In July 2022, data on the gene expression of ESCA tissues and normal esophageal tissues were obtained from The Cancer Genome Atlas (TCGA) and Genotype-Tissue Expression (GTEx) databases. Among them, 162 ESCA tissue samples and 11 tumor-adjacent normal esophageal epithelial tissue samples were obtained from TCGA database and 653 normal esophageal tissue samples were obtained from the GTEx database. The data on the clinical characteristics and survival information of 173 patients with ESCA were also downloaded from TCGA database. The data on ESCA included transcripts per million (TPM) types and fragments per kilobase per million (FPKM) types. The level of expression of the DPY30 mRNA in ESCA tissues and normal esophageal tissues was determined using the aforementioned data.

In total, 57 patients with ESCA underwent surgical resection at Union Hospital of Huazhong University of Science and Technology, Tongji Medical College (Wuhan, China) between 2013 and 2015. Paraffin specimens (57 pairs) from these patients were obtained by surgical resection. The clinical features of the patients are presented in Table SI. All ESCA tissues and paired paracancerous tissues (diameter of 1.5 mm; selected by a pathologist) from each paraffin specimen were extracted and used to constructed a tissue microarray (TMA) paraffin block. After checking the quality of the specimens, the TMA was cut into sections (3 mm thick) for immunohistochemistry analysis. The study was approved by the Institutional Ethics Committee of Tongji Medical College, Huazhong University of Science and Technology [approval number (2021) 0158]. Written informed consent was obtained from all patients before they underwent surgical procedures.

Immunohistochemistry (IHC) staining was performed for the TMA with 57 pairs of tissues by two professional pathologists. The primary rabbit anti-DPY30 antibody (dilution 1:100; incubation at 37°C for 30 min; cat. no. CSB-PA861193LA01HU) was purchased from Cusabio Technology LLC. Using the CaseViewer 2.4 software (3DHISTECH Ltd.) and HALO image analysis software (v2.1.1637.26; Indica Labs), images were captured and analyzed. Each specimen was first assigned a score based on the staining intensity and the extent of stained cells (no staining=0, weak staining=1, moderate staining=2 and strong staining=3) and (0–5%=0, 5–25%=1, 26–50%=2, 51–75%=3 and 76–100%=4). Based on the AI pathwell v2 image analysis software (Servicebio) using the principles of AI deep learning, the percentage of weak, moderate and strong intensity, percentage of positive area, mean density and area density was calculated (Table SII). Then, two professional pathologists validated the results of the immunohistochemical staining analysis and the results were automatically calculated for each specimen based on the original basic data and formulae. The final IHC scores were calculated using the formula ∑(pi × i)=(percentage of weak intensity ×1) + (percentage of moderate intensity ×2) + (percentage of strong intensity ×3). Here, ‘pi’ represents the ratio of the positive signal pixel area and ‘i’ represents the staining intensity; the IHC scores (area) are data between 0–300 (17).

By performing the Receiver Operating Curve (ROC) analysis in TCGA database, the diagnostic value of DPY30 expression in ESCA was assessed using the area under the curve (AUC). Kaplan-Meier survival analysis was performed to determine the relationship between DPY30 expression and overall survival (OS) in patients with ESCA. The groups were formed based on the median value of the expression of DPY30. Based on the clinical characteristics of the patients with ESCA in TCGA database, the relationship between clinicopathological characteristics and DPY30 mRNA expression levels was first determined to establish a prognostic nomogram. Then, according to the collected clinicopathological characteristics from 57 patients with ESCA, the relationship between the clinicopathological characteristics and DPY30 protein expression levels was confirmed.

Based on the sequencing data of tumor tissues from patients with ESCA in TCGA, the DPY30-associated genes were analyzed by Spearman's method. The co-expressed genes of DPY30 were screened with correlation coefficients greater than 0.4 and P-values less than 0.01. The genes were then divided into two groups based on the median value of the expression level of DPY30. The genes that were significantly correlated (positively and negatively) with DPY30 were obtained by performing the Gene Set Enrichment Analysis (GSEA). A protein-protein correlation network was constructed based on the co-expressed genes using the STRING database (http://string-db.org/) and Cytoscape software (version 3.7.2; www.cytoscape.org). The Gene Ontology (GO)/Kyoto Encyclopedia of Genes and Genomes (KEGG) function analysis was conducted to determine the role of DPY30 in ESCA.

ESCA tissues from TCGA database were evaluated by performing immune scoring via the Estimation of Stromal and Immune cells in Malignant Tumor tissues using the Expression data (ESTIMATE) and ssGSEA algorithm (18,19). The relationship between the DPY30 expression level and ESCA infiltrating immune cells was investigated using Spearman's correlation.

Statistical analysis was performed using SPSS v22.0 statistical software (IBM Corp.). The significant differences between groups were determined by performing unpaired and paired Student's t-tests. The association between DPY30 expression levels with the clinicopathological features of patients with ESCA was evaluated by performing a one-way ANOVA. If the association was significant, it was evaluated by Scheffe post hoc test (20). All quantitative results were expressed as the mean ± SD. P<0.05 was considered to indicate a statistically significant difference.

