Pan-cancer analysis of oncogenic role of CEP55 and experiment validation in clear cell renal cell carcinoma

Abstract

Immunotherapy has emerged as a vital component in the contemporary landscape of cancer treatment. Recent studies have indicated that CEP55 plays an oncogenic role; however, its specific mechanisms in promoting tumor proliferation and its potential value in prognosis and immunotherapy prediction across various cancers remain to be elucidated. CEP55 was significantly overexpressed in 22 cancer types compared with their adjacent normal tissues. Elevated CEP55 expression was positively correlated with younger onset age, worse tumor stage, lower response rate to the first treatment, lower tumor-free survival rate, and poorer overall survival (OS) and disease-free survival (DFS) prognosis in most cancers. Moreover, CEP55 expression was positively correlated with its binding and related genes, such as KIF11 (R = 0.83, P < 0.001), CDK1 (R = 0.77, P < 0.001) and CCNA2 (R = 0.76, P < 0.001), and the classic proliferation markers, including MKI67 and PCNA. Enrichment analyses indicated that CEP55 was predominantly associated with cell division, cell cycle activities and proliferation. Immune cell infiltration analysis by TIMER2.0 revealed that CEP55 expression was positively correlated with many kinds of infiltrating cells, such as Th2 cells and some CD4⁺ T cell subsets. The CEP55 expression was positively associated with increased MSI and TMB in various cancers. Our analyzation indicated that the CEP55 expression level in patients with complete remission (CR) or partial remission (PR) to anti-PDL1 therapy was significantly higher than patients with stable disease (SD) or progressive disease (PD) based on IMvigor210 cohort. We also used Gene Set Cancer Analysis (GSCA) to predict a serious of small molecule CEP55 targeted drugs, such as AZ628, SB52334, SB590885, A-770,041, AZD7762, Elesclomol, panobinostat, BRD-A94377914, and LRRK2-IN-1. Furthermore, the patients with high level of CEP55-posivie tumor epithelial cells had inferior overall survival in ccRCC according to single-cell analysis. Finally, our wet lab experiments verified that the CEP55-positive rate in ccRCC tissues (19/30, 63.3%) was significantly higher than that in renal adjacent tissues (10/30, 33.3%). The clinicopathologic analysis revealed that CEP55 protein level was significantly associated with tumor size (P = 0.044), histology grade (P < 0.001) and stage (P = 0.034). Our study indicated that CEP55 overexpression in most caner types was associated with poor prognosis. Notably, CEP55 was closely relevant to immune cell infiltration and impacted the response to immunotherapy and small molecule drugs against cancers.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-024-80057-2.

Introduction

Cancer has been one of the top killers worldwide, and the number of cancer patients continues to increase. Early diagnosis and targeted cancer therapies would help reduce cancer mortality¹. Immunotherapy has been established as a pillar of cancer therapy improving the outcomes of a broad variety of cancers². According to the comprehensive survey of the global immuno-oncology landscape, novel immunotherapies, including chimeric antigen receptor (CAR) T cell therapies, anti-PD1, anti-PDL1 or anti-CTLA4 monoclonal antibody, CD3-targeted bispecific antibody and so on, have become the front-line treatments or the standard of care for most cancer types since 2011³. Despite the remarkable efficacy of cancer immunotherapies, obstacles and challenges still exist and hinder the further development and clinical application of immunotherapies. The common challenges include limited response rates and immune-mediated toxicities, such as cytokine storm and immune effector cell-associated neurotoxicity syndrome⁴. Therefore, to expect higher anti-tumor efficacy and better survival outcomes in malignancies, looking for potential immune biomarkers predicting better response are being demanded.

Centrosomal Protein 55 (CEP55), called c10orf3 and FLJ10540, is originally discovered as the mitotic phosphoprotein. CEP55 has an important effect on final cytokinesis stage and promotes abscission⁵. CEP55 activates PI3K/AKT and FOXM1-dependent pathways, which interact with the proteins related with angiogenesis, proliferation and metastasis in tumorigenesis⁶. CEP55 is closely related to tumor stage, tumor invasiveness, poor prognosis, and metastasis⁷. In addition, growing evidence shows that CEP55 over-expresses within various human cancers, such as liver cancer⁸, breast cancer⁹, renal cell carcinoma¹⁰ and so on. CEP55 overexpression promotes genomic instability suggesting a potential role in tumor immune microenvironment (TIME)¹¹. Furthermore, increased CEP55 expression is related to poor survival, and can determine immune infiltration and prognosis of liver cancer cases¹². Owing to the limited studies on the influence of CEP55 on TIME and its immunotherapy value in multiple tumors, this work was conducted to investigate CEP55’s function in pan-cancers.

