Integrative analysis of HMGB3 in esophageal carcinoma: expression profile, prognostic significance, and immune microenvironment associations

Introduction

Globally, esophageal carcinoma (ESCA) ranks among the most aggressive tumors of the gastrointestinal system (1,2). In China, ESCA is the seventh most frequently diagnosed malignancy and fifth most common cause of cancer mortality, being responsible for an estimated 224,000 new cases and 187,500 deaths in 2022 (3). The disease is mainly divided into two histopathological forms: esophageal squamous cell carcinoma (ESCC) and esophageal adenocarcinoma (EAC) (4). Therapeutic approaches for ESCA typically encompass endoscopic excision, surgical intervention, systemic chemotherapy, radiotherapy, targeted molecular therapy, and immunotherapy (5). Because of the insidious clinical symptoms of early-stage ESCA, most patients are diagnosed at intermediate or late stages. Even with advancements in available treatments, the outlook for ESCA remains dismal, as the 5-year survival rate ranges from 15% to 25% (6). Early detection is critical for better clinical outcomes; however, the absence of effective diagnostic markers and prognostic indicators remains a major obstacle. Consequently, the identification and development of innovative biomarkers are vital for improving the early diagnosis and outcome of ESCA.

The high mobility group box (HMGB) protein family, comprising HMGB1, HMGB2, HMGB3, and HMGB4, is primarily localized to the nucleus. These proteins are essential for regulating various cellular processes, including DNA replication, transcription, recombination, and repair (7-9). In addition, these proteins operate as cytokines that participate in immune and inflammatory responses triggered by infection or injury (10,11). HMGB3, encoded in the Xq28 region of the X chromosome, serves as a chromatin-binding protein (9). It exhibits predominant expression in embryonic tissues and hematopoietic stem cells of the bone marrow, whereas it is minimally expressed or absent in most other normal tissues (12,13). HMGB3 is involved in tumor initiation, development, metastasis, and immune escape (14,15). Elevated HMGB3 expression has been observed in various cancers, with research linking high HMGB3 expression to poorer patient prognosis. As an example, ovarian cancer exhibits pronounced HMGB3 overexpression, which is correlated with adverse clinical outcomes. HMGB3 drives cell growth, promotes metastatic spread, and maintains stemness by engaging the MAPK/ERK signaling cascade (16). In breast cancer, HMGB3 inhibition markedly suppresses proliferation and increases responsiveness to paclitaxel (17). In colorectal cancer, HMGB3 functions as an oncogene, driving both proliferation and migration (18). In a similar context, Liu and colleagues demonstrated that HMGB3 is an important contributor to small-cell lung cancer progression, as it enhances migratory capacity and serves as a robust prognostic marker (19). Despite these findings, its involvement in ESCA remains unclear, necessitating further investigation.

To address the need for effective biomarkers for early detection and prognostic stratification in ESCA and the incompletely defined role of HMGB3 in this malignancy, this study systematically evaluated HMGB3 expression patterns, clinical significance, and associations with the immune microenvironment using bioinformatic approaches, followed by in vitro validation. We aimed to improve clinical risk assessment and prognostic prediction in ESCA. We present this article in accordance with the MDAR reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1-0362/rc).

Methods

HMGB3 gene expression

HMGB3 expression in tumor and normal tissues was assessed using the Tumor Immune Estimation Resource (TIMER) database (https://cistrome.shinyapps.io/timer/). Messenger RNA (mRNA) expression profiles and clinical data for ESCA, including 184 tumor and 13 normal samples, were retrieved from The Cancer Genome Atlas (TCGA, https://portal.gdc.cancer.gov/). The expression data were converted to transcripts per million and log2-transformed for subsequent analyses. Violin plots with overlaid boxplots were generated using the ggpubr package to visualize differences in expression between ESCA tumors and normal tissues, and paired scatter plots for matched tumor-normal samples were generated using the ggplot2 package in R (The R Foundation for Statistical Computing, Vienna, Austria).

Analysis of HMGB3 expression and clinicopathological features

To explore the association between HMGB3 expression and clinicopathological parameters in ESCA, the University of Alabama at Birmingham Cancer (UALCAN) database (http://ualcan.path.uab.edu/), which is based on TCGA RNA-seq expression and corresponding clinical data, was utilized. HMGB3 mRNA expression was compared across subgroups stratified by sex, tumor grade, nodal metastasis status, clinical stage, pathological type, and TP53 mutation status. The Human Protein Atlas (HPA; https://www.proteinatlas.org/) was used to characterize HMGB3 mRNA expression in ESCA cell models, with normalized transcript levels extracted to assess its expression profiles across multiple ESCA cell lines. In addition, genomic alterations of the HMGB3 gene were analyzed using cBioPortal (http://www.cbioportal.org/) based on TCGA-ESCA data, enabling characterization of alteration frequency, mutation spectra, and copy number variation profiles to define the genomic alteration landscape of HMGB3 in ESCA.

Analysis of the diagnostic value of HMGB3

The diagnostic performance of HMGB3 was evaluated using ESCA tumor and normal samples from the TCGA-ESCA dataset. HMGB3 expression data were normalized using the voom function in the limma package, and receiver operating characteristic (ROC) curves were generated with the pROC package in R. The area under the curve (AUC) and corresponding 95% confidence intervals (CIs) were calculated to quantitatively assess the diagnostic accuracy of HMGB3, and an AUC greater than 0.7 was considered indicative of acceptable diagnostic performance.

Survival analysis

Time-dependent ROC curves were generated via the time ROC package in R, and AUCs were calculated at different time intervals. Patients were grouped into high and low HMGB3 expression categories based on the median expression level. Overall survival (OS) differences between the two groups were evaluated using the survival and survminer packages in R, with the log-rank test employed to determine statistical significance. Cox proportional hazards models, both univariate and multivariate, were fitted using the coxph function in the survival package. Pearson’s Chi-squared test was used to assess associations between HMGB3 expression and clinicopathological features. The Gene Expression Profiling Interactive Analysis (GEPIA) database (http://gepia.cancer-pku.cn) was additionally used to evaluate the effect of HMGB3 expression on OS and disease-free survival (DFS) in ESCA.

