced by ¹²⁵I (Fig. 2D).

Fig. 2.

PIK3CB loss enhances HCC cell response to ¹²⁵I irradiation. A) The protein expression of PIK3CB in Huh-7 and PLC/PRF/5 cells was examined using Western blot analysis. B) The colony-forming ability of HCC cells at different doses of ¹²⁵I irradiation was examined using colony-formation assays. C) γ-H2AX foci accumulation in cells at different time points (0.5, 6, 12, 24 h) following 4 Gy of ¹²⁵I was examined using immunofluorescence. D) Apoptosis of HCC cells at 4 Gy.¹²⁵I was examined using TUNEL. The experiments were repeated at least five times independently with similar results. The data are presented as the mean ± SEM, and p values were calculated using ANOVA. ** p < 0.01, *** p < 0.001, **** p < 0.0001

Loss of PIK3CB hinders HCC tumor growth and improves the efficacy of ¹²⁵I seeds in mice

We then established subcutaneous xenografts in vivo. To minimize animal use, we administered the Huh-7 cells infected with PIK3CB gRNA 1# into mice. Analysis of tumor growth in mice treated with or without ¹²⁵I seed showed that the loss of PIK3CB inhibited tumor growth and reduced tumor weight. In addition, ¹²⁵I seed suppressed tumor growth, and the tumor suppression rate was higher in cells lacking PIK3CB when treated with ¹²⁵I (Fig. 3A, B). Analysis of γ-H2AX expression levels in tumor tissue also confirmed that the loss of PIK3CB promoted sensitivity to ¹²⁵I, whereas γ-H2AX was virtually undetectable in tumor tissues that had not been treated with ¹²⁵I seed brachytherapy (Fig. 3C). IHC showed that both ¹²⁵I seed brachytherapy and PIK3CB deletion suppressed KI67 expression in tumor tissue, and the combination of PIK3CB deletion and ¹²⁵I seed brachytherapy had the most pronounced inhibitory effect on KI67 expression (Fig. 3D). However, ¹²⁵I seed brachytherapy increased PIK3CB expression in tumor tissues in the KO-NC group, which may be associated with acquired resistance to ¹²⁵I treatment (Fig. 3E). As revealed by the TUNEL assay in Fig. 3F, ¹²⁵I seed brachytherapy induced apoptotic cells in tumor tissue, and this effect was more pronounced in the absence of PIK3CB.

Fig. 3.

Depletion of PIK3CB hinders HCC tumor growth and improves ¹²⁵I efficacy in vivo. A-B) Tumor growth A) And tumor weight B) Formed by Huh-7 cells in mice treated with or without.¹²⁵I. C) Immunofluorescence detection of the number of γ-H2AX positive cells in tumor tissues of mice. D-E) IHC analysis of KI67 D) And PIK3CB E) Expression in tumor tissues of mice. F) TUNEL analysis of apoptosis in tumor tissues of mice. n = 6 mice per group. The data are presented as the mean ± SEM, and p values were calculated using ANOVA. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001

PIK3CB was conversely correlated with the infiltration of CD8⁺ T cells and NK cells in HCC

As shown in Fig. S2A, PI3K, to which PIK3CB serves as a subunit, activates the AKT signaling to mediate downstream NF-κB activation and PD-L1 expression, thereby suppressing T cell function to enable immune evasion. A negative correlation was identified between the expression of PIK3CB and the infiltration of CD8⁺ T cells and NK cells in HCC (Fig. 4A). Moreover, SD patients had significantly fewer CD8⁺ T cells and NK cells in their tumors than CR/PR patients (Fig. 4B). PIK3CB expression in tumor tissue was negatively correlated with the expression of CD8⁺ T cell and NK cell markers (Fig. 4C).

Fig. 4.