To this end, CEP55 expressions within pan-cancers were analyzed from several databases, besides, the relation of CEP55 expression with tumor clinical features and survival was validated. Moreover, CEP55’s molecular biological activity was analyzed based on enrichment analysis, and the relation of CEP55 with immune cell infiltration, microsatellite instability (MSI), tumor mutational burden (TMB), immunomodulatory genes, immunotherapy response and sensitive drugs was further examined. Especially, we verified CEP55 expression and its relationship with CEP55 clinical characteristics within clear cell renal carcinoma (ccRCC) samples through immunohistochemistry. Our results revealed that the upregulation of CEP55 predicted poor prognosis and affected the cancer immunotherapeutic sensitivity and immune microenvironment.

Materials and methods

Data extraction

Transcriptome data from 33 tumor as well as matched clinical characters were obtained via UCSC Xena (https://xena.ucsc.edu/). After eliminating duplicate and missing data, we converted expression data into log2 (TPM).

CEP55 protein expression and localization

The CEP55 total protein levels within 12 tumors were conducted by the model of CPTAC (Clinical proteomic tumor analysis consortium) in UALCAN portal (http://ualcan.path.uab.edu/analysis-prot. html)^13,14. The Human Protein Atlas (HPA; www.proteinatlas. org) was utilized to analyze CEP55’s subcellular localization in two human cancer cells (A-431 and U251) by immunofluorescence staining images¹⁵.

Survival analysis

“Survival Analysis” and “Survival Map” modules of Gene Expression Profiling Interactive Analysis 2 (GEPIA2) database (http://gepia2.cancer-pku.cn) were utilized for examining relationship between CEP55 level and overall survival (OS) as well as disease-free survival (DFS) in distinct tumor patients¹⁶. Patients in each cancer type were divided into CEP55 low- and high-expression groups according to its median expression. P < 0.05 stood for significance.

Enrichment and function analysis

STRING database (https://string-db.org/) was searched using one individual protein name (“CEP55”) and organism (“Homo sapiens”). Then, the main parameters were set, including active interaction sources (“co-expression” and “experiments”), minimal required interaction score [“high confidence (0.700)”] and maximal interactors for display [“custom value” and “100”]. As a result, the CEP55-binding proteins were retrieved. Next, we acquired “Similar Gene Detection” module in GEPIA2 for obtaining 100 most significant CEP55-related genes according to pan-cancer data. Venny, an interactive online software, was used to compare the intersected genes between the CEP55-binding proteins and CEP55-related genes (https://bioinfogp.cnb.csic.es/tools/venny/index.html). Furthermore, the intersected genes were applied for GO (Gene ontology) as well as KEGG (Kyoto encyclopedia of genes and genomes) analysis with R packages “clusterProfiler”.

We employed CancerSEA (http://biocc.hrbmu.edu.cn/CancerSEA/) for evaluating following 14 CEP55 functional states within various cancers, such as stemness, metastasis, invasion, proliferation, EMT, apoptosis, angiogenesis, differentiation, cell cycle, DNA repair, DNA damage, hypoxia, quiescence, and inflammation¹⁷.

Immune cell infiltration analysis of CEP55

Immune cell infiltration fractions were estimated using the database-driven web of Tumor Immune Estimation Resource 2.0 (TIMER2.0, http://timer.cistrome.org/) through various quantification methods¹⁸. The CEP55-associated immune cell infiltration correlations were visualized by entering the gene name and selecting the certain immune cell by “Gene” function in “Immune Association”. Totally, 21 immune cell subpopulations were evaluated, namely, cancer-associated fibroblast (CAF), progenitors of lymphoid, neutrophils, B cells, CD4⁺ T cells, hematopoietic stem cells (HSC), progenitors of myeloid, progenitors of monocytes, eosinophil (Eos), endothelial cells (Endo), regulatory T cells (Tregs), NK T cells, T cell follicular helper, γ/δ T cells, macrophages, monocytes, CD8⁺ T cells, dendritic cells, NK cells, and mast cells.

Immunotherapy prediction analysis

For analyzing the relationship of CEP55 with TMB, MSI and immunomodulators across pan-cancer, we used Spearman correlation. For exploring the relation of CEP55 with response to immune checkpoint blockade (ICB) treatment, IMvigor210 cohort was applied for validating CEP55’s capacity in predicting response to immunotherapy, which includes 298 urological cancer cases receiving atezolizumab (anti-PDL1) treatment¹⁹.

Drug sensitivity analysis

CEP55 expression profiling and drug sensitivity data in tumor cells from Therapeutics Response Portal (CTRP) and Genomics of Drug Sensitivity in Cancer (GDSC) were obtained in Gene Set Cancer Analysis (GSCA) platform and integrated for investigation²⁰. Spearman correlation was conducted to evaluate relation of gene expression drug, with positive correlation indicating resistance of gene up-regulation to drug.

Collection of scRNA-seq data

The scRNA-seq data of ccRCC were downloaded from GSE159115 in the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo), which included 8 renal tumor specimens and 6 benign human kidney specimens from ccRCC and chromophobe renal cell carcinoma (chRCC) patients. Then, we collected four paired ccRCC and normal kidney tissues for the following study and the sample information was presented in Table S1.