Protein-protein interaction (PPI) network construction and functional enrichment analysis

In the TCGA-ESCA cohort, patients were divided into high- and low-HMGB3 expression groups based on the median expression level. Differentially expressed genes (DEGs) between the two groups were identified using the limma package in R without adjustment for additional clinical covariates. The limma package employs a linear modeling framework with weighted least squares to estimate expression differences between groups. Variance estimates were stabilized using empirical Bayes methods to generate moderated t-statistics, and P values were corrected for multiple comparisons using the false discovery rate (FDR) method. Genes with |log₂ fold change| >1 and FDR <0.05 were considered significantly differentially expressed.

The PPI network was constructed by importing the DEGs into the STRING database (https://cn.string-db.org/), with the species restricted to Homo sapiens and a high-confidence interaction score threshold set at >0.7. Only interactions meeting this criterion were retained. After excluding isolated nodes, the remaining genes and their interactions were visualized using Cytoscape (version 3.9.1) to facilitate exploration of the overall interaction structure and to identify potential hub genes and key subnetworks.

For functional annotation of DEGs, the clusterProfiler package in R was employed to perform Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses. Over-representation analysis was conducted with multiple testing correction using the Benjamini-Hochberg method, and GO terms or KEGG pathways with FDR <0.05 were considered significantly enriched. In addition, gene set enrichment analysis (GSEA) based on KEGG gene sets was applied to identify pathways significantly associated with HMGB3 expression (FDR <0.05). The top 10 enriched pathways were visualized using the enrichplot package.

Associations of HMGB3 with immune cell infiltration and immune checkpoint expression

To improve the robustness of immune infiltration assessment, we used a multi-algorithm strategy integrating Microenvironment Cell Populations-counter (MCP-counter), Cell-type Identification by Estimating Relative Subsets of RNA Transcripts (CIBERSORT), Estimation of STromal and Immune cells in MAlignant Tumor tissues using Expression data (ESTIMATE), and TIMER. These complementary approaches provided information on immune cell abundance, stromal and immune components, and immune cell-related marker expression. Associations between HMGB3 expression and the tumor microenvironment (TME) in ESCA were therefore interpreted across these methods rather than inferred from a single algorithm. In TCGA-ESCA samples, immune infiltration by 10 distinct cell types was quantified using the MCP-counter package in R, with correlations between HMGB3 expression and immune cell proportions calculated using the cor function. The relative abundance of 22 immune cell populations was assessed via the CIBERSORT algorithm. Group-wise differences in immune cell composition between the high and low HMGB3 expression groups were visualized using ggplot2, and scatter plots were produced to depict links among HMGB3 expression, immune cell infiltration, and immune checkpoint markers. The ESTIMATE algorithm was used to obtain immune and stromal scores, whereas the TIMER database was consulted to investigate the associations of HMGB3 with immune cell markers, including B cells, T cells, CD8⁺ T cells, monocytes, tumor-associated macrophages (TAMs), M1 macrophages, M2 macrophages, neutrophils, natural killer (NK) cells, and dendritic cells (DCs).

HMGB3 and drug sensitivity

Using the Gene Set Cancer Analysis (GSCA) platform, which integrates information from the Genomics of Drug Sensitivity in Cancer project, we analyzed the relationships of HMGB3 mRNA levels with the sensitivity to multiple anticancer agents. A positive correlation suggested that increased HMGB3 expression is associated with drug resistance, whereas a negative correlation suggested enhanced drug sensitivity.

Cell culture and transfection

KYSE510 and KYSE140 human ESCA cells were cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS) and penicillin/streptomycin. The Het-1A human esophageal epithelial cell line was maintained in a serum-free medium containing the necessary supplements and penicillin/streptomycin. Cells were cultured at 37 ℃ under humidified conditions with 5% CO₂. Lentiviral vectors were supplied by Jikai (Shanghai, China). KYSE510 and KYSE140 cells were infected with HMGB3 short hairpin RNA (shRNA) or negative-control (NC) lentivirus. Twelve hours after transduction, the medium was replaced with fresh complete medium, and cells were further cultured. At 72 h, fluorescence microscopy was used to evaluate cell morphology and infection efficiency. Stable cell lines were generated by puromycin selection at 2 µg/mL in preparation for downstream assays.

Reverse transcription quantitative polymerase chain reaction (RT-qPCR)

RNA was isolated from cells using TRIzol (), after which complementary DNA (cDNA) was generated via reverse transcription. RT-qPCR was conducted using SYBR Green Master Mix on the Roche LightCycler 480 II platform (). ACTB served as the reference gene, and relative expression values were calculated using the 2^−∆∆Ct approach.

Cell Counting Kit-8 (CCK-8) cell proliferation assay

For the proliferation assay, 3,000 cells per well were seeded into 96-well plates with three replicate wells per condition. After 24, 48, 72, 96, and 120 h of culture, 10 µL of CCK-8 reagent was introduced into each well, followed by incubation for 3 h. The absorbance of each well at 450 nm was then measured.

Cell cycle analysis

In total, 2×10⁵ cells were seeded into each well of 6-well plates (three wells per condition). Cells were harvested, fixed in 70% ethanol at 4 ℃ for 2 h, treated with RNase A, and stained with propidium iodide. The cell cycle distribution was subsequently assessed using a BD flow cytometer.

Transwell assay

Matrigel was placed in a Transwell chamber and incubated at 37 ℃ for 6 h. Then, 200 µL of the cell suspension were added to the upper chamber, and 20% FBS-supplemented RPMI 1640 medium was added to the lower chamber. After 36 h of incubation, the cells were fixed in methanol, stained with crystal violet, washed with PBS, and photographed under a microscope in randomly selected fields. Each condition was assessed in triplicate.

Western blotting

Total protein was extracted by lysing cells in radioimmunoprecipitation buffer supplemented with phenylmethanesulfonyl fluoride (Pioneer Technology, Xi’an, China). Protein concentrations were measured using the bicinchoninic acid assay (Beyotime Biotechnology, Shanghai, China). Proteins were separated on 10% sodium dodecyl sulfate-polyacrylamide gels and transferred to polyvinylidene fluoride membranes (Millipore, Billerica, MA, USA). Membranes were blocked for 20 min at room temperature and then incubated overnight at 4 with anti-HMGB3 (1:1,000, ab75782, Abcam, Cambridge, UK) and anti-GAPDH (1:10,000, ab181602, Abcam). After washing, the membranes were incubated with anti-rabbit secondary antibody (1:2,000, #7074, Cell Signaling Technology, Danvers, MA, USA) for 1.5 h at room temperature. Visualization was achieved using enhanced chemiluminescence, and quantification was conducted using ImageJ software (US National Institutes of Health, Bethesda, MD, USA).