PIK3CB expression was significantly and negatively correlated with CD8⁺ T cell and NK cell infiltration in HCC. A) The correlation of PIK3CB expression with CD8⁺ T cell and NK cell infiltration in HCC (n = 363) was analyzed using the Sangerbox tool. B) IHC analysis of CD8 and CD56 expression in tumor tissues of patients with CR/PR (n = 21) after ¹²⁵I implantation versus patients with SD (n = 24). C) The correlation between PIK3CB expression and CD8 or CD56 expression in tumor tissues of patients with CR/PR (n = 21) after ¹²⁵I implantation versus patients with SD (n = 24) was analyzed using Pearson’s correlation analysis. D) Crystal violet staining analysis of HCC cell survival after co-culture of T cells with HCC cells with PIK3CB knockout. E) Cytotoxicity in the co-culture system after co-culture of T cells with HCC cells with PIK3CB knockout. F) ELISA detection of PFN, GZMB, and TNF-α secreted by T cells in the co-culture system after co-culture of T cells with HCC cells with PIK3CB knockout. G) Flow cytometry detection of IFN-γ⁺CD8⁺ activation in T cells after co-culture with HCC cells with PIK3CB knockout. H) Assessment of the chemotactic migration of CD8⁺ T cells induced by HCC cell culture supernatant. I) Crystal violet staining analysis of HCC cell survival after co-culture of NK cells with HCC cells with PIK3CB knockout. J) Cytotoxicity in the co-culture system after co-culture of NK cells with HCC cells with PIK3CB knockout. K) ELISA detection of IFN-γ and TNF-α secreted by NK cells in the co-culture system after co-culture of NK cells with HCC cells with PIK3CB knockout. L) Flow cytometry detection of NKG2D⁺CD56.⁺ NK cells after co-culture with HCC cells with PIK3CB knockout. M) Assessment of NK cell migration induced by HCC cell culture supernatant using a chemotaxis assay. N) γH2AX expression in HCC cells with PIK3CB knockout was examined using immunofluorescence staining. The experiments were repeated at least five times independently with similar results. The data are presented as the mean ± SEM, and p values were calculated using an unpaired t-test or ANOVA. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001

Co-culture systems comprising HCC cells and CD8⁺ T cells or NK cells were developed to analyze the immune evasion in vitro. In the co-cultured model of HCC cells and CD8 + T cells, the number of surviving HCC cells decreased after the loss of PIK3CB (Fig. 4D) and an increase in cytotoxicity (Fig. 4E). Depletion of PIK3CB in HCC cells induced the release of PFN, GZMB, and TNF-α (Fig. 4F). Flow cytometry showed that loss of PIK3CB in HCC enhanced the number of INF-γ⁺CD8⁺ T cells (Fig. 4G). PIK3CB knockout enhanced the ability of HCC cell culture supernatant to recruit CD8⁺ T cells (Fig. 4H).

In the co-culture model of HCC cells and NK cells, PIK3CB loss in HCC cells decreased the number of surviving cells (Fig. 4I) and increased cytotoxicity (Fig. 4J). The loss of PIK3CB in HCC cells also promoted the levels of INF-γ and TNF-α (Fig. 4K). Depletion of PIK3CB in HCC activated NK cells (NKG2D⁺CD56⁺) (Fig. 4L). Similarly, PIK3CB knockout enhanced the ability of culture supernatants produced by HCC cells to recruit NK cells (Fig. 4M). Immunofluorescence staining for γH2AX revealed that PIK3CB knockout did not induce a DNA damage response (Fig. 4N), suggesting that the suppression of immune evasion observed following PIK3CB knockout was independent of the DNA damage response.

The loss of PIK3CB in HCC decreased AKT and NF-κB phosphorylation levels, accompanied by a reduction in PD-L1 expression (Fig. S2B). We then induced HCC cells to overexpress PIK3CB (Fig. S2C), and a Western blot showed that AKT and NF-κB phosphorylation levels and PD-L1 expression increased with PIK3CB upregulation in HCC cells (Fig. S2D). Subsequently, we used HCC cells overexpressing PIK3CB to co-culture with CD8⁺ T cells and NK cells, respectively, and added Atezolizumab for PD-L1/PD-1 binding inhibition. First, the analysis of HCC cell survival after co-culture showed that the killing of HCC cells by CD8⁺ T cells and NK cells was weakened, and cell survival was promoted after overexpression of PIK3CB. The inhibition of PD-L1/PD-1 binding decreased the number of surviving cells, while cytotoxicity was instead increased (Fig. S2E, F). Overexpression of PIK3CB inhibited the activation of CD8⁺ T cells and NK cells, whereas blockade of PD-L1/PD-1 binding reversed this effect and exacerbated their activation (Fig. S2G).