Single-cell quality control and cell type annotation

The Seurat R package was employed for quality control (QC)²¹. Cells with mitochondrial gene expression above 15% were excluded, and low-quality cells were filtered based on QC thresholds (nFeature-RNA < 200 or nFeature-RNA > 3000). Data normalization was conducted using the “SCTransform” function in Seurat. To generate cell clusters, the number of principal components (PCs) was set to 20. Clusters were identified using the FindClusters function (resolution = 0.4) and visualized through uniform manifold approximation and projection (UMAP)²¹. Differentially expressed genes within clusters were determined using the FindAllMarkers function with Wilcoxon rank-sum tests. Each cluster was annotated according to classical marker genes: epithelial cells (EPCAM, KRT19), stromal cells (PECAM1, VWF, ACTA2), immune cells (CD68, PTPRC, JCHAIN), and tumor epithelial cells (NDUFA4L2, CA9, NNMT).

Survival analysis of CEP55-positive and CEP55-negative tumor epithelial cells based on the TCGA kidney clear cell carcinoma (KIRC) cohort

FPKM-normalized gene expression values of 610 KIRC cases were downloaded from UCSC Xena and log2 transformed. Marker genes from CEP55-positive and CEP55-negative tumor epithelial cells were used as gene sets to perform single-sample gene set enrichment analysis (ssGSEA)²² on TCGA KIRC gene expression data using the R package “GSVA”²³. Associations between enrichment scores (ES) from ssGSEA were calculated using the R package “corrplot”. Next, patients were grouped by dichotomization of ssGSEA ES (> or ≤ median). Survival data of the TCGA KIRC cohort was downloaded from UCSC Xena. Survival curves and log-rank statistics were calculated using the R packages “survival” and “survminer”.

Tissue samples

Tissue chip (HkidE180su02) that included 30 non-carcinoma and 150 ccRCC samples was purchased from Shanghai Outdo Biotech Company. Surgical procedures were performed from February 2008 to March 2010. All patients were followed up till August 2015 for 5.5–7.5 years. Patients in this chip had complete clinical characteristics and follow-up data. The present work has gained approval from the Ethics Committee. Informed consent (KY2019YJ029, SHYJS-CP-1510001) was obtained from each patient. The research was performed in accordance with relevant guidelines and informed consent was obtained from all participants. The tissue chip was deparaffinized, hydrated, and immune stained using an antibody against human CEP55 (1:3000 dilution, Abcam, UK). Each tumor was scored for the staining intensity and extent in accordance with the criterion of the previous study²⁴. We then designated total score < 8 as low expression, whereas total score ≥ 8 as high expression. Immunohistochemical assessments were performed blindly by two independent pathologists.

Statistical analysis

This study performed statistical analysis with Graphpad Prism 7.0 and R 4.1.2 software. Protein and mRNA expressions within ccRCC and corresponding non-carcinoma kidney tissues were examined through Student’s t-test. One-way ANOVA for multiple groups was conducted to evaluate differences. The CEP55 expression was examined across diverse ccRCC clinicopathological factors with Mann-Whitney test. Data were indicated by mean ± SD of every group. P < 0.05 stood for statistical significance.

Results

CEP55 levels within pan-cancer

CEP55 expressions within 33 cancer types were examined across TCGA database. As a result, its expression was highest in endocervical adenocarcinoma (CESC) and cervical squamous cell carcinoma, but lowest in kidney chromophobe (KICH) (Fig. 1A). The levels of CEP55 expression in tumor and matched non-carcinoma samples were markedly upregulated within 22 cancers (Fig. 1B). Furthermore, CEP55 protein level was notably overexpressed within breast cancer (P = 0.001), ovarian cancer (P < 0.001), kidney renal clear cell carcinoma (KIRC, P < 0.001), uterine corpus endometrial carcinoma (UCEC, P < 0.001), lung adenocarcinoma (P < 0.001), head and neck squamous cell carcinoma (HNSC, P < 0.001), pancreatic adenocarcinoma (P < 0.05), and glioblastoma multiforme (P < 0.001), when relative to adjacent non-carcinoma samples (Fig. 1C). The CEP55 protein showed major centriolar satellite and plasma membrane location, as visualized in the A-431 and U251 cells using immunofluorescence staining (Fig. 1D). The above results suggested that CEP55 was an oncogene for many cancers, which was probably related to tumorigenesis and development.

Fig. 1.

Differential expression of CEP55. (A) CEP55 mRNA expression in normal tissues from TCGA data. (B) Differential CEP55 mRNA expression between cancers and normal tissues. The red column represents cancer samples, and the blue column represents normal samples. Normal group was normal tissue in TCGA and GTEX database. (C) CEP55 protein expression in different cancer types in CPTAC. (D) The immunofluorescence images of CEP55 protein, nucleus, endoplasmic reticulum (ER), microtubules and the merged images in A-431 and U251 cell lines from HPA database. (*P < 0.05, **P < 0.01, and ***P < 0.001).