Clinical samples and immunohistochemistry (IHC)

The ESCA tissue chip (HEsoS150CS03) was obtained from Shanghai Outdo Biotech (Shanghai, China), and it includes both cancer tissue and normal esophageal tissue. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. Paraffin sections were subjected to dewaxing, hydration, antigen retrieval, inhibition of endogenous peroxidase activity, and serum blocking. The sections were incubated overnight at 4 ℃ with the primary anti-HMGB3 antibody (1:2,000, ab75782, Abcam), followed by exposure to an HRP-linked secondary antibody (#8114, Cell Signaling Technology). Diaminobenzidine was applied to reveal antibody staining, followed by hematoxylin counterstaining. Staining intensity scores were assigned as 0 for negative, 1 for weak, 2 for moderate, and 3 for strong. The proportion of positively stained cells was scored as 0 (0%), 1 (1–25%), 2 (26–50%), 3 (51–75%), or 4 (>75%). The final HMGB3 IHC score was obtained by multiplying the staining intensity score by the staining positivity rate score.

Statistical analysis

Statistical analyses were performed using R software (version 4.1.10). t-tests were applied to compare continuous variables, whereas the Chi-squared test was used to evaluate the associations of HMGB3 expression with clinicopathological parameters. Relationships among the variables were analyzed using Spearman’s or Pearson’s correlation, and P<0.05 was regarded as statistically significant.

Results

HMGB3 is widely upregulated in multiple cancers and markedly overexpressed in ESCA

The TIMER database was used to analyze HMGB3 mRNA expression in tumor and normal tissues. As presented in Figure 1A, HMGB3 expression was significantly higher in tumor tissues than in normal tissues in numerous cancers, such as ESCA, bladder urothelial carcinoma (BLCA), breast invasive carcinoma (BRCA), cholangiocarcinoma, colon adenocarcinoma, head and neck squamous cell carcinoma (HNSC), kidney chromophobe, kidney renal clear cell carcinoma, kidney renal papillary cell carcinoma, liver hepatocellular carcinoma (LIHC), lung adenocarcinoma (LUAD), lung squamous cell carcinoma (LUSC), prostate adenocarcinoma (PRAD), rectum adenocarcinoma, stomach adenocarcinoma, and uterine corpus endometrial carcinoma (UCEC). These results indicate that HMGB3 is broadly upregulated across multiple cancer types, suggesting a potential role as a pan-oncogene involved in tumorigenesis and cancer progression.

Figure 1 Pan-cancer analysis of HMGB3 expression and diagnostic value in ESCA. (A) Comparative HMGB3 expression in 32 tumor types and normal tissues. (B,C) Elevated HMGB3 expression in TCGA-ESCA and paired normal tissues. (D) ROC curves evaluating HMGB3’s diagnostic performance for ESCA. (E) Higher HMGB3 expression in ESCA than in normal tissues in the UALCAN database. ***, P<0.001. AUC, area under the curve; CI, confidence interval; ESCA, esophageal carcinoma; HMGB3, high mobility group box 3; ROC, receiver operating characteristic; TCGA, The Cancer Genome Atlas; TPM, transcripts per million; UALCAN, University of Alabama at Birmingham Cancer.

Analysis of the TCGA-ESCA dataset revealed that HMGB3 expression was markedly elevated in both unpaired and paired ESCA tissues compared with that in normal samples (Figure 1B,1C). ROC analysis demonstrated an exceptional AUC of 0.957 for HMGB3 (Figure 1D). Consistently, data from UALCAN online database confirmed that HMGB3 mRNA expression was increased in ESCA tissues (Figure 1E).

HMGB3 overexpression is correlated with clinicopathological characteristics of ESCA

We employed the UALCAN database to analyze HMGB3 mRNA expression in ESCA patients stratified by clinicopathological features. Compared with normal esophageal tissue, HMGB3 upregulation in ESCA was particularly evident across several clinicopathological factors, including sex, tumor grade, histological subtype, lymph node metastasis, and clinical stage (Figure 2A-2F). Notably, HMGB3 expression was higher in male patients (P=0.003) than in female patients and higher in ESCC (P<0.001) than in EAC (Figure 2A,2E). A notable increase in HMGB3 expression was detected in TP53-mutated tumors relative to that in wild-type TP53 tumors (P<0.001, Figure 2F).

Figure 2 HMGB3 is overexpressed in ESCA, and it exhibits distinct associations with clinicopathological features and tumor progression. HMGB3 expression in ESCA based on UALCAN database analysis stratified by gender (A), tumor grade (B), nodal metastasis status (C), cancer stage (D), tumor histology (E), and TP53 mutation status (F). (G,H) IHC staining of HMGB3 in in ESCA and normal tissues (magnification). (I-K) The associations of HMGB3 expression with T stage (I), N stage (J), and pathological stage (K). *, P<0.05; **, P<0.01; ***, P<0.001. ESCA, esophageal carcinoma; HMGB3, high mobility group box 3; IHC, immunohistochemical; N, lymph node; ns, not significant; T, tumor; TCGA, The Cancer Genome Atlas; UALCAN, University of Alabama at Birmingham Cancer.

To comprehensively validate HMGB3 dysregulation at the protein level, IHC was performed using 46 paired ESCA and adjacent noncancerous tissues. The results revealed a pronounced and statistically significant upregulation of HMGB3 in tumor tissues (Figure 2G,2H). Moreover, HMGB3 expression displayed a stepwise increase in parallel with T stage, N stage, and overall clinical stage progression (Figure 2I-2K and Table 1), indicating a close association between HMGB3 overexpression and disease progression.

Genetic alterations of HMGB3

Using cBioPortal, we found that HMGB3 mutations in ESCA predominantly involved gene amplification, increased mRNA levels, and multiple alterations (Figure 3A). As illustrated in Figure 3B, gain and amplification were the most common copy number variations. Given that gene amplification typically results in increased transcript levels, this finding provides a potential genetic explanation for the elevated HMGB3 expression observed in ESCA.

Figure 3 Genetic alterations of HMGB3 in ESCA, including mutation types (A) and copy number variations (B). CNA, copy number alteration; ESCA, esophageal carcinoma; HMGB3, high mobility group box 3; mRNA, messenger RNA; RNA-seq, RNA-sequencing; RSEM, RNA-Seq by Expectation Maximization; VUS, variant of unknown significance.