ZBED4 mediates transcriptional activation of PIK3CB in HCC

Two transcription factor binding sites were present in the PIK3CB promoter (chr3: 138834749–138835095): ZBED4 and ZNF93 (Fig. S3A). It was observed in the GEPIA database that it was ZBED4, but not ZNF93, that had remarkable prognostic significance for OS and RFS in HCC patients (Fig. S3B). The GEPIA database analysis learned that the correlation between PIK3CB expression and ZBED4 expression in HCC was stronger than that of ZNF93. Therefore, ZBED4 was selected for further analysis (Fig. S3C). In accordance with the predictions made by CIBERSORT, ZBED4 expression was found to be inversely associated with the infiltration of CD8⁺ T cells and NK cells in HCC (Fig. S3D). Whether it mediates the response to ¹²⁵I irradiation and immune evasion through transcriptional activation of PIK3CB remains to be revealed.

Firstly, the expression of ZBED4 was enhanced in SD patients (Fig. 5A). In tumor tissues from SD patients and all HCC patients, ZBED4 expression levels were positively correlated with PIK3CB expression (Fig. 5B). ZBED4 expression was knocked out in HCC cells (Fig. 5C). The expression of PIK3CB was reduced with the downregulation of ZBED4 (Fig. 5D, E). Next, we constructed WT and Mut luciferase reporter vectors and analyzed the effect of the knockout of ZBED4 on the transcriptional activity of PIK3CB. ZBED4 knockout led to a reduction in luciferase activity when expressed in the WT reporter vector. However, no significant impact was observed on the Mut type (Fig. 5F). Finally, we performed ChIP experiments and found that the product immunoprecipitated by anti-ZBED4 showed a large enrichment of the PIK3CB promoter fragment (Fig. 5G).

Fig. 5.

ZBED4 mediates transcriptional activation of PIK3CB in HCC. A) IHC score of ZBED4 in tumor tissues of patients with CR/PR (n = 21) versus patients with SD (n = 24) after.¹²⁵I implantation. B) The correlation between ZBED4 and PIK3CB expression in tumor tissues of SD patients and all patients (CR/PR + SD) was analyzed using Pearson’s correlation analysis. C) Western blot detection of protein expression of ZBED4 after knockout of ZBED4 in HCC cells. D-E) The mRNA D) and protein E) expression of PIK3CB after knockout of ZBED4 in HCC cells was analyzed using RT-qPCR and Western blot analysis. F) The transcriptional activity of PIK3CB after the knockout of ZBED4 in HCC cells was analyzed using a dual-luciferase reporter assay. G) ZBED4 binding at the PIK3CB promoter in HCC cells was analyzed using a ChIP-qPCR assay. The experiments were repeated at least five times independently with similar results. The data are presented as the mean ± SEM, and p values were calculated using an unpaired t-test or ANOVA. *** p < 0.001, **** p < 0.0001

ZBED4 loss enhances HCC sensitivity to ¹²⁵I irradiation and inhibits immune evasion through inhibition of PIK3CB

In ZBED4-knockout cells, a combined genetic intervention involving PIK3CB overexpression was performed, and Western blot analysis confirmed increased PIK3CB expression (Fig. 6A). ZBED4 deletion decreased the colony formation capacity of HCC under ¹²⁵I irradiation, whereas combined overexpression of PIK3CB restored the colony number (Fig. S1B). The absence of ZBED4 reduced levels of D₀, Dq, N, and SF2, which were enhanced by overexpression of PIK3CB (Fig. 6B). ZBED4 deficiency increased nuclear foci of γ-H2AX in cells, and a large number of γ-H2AX-positive cells persisted 24 h after ¹²⁵I irradiation. Nevertheless, the concomitant expression of PIK3CB reduced the formation of γ-H2AX nuclear foci (Fig. 6C). The apoptosis induced by ZBED4 deficiency was significantly attenuated following combined overexpression of PIK3CB (Fig. 6D).