CEP55 clinical characteristics within pan-cancer

The correlation between CEP55 mRNA level and the clinical characteristics in pan-cancer was explored. Higher expression of CEP55 was found in age ≥ 65 group in lower grade glioma (LGG, P < 0.001), prostate adenocarcinoma (PRAD, P < 0.05), and stomach adenocarcinoma (STAD, P < 0.01, Fig. 2A). The expression of CEP55 mRNA in adrenocortical carcinoma (ACC), breast invasive carcinoma (BRCA), colon adenocarcinoma (COAD), KICH, KIRC, kidney renal papillary cell carcinoma (KIRP), liver hepatocellular carcinoma (LIHC), lung adenocarcinoma (LUAD), lung squamous cell carcinoma (LUSC), testicular germ cell tumors (TGCT) and thyroid carcinoma (THCA) was significantly associated with tumor stage (Fig. 2B). Then, the response to the first therapy in pan-cancer patients was further conducted. The results showed that CEP55 expression was related to tumor treatment response in ACC, bladder urothelial carcinoma (BLCA), KICH, KIRP, LGG, pheochromocytoma and paraganglioma (PCPG) and PRAD (Fig. 2C). Moreover, the overexpression of CEP55 was significantly associated with tumor status in ACC, BLCA, KICH, KIRC, KIRP, LGG, LIHC, LUAD, pancreatic adenocarcinoma (PAAD), PCPG, PRAD, THCA, UCEC and uveal melanoma (UVM) (Fig. 2D). These results indicated that the overexpression of CEP55 was positively correlated with worse clinical features, which might be helpful to the judgment of tumor progression or prognosis.

Fig. 2.

The relationship between CEP55 mRNA and clinical characteristics. (A) CEP55 mRNA and age. (B) CEP55 mRNA and stage. (C) CEP55 mRNA and response. (D) CEP55 mRNA and tumor status. (*P < 0.05, **P < 0.01, and ***P < 0.001).

Prognostic significance of CEP55 for Pan-cancer

Cancer samples were classified as two groups based on median CEP55 expression and correlation of CEP55 level with OS and DFS of diverse cancers was conducted with GEPIA2 database. Patients with ACC (P < 0.001), KIRC (P = 0.003), KIRP (P = 0.001), LIHC (P < 0.001), LGG (P < 0.001), LUAD (P = 0.008), mesothelioma (MESO, P < 0.001), and PAAD (P = 0.005) had significantly poorer OS in the high expression of CEP55 (Fig. 3A). However, CEP55 down-regulation was associated with dismal OS of STAD (P = 0.021, Fig. 3A). DFS results suggested patients with CEP55 up-regulation manifested poor DFS in ACC (P < 0.001), KIRC (P = 0.006), KIRP (P < 0.001), LGG (P = 0.003), LIHC (P < 0.001), MESO (P = 0.028), PAAD (P = 0.034), PRAD (P = 0.006), sarcoma (SARC, P = 0.013), UVM (P = 0.002) and THCA (P = 0.007) Fig. 3B). The above results showed that CEP55 up-regulation was the risk factor related to poor prognosis in most malignant tumors.

Fig. 3.

Relationship between CEP55 mRNA and prognosis of pan-cancer based on GEPIA2. The survival map and Kaplan-Meier curves were used to display the overall survival (A) and disease-free survival (B) results of different tumors with significant differences.

Enrichment and mechanism analysis of CEP55 within pan-cancer

For further investigating CEP55’s effect and molecular mechanism on carcinogenesis, its binding or related genes were screened out for enrichment analyses. 85 CEP55-binding proteins were obtained by STRING database. Then, GEPIA2 database was utilized to aggregate all tumor expression profiles, and those 100 most significant genes associated with CEP55 expression were selected out. As shown in the Venn diagram, 53 overlapping genes were shared between the two groups (Fig. 4A). Figure 4B displayed the protein interaction network. KIF11, CDK1, and CCNA2 were among those five most significant genes for the two groups. CEP55 expression showed positive relation to KIF11 (R = 0.83, P < 0.001, Fig. 4C), CDK1 (R = 0.77, P < 0.001, Fig. 4D) and CCNA2 (R = 0.76, P < 0.001, Fig. 4E). The correlation heatmap showed CEP55 level was significantly positively related to KIF11, CDK1, and CCNA2 of all cancers (Fig. 4F). Furthermore, we conducted GO and KEGG analyses based on those 53 selected genes. The enriched biological processes (BP) mainly included organelle fission, nuclear division, mitotic nuclear division, and chromosome segregation (Fig. 4G). The main cellular components (CC) included spindle, chromosome region and chromosome, centromeric region (Fig. 4G). Those associated molecular functions (MF) included microtubule binding, tubulin binding and protein serine kinase activity (Fig. 4G). Additionally, those associated pathways included oocyte meiosis, cell cycle, progesterone-mediated oocyte maturation, cellular senescence, and p53 pathway, according to the KEGG analysis (Fig. 4H).