Elevated HMGB3 expression indicates poor prognosis in ESCA

In the TCGA-ESCA dataset, ROC curve analyses were performed to evaluate the prognostic utility of HMGB3 expression. The AUCs for predicting 1-, 3-, and 5-year survival were 0.57, 0.65, and 0.78, respectively (Figure 4A). Patients with ESCA were stratified into high- and low-HMGB3 expression groups based on the median expression level. Patients exhibiting higher HMGB3 expression had markedly worse survival than those with lower expression (Figure 4B). Subgroup analysis of patients aged ≥55 years (P<0.001) and those with tumors (P=0.03) indicated that OS was significantly associated with high HMGB3 expression (Figure 4C,4D). GEPIA database analysis illustrated that HMGB3 expression was inversely associated with both OS and DFS in ESCA (Figure 4E,4F).

Figure 4 Prognostic value of HMGB3 in ESCA. (A) ROC curves for 1-, 3-, and 5-year survival prediction. (B) Kaplan-Meier survival analysis for the high and low HMGB3 expression groups in TCGA-ESCA. (C,D) Subgroup survival analysis in TCGA-ESCA. (E,F) Correlation of HMGB3 expression with OS and DFS in GEPIA. AUC, area under the curve; CI, confidence interval; DFS, disease-free survival; ESCA, esophageal carcinoma; GEPIA, Gene Expression Profiling Interactive Analysis; HMGB3, high mobility group box 3; HR, hazard ratio; OS, overall survival; ROC, receiver operating characteristic; TCGA, The Cancer Genome Atlas; TPM, transcripts per million.

Univariate Cox regression identified cancer status [hazard ratio (HR) = 2.871, P<0.001], N stage (HR =1.431, P=0.004), M stage (HR =1.564, P=0.01), tumor-node-metastasis (TNM) stage (HR =2.297, P<0.001), and HMGB3 expression (HR =1.370, P=0.01) as factors influencing prognosis (Table 2). In multivariate analysis, TNM stage (HR =1.738, P=0.01) and cancer status (HR =2.083, P=0.01) remained independent predictors. These results indicate that HMGB3 expression might assist in forecasting clinical outcomes in ESCA.

HMGB3-related differential genes and pathway enrichment analyses in ESCA

Patients from the TCGA-ESCA cohort were divided into groups with elevated or reduced HMGB3 expression to investigate its possible roles in ESCA. From this analysis, 787 DEGs were obtained, as visualized in the volcano plot and heatmap (Figure 5A,5B). These genes were further used to generate a PPI network via the STRING database (Figure 5C). We then conducted a series of functional enrichment analyses. GO analysis linked the deregulated genes primarily to the immune response, localization in the apical region of cells, and anion membrane transporter activity (Figure 5D). In parallel, KEGG pathway analysis highlighted their involvement in cell adhesion and metabolic processes related to cytochrome P450, specifically the metabolism of xenobiotics and drugs (Figure 5E). In the high HMGB3 expression group, GSEA identified marked enrichment of pathways involved in antigen processing and presentation, cell adhesion, the cell cycle, DNA replication, and Hedgehog and Wnt signaling (Figure 5F).

Figure 5 PPI network of HMGB3 and enrichment analyses. (A) The heatmap and (B) volcano plot of DEGs. (C) PPI network of HMGB3. (D) GO enrichment analysis of HMGB3-related genes. (E) KEGG analysis of HMGB3-related genes. (F) GSEA analysis of HMGB3 in ESCA. BP, biological process; CC, cellular component; DEG, differentially expressed gene; ESCA, esophageal carcinoma; GO, Gene Ontology; GSEA, gene set enrichment analysis; HMGB3, high mobility group box 3; KEGG, Kyoto Encyclopedia of Genes and Genomes; MF, molecular function; NADP, nicotinamide adenine dinucleotide phosphate; NS, not significant; PPI, protein-protein interaction.

HMGB3 overexpression is correlated with altered immune cell infiltration and immune checkpoint dysregulation in ESCA

To investigate the relationship between HMGB3 expression and immune cell infiltration in ESCA, we utilized the MCP-counter package in R to quantify immune cell populations based on data from TCGA-ESCA tumor samples. Negative associations were identified between HMGB3 expression and multiple immune cell subsets, such as T cells (r=−0.35, P<0.001), CD8⁺ T cells (r=−0.25, P<0.001), B lineage cells (r=−0.36, P<0.001), NK cells (r=−0.22, P=0.003), myeloid DCs (r=−0.20, P=0.008), neutrophils (r=−0.27, P<0.001), and endothelial cells (r=−0.35, P<0.001, Figure 6).

Figure 6 Correlation between HMGB3 and immune/stromal cells as assessed by MCP-counter. (A) Association of HMGB3 with 10 immune and stromal cell types. Correlations of HMGB3 expression with (B) T cells, (C) CD8⁺ T cells, (D) B lineage cells, (E) NK cells, (F) myeloid dendritic cells, (G) neutrophils, and (H) endothelial cells. **, P<0.01; ***, P<0.001. Cor, correlation coefficient; HMGB3, high mobility group box 3; MCP, Microenvironment Cell Populations; NK, natural killer.

To further profile immune cell composition, we applied the CIBERSORT algorithm to estimate the relative abundance of 22 immune cell subsets in ESCA (Figure 7A). Comparisons between the HMGB3-high and HMGB3-low groups revealed marked reductions in the abundance of naïve B cells and resting CD4⁺ memory T cells alongside an increased proportion of M0 macrophages in the high expression group (Figure 7B). Correlation analysis uncovered negative associations of HMGB3 expression with the abundance of naïve B cells (r=−0.29, P<0.001) and resting CD4⁺ memory T cells (r=−0.32, P<0.001, Figure 7C,7D), but positive associations with that of M0 (r=0.20, P=0.006) and M2 macrophages (r=0.20, P=0.007, Figure 7E,7F).

Figure 7 Correlation between HMGB3 and immune cell subsets as assessed by CIBERSORT. (A) Proportions of 22 immune cells in ESCA. (B) Distribution differences between the HMGB3-high and HMGB3-low groups. (C-F) Correlations of HMGB3 expression with (C) naïve B cells, (D) resting CD4⁺ memory T cells, (E) M0 macrophages, and (F) M2 macrophages. *, P<0.05; **, P<0.01; ***, P<0.001. CIBERSORT, cell-type identification by estimating relative subsets of RNA transcripts; ESCA, esophageal carcinoma; HMGB3, high mobility group box 3; NK, natural killer; ns, not significant.