Fig. 6.

Deletion of ZBED4 enhances HCC sensitivity to ¹²⁵I irradiation and inhibits immune evasion through the inhibition of PIK3CB. A) The protein expression of PIK3CB after knockout of ZBED4 and overexpression of PIK3CB in HCC cells was analyzed using Western blot analysis. B) Colony-forming ability under different doses of ¹²⁵I irradiation after knockout of ZBED4 in HCC cells. C) Percentage of HCC cells positive for γ-H2AX nuclear foci at different time points (0.5, 6, 12, 24 h) following 4 Gy ¹²⁵I irradiation was detected by immunofluorescence. D) Apoptosis in HCC cells with knockout of ZBED4 and overexpression of PIK3CB at 4 Gy ¹²⁵I was examined using the TUNEL assay. E) Crystal violet staining analysis of cell survival of HCC cells with knockout of ZBED4 and overexpression of PIK3CB after co-culture with CD8⁺ T cells or NK cells, respectively. F) The cytotoxicity of HCC cells with knockout of ZBED4 and overexpression of PIK3CB in the co-culture system after co-culture with CD8⁺ T cells or NK cells, respectively, was examined using the LDH content assay. G) Cytotoxic activation of CD8⁺ T cells and NK cells after co-culture of HCC cells with knockout of ZBED4 and overexpression of PIK3CB was examined using flow cytometry. H) Western blot detection of PD-L1 expression in HCC cells with knockout of ZBED4 and overexpression of PIK3CB. The experiments were repeated at least five times independently with similar results. The data are presented as the mean ± SEM, and p values were calculated using ANOVA. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001

After co-culturing ZBED4-deficient HCC cells with CD8⁺ T cells and NK cells, respectively, HCC survived in smaller numbers (Fig. 6E) and had increased cytotoxicity (Fig. 6F), along with increased secretion of cytotoxic cytokines by CD8⁺ T cells and NK cells, respectively (Fig. S3E). ZBED4 deletion resulted in the activation of cytotoxic CD8⁺ T and NK cells (Fig. 6G). Finally, reduced PD-L1 expression was found in ZBED4-deficient cells (Fig. 6H). Combined overexpression of PIK3CB significantly reversed the immunosuppressive effects of ZBED4 deficiency.

ASB2-mediated ubiquitination modification of ZBED4 induces degradation of ZBED4

Transcriptome sequencing results revealed that the transcript levels of ZBED4 showed insignificant change in the SD and CR/PR samples, so we focused on the post-translational modifications of ZBED4. In UbiBrowser, we predicted E3 ubiquitin ligases with ZBED4 as a substrate (Fig. S4A). Cross-screening the E3 ligase with DEGs screened by transcriptome sequencing resulted in two intersections (Fig. S4B): ASB2 and MYLIP. The HCC patients with high ASB2 transcription had better DFS and generally had greater OS (Fig. S4C). By contrast, the expression of MYLIP did not have prognostic significance for HCC patients (Fig. S4D). Sequencing results showed that ASB2 expression was reduced in tumor tissues of SD patients (Log₂FoldChange = −2.49110306321521). Whether loss of ASB2 reduces degradation of ZBED4 protein, which leads to ZBED4-mediated induction of PIK3CB and subsequent ¹²⁵I nonresponse and immune evasion, remains to be revealed.

IHC detection of ASB2 expression in tumor tissues of patients with CR/PR or SD before ¹²⁵I showed that ASB2 expression was reduced in tumor tissues of SD patients relative to patients with CR/PR (Fig. 7A). Then, we performed the overexpression of ASB2 in HCC cells (Fig. 7B). With the elevated protein expression of ASB2, the expression of ZBED4 and PIK3CB in the cells was reduced (Fig. 7C). Overexpression of ASB2 also exacerbated the degradation of ZBED4 and shortened the protein half-life (Fig. 7D). Co-IP experiments detected an interaction between ASB2 and ZBED4 in HCC cells. When the ubiquitin–proteasome degradation pathway was blocked by MG132, overexpression of ASB2 increased the level of ZBED4 ubiquitination but did not affect ZBED4 protein expression (Fig. 7E). Using a tag-fusion plasmid, we found that the interaction between ASB2 and ZBED4 induced ZBED4 ubiquitination levels in HEK293T cells (Fig. 7F).