Based on TCGA and GTEx databases, the results demonstrated that the expression of DPY30 mRNA was significantly higher in unpaired ESCA tissues compared with normal esophageal tissues (Fig. 1A-C). For one-to-one matching, the expression of DPY30 mRNA in paired ESCA tissues was significantly higher compared with normal esophageal tissues (Fig. 1D and E). Additionally, the results of the immunohistochemistry analysis demonstrated that the protein of DPY30 was significantly overexpressed in ESCA tissues compared with that in paracancerous tissues (Figs. 1F and S1). For the DPY30 protein expression in the clinical TMA from 57 paired ESCA (ESCA) tissues, the representative immunohistochemical images of the high, medium and low protein expression of DPY30 are shown in Fig. 2. Thus, DPY30 was overexpressed in ESCA patients and might influence the diagnosis and prognosis of ESCA.

The diagnostic value of DPY30 in ESCA was determined based on the ROC curves. The TPM type of TCGA database analysis demonstrated that the area under the ROC curve (AUC) value of DPY30 was 0.938 in ESCA (Fig. 3A). By analyzing the TPM type of TCGA and GTEx databases, the AUC value of DPY30 was calculated to be 0.858 in ESCA (Fig. 3B). By analyzing the FPKM type of TCGA database, the AUC value of DPY30 was calculated to be 0.909 in ESCA (Fig. 3C). These results suggested that DPY30 plays an important role in the diagnosis of ESCA.

The prognostic value of DPY30 was determined using different types of TCGA databases. Based on the TPM and FPKM types of TCGA databases, the results of the analysis revealed a significant correlation between the high expression of DPY30 and a poor prognosis in ESCA (Fig. 4). High expression of DPY30, predicted worse OS, disease-specific survival (DSS) and progression-free survival (PFS) based on different data types in patients with ESCA. The results of the prognostic analysis of patients with ESCA demonstrated that the high expression of DPY30 predicted poor overall survival.

The correlation between the expression level of the DPY30 mRNA and the clinicopathological features of patients with ESCA was analyzed from TCGA database. The results demonstrated that DPY30 overexpression was associated with the advanced pathologic T stage and advanced pathologic N stage (Fig. 5A and B). The expression levels of DPY30 were higher in esophageal squamous carcinoma than that in esophageal adenocarcinoma (Fig. 5C). More notably, the overexpression of DPY30 suggested a worse outcome for the neoadjuvant treatment of ESCA (Fig. 5D). The results of the analysis of the clinical characteristics and the expression level of the DPY30 protein in 57 patients with ESCA (Table I) demonstrated that DPY30 overexpression was associated with the advanced pathologic TNM stage and OS (Fig. 6).

To find the prognostic risk factor for patients with ESCA in TCGA database, univariate Cox regression analysis was performed to show that pathologic N stage, pathologic M stage, pathological stage, treatment outcome and DPY30 mRNA expression level might be the risk factors for OS (Table II). Based on univariate Cox regression was performed to analyze the clinical characteristics of 57 patients with ESCA. The results demonstrated that the overexpression of DPY30 was associated with the advanced pathologic N stage, pathologic TNM stage and the expression level of the DPY30 protein (Table III). The results of the multivariate Cox regression analysis did not show independent risk factors for ESCA in patients with ESCA. An accurate prognostic nomogram based on the clinicopathological characteristics and DPY30 expression level in ESCA was constructed (Fig. 7).

Based on the correlation analysis of the DPY30-related genes, 669 genes were found to be co-expressed with DPY30 (Table SIII). The top 12 genes with positive and negative correlations with DPY30 in ESCA are shown in Figs. 8 and S2. Based on GO analysis, it was found that DPY30 and its co-expressed genes were involved in RNA and DNA modification, regulation of mitochondria and lysosome function and interference of the cell cycle (Fig. 9A-C). The results of the KEGG analysis demonstrated that these genes were associated with RNA transport and degradation, spliceosome, ribosome, oxidative phosphorylation, the cell cycle and other mechanisms (Fig. 9D and Table SIV). To visualize the relationship of the DPY30 co-expressed genes, the PPI network was constructed (Fig. 10).

Based on the diagnostic and prognostic value of DPY30 in ESCA, the ESTIMATE algorithm was used to further determine the value of DPY30 in the tumor microenvironment of ESCA. Based on the result of the Stromal Score, Immune Score and ESTIMATE Score, the high expression of DPY30 was found to be related to low immune scores in the tumor microenvironment of ESCA (Fig. 11A-C). DPY30 expression was negatively correlated with the overall immune infiltration, stromal content and the combined score of both (Fig. 11D and E). The analysis of the correlation between DPY30 expression and immune cells in ESCA demonstrated that DPY30 expression was significantly correlated with the number of types of immune cells (Table SV; Figs. S3 and S4). Among them, mast cells, effector memory T cells, eosinophils, neutrophils, central memory T cells, dendritic cells, T cells, T follicular helper cells and B cells demonstrated a significant negative correlation with the expression of DPY30 (Fig. 12).