Fig. 4.

Enrichment and Mechanism Analysis of CEP55 in Pan-Cancer. (A) An intersection analysis of the CEP55-binding genes from STRING database and correlated genes from GEPIA2 database. (B) Protein–protein interaction network based on STRING results. Correlation between CEP55 and KIF11 (C), CCNA2 (D), and CDK1 (E) according to GEPIA2 database. (F) The heat map of correlation between CEPP55 and KIF11, CCNA2, CDK1, proliferation and MMR markers in different types of cancer. (G) GO enrichment analysis. (H) KEGG enrichment analysis. (I) A heatmap indicated the relationship between CEPP55 expression and cancer functional states from CancerSEA.

To better understand the functional states of CEP55 expression in pan-cancer, CancerSEA database was employed at a single-cell level within 9 cancers, like ALL (acute lymphoblastic leukemia), AML (acute myeloid leukemia), GBM (glioblastoma), LUAD, MEL (melanoma), NSCLC (non-small cell lung cancer), BRCA, RCC (renal cell carcinoma), PC (prostate cancer), CRC(colorectal cancer), HNSCC (head and neck squamous cell carcinoma), UM (uveal melanoma), and RB (retinoblastoma). CEP55 was significantly positively related to cell proliferation and cell cycle in AML, LUAD, MEL, BRCA and PC. CEP55 was positively related to DNA damage and DNA repair of LUAD, MEL, BRCA, but negatively of RB and UM (Fig. 4I).

Then the candidate relation of CEP55 with typical proliferation markers, such as PCNA and MKI67, began to be focused on²⁵. From the heat map, CEP55 exhibited positive relations to PCNA and MKI67 expression within almost all tumors (Fig. 4F). In addition, we also found that CEP55 displayed positive relation to mismatch repair (MMR) marker levels in many cancers, in particular for BRAC-Basal, PRAD, rectum adenocarcinoma (READ), and UCEC (Fig. 4F). Consequently, CEP55 might be involved in the development and progression of tumors by regulating cell cycle genes.

CEP55 immune cell infiltration within pan-cancer

Malignant solid tumor samples included cancer cells as well as tumor-associated cells, including immune, stromal, epithelial cells. Tumor-associated immune and stromal cells were essential for modulating tumor proliferation, drug resistance and metastasis^26,27. ESTIMATE algorithm²⁸ was employed in the present work for calculating the possible relation of infiltrating immune and stromal cells with CEP55 expression. Clearly, CEP55 level had markedly positive relation to immune/stromal/ESTIMATE scores of KIRC, LGG and THCA (P < 0.01, Fig. 5A). By contrast, CEP55 expression exhibited negative relation to immune/stromal/ ESTIMATE scores of CESC, HNSC, LUSC, PAAD, STAD, and UCEC (P < 0.01, Fig. 5A). Therefore, CEP55 might regulate tumor microenvironment (TME).

Fig. 5.

Correlation analysis between CEP55 expression and immune cells. (A) The potential associations between infiltrating stromal and immune cells and CEP55 expression level were explored by ESTIMATE algorithm. (B) The correlations of CEP55 expression and the infiltration of immune cells in pan-cancer. Red and blue represented positive and negative correlation, respectively.

Furthermore, the relation of CEP55 expression with immune cell infiltration within pan-cancer was further analyzed using the TIMER2.0 database (Fig. 5B). CEP55 was closely related to immune cell infiltration levels, like MDSC, progenitors of lymphoid, neutrophil, CD8⁺ T cell as well as some CD4⁺ T cell subsets of diverse cancer types. Noteworthily, Th2 cell infiltration showed positive relation to CEP55 expression within all cancer types. CEP55 displayed positive relation to immune infiltration degrees of memory CD4⁺ T (Tmem) cells and MDSCs, but negatively correlated with central memory CD4⁺ T (Tcm) and effector memory CD4⁺ T (Tem) cells in various cancers. Particularly, CEP55 in THCA and thymic carcinoma (THYM) was evidently related to macrophages, B cells, CAF, Endo, Treg, and CD8⁺ T cells infiltration (Fig. S1). However, because of additional immune levels in some tumors, these correlations were slightly different.

MSI, TMB and immunomodulators analysis of CEP55 in pan-cancer

Associations of CEP55 mRNA expression with MSI/TMB/immunomodulators in 33 cancers based on the TCGA data were further analyzed. CEP55 exhibited positive relations to the increased MSI of LIHC, COAD, LUSC, READ, SARC, STAD, uterine carcinosarcoma (UCS), and UCEC, but negative relation to MSI of diffuse large B-cell lymphoma (DLBC) (Fig. 6A). Besides, CEP55 expression was positively related to TMB of ACC, BLCA, CESC, COAD, HNSC, KIRC, KICH, LGG, LUSC, LUAD, PAAD, READ, PRAD, SARC, STAD, skin cutaneous melanoma (SKCM), and UCEC, but negatively related to TMB of THYM (Fig. 6B). Furthermore, the correlation of CEP55 level with immunomodulatory genes (such as immune cell markers and immune checkpoint genes) with pan-cancer was analyzed. The results showed that CEP55 displayed positive relations to many immunomodulatory genes in KIRC, LGG, LIHC, PRAD, THCA, and THYM, whereas negative relations in ESCA and LUSC (Fig. 6C).