Fig. 7.

ASB2-mediated ubiquitination modification of ZBED4 induces degradation of ZBED4. A) IHC score of ASB2 in tumor tissues of patients with CR/PR (n = 21) versus patients with SD (n = 24) after ¹²⁵I implantation. B) RT-qPCR detection of ASB2 mRNA expression after overexpression of ASB2 in HCC cells. C) Western blot detection of ASB2, ZBED4, and PIK3CB protein expression after overexpression of ASB2 in HCC cells. D) The effect of ASB2 overexpression on ZBED4 protein stability was analyzed using a CHX-mediated protein half-life assay. E) Co-IP analysis of the interaction between ASB2 and ZBED4 proteins in HCC cells, and the effects of ASB2 overexpression on ZBED4 protein ubiquitination. F) The alteration of ZBED4 ubiquitination modification by the interaction of ASB2 with ZBED4 in HEK293T cells was analyzed using the ubiquitination immunoprecipitation assay. The experiments were repeated at least five times independently with similar results. The data are presented as the mean ± SEM, and p values were calculated using an unpaired t-test or ANOVA. ** p < 0.01, *** p < 0.001, **** p < 0.0001

ASB2 overexpression enhances the anti-tumor effects of ¹²⁵I, and upregulation of ZBED4 or PIK3CB reverses the role of ASB2

Furthermore, we overexpressed ZBED4 or PIK3CB in HCC cells overexpressing ASB2 (Fig. 8A, B). Overexpression of ASB2 repressed the surviving fraction (Fig. S1C) and enhanced γ-H2AX nuclear foci in the cells, and the rate at which these foci disappeared was slowed. Combined overexpression of ZBED4 or PIK3CB limited ¹²⁵I response (Fig. 8C, D). Overexpression of ASB2 suppressed the evasion of HCC cells from CD8⁺ T cells and NK cells, with a decreased number of surviving cells after co-culture (Fig. 8E), an increase in cytotoxicity (Fig. 8F), and activation of CD8⁺ T cells and NK cells (Fig. 8G). When we overexpressed ZBED4 or PIK3CB in the presence of enforced ASB2 expression, the immune evasion of HCC cells was enhanced, reversing the effect of ASB2 overexpression.

Fig. 8.

ASB2 overexpression enhances HCC cell response to ¹²⁵I and inhibits immune evasion. A-B) The mRNA A) and protein B) expression of ZBED4 and PIK3CB in cells after overexpression of ZBED4 or PIK3CB in HCC overexpressing ASB2 was examined using RT-qPCR and Western blot analysis. C) Colony-forming ability under different doses of ¹²⁵I irradiation in HCC cells. D) Percentage of HCC cells positive for γ-H2AX nuclear foci at different time points (0.5, 6, 12, 24 h) following 4 Gy ¹²⁵I irradiation was detected by immunofluorescence. E) Crystal violet staining analysis of cell survival of HCC cells after co-culture with CD8⁺ T cells or NK cells, respectively. F) The cytotoxicity of HCC cells in the co-culture system after co-culture with CD8⁺ T cells or NK cells was examined using the LDH content assay. G) Cytotoxic activation of CD8⁺ T cells and NK cells after co-culture of HCC cells was examined using flow cytometry. The experiments were repeated at least five times independently with similar results. The data are presented as the mean ± SEM, and p values were calculated using an unpaired t-test or ANOVA. ** p < 0.01, *** p < 0.001, **** p < 0.0001

ASB2 regulates immunosurveillance through the ZBED4-PIK3CB axis

Analysis of tumor growth in subcutaneous xenograft tumors formed by Huh-7 treated with ¹²⁵I showed that overexpression of ASB2 inhibited tumor growth and weight, whereas overexpression of ZBED4 and PIK3CB enhanced tumor growth and weight (Fig. 9A, B). ASB2 overexpression induced the number of γ-H2AX positive cells in tumor tissues, which was instead reduced after overexpression of ZBED4 or PIK3CB (Fig. 9C). TUNEL assay also corroborated that the pro-apoptotic effect of ASB2 overexpression on HCC cells was reversed by overexpression of ZBED4 or PIK3CB (Fig. 9D).

