UBE2S and HIF1α expression patterns and stratified analysis reveal prognostic value in esophageal squamous cell carcinoma

Patient and tissue samples

A retrospective cohort of 356 formalin-fixed paraffin-embedded (FFPE) esophageal squamous cell carcinoma (ESCC) specimens with matched adjacent non-tumorous tissues (collected from January 2012 to December 2021) was obtained from the Department of Pathology, First Affiliated Hospital of Xinjiang Medical University. All participants provided informed consent. The study protocol was approved by the Ethics Committee of the First Affiliated Hospital of Xinjiang Medical University (Approval No. K202207-34) and conducted in accordance with the Declaration of Helsinki. Clinical medical records were obtained, and telephone follow-ups were conducted until December 2023. Two senior pathologists reviewed all slides to confirm the pathological diagnosis. The inclusion criteria were as follows: (1) curative resection specimens of primary ESCC; (2) no history of neoadjuvant therapy; (3) complete clinicopathological records. The exclusion criteria were: (1) biopsy-only specimens; (2) non-squamous histology; (3) preoperative radiotherapy or chemotherapy; (4) metastatic or secondary esophageal tumors; (5) incomplete follow-up data. Based on these criteria, 259 specimens were included in the final analysis. For bioinformatics validation, data were obtained from The Cancer Genome Atlas (TCGA) Esophageal Carcinoma (ESCA) cohort and the Gene Expression Omnibus (GEO) dataset GSE161533. Given the limited number of normal tissue samples (n = 3) in the TCGA-ESCA cohort, RNA-seq data from normal esophageal mucosa were sourced from the Genotype-Tissue Expression (GTEx) project to construct a robust normal tissue reference. The TCGA and GTEx datasets were integrated, and batch effects were corrected using the ComBat algorithm from the sva R package prior to comparative analysis.

Immunohistochemical staining and evaluation

Immunohistochemical (IHC) staining was performed on tissue microarrays using the streptavidin-peroxidase (SP) method. The primary antibodies used were: UBE2S (rabbit polyclonal, 14115-1-AP, 1:360; Proteintech), HIF1α (rabbit monoclonal, bs-0737R, 1:400; BIOSS), and p110α (rabbit polyclonal, C73F8, 1:250; ZSGB-Bio). The IHC procedure was conducted as follows. Briefly, 4-µm FFPE sections were baked at 65 °C for 100 min, deparaffinized in xylene and rehydrated through an ethanol series. Heat-induced antigen retrieval was performed in citrate buffer (pH 6.0) at 95 °C for 20 min. Endogenous peroxidase activity was blocked with 3% H₂O₂ for 10 min at room temperature. Sections were then incubated with primary antibodies at 4 °C overnight, followed by incubation with an HRP-conjugated polymer secondary antibody for 30 min at room temperature. Staining was visualized with DAB chromogen, counterstained with hematoxylin, and mounted with a resin-based medium. Two senior pathologists, blinded to clinical outcomes, independently evaluated the IHC staining. Ten non-overlapping fields were selected per section under 400x magnification, and 100 cells were counted per field. A scoring system integrating staining intensity (0: negative; 1: weak; 2: moderate; 3: strong) and the percentage of positive cells (0: ≤1%; 1: 1–20%; 2: 21–50%; 3: >50%) was used. The final score was calculated by multiplying the intensity and percentage scores. A score of ≥4 was defined as positive expression. Interobserver agreement was assessed using Cohen’s kappa coefficient, with a value >0.75 indicating excellent agreement.

Statistical analysis

Statistical analyses were performed with SPSS 27.0 and R (version 4.2.2). Paired t-tests were used for normally distributed continuous variables in paired samples; the Wilcoxon rank-sum test (Mann–Whitney U) and Kruskal–Wallis test were used for non-normally distributed data in two or multiple groups, respectively. Pearson’s correlation analysis was used to assess correlations between protein expression levels. Survival analysis was performed using the Kaplan–Meier method with log-rank tests for group comparisons. Univariate Cox proportional hazards regression was used to screen prognostic variables (p < 0.1 for inclusion), followed by multivariate Cox regression (stepwise method) to identify independent prognostic factors. A two-tailed p-value <0.05 was considered statistically significant.