Fig. 6.

Relationship between CEP55 expression and MSI, TMB, and immunomodulatory genes in pan-cancer. (A) Correlations between CEP55 expression and microsatellite instability (MSI). (B) Correlations between CEP55 expression and tumor mutation burden (TMB). (C) The Spearman correlation heatmap shows the correlation between the expression of CEP55 and immune regulators. Red represents positive correlation and blue represents negative correlation.

Response to immunotherapy and forecasting of sensitive drugs according to CEP55 expression

For exploring CEP55’s role as the new immune target within pan-cancer, response to immunotherapy and forecasting of sensitive drugs were performed according to CEP55 level. We loaded IMvigor210 data from previous study for analyzing the sensitivity to anti-PDL1 therapy²⁹. The CEP55 expression level in patients with CR/PR to anti-PDL1 therapy markedly increased relative to SD/PD patients (P = 0.001, Fig. 7A, B). Moreover, based on the relation of CEP55 expression with drug sensitivity according to GDSC dataset, AZ628, SB52334, and SB590885 represented three drugs with highest positive correlation with CEP55 level (Fig. 7C, P < 0.05). In contrast, A-770,041, AZD7762, and elesclomol represented 3 drugs with highest negative correlation with CEP55 level (Fig. 7C, P < 0.05). As revealed by the relation of CEP55 expression with drug sensitivity according to CTRP dataset, panobinostat, BRD-A94377914, and LRRK2-IN-1 represented drugs with highest positive correlation with CEP55 level (Fig. 7D, P < 0.0001). Based on the above findings, we predicted a series of small molecule targeted drugs that can be used for cancer treatment, and CEP55 expression level provided a direction for targeting and immunotherapy for pan-cancer.

Fig. 7.

The value of CEP55 in predicting the sensitivity to immunotherapy and drugs. (A) The different CEP55 expressions between the CR/PR and SD/PD groups. (B) The percent of CR/PR and SD/PD in high- and low-CEP55 expression groups. (C) Correlation between GDSC drug sensitivity and CEP55 mRNA expression. (D) Correlation between CTRP drug sensitivity and CEP55 mRNA expression.

Function and prognostic prediction analysis CEP55 based on ScRNA Data of ccRCC

To explore the function of CEP55 and its prognostic potential in ccRCC, an in-depth analysis was performed based on single-cell RNA sequencing data. Four paired ccRCC and adjacent normal kidney samples were selected from the scRNA-seq dataset GSE159115, yielding 20,850 high-quality transcriptomes post-quality control and filtering (Fig. 8A). Using the UMAP method, three major cell types—epithelial, immune, and stromal cells—were identified and visualized based on the expression of specific marker genes (Fig. 8B-D). Notably, immune cells, stromal cells, and epithelial cells displayed interpatient heterogeneity (Fig. 8E). Subsequently, tumor epithelial cells were categorized into two groups according to CEP55 expression levels (Fig. 8F, G), revealing distinct gene expression profiles between CEP55-positive and CEP55-negative tumor epithelial cells. GO enrichment analysis indicated that these genes were associated with intrinsic apoptotic signaling pathways, ribosomes, focal adhesion, cell-substrate junctions, and other biological processes (Fig. 8H). According to KEGG pathway analysis, these differential expression genes were linked to pathways such as chemical carcinogenesis involving reactive oxygen species, protein processing in the endoplasmic reticulum, and oxidative phosphorylation (Fig. 8I). In the TCGA KIRC cohort, patients were stratified into high and low CEP55 expression groups based on single-cell analysis results. Patients with high CEP55-positive tumor epithelial cells demonstrated significantly poor overall survival compared to those with low CEP55-positive tumor epithelial cells (Fig. 8J). Conversely, individuals with high CEP55-negative tumor epithelial cells showed improved overall survival relative to their low CEP55-negative counterparts (Fig. 8K). Furthermore, CEP55-positive tumor epithelial cells were significantly associated with metabolism-related pathways like steroid hormone biosynthesis, oxidative phosphorylation, nitrogen metabolism, and ascorbate and aldarate metabolism by scMetabolism method³⁰ (Fig. 8L). These results identified that CEP55 might promote the occurrence and development of ccRCC by regulating oxidative phosphorylation and was associated with ccRCC prognosis.

Fig. 8.