Fig. 9.

ASB2 overexpression inhibits the growth of subcutaneous xenograft tumors and orthotopic tumors. A) Growth of subcutaneous xenograft tumors formed by Huh-7 cells after overexpression of ASB2 alone or combined with ZBED4 or PIK3CB overexpression in BALB/c nude mice treated with ¹²⁵I. B) Weight of subcutaneous xenograft tumors. C) Immunofluorescence detection of the number of γ-H2AX-positive cells in subcutaneous xenografts of mice. D) TUNEL analysis of apoptosis in tumor tissues of mice. E) Observation of growth of orthotopic tumors formed by Hepa1-6 in C57 mice, with and without ¹²⁵I. IHC detection of ASB2 (F), ZBED4 (G), PIK3CB (H) expression in orthotopic tumors. I) Infiltration of GZMB⁺CD8⁺T cells in orthotopic tumors was detected by dual-labeling immunofluorescence. J) NKG2D⁺CD56⁺ NK cell infiltration in orthotopic tumors was detected by dual-labeling immunofluorescence. K) IHC detection of PD-L1 expression in orthotopic tumors. n = 6 mice per group. The data are presented as the mean ± SEM, and p values were calculated using ANOVA. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001

Subsequently, in the orthotopic tumor model established using Hepa1-6 cells, we observed multiple tumor nodules that may have resulted from intrahepatic spread. Compared with the group that did not receive ¹²⁵I, tumor growth was significantly inhibited in all groups that received ¹²⁵I. Overexpression of ASB2 inhibited tumor growth in liver tissue, while overexpression of ZBED4 and PIK3CB exacerbated the tumor load of liver tissues (Fig. 9E). IHC confirmed that overexpression of ASB2 (Fig. 9F) suppressed the expression of ZBED4/PIK3CB in tumor tissue, and both ZBED4 overexpression and PIK3CB overexpression restored PIK3CB expression in tumor tissue (Fig. 9G, H). Compared with the group that did not receive ¹²⁵I, tumor tissue from mice that underwent ¹²⁵I showed reduced ASB2 expression (Fig. 9F) and increased ZBED4/PIK3CB expression (Fig. 9G, H), which may be associated with acquired resistance. Dual-labeling immunofluorescence staining revealed that ¹²⁵I increased the number of CD8⁺GZMB⁺ cells and CD56⁺NKG2D⁺ cells in tumor tissue. Overexpression of ASB2 also increased the proportion of antitumor immune cells, whereas overexpression of ZBED4 and PIK3CB impaired the antitumor immune response in tumor tissue (Fig. 9I, J). Lastly, ¹²⁵I enhanced PD-L1 expression in tumor tissues, and ASB2 overexpression repressed PD-L1 expression. Overexpression of ZBED4 or PIK3CB promoted PD-L1 expression in tumor tissues (Fig. 9K).

Virtual screening to identify potential compounds targeting ASB2

Finally, to develop a novel combination therapy targeting ASB2 to enhance ¹²⁵I sensitivity, we conducted virtual screening within an FDA-approved compound library using the ASB2 3D structure predicted by the AlphaFold database (https://alphafold.ebi.ac.uk/). Based on the predicted score ranking (Fig. 10A), we performed molecular docking using Autodock Vina for the top 5 ranked drugs. Among the top 5 ranked drugs, terazosin hydrochloride (Mol ID: 63,016) exhibits the strongest binding affinity with ASB2 (Vina score = −5.26) (Fig. 10B).

Fig. 10.

