We simultaneously extracted APOC1 expression and clinical data from the TCGA (https://portal.gdc.cancer.gov/) and GTEx (https://gtexportal.org/) databases in ESCA tissues, adjacent tumor tissues, and normal tissues to analyze the differences in APOC1 expression between tumor and normal tissues and the relationship between the expression of APOC1 and clinical characteristics. The data sets GSE161533 (84 paired normal tissues, para-tumor tissues, and tumor tissues from 28 ESCC patients microarray experiments performed by microarray analysis), GSE20347 (RNA was extracted from 17 paired tumor tissues and matched normal adjacent tissue from ESCC patients using the PureLink Micro-to-Midi Total RNA Purification System), and GSE29001(Gene expression from twelve cases of patient-matched normal basal epithelial cells, normal differentiated squamous epithelium, and cancer by expression microarrays.) were downloaded from the GEO (https://www.ncbi.nlm.nih.gov/gds/) database for the analysis of APOC1 expression differences between tumor and normal tissues and between tumor and adjacent tissues, respectively.

After analyzing the difference between APOC1 expression in ESCA tissues and normal tissues, we performed a ROC curve analysis to detect the diagnostic ability of APOC1 for ESCA.

Following the analysis of the diagnostic value of APOC1 for ESCA, we used Kaplan–Meier (KM) curves to analyze the influence of APOC1 expression status on the prognosis of patients with ESCA.

We compared the relationship between APOC1 expression status and the overall prognosis of ESCA patients by univariate and multifactorial analyses. We also analyzed whether APOC1 expression is an independent risk factor for the poor prognosis of ESCC patients. The significant criterion for COX analysis was set at p < 0.05.

Differential analysis of APOC1 expression data (|log2 (fold change|) ≥1, adj. p < 0.05) was performed using the R (version 3.6.3) (R Core Team (2021). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/) limma package. Gene Ontology/Kyoto Encyclopedia of Genes and Genome (GO/KEGG) analysis was used to obtain the correlation between APOC1 and ESCC, (adj. p < 0.05 was considered statistically significant), and the above results were visualized by applying the ggplot2 package (https://CRAN.R-project.org/package=ggplot2).

Gene Set Enrichment Analysis (GSEA) was performed using the R (version 3.6.3) cluster profiler and ggplot2 packages. Statistical analysis with |NES| ((normalized enrichment score))>1, p < 0.05, and FDR (false discovery rate) > 0.25 was considered to indicate statistically significant differences.

In the protein interaction analysis, we used STRING online tool (31) to study and analyze the protein-protein binding to APOC1 protein-protein interaction (PPI) with APOC1. The network was operated to obtain 50 proteins that bind to APOC1, and the genes encoding these 50 proteins were intersected with the differentially expressed genes associated with APOC1 in ESCA tissues. We then analyzed the correlation analysis of APOC1 with these intersecting genes separately using Spearman’s relationship analysis. Statistical significance was considered at p < 0.05.

We used the TISIDB (32) (http://cis.hku.hk/TISIDB/browse.php?gene=APOC1) web tool and Gene Set Variation Analysis (GSVA) method to analyze the relationship between the expression of APOC1 and tumor-infiltrating lymphocytes (TILs). The correlation between APOC1 and immune checkpoint genes in ESCA tissues was also analyzed, and the link between APOC1 expression and migration of immune cells was analyzed by chemokine and chemokine receptor modules.

We performed statistical analysis using the R (version 3.6.3) and plotted receiver operating characteristic (ROC) curves. The correlation between APOC1 mRNA expression and clinical characteristics of ESCA patients was tested using a chi-square test.

We extracted clinical data from the TCGA-GTEx database for 173 patients (162 patients with tumors and 11 healthy controls). The clinical information is shown in Table 1.