TIMER analysis, adjusted for tumor purity, further indicated that HMGB3 expression in ESCA was significantly correlated with the monocyte marker CD86; TAM marker CD68; M1 macrophage marker NOS2; M2 macrophage marker VSIG4; neutrophil marker CCR7; and DC markers HLA-DPB1, HLA-DQB1, HLA-DRA, HLA-DPA1, and NRP1 (Figure 8A-8F and Table 3). In addition, stromal, immune, and ESTIMATE scores were all significantly lower in the HMGB3-high group than in the HMGB3-low group (Figure 8G). Collectively, these analyses consistently demonstrate that high HMGB3 expression is associated with reduced infiltration of antitumor immune cells and increased recruitment of protumorigenic immune cells, highlighting a potential role for HMGB3 in shaping an immunosuppressive TME in ESCA.

Figure 8 Analysis of correlations between HMGB3 expression and immune cell markers, as well as ESTIMATE score in ESCA. Correlation of HMGB3 with immune cell markers for monocytes (A), TAMs (B), M1 macrophages (C), M2 macrophages (D), neutrophils (E), and DCs (F). (G) Stromal, immune, and ESTIMATE scores in the HMGB3 groups. *, P<0.05; ***, P<0.001; ****, P<0.0001. Cor, correlation coefficient; DCs, dendritic cells; ESCA, esophageal carcinoma; ESTIMATE, Estimation of STromal and Immune cells in MAlignant Tumor; HMGB3, high mobility group box 3; TAMs, tumor-associated macrophages; TPM, transcripts per million.

In addition, HMGB3 expression was negatively correlated with programmed cell death protein 1 (PDCD1; r=−0.21, P=0.004), cytotoxic T-lymphocyte-associated protein 4 (CTLA4; r=−0.20, P=0.007), and CD47 expression (r=−0.18, P=0.02), but positively correlated with CD276 expression (r=0.39, P<0.001) (Figure 9).

Figure 9 HMGB3 and immune checkpoints in ESCA. (A) Heatmap of immune checkpoint expression. (B-H) Correlations of HMGB3 expression with (B) CD276, (C) CD47, (D) S1PR1, (E) CX3CR1, (F) CSF1R, (G) CTLA4, and (H) PDCD1. *, P<0.05; **, P<0.01; ***, P<0.001. CTLA4, cytotoxic T-lymphocyte-associated protein 4; ESCA, esophageal carcinoma; ESTIMATE, Estimation of STromal and Immune cells in MAlignant Tumor tissues using Expression data; HMGB3, high mobility group box 3; PDCD1, programmed cell death protein 1.

Association of HMGB3 with drug sensitivity

We further examined the impact of HMGB3 expression on drug sensitivity using the GSCA database. Comprehensive pharmacogenomic profiling demonstrated a distinct drug sensitivity spectrum associated with HMGB3 expression in ESCA. Specifically, elevated HMGB3 expression was significantly correlated with heightened susceptibility to multiple small-molecule inhibitors, including 5Z-7-oxozeaenol, CI-1040, dabrafenib, PD-0325901, RDEA119, selumetinib, and trametinib, whereas inverse correlations were observed with agents such as CP724714 and WZ3105 (Figure 10). These findings suggest that HMGB3 expression might be associated with differential responses to specific targeted agents in ESCA, and they could provide a basis for future studies exploring HMGB3-related drug sensitivity.

Figure 10 Correlation of HMGB3 expression and drug sensitivity using the GSCA database. FDR, false discovery rate; GDSC; GSCA, Gene Set Cancer Analysis; HMGB3, High mobility group box 3; mRNA, messenger RNA.

HMGB3 knockdown inhibits ESCA cell proliferation and invasion and induces cell cycle arrest

Transcriptomic profiling of 27 ESCA cell lines from the HPA database confirmed consistently high HMGB3 expression, with peak levels detected in KYSE140, KYSE510, TE-10, TE-15, and TE-4 cells (Figure 11A). These bioinformatics observations were corroborated by western blotting, which demonstrated significantly higher HMGB3 protein expression in KYSE510 (t=7.06, P=0.002) and KYSE140 cells (t=5.10, P=0.007) than in Het-1A cells (Figure 11B). To elucidate the contribution of HMGB3 to ESCA cell phenotypes, three shRNA constructs were designed to selectively target HMGB3 transcripts. RT-qPCR identified shRNA2 as achieving the highest knockdown efficiency (Figure 11C), prompting its incorporation into a lentiviral expression system for stable knockdown. Subsequent Western blotting confirmed a marked reduction in HMGB3 protein expression in KYSE510 and KYSE140 cells relative to that in NC cells (Figure 11D). Functionally, HMGB3 knockdown attenuated proliferative capacity and invasive potential in both cell lines (Figure 11E,11F). Moreover, cell cycle analysis revealed a pronounced accumulation of cells in G0/G1 phase accompanied by a reduction in the G2/S fraction, indicative of cell cycle arrest (Figure 11G).

Figure 11 Effect of HMGB3 knockdown on proliferation, invasion, and cell cycle in ESCA. (A) HMGB3 expression was assessed across 27 ESCA cell lines in the HPA database. (B) The expression of HMGB3 protein was assessed in esophageal epithelial and ESCA cell lines. (C) RT-qPCR and (D) western blotting revealing the knockdown efficiency. (E) CCK-8 assay to assess proliferation. (F) Transwell assay to assess invasion (crystal violet staining; magnification). (G) Flow cytometry to assess cell cycle distribution. For experiments presented in panels E-G, each condition was assessed in triplicate, and data are presented as mean ± standard deviation. *, P<0.05; **, P<0.01; ***, P<0.001. CCK-8, Cell Counting Kit-8; ESCA, esophageal carcinoma; HMGB3, high mobility group box 3; HPA, Human Protein Atlas; NC, negative control; nTPM, normalized transcripts per million; OD, optical density; RT-qPCR, reverse transcription quantitative polymerase chain reaction; sh, short hairpin RNA.