The function and prognosis of CEP55 based on single cell analysis. (A) The range of detected gene numbers, the depth of sequencing, the percentage of mitochondrial content and percent of reads in poly-A region in each sample after filtering. (B) UMAP visualization of the cell populations. (C) Marker genes expression for each cell type. (D) UMAP visualization of the main cell types. (E) Bar plots of the proportion of different main cell types in each tissue. (F) CEP55 expression for each cell type of tumor epithelial cells. (G) UMAP visualization of different expression levels of CEP55 in tumor epithelial cells. (H) GO analysis of different genes between CEP55-positive and CEP55-negative tumor epithelial cells. (I) KEGG analysis of different genes between CEP55-positive and CEP55-negative tumor epithelial cells. (J) Kaplan-Meier plots to depict the survival of patients with high and low CEP55-positive tumor epithelial cells. (K) Kaplan-Meier plots to depict the survival of patients with high and low CEP55-negative tumor epithelial cells. (L) Metabolism-related pathways in CEP55-positive and CEP55-negative tumor epithelial cells.

Validation of CEP55 protein expression and its relationship with clinical factors in ccRCC

To demonstrate bioinformatic analysis findings, CEP55 protein level and its relationship with clinical characteristics were further confirmed in 150 ccRCC samples and 30 renal adjacent tissues by immunohistochemistry (Fig. 9A,B). CEP55 staining was positive in ccRCC and renal adjacent tissues and mainly located in the cytoplasm. According to the 30 paired tissues, the CEP55-positive rate is 63.3% (19/30), which was higher than that of renal adjacent tissues (33.3%,10/30). Finally, the correlation between CEP55 protein level and clinical characteristics indicated that the CEP55 protein level was apparently related to tumor size (P = 0.044), histology grade (P < 0.001) and stage (P = 0.034, Table 1).

Fig. 9.

The representative IHC staining images of CEP55 protein in ccRCC. (A) Adjacent tissue. (B) The CEP55 score less than 8 in tumor tissue. (C) The CEP55 score greater than or equal to 8 in tumor tissue.

Table 1.

Relationship between CEP55 expression and clinical characteristics of ccRCC patients.

Characteristics	Number of cases	CEP55 expression		χ2 value	P value
		High (%)	Low (%)
Age (y)				0.677	0.411
≥ 60	63	44 (69.8%)	19 (30.1%)
< 60	87	66 (75.9%)	21 (24.1%)
Gender				1.014	0.314
Male	107	76 (71.0%)	31 (29.0%)
Female	43	34 (79.1%)	9 (20.9%)
Tumor size				4.073	0.044
> 7 cm	22	20 (90.9%)	2 (9.1%)
≤ 7 cm	128	90 (70.3%)	38 (29.7%)
Histology Grade				34.097	< 0.001
G1	26	8 (30.8%)	18 (69.2%)
G2	84	64 (76.2%)	20 (23.8%)
G3	36	34 (94.4%)	2 (5.5%)
G4	4	4 (100%)	0 (0%)
Stage				4.48	0.034
I	122	85 (69.7%)	37 (30.3%)
II–VI	28	25 (89.3%)	3 (10.7%)
Vital status				0.483	0.487
Alive	122	88 (72.1%)	34 (27.9%)
Dead	28	22 (78.5%)	6 (21.4%)

Discussion

CEP55, the coiled-coil centrosomal protein, was initially identified to be important for cytokinetic abscission in the anaphase of mitosis³¹. In the past decade or so, CEP55 overexpression has been widely suggested to be related to tumor cell initiation, growth, survival, proliferation, invasion, and migration in human cancers^6,11. Our previous bioinformatics analysis demonstrated that CEP55 up-regulation is closely related to sex, stage, histological grade, TNM classification in ccRCC patients³². Furthermore, present immunohistochemistry and single-cell analysis have further verified these results. Based on our results, CEP55 likely has an oncogenic effect on ccRCC. However, the expression levels, molecular mechanisms and the relationship with immune microenvironment of CEP55 in pan-cancer remain unclear. Consequently, it is necessary to conduct integrated pan-cancer analysis to determine the biological activity and immune therapeutic value of CEP55 from the perspective of overall tumors.

According to our analysis, CEP55 was overexpressed in 22 cancer types and existed significant correlations with clinical characteristics. Elevated CEP55 expression was associated with young onset age, poor staging of tumor, low response rate to the first treatment, low tumor-free survival rate, and poor OS and DFS prognosis in most cancers. Previous significant in vitro studies have confirmed that CEP55 promotes tumorigenesis and can predict the poor prognosis in single tumor³³, these were also consistent with our above findings. However, the causal role of CEP55 in vivo tumorigenesis and specific molecular mechanism remains elusive. Our further analyses indicated that CEP55 was predominantly associated with cell cycle and proliferation related genes. As first demonstrated by Sinha D et al., CEP55 up-regulation induces spontaneous tumorigenesis¹¹. Several studies also implied that CEP55 induced the proliferation of tumor cells through regulating diverse cellular signaling networks, such as PI3K/AKT pathway hyperactivation^34,35. Hence, CEP55 may play critical roles in cancer proliferation by regulating cell cycle-related genes. However, due to the heterogeneity of tumors, the above characteristics vary with different types of cancers. The immune cell infiltration, response to immunosuppressive therapy and drug sensitivity of CEP55 expression in various cancers depend on case-by-case analyses.