Virtual screening to identify potential compounds targeting ASB2. A) The top five compounds targeting ASB2 based on virtual screening predicted scores. B) The visualization of five compounds targeting ASB2 using Autodock Vina. C) The effects of terazosin hydrochloride at 1, 5, 10, 20 μg/mL on ASB2 protein expression in HCC cells after 24 h of treatment were analyzed using Western blot analysis. D) Colony-forming ability under different doses of ¹²⁵I irradiation after terazosin hydrochloride (10 μg/mL for 24 h) in HCC cells. E) Percentage of HCC cells positive for γ-H2AX nuclear foci at different time points (0.5, 6, 12, 24 h) following 4 Gy ¹²⁵I irradiation and terazosin hydrochloride (10 μg/mL for 24 h) was detected by immunofluorescence. F) Crystal violet staining analysis of cell survival of HCC cells with terazosin hydrochloride (10 μg/mL for 24 h) after co-culture with CD8⁺ T cells or NK cells, respectively. The experiments were repeated at least five times independently with similar results. The data are presented as the mean ± SEM, and p values were calculated using an unpaired t-test or ANOVA. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001

Terazosin hydrochloride enhanced ASB2 protein expression in HCC cells in a concentration-dependent manner. When the concentration reached 10 μg/mL, further increases in concentration did not result in any additional increase in ASB2 expression (Fig. 10C). Terazosin hydrochloride at 10 μg/mL enhanced the sensitivity of HCC cells to ¹²⁵I (Fig. 10D, E). Treatment with terazosin hydrochloride also suppressed immune evasion in HCC cells, resulting in reduced cell survival following co-culture with T cells or NK cells (Fig. 10F).

Discussion

In the present study, a series of bioinformatic analyses and transcriptome sequencing were conducted to confirm the prognostic role of PIK3CB in predicting the poor outcomes of HCC patients. Subsequent analyses revealed the central role of PIK3CB expression in conferring the nonresponse to ¹²⁵I irradiation and evasion of HCC cells from CD8⁺ and NK cells-mediated immune surveillance.

Here, we linked the poor outcome of HCC patients predicted in the online database to the resistance to ¹²⁵I irradiation and immune evasion. By using two gRNAs targeting PIK3CB, we observed the enhanced γ-H2AX-positivity and apoptosis of HCC cells under ¹²⁵I at 4 Gy. Interestingly, programmed cell death 4 interacted with PIK3CB to promote the apoptosis of multiple myeloma cells (Liu et al. 2024). The maintenance of immune surveillance is achieved through the collaborative efforts of cytotoxic CD8⁺ T cells and NK cells (Chen et al. 2023). CD8⁺ T cells execute the elimination of target cells by way of granular exocytosis and apoptosis, producing IFN-γ and TNF-α to elicit cancer cell death (Hao et al. 2021). NK cells, defined by the CD56 expression, are an important kind of lymphocyte, and activated NKG2D upon binding to its ligand MICA/B enhanced the cytotoxicity of NK cells (Sajid et al. 2022). The activation and cytotoxicity of these two lymphocytes were found here to be significantly enhanced after co-culture with HCC cells with PIK3CB knockout. Comprehending the intricacies of PD-L1 regulation is paramount to elucidating the immunosuppressive microenvironment that fosters antitumor immunity (Han et al. 2025a, b). The findings of this study suggest that the impaired activation and toxicity of CD8⁺ T cells and NK cells induced by PIK3CB expression in HCC cells in the co-culture system were reversed by the suppression of PD-1/PD-L1 binding.

Pridham et al. showed that connexin 43 selectively bound to the PIK3CB, thereby conferring temozolomide resistance in glioblastoma (Pridham et al. 2022), which intrigued us to expound on the upstream modifier of PIK3CB in HCC. ZBED1 regulates the expression of cell proliferation-, cell cycle-, and apoptosis-associated genes (Jin et al. 2020). Moreover, overexpression of ZBED3 promoted lung cancer cell proliferation and invasiveness (Shi et al. 2019). Elevated ZBED4 expression in HCC patients was linked to a diminished response to immune checkpoint blockade, but an enhanced response to chemotherapy and the majority of targeted therapies (Ding et al. 2025). Here, by using two gRNAs targeting ZBED4, we found that the loss of ZBED4 enhanced the HCC response to ¹²⁵I and T cell and NK cell-mediated cytotoxicity. It has been demonstrated that the RNA 5-methylcytosine writer NSUN5 is responsible for promoting HCC cell proliferation via a mechanism that is dependent on ZBED3 (Gu et al. 2024). However, the transcription of ZBED4 was found unchanged in our sequencing results, which made us shift our attention to post-translational modification.