Discussion

ESCA represents a highly aggressive neoplasm of the digestive tract, and it is distinguished by its considerable global disease burden, worrisome incidence trends, and persistently unfavorable survival outcomes. The identification of reliable biomarkers is essential for improving early detection and guiding effective therapeutic strategies. HMGB3, which is dysregulated in various cancers, has been linked to the prognosis of cancers such as lung (19,20), gastric (21), and breast cancers (22,23), suggesting its potential clinical relevance. However, its role in ESCA development remains unclear. This study systematically examined the diagnostic and prognostic value, immune associations, and potential biological functions of HMGB3 in ESCA using multiple databases, and further investigated its effects on ESCA cell behavior.

Pan-cancer assessment revealed that HMGB3 is markedly overexpressed in a spectrum of malignancies, with particularly high levels in ESCA. In the TCGA-ESCA cohort, ROC curve analyses yielded an AUC of 0.957, highlighting its strong diagnostic value in distinguishing ESCA tumor tissues from adjacent non-tumorous tissues. Consistently, IHC of clinical ESCA specimens confirmed a pronounced increase in HMGB3 protein expression in tumor tissues and demonstrated a significant association with advanced TNM stage. Collectively, these results suggest that HMGB3 might serve as a valuable supplementary biomarker for ESCA detection and pathological assessment, particularly in light of the current limitations in early screening modalities.

Moreover, survival analyses revealed a correlation between elevated HMGB3 expression and poorer patient outcomes in ESCA. Univariate Cox regression analysis identified HMGB3 expression, together with N stage, M stage, TNM stage, and cancer status, as factors influencing prognosis. Although HMGB3 did not retain independent prognostic significance in multivariate analysis, its effect in univariate models and its close association with advanced disease suggest that it might still provide incremental prognostic information. Notably, previous studies in ESCC identified HMGB3 as an independent prognostic determinant of OS, further underscoring its potential clinical relevance (24). Nevertheless, the independent prognostic value of HMGB3 in ESCA warrants validation in larger, prospective, and multicenter cohorts.

GSEA showed that HMGB3-associated pathways were significantly enriched in cell adhesion, cell cycle regulation, DNA replication, Hedgehog signaling, and Wnt signaling. These results suggest that HMGB3 may have broad effects on ESCA development and progression by participating in several tumor-related biological processes. The enrichment of Wnt and Hedgehog signaling in the HMGB3-high group is particularly noteworthy, as both pathways have been implicated in the proliferation, invasion, and malignant progression of esophageal cancer (25-27). This observation is consistent with our in vitro findings showing that HMGB3 knockdown markedly suppressed ESCA cell proliferation and invasion. In addition, enrichment of cell cycle- and DNA replication-related pathways is consistent with the cell cycle arrest observed after HMGB3 knockdown, suggesting that HMGB3 may support sustained ESCA cell proliferation by influencing transcriptional programs involved in cell cycle transition and DNA replication. Given that HMGB family proteins function as chromatin-binding factors that modulate DNA conformation and transcriptional activity, HMGB3 may contribute to ESCA progression, at least in part, by regulating key cell cycle genes, DNA replication machinery, and invasion-related genes. However, the direct molecular links between HMGB3 and the Wnt and Hedgehog pathways, as well as the downstream regulatory networks involved, were not mechanistically validated in this study and require further investigation.

The TME, comprising tumor cells, immune cells, stromal components, and cytokines, plays pivotal roles in cancer initiation, progression, and metastasis, largely through the balance between antitumor and protumor immune components (28-30). Within this immunological landscape, effector populations such as cytotoxic CD8⁺ T cells, CD4⁺ T cells, DCs, and NK cells exert tumor-restraining functions (31-35), whereas immunosuppressive subsets such as regulatory T cells (Tregs), myeloid-derived suppressor cells (MDSCs), alternatively activated macrophages, and protumor neutrophils facilitate immune escape and tumor propagation (36-38). Emerging evidence underscores a pivotal role of the HMGB family members in modulating immune cell activation and function, thereby influencing therapeutic responsiveness, particularly in the context of cancer immunotherapy. Research by Hubert and colleagues (39) demonstrated that inhibiting HMGB1 can remodel the TME by decreasing MDSC and Treg accumulation, increasing the M1/M2 macrophage ratio, and activating DCs and plasmacytoid DCs, ultimately enhancing the antitumor effects of PD-1 blockade. In glioma, HMGB3 has been implicated in promoting the polarization of M2 macrophages and NLRP3-mediated pyroptosis, which contributes to tumor progression and immune modulation (40). These studies suggest that HMGB proteins can shape the immune landscape and influence responses to immunotherapy. Notably, the role of HMGB3 in the TME of ESCA has not been previously studied. In this study, we observed that HMGB3 expression displayed negative correlations with key antitumor immune cells, including CD8⁺ T cells, CD4⁺ T cells, myeloid DCs, and NK cells, but a positive correlation with M2 macrophages. In parallel, GO analysis indicated that HMGB3-related genes were enriched in immune-related biological processes, such as activation of immune cell surface receptor signaling, signal transduction, and humoral immune response. These findings, together with the lower stromal and immune scores observed in HMGB3-high tumors, support the notion that elevated HMGB3 expression is associated with a more immunologically “cold” and immunosuppressive TME by influencing immune cell infiltration and functional composition.

In addition, we identified a distinct pattern of associations between HMGB3 expression and multiple immune checkpoint molecules in ESCA. HMGB3 expression was positively correlated with CD276 but negatively correlated with PD-1, CTLA-4, and CD47. CD276 has been widely implicated in tumor immune evasion and poor clinical outcomes, and it is frequently expressed on tumor and stromal cells, in which it can dampen T cell-mediated antitumor responses (41-44). Notably, a recent study in ESCC showed that CD276 promotes immune escape by upregulating CXCL1, activating the CXCL1-CXCR2 axis, inducing neutrophil recruitment and neutrophil extracellular trap formation, and ultimately suppressing NK-cell infiltration and activity (41). Moreover, KRT15 has been reported to enhance GSK3β/β-catenin/CD276 signaling, thereby increasing CD276 transcription and impairing NK-cell function in ESCC (45). These findings are consistent with our observation that HMGB3 expression was negatively correlated with NK-cell infiltration in ESCA, raising the possibility that HMGB3-high tumors preferentially engage CD276-related suppressive circuits rather than relying primarily on canonical PD-1/CTLA-4-mediated immune exhaustion. Our data further suggest that HMGB3 may be linked to macrophage polarization-related immune marker patterns. CIBERSORT analysis showed a positive correlation between HMGB3 expression and M2 macrophage abundance, whereas tumor purity-adjusted TIMER analysis showed a negative correlation with the M1 macrophage marker NOS2 and a positive correlation with the M2 macrophage marker VSIG4. Consistently, HMGB3 has been reported to promote macrophage infiltration and M2 polarization in glioma; in that setting, exosomal HMGB3 increased M2-associated markers, including CD206 and ARG1, and macrophage infiltration-related markers, including CD68 and CCL2, thereby contributing to an immunosuppressive microenvironment (40). Together, these findings suggest that HMGB3 may facilitate immune suppression in ESCA by coordinating immune checkpoint signaling and shaping macrophage polarization and immune cell infiltration. However, the present study did not directly determine whether HMGB3 transcriptionally regulates CD276 or macrophage polarization-related genes in ESCA; these mechanistic links require further experimental validation.