Baseline characteristics of patients

This study included a total of 259 samples; the detailed clinical features are listed in Table 1 and Table S1. Among the samples, 91 cases (35.1%) were 60 years old or younger, while 168 cases (64.9%) were older than 60. There were 180 males (69.5%) and 79 females (30.5%). There were 135 Han cases (52.1%) and 124 Kazak cases (47.9%).

Regarding tumor location, there were 13 cases (5.0%) in the upper esophagus, 140 cases (54.1%) in the middle esophagus, and 106 cases (40.9%) in the lower esophagus. For tumor size, 180 cases (69.5%) had tumors measuring ≥3 cm, while 79 cases (30.5%) had tumors <3 cm. In terms of tumor differentiation, 64 cases (24.7%) were well differentiated, 145 cases (56.0%) were moderately differentiated, and 50 cases (19.3%) were poorly differentiated. There were 89 cases (34.4%) with lymph node metastasis and 170 cases (65.6%) without lymph node metastasis.

UBE2S and HIF1α are highly expressed in ESCC

IHC results showed that UBE2S, HIF1α, and p110α were expressed in the cytoplasm and nucleus of tumor cells. Of these cases, 232 (89.6%) were UBE2S-positive, while 27 (10.4%) were UBE2S-negative. UBE2S and HIF1α were either not expressed or only weakly expressed in the basal cells of cancer-adjacent normal mucosa. The staining intensity could be categorized as colorless, weak, moderate, or strong (Figs. 1A–1D, 1F–1I 1K–1N).

In addition, UBE2S was found to be co-expressed at the sites where HIF1α was expressed, suggesting that there may be an interaction between UBE2S and HIF1α. According to our established IHC scoring system, histograms demonstrated that IHC scores for UBE2S, HIF1α, and p110α were significantly higher in cancer tissues than in normal tissues (Figs. 1E, 1J, 1O; p < 0.001, p = 0.004, p = 0.039).

The positive rates of UBE2S and HIF1α in ESCC tissues were 89.6% (232/259) and 87.2% (226/259), respectively-much higher than those in normal esophageal mucosa (15.7%, 31/198; 9.7%, 18/186). These results indicated that the expression levels of UBE2S and HIF1α in ESCC were significantly higher than those in normal tissues.

Relationship between UBE2S/HIF1α/p110α expression and clinicopathological features

The associations between the expression of UBE2S, HIF1α, and p110α and clinicopathological features are summarized in Table 1 and Table S1. HIF1α expression showed a significant correlation with patient gender (p = 0.040), with a higher positive rate observed in male compared to female patients. Similarly, the positive rate of UBE2S was significantly higher in Kazak than in Han patients (p = 0.001). UBE2S expression was also significantly associated with the presence of vessel invasion (p = 0.020). A trend toward significance was noted for the correlation between HIF1α expression and depth of invasion (p = 0.050). Furthermore, a significant positive correlation was found between UBE2S and HIF1α expression (p = 0.013). We further assessed the clinical relevance of p110α. Analysis revealed no significant associations between p110α expression and standard clinicopathological parameters (Table S2). Notably, however, its correlations with UBE2S and HIF1α revealed a divergent pattern: a statistically significant positive correlation was observed with HIF1α (p = 0.018), whereas the correlation with UBE2S was not significant. This specific correlation pattern may suggest a closer functional relationship between HIF1α and p110α under the present conditions.

UBE2S as an independent prognostic factor for OS

To evaluate the impact of UBE2S and HIF1α expression on patient survival, we performed univariate and multivariate Cox proportional hazards analyses on 259 ESCC samples. Univariate analysis identified UBE2S expression, nationality, lymph node metastasis, infiltration depth, American Joint Committee on Cancer (AJCC) stage, and nerve invasion as factors significantly associated with overall survival (OS) (Fig. 2A). Subsequent multivariate analysis confirmed that UBE2S expression, nationality, lymph node metastasis, and nerve invasion remained independently associated with OS (Fig. 2B). For progression-free survival (PFS), univariate analysis showed significant associations for UBE2S expression, nationality, lymph node metastasis, AJCC stage, and nerve invasion (Fig. 2C). In the multivariate model, however, only nationality retained a significant independent correlation with PFS (Fig. 2D). Collectively, these results indicate that UBE2S expression serves as an independent prognostic factor for OS in ESCC patients.