To understand the expression status of APOC1 in tumor tissues, adjacent tumor tissues, and normal tissues, we analyzed the expression of APOC1 in various types of tumors from the TCGA(10534 patients) and GTEx (15776 patients) databases. As shown in Figures 1A, B, the expression of APOC1 was significantly upregulated in tumor tissue compared with normal tissue in Breast invasive carcinoma (BRCA), Cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC), Colon adenocarcinoma (COAD), ESCA, Kidney Chromophobe (KICH), Kidney renal clear cell carcinoma (KIRC), Kidney renal papillary cell carcinoma (KIRP), and Ovarian serous cystadenocarcinoma (OV). However, the expression level of APOC1 in tumors was significantly lower than in normal tissues in Adrenocortical carcinoma (ACC). APOC1 expression was not significantly different between Lung adenocarcinoma (LUAD) and normal tissues. We also analyzed APOC1 expression in various tumor tissues and tumor-adjacent tissues. APOC1 was more highly expressed in tumor tissues than tumor adjacent in Bladder Urothelial Carcinoma (BLCA), COAD, Head and Neck squamous cell carcinoma (HNSC), KICH, KIRC, KIRP, Prostate adenocarcinoma (PRAD), Stomach adenocarcinoma (STAD), Thyroid carcinoma (THCA), and Uterine Corpus Endometrial Carcinoma (UCEC), while it was less expressed in tumor tissues than tumor adjacent tissues in Cholangiocarcinoma (CHOL), LUAD, and LUSC(lung squamous cell carcinoma), with no significant difference in expression in reading (Figures 1C, D). Next, we analyzed the transcription levels of APOC1 with the TCGA, GTEx, and GEO databases. The analysis showed that APOC1 mRNA expression was significantly upregulated in ESCA tissues compared with normal tissues (p < 0.001) (Figure 2A), and the expression of APOC1 mRNA was also significantly increased in ESCA tissues compared with tumor adjacent tissues (p < 0.01) (Figure 2B).

The diagnostic value of APOC1 mRNA expression in ESCA was analyzed in our study using the ROC method. Our results showed that the area under the curve (AUC) was 0.887. We also analyzed the diagnostic value of APOC1 mRNA expression in different stages, grading, lymph node metastasis, and distant metastasis. The AUC values were 0.808, 0.895, 0.926, and 0.898 by stage I, stage II, stage III, and stage IV, respectively; 0.862, 0.899, 0.900, and 1.000 by T stage T1, T2, T3, and T4, respectively; 0.884, 0.890, 0.949, and 0.955 by lymph node metastasis stages N0, N1, N2, and N3, respectively; and 0.898 for both M0 and M1 according to M staging (Supplementary Figure S1).

We applied the K-M survival curve to analyze the prognostic impact of APOC1 mRNA expression on patients with ESCA. The results showed that APOC1 mRNA expression had no significant effect on all tumors including EAC (p = 0.052) (Figure 3A) and EAC alone (p = 0.09) (Figure 3B), while patients with ESCC who overexpressed APOC1 mRNA in their tissues had a worse survival prognosis (p = 0.043) (Figure 3C). Overexpression of APOC1 mRNA also significantly affected the OS of patients with ESCC at T-stage (T1 and T2) (p = 0.023) (Figure 3D), pathological stage (TNM-stage, AJCC prognostic stage) (stage I and stage II) (p = 0.039) (Figure 3E), tumor histological grade (grade 1 and grade 2) (p = 0.009) (Figure 3F), and columnar metaplasia (p = 0.027) ESCC (Figure 3G), while the effect was not significant in T-stage (T3 and T4) (p = 0.282) (Figure 3H) and pathological stage (TNM-stage, AJCC prognostic stage) (Stage III and Stage IV) (p = 0.544) (Figure 3I) as shown in the subgroup analysis. Univariate and multifactorial analyses showed that high expression of APOC1mRNA, lymph node metastasis, and sex were associated with OS of patients with ESCC, and APOC1 mRNA was likely an independent prognostic risk factor for OS in patients with ESCC (Table 2; Figure 4).