Although our study offers meaningful insights, it has several limitations that should be addressed in future research. First, despite the inclusion of multiple public datasets and a cohort of 46 paired ESCA specimens, the overall sample size remains relatively modest. Further research involving large-scale ESCA cohorts is crucial to validate the diagnostic and prognostic value of HMGB3 and determine whether it can serve as an independent prognostic factor. Second, although our analyses revealed associations among HMGB3 expression, immune infiltration, and immune checkpoint molecules, they do not establish causality. Additional mechanistic studies are therefore required to elucidate how HMGB3 interacts with the tumor immune microenvironment, whether its expression can predict responses to immune checkpoint blockade in ESCA, and whether modulation of HMGB3 could potentially convert immunologically “cold” tumors into “hot” ones.

Conclusions

In summary, our research underscores the significant overexpression of HMGB3 in ESCA, which is linked to tumor progression, an immunosuppressive microenvironment, and adverse patient outcomes. Functional assays confirm that HMGB3 promotes malignant phenotypes in ESCC cells. The study findings support that HMGB3 as a promising biomarker for diagnosis and prognosis and suggest that it might represent a rational candidate for therapeutic targeting in ESCA, warranting further mechanistic and translational investigation.

Acknowledgments

We thank Medjaden Inc. for scientific editing of this manuscript.

Footnote

Reporting Checklist: The authors have completed the MDAR reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1-0362/rc

Data Sharing Statement: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1-0362/dss