Bioinformatics analysis reveals elevated expression and co-expression of UBE2S and HIF1α in esophageal carcinoma

Initial pan-cancer analysis of TCGA data indicated that UBE2S and HIF1α were significantly overexpressed in multiple malignancies, including ESCA (Figs. 3A, 3B). To further explore this observation, we examined their expression in a larger ESCC cohort. Given the limited number of normal tissue samples (n = 3) in the TCGA-ESCA dataset, we integrated RNA-seq data from the Genotype-Tissue Expression (GTEx) project to establish a more robust normal tissue reference. After batch effect correction and normalization of the combined dataset, comparative analysis confirmed significantly elevated mRNA expression of both UBE2S and HIF1α in tumor tissues compared to the pooled normal controls (Figs. 3C, 3D; p < 0.001). For independent validation, we examined the GEO dataset GSE161533. Analysis of this cohort revealed that UBE2S expression was significantly higher in ESCC tissues relative to paired adjacent non-tumorous tissues (Fig. 3E; p < 0.001). Similarly, HIF1α expression was also significantly elevated in tumors compared to matched normal controls within the same dataset (Fig. 3F; p < 0.001). Furthermore, correlation analysis within the TCGA-ESCA cohort revealed a significant positive correlation between UBE2S and HIF1α mRNA expression levels (Fig. 3G; R = 0.42, p = 0.0013). Collectively, these cross-platform bioinformatics analyses support the notion that UBE2S and HIF1α are consistently overexpressed in esophageal carcinoma. The observed weak-to-moderate positive correlation between their expression levels may point to a potential functional interplay in ESCC, providing preliminary evidence for further mechanistic investigation.

Prognostic impact of UBE2S and HIF1α expression and their synergistic effect

Having established the association between protein expression and clinicopathological features, we next evaluated the prognostic impact of UBE2S and HIF1α. Survival analysis was performed by categorizing patients based on UBE2S and HIF1α IHC expression scores. High expression of UBE2S was significantly associated with poorer overall survival (OS; p = 0.006) and progression-free survival (PFS; p = 0.034) (Figs. 4A–4B). Similarly, high expression of HIF1α was associated with reduced OS (p < 0.001) and PFS (p = 0.002) (Figs. 4C–4D). Taken together, these observations may point to the potential of both UBE2S and HIF1α as biomarkers for adverse prognosis in ESCC. To further investigate the combined effect of UBE2S and HIF1α, a key focus of this study, by classifying patients into four groups based on co-expression status. Kaplan–Meier analysis revealed significant stratification of both OS (p < 0.001) and PFS (p = 0.005) among these groups (Figs. 4E–4F). Patients with dual-positive expression (Group 4) had the poorest outcomes, while those with dual-negative expression (Group 1) had the most favorable prognosis. Groups with discordant expression (Groups 2 & 3) showed intermediate survival, suggesting a potential synergistic adverse effect when both proteins are highly expressed.

Exploratory stratified analysis of UBE2S prognostic value

To assess the prognostic impact of UBE2S across different patient subgroups, an exploratory stratified analysis was performed. UBE2S-positive expression was associated with reduced OS in subgroups including patients aged >60 years (Fig. 5A, p = 0.047), males (Fig. 5B, p = 0.022), Kazak (Fig. 5C, p = 0.014), tumor size ≥ three cm (Fig. 5D, p = 0.025), mid-esophageal tumors (Fig. 5E, p = 0.005), and moderately to poorly differentiated histology (Figs. 5F–5G, p = 0.017, p = 0.033). Shorter PFS was specifically observed in UBE2S-positive patients with mid-esophageal tumors (Fig. 5H, p = 0.035) and Kazak (Fig. 5I, p = 0.050). It is important to note that these stratified analyses were exploratory. After Bonferroni correction for multiple comparisons, the subgroup p-values did not retain formal significance, and interaction tests were non-significant. Thus, these findings should be interpreted as hypothesis-generating trends requiring validation in future studies.