GO functional annotation analysis revealed that APOC1 co-expressed genes in ESCC were mainly involved in the following aspects: regulation of T cell activation, cornification, cytolysis, external side of plasma membrane, MHC protein complex, MHC class II protein complex, serine−type peptidase activity, serine-type endopeptidase activity (Supplementary Table S1; Figure 5A). The KEGG pathway enrichment analysis showed that co-expressed genes of APOC1 were mainly enriched in Staphylococcus aureus infection, antigen processing and presentation, and graft-versus-host disease pathways (Supplementary Table S2; Figure 5B).

The signaling pathways associated with ESCC were analyzed by the GSEA method. The results showed that immunoregulatory interactions between a lymphoid and a non-lymphoid cell and FCERI-mediated NF-KB activation were significantly enriched in APOC1-related phenotypes were significantly enriched (Figure 6).

Through the establishment of protein interactions signaling network, we can understand the metabolic and molecular mechanism of ESCC. Therefore, we analyzed the interaction of APOC1 in ESCC through the STRING website protein PPI network. First, on the STING website, we set the parameters for text mining and experimental evidence identification, and then we obtained the APOC1-related protein interactions network (Figure 7A). In addition, we also took the proteins interacting with APOC1 and the differential proteins associated with APOC1 expression in ESCC for intersection analysis and screened the common proteins as APOE, SAA1, APOF, LPL, APOC2, respectively (Figure 7B). Next, we analyzed the correlation between the expression of APOC1 and that of these common proteins, and the results showed that APOC1 had a significant positive correlation with the expression of APOE (r = 0.324, p = 0.003); APOF (r = 0.932, p < 0.001); LPL (r = 0.550, p < 0.001); and APOC2 (r = 0.729, p < 0.001) (Figures 7C–F).

The results of this study showed that in ESCC tissue, a variety of immune infiltrating cells were positively correlated with APOC1 expression such as T cells (r = 0.376, p < 0.001); aDC (antibody-dependent cytotoxicity) (r = 0.411, p < 0.001); CD8T cells (r = 0.402, p < 0.001), cytotoxic cells (r = 0.463, p < 0.001), iDC (immature dendritic cell) (r = 0.507, p < 0.001), macrophages (r = 0.551, p < 0.001), NK CD56 dim cells (r = 0.381, p < 0.001), Th1 cells (r = 0.346, p = 0.002), and Treg (r = 0.460, p < 0.001) (Figures 8A, B). Moreover, cell grouping analysis and individual cell analysis showed a significant positive correlation between APOC1 and immune infiltrating cells (Figures 8C–K). To further explore the relationship between APOC1 expression and tumor-infiltrating lymphocytes (TILs), we examined the relationship between TILs and APOC1 expression through the TISIDB website. The results showed a significant positive correlation between APOC1 and multiple TILs cells in ESCA (Supplementary Figure S2). Given that current targeted drug therapy against tumors benefits patients with a wide range of tumors (33), in this study we also analyzed the relationship between APOC1 and various tumor control-related genes. Our results showed that tumor targeting-related immune checkpoint genes such as CD274 (p = 0.0137), TGFB1 (p = 0.0175), PDCD1 (p < 0.001), CTLA4 (p < 0.001), and LAG3 (p < 0.001) were significantly associated with APOC1 (Figure 9). These significant correlations may be related to the poorer prognosis of patients with ESCC owing to high expression of APOC1. Interestingly, APOC1 expression was also significantly correlated with inflammatory factors such as IL10 (p < 0.001) (Figure 9). In addition, we also analyzed the relationship between APOC1 expression and chemokines (Supplementary Figure S3) and chemokine receptor (Supplementary Figure S4). The analysis showed a significant positive correlation between APOC1 and a large variety of immune chemokines and chemokine receptors. Combined with the results of correlation analysis of APOC1 expression with target molecules, immune-related molecules, and inflammatory chemokines, AOPC1 may be involved in immune infiltration and inflammatory responses in the tumor microenvironment and tumorigenesis and development.