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1-0362/prf

Funding: This work was supported by the Fundamental Research Funds for the Central Universities (No. xzy012024122); Key Research and Development Program of Shaanxi Province (No. 2024SF-YBXM-121); Shaanxi Provincial Health Commission (No. 2021B007); Xi’an Science and Technology Plan Project (No. 24YXYJ0184); and Key Research and Development Program of Shaanxi Province (No. 2022SF-495).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1-0362/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74:229-63. [Crossref] [PubMed]
2. Ding Y, Ren L, Geng Y, et al. The Current Status and Prospects of Early Diagnosis and Treatment of Esophageal Cancer in China. Cancer Screen Prev 2024;3:106-12.
3. Han B, Zheng R, Zeng H, et al. Cancer incidence and mortality in China, 2022. J Natl Cancer Cent 2024;4:47-53. [Crossref] [PubMed]
4. Arnold M, Ferlay J, van Berge Henegouwen MI, et al. Global burden of oesophageal and gastric cancer by histology and subsite in 2018. Gut 2020;69:1564-71. [Crossref] [PubMed]
5. Yang H, Wang F, Hallemeier CL, et al. Oesophageal cancer. Lancet 2024;404:1991-2005. [Crossref] [PubMed]
6. Zeng H, Chen W, Zheng R, et al. Changing cancer survival in China during 2003-15: a pooled analysis of 17 population-based cancer registries. Lancet Glob Health 2018;6:e555-67. [Crossref] [PubMed]
7. Starkova T, Polyanichko A, Tomilin AN, et al. Structure and Functions of HMGB2 Protein. Int J Mol Sci 2023;24:8334. [Crossref] [PubMed]
8. Guo L, Wang D, Jiang X, et al. HMGB1: From Molecular Functions to Clinical Applications in Cancer and Inflammatory Diseases. Med Res Rev 2026;46:408-44. [Crossref] [PubMed]
9. Chikhirzhina E, Tsimokha A, Tomilin AN, et al. Structure and Functions of HMGB3 Protein. Int J Mol Sci 2024;25:7656. [Crossref] [PubMed]
10. Yanai H, Ban T, Wang Z, et al. HMGB proteins function as universal sentinels for nucleic-acid-mediated innate immune responses. Nature 2009;462:99-103. [Crossref] [PubMed]
11. Qin X, Niu Z, Chen H, et al. Macrophage-derived exosomal HMGB3 regulates silica-induced pulmonary inflammation by promoting M1 macrophage polarization and recruitment. Part Fibre Toxicol 2024;21:12. [Crossref] [PubMed]
12. Nemeth MJ, Curtis DJ, Kirby MR, et al. Hmgb3: an HMG-box family member expressed in primitive hematopoietic cells that inhibits myeloid and B-cell differentiation. Blood 2003;102:1298-306. [Crossref] [PubMed]
13. Nemeth MJ, Kirby MR, Bodine DM. Hmgb3 regulates the balance between hematopoietic stem cell self-renewal and differentiation. Proc Natl Acad Sci U S A 2006;103:13783-8. [Crossref] [PubMed]
14. Zhang J, Sun Y, Wang L, et al. HMGB3: A pivotal orchestrator of therapy resistance and cancer stemness in human malignancies Oncol Rep 2026;55:104. (Review). [Crossref] [PubMed]
15. Mirzaee F, Abbaszade-CheragheAli A, Khamoushi A. Overview of high mobility group box 3 (HMGB3 protein. Mol Genet Genomics 2025;300:59. [Crossref] [PubMed]
16. Ma H, Qi G, Han F, et al. HMGB3 promotes the malignant phenotypes and stemness of epithelial ovarian cancer through the MAPK/ERK signaling pathway. Cell Commun Signal 2023;21:144. [Crossref] [PubMed]
17. Gu J, Xu T, Huang QH, et al. HMGB3 silence inhibits breast cancer cell proliferation and tumor growth by interacting with hypoxia-inducible factor 1α. Cancer Manag Res 2019;11:5075-89. [Crossref] [PubMed]
18. Gong W, Guo Y, Yuan H, et al. HMGB3 is a Potential Therapeutic Target by Affecting the Migration and Proliferation of Colorectal Cancer. Front Cell Dev Biol 2022;10:891482. [Crossref] [PubMed]
19. Liu Q, Zhang J, Guo C, et al. Proteogenomic characterization of small cell lung cancer identifies biological insights and subtype-specific therapeutic strategies. Cell 2024;187:184-203.e28. [Crossref] [PubMed]
20. Cui H, Yang Y, Li S, et al. HMGB3 promotes brain metastasis of lung adenocarcinoma by recruiting SSBP1 for nuclear translocation to remodel mitochondrial metabolism. Cancer Commun (Lond) 2025;45:1676-705. [Crossref] [PubMed]
21. Fang J, Ge X, Xu W, et al. Bioinformatics analysis of the prognosis and biological significance of HMGB1, HMGB2, and HMGB3 in gastric cancer. J Cell Physiol 2020;235:3438-46. [Crossref] [PubMed]
22. Zhou X, Zhang Q, Liang G, et al. Overexpression of HMGB3 and its prognostic value in breast cancer. Front Oncol 2022;12:1048921. [Crossref] [PubMed]
23. Tang M, Huang T, Zhang W, et al. Constructing an immune-related prognostic model and exploring the function of HMGB3, TNFSF4, and CORO2A in breast cancer. Transl Cancer Res 2025;14:7071-88. [Crossref] [PubMed]
24. Gao J, Zou Z, Gao J, et al. Increased expression of HMGB3: a novel independent prognostic marker of worse outcome in patients with esophageal squamous cell carcinoma. Int J Clin Exp Pathol 2015;8:345-52.
25. Zang W, Zhu W, Jing F, et al. The role of the hedgehog signaling pathway in the regulation of gastrointestinal cancer cell death. Cell Signal 2026;141:112384. [Crossref] [PubMed]
26. Zhang Y, Shi C, Zhou Y, et al. H2BC9 lactylation modulates esophageal squamous cell carcinoma progression via the Wnt/β-catenin signaling pathway. J Transl Med 2025;23:1085. [Crossref] [PubMed]
27. Romanowicz A, Łukaszewicz-Zając M. The Significance of the Wnt/β-Catenin Pathway and Related Proteins in Gastrointestinal Malignancies. Int J Mol Sci 2025;26:8130. [Crossref] [PubMed]
28. de Visser KE, Joyce JA. The evolving tumor microenvironment: From cancer initiation to metastatic outgrowth. Cancer Cell 2023;41:374-403. [Crossref] [PubMed]
29. Bayat M, Nahand JS. Battlegrounds of treatment resistance: decoding the tumor microenvironment. Naunyn Schmiedebergs Arch Pharmacol 2025;398:11179-97. [Crossref] [PubMed]
30. Fioretzaki R, Trifylli EM, Sarantis P, et al. The Interplay Between Esophageal Adenocarcinoma and Its Tumor Microenvironment: Toward Innovative Therapies. Cells 2025;14:1895. [Crossref] [PubMed]
31. Verneau J, Sautés-Fridman C, Sun CM. Dendritic cells in the tumor microenvironment: prognostic and theranostic impact. Semin Immunol 2020;48:101410. [Crossref] [PubMed]
32. Pan Y, Yu Y, Wang X, et al. Tumor-Associated Macrophages in Tumor Immunity. Front Immunol 2020;11:583084. [Crossref] [PubMed]
33. Speiser DE, Chijioke O, Schaeuble K, et al. CD4+ T cells in cancer. Nat Cancer 2023;4:317-29. [Crossref] [PubMed]
34. Reina-Campos M, Scharping NE, Goldrath AW. CD8(+) T cell metabolism in infection and cancer. Nat Rev Immunol 2021;21:718-38. [Crossref] [PubMed]
35. Liu S, Galat V, Galat Y, et al. NK cell-based cancer immunotherapy: from basic biology to clinical development. J Hematol Oncol 2021;14:7. [Crossref] [PubMed]
36. Mehdizadeh R, Shariatpanahi SP, Goliaei B, et al. Targeting myeloid-derived suppressor cells in combination with tumor cell vaccination predicts anti-tumor immunity and breast cancer dormancy: an in silico experiment. Sci Rep 2023;13:5875. [Crossref] [PubMed]
37. Xia Y, Rao L, Yao H, et al. Engineering Macrophages for Cancer Immunotherapy and Drug Delivery. Adv Mater 2020;32:e2002054. [Crossref] [PubMed]
38. Shan F, Somasundaram A, Bruno TC, et al. Therapeutic targeting of regulatory T cells in cancer. Trends Cancer 2022;8:944-61. [Crossref] [PubMed]
39. Hubert P, Roncarati P, Demoulin S, et al. Extracellular HMGB1 blockade inhibits tumor growth through profoundly remodeling immune microenvironment and enhances checkpoint inhibitor-based immunotherapy. J Immunother Cancer 2021;9:e001966. [Crossref] [PubMed]
40. Hu P, Yan T, Lv S, et al. Exosomal HMGB3 released by glioma cells confers the activation of NLRP3 inflammasome and pyroptosis in tumor-associated macrophages. Tissue Cell 2024;88:102406. [Crossref] [PubMed]
41. Xiong G, Chen Z, Liu Q, et al. CD276 regulates the immune escape of esophageal squamous cell carcinoma through CXCL1-CXCR2 induced NETs. J Immunother Cancer 2024;12:e008662. [Crossref] [PubMed]
42. Corrigan DT, Tanwar A, Du M, et al. The B7-H3 (CD276) pathway: emerging biology and clinical therapeutics. Trends Pharmacol Sci 2025;46:975-88. [Crossref] [PubMed]
43. Liu HJ, Du H, Khabibullin D, et al. mTORC1 upregulates B7-H3/CD276 to inhibit antitumor T cells and drive tumor immune evasion. Nat Commun 2023;14:1214. [Crossref] [PubMed]
44. Cai D, Li J, Liu D, et al. Tumor-expressed B7-H3 mediates the inhibition of antitumor T-cell functions in ovarian cancer insensitive to PD-1 blockade therapy. Cell Mol Immunol 2020;17:227-36. [Crossref] [PubMed]
45. Yang C, Zhong Z, Li C, et al. KRT15 drives immunosuppression in esophageal squamous cell carcinoma through GSK3β/β-catenin/CD276 signaling. Exp Cell Res 2026;458:114983. [Crossref] [PubMed]