Ubiquitination is an important post-translational modification that empowers proteins to realize their biological functions and sustain their expression stability, and disruptions in protein ubiquitination and deubiquitination often accompany the HCC progression (Zhang et al. 2024a). By combining the DEGs screened out from transcriptome sequencing, the E3 ligases of ZBED4, and prognostic values, we identified ASB2. ASB2 knockdown has been demonstrated to impede the migration of human NK cells (Shin et al. 2021). In addition, preventing ASB2-mediated ubiquitination enhanced MLL1 stability and increased the tumoral expression of PD-L1/IDO-1, contributing to T-cell dysfunction and facilitating immune evasion and tumor development (Liu et al. 2023). These findings suggest an association between ASB2 and CD8⁺ T cells and NK cells in the TME. Still, its regulatory function in irradiation response has not been revealed. Here, the irradiation-inducing and immunomodulatory functions of ASB2 were both diminished by the overexpression of ZBED4 or PIK3CB. Gene expression of ASB2 was found to increase in samples of HCC tumor tissues from mice subjected to a high-dose regimen of ascorbate (Zhang et al. 2019). Here, we identified terazosin hydrochloride as the compound with the highest binding potential with ASB2 using virtual screening. The alpha1-adrenoceptor antagonist terazosin has been reported to induce prostate cancer cell death (Xu et al. 2003). Here, we found that terazosin hydrochloride at 10 μg/mL enhanced the sensitivity of HCC cells to ¹²⁵I and suppressed immune evasion in HCC cells.

However, this study has some limitations. Firstly, despite the fact that each experiment was independently repeated a minimum of five times and yielded consistent results, the modest sample size may constrain the generalizability of the findings. This is acknowledged as a limitation, and it is noted that future studies with expanded sample sizes will be important to further strengthen the accuracy and credibility of the conclusions. Furthermore, the mechanistic model under investigation did not incorporate the phenomenon of hypoxia, a key driver of radioresistance in solid tumors (Shi et al. 2024). Therefore, future studies should integrate hypoxia-related mechanisms to provide a more comprehensive understanding of radiosensitivity in HCC. Moreover, endothelial cell PI3Kβ inactivation with angiogenesis inhibitor reduces microvessel turnover and decreases heterogeneity of the TEM (Azad et al. 2020), while angiogenesis has been implicated in the HCC progression (Han et al. 2025b; Han et al. 2024). Therefore, future research should explore the broader functions of PIK3CB in angiogenesis, which might provide a more comprehensive understanding of its contributions to HCC progression. Lastly, while the present study provides direct evidence from cell-based experiments that teradosin hydrochloride enhances ASB2 activation and increases the radiosensitivity of HCC cells to ¹²⁵I, it is acknowledged that the absence of animal experiments limits the translational scope of these findings. In vivo validation will be essential to confirm efficacy, safety, and therapeutic relevance in a physiological context, and it represents an important direction for future research.

Conclusion

In summary, we elucidated the resistance mechanism of the ASB2/ZBED4/PIK3CB axis to ¹²⁵I irradiation and validated their function in T cell- and NK cells-mediated immune evasion. ASB2 regulates ZBED4 stability, and ZBED4 induces the expression of its target gene PIK3CB in HCC (Fig. 11).

Fig. 11.

Graphical abstract

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

YRG designed the study and performed all the experiments and drafted the manuscript. RHA participated in the analysis of the data. WBW performed the experiments and acquired the data. YLC supervised the study and revised the manuscript. All authors read and approved the final manuscript.

Funding

Financial support was provided by the National Natural Science Foundation of China: Acceptance number 8227071164.

Data availability

All data supporting the findings of this study are available within the paper and its Supplementary Information.

Declarations

Ethical approval

This research followed the Declaration of Helsinki. The use of human specimens has been approved by the ethics board of Harbin Medical University Affiliated Cancer Hospital (approval number: JS2024-68). Informed consent was gained from the patients. All animal studies and experiments were approved by the Animal Research Committee of Harbin Medical University (approval no. SYDW2023-080) and comply with the ARRIVE guidelines.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.