High Expression of CEMIP Correlates Poor Prognosis and the Tumur Microenvironment in Breast Cancer as a Promisingly Prognostic Biomarker

Abstract

Cell migration-inducing hyaluronidase 1 (CEMIP), a Wnt-related protein and also known as KIAA1199, is implicated in the process of metastatic colonization in a variety of malignant tumors, including breast cancer (BC), which is one of the most frequently diagnosed tumors in women worldwide. In this study, multiple public databases, online analytical tools, and bioinformatics approaches were applied to explore the expression levels, regulatory mechanisms, and biological functions of CEMIP in BC. We illustrated that CEMIP was highly expressed in various kinds of carcinomas, including BC, especially advanced subtypes, and predicted less favorable prognosis (negatively associated with overall survival) in BC patients, which might be an independent prognostic factor. Then, we revealed that the mutation and high expression of CEMIP might lead to it as an oncogene. We also demonstrated that TP53 mutation, DNA hypo-methylation, and the expression changes of three potential upstream transcription factors (EZH2, EGR1, and JUN) of CEMIP were likely to cause the hyperexpression of CEMIP in BC. Moreover, our findings suggested that CEMIP might exert its carcinogenic roles in the tumor microenvironment via participation in the extracellular matrix formation, increasing cancer-associated fibroblast (CAF), M2 macrophage, and neutrophil infiltration and decreasing CD8⁺ T cell infiltration. In summary, our study provided more solid evidence for CEMIP as a prognostic and metastatic biomarker and a potential therapeutic target in BC. Of course, these findings also need more confirmations of basic experiments and further clinical trials in the future.

Keywords: CEMIP; biomarker; breast cancer; prognosis; tumor microenvironment.

FIGURE 1

Expression analysis of CEMIP across a variety of cancers and clinical features of BC. (A) Expression profile of CEMIP in various cancers analyzed by TIMER2.0 based on the TCGA database. *p < 0.05; **p < 0.01; ***p < 0.001. (B,C) Expression analysis of CEMIP performed by the UALCAN web based on sample types and TP53 mutation status with the TCGA database. ***p < 0.001. (D–F) Expression analysis of CEMIP carried out through the Breast Cancer Gene-Expression Miner v4.7 resource based on BC subtype, patient age, and TP53 mutation status with 11,359 DNA microarrays of bc-GenExMiner data. bc-GenExMiner, the Breast Cancer Gene-Expression Miner; TCGA, The Cancer Genome Atlas.

FIGURE 2

Survival analysis. (A–C) OS, RFS, and DMFS plots of CEMIP in BC with the Kaplan-Meier plotter. (D–F) OS, DFS, and DMFS plots of CEMIP in all BC subtypes with the bc-GenExMiner. (G) OS plot of CEMIP in ER/PR⁺ BC with bc-GenExMiner. Log rank p < 0.05 and p < 0.05 were considered to be statistically significant. OS, overall survival; RFS, elapse-free survival; DMFS, distant metastasis-free survival; DFS, disease-free survival.

FIGURE 3

Promoter methylation analysis using the UALCAN web and genetic alteration analysis performed by the cBioPortal web based on TCGA. (A) Promoter methylation level of CEMIP in BC compared with normal breast samples. ***p < 0.001. (B) Promoter methylation level of CEMIP in BC subtypes. *p < 0.05; ***p < 0.001. (C) Genetic alteration frequency of CEMIP in multiple malignancies. (D) Mutation site of CEMIP in BC.

FIGURE 4

Upstream transcription factor and kinase interaction analysis. (A) Differentially expressed predicted upstream TFs of CEMIP and kinases interacted with CEMIP identified in BC based on DEGs detected from the GSE42568 dataset. (B–F) Expression correlations of CEMIP and validated TFs in BC. (G) Functions and pathways enriched by kinases interacted with CEMIP and differentially expressed in BC. DEGs, differentially expressed genes.

FIGURE 5

Expression, survival, and correlation analysis of upstream TFs of CEMIP in BC. (A–C) Expression levels of EZH2, EGR1, and JUN among subtypes of BC. (D–F) OS, DFS, and DMFS analysis of EZH2 in BC patients. (G–I) OS, DFS, and DMFS analysis of EGR1 in BC patients. (J–L) OS, DFS, and DMFS analysis of JUN in BC patients. (M–O) Correlations between EZH2 and EGR1, EZH2 and JUN, and EGR1 and JUN in BC.

FIGURE 6

PPI network construction and enrichment analysis of CEMIP-related genes. (A) PPI network of CEMIP-related genes generated by the STRING tool. The red nodes represent up-regulated genes in BC compared with normal samples, while the blue nodes represent down-regulated genes. (B) Top 10 BP terms of GO enriched by CEMIP-related genes. (C) Top 10 CC terms of GO enriched by CEMIP-related genes. (D) Top 10 MF terms of GO enriched by CEMIP-related genes. (E) Top 10 KEGG pathways enriched by CEMIP-related genes. PPI, protein–protein interaction; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; BP, biological processes; CC, cellular components; MF, molecular functions.

FIGURE 7

Correlation of CEMIP expression with CAF infiltration level explored by TIMER2.0 web. (A) Correlations between CEMIP expression and the infiltration level of CAF across various types of cancers analyzed by four algorithms provided (EPIC, MCPCOUNTER, XCELL, and TIDE). The red squares represent positive correlations, while the blue squares represent negative correlations with statistical significance (p < 0.05). (B–E) Correlations between CEMIP expression and the infiltration level of CAF among four subtypes of BC (basal-like/TNBC, HER2⁺, luminal A, and luminal B) using the EPIC algorithm. CAF, cancer-associated fibroblast.

FIGURE 8

Gene Set Enrichment Analysis based on the GSE42568 dataset and the gene sets of KEGG pathways and HALLMARK. (A,B) Top two KEGG pathways enriched by the CEMIP high expression group with the highest NES and NOM p-val < 0.05. (C,D) Two HALLMARK terms enriched by the CEMIP high-expression group with NOM p-value < 0.05. NES, normalized enrichment score; NOM p-value, nominal p-value.