Then, we noted that the top three genes related to CEP55 have been described in previous studies. KIF11 is a motor protein, while CDK1 and CCNA2 belong to the cell cyclin family. The expression of these genes enhances cell division, proliferation, migration and the epithelial-mesenchymal transition (EMT) process with decreased apoptosis rate in RCC^36,37. As verified by our KEGG analysis, the high correlation between CEP55 and these genes indicates that the biological function of CEP55 is to participate in cell cycle activity and proliferation. Additionally, CEP55 expression showed high relation to TMB within KICH, KIRC, BLCA, PRAD and ACC, this could improve the predictive accuracy for cancer immunotherapy outcomes³⁸. For exploring CEP55’s potential application as the new immunotherapy target for pan-cancer, therapy response was predicted, and sensitive drugs were also forecast according to CEP55 level. We analyzed the sensitivity to anti-PDL1 therapy in urological cancer cases, as a result, CEP55 expression in cases achieving CR or PR significantly increased compared with those developing SD or PD. Among the sensitive drugs we screened out, it has been reported that AZD7762 strongly sensitizes BLCA to gemcitabine treatment³⁹, and AZ628 can improve the prognosis of patients with KIRC⁴⁰, and elesclomol directly target renal carcinoma cells and participate in regulating TME characteristics, disease progression, and long-time ccRCC prognosis⁴¹. Further research on whether the expression of CEP55 is involved in the anticancer effect of these drugs will shed new lights on tumor-targeted therapies and overcoming tumor resistance.

For lung cancer (LC), existing studies indicated that CEP55 promotes LC cell proliferation⁴², which is related to dismal clinical outcomes and prognosis in LC patients⁴³. However, few studies have compared the similarities and differences between different types of LC. In the present study, we found several common characteristics both in LUAD and LUSC. CEP55 was overexpressed in LC, had decreased expression in elderly patients and was proportional to clinical stage. Besides, CEP55 expression was related to TMB in both LC. Nevertheless, we also found out the differences between them. CEP55 up-regulation was related to tumor status and manifested poor OS prognosis of LUAD rather than LUSC. These results implied that CEP55 was probably the biomarker used to diagnose LUAD and LUSC and predict the prognosis of LUAD instead of LUSC. The deeper single-cell analysis within LUAD and other cancers also indicated correlation between CEP55 and proliferation and cell cycle, DNA damage and repair. Furthermore, correlation analysis verified that CEP55 was positively correlated with typical proliferation markers, such as PCNA and MKI67, and the MMR markers among almost all types of cancers. According to the above findings, the upregulation of CEP55 had an important effect on cancer cell growth via regulating cell cycle-related genes. Consequently, further molecular biology research was required to confirm the specific mechanism of CEP55 within diverse cancers.

Then, this study focused on digestive system neoplasms, including ESCA, COAD, READ, STAD, LIHC, CHOL, and PAAD. CEP55 expression increased within these cancers, besides, the correlation with clinical features and prognosis was the same as the general trend, except STAD. Combined with literature learning, we found that through the PI3K/AKT signal pathway, KIF11 gene promotes cell proliferation in CHOL⁴⁴ and the SB590885 drug increases impairment of proliferation and LIHC tumor inhibition via the same pathway⁴⁵. As mentioned above, CEP55 is also involved in activating this pathway, nonetheless, the specific connection among them in digestive system tumors needs to be deeply explored. To our surprise, we found CEP55 expression shows completely opposite characteristics in STAD. Specifically, CEP55 overexpression exhibited positive relations to older age, higher tumor-free survival rate and better OS prognosis. CEP55 seems to be the protective factor of STAD. However, abnormal CEP55 high expression contributes to gastric cancer progression⁴⁶. Consequently, larger sample sizes are required to clarify such a contradictory conclusion.

In addition to the above example analyses, we found the common biological function and immune infiltration pattern of CEP55 in most tumors. Enrichment analyses revealed that CEP55 up-regulation was markedly linked to cell cycle progression, cell division, and p53 signaling pathway. CEP55 is required for mitotic exit and cytokinesis, however, CEP55 up-regulation may result in cytokinesis disorder and the elevation of multinucleated cells, the carcinogenic characteristics of carcinogenesis⁴⁷. Previous studies have investigated that CEP55 expression levels can be adjusted by genes associated with cell cycle progression in a p53-dependent manner⁴⁸. P53 is a tumor suppressor, which has a wide range of functions in triggering cell-cycle checkpoint in response to DNA damage⁴⁹. Further molecular biology research is required to clarify the more exact crosstalk between CEP55 and p53 signaling pathways.

In the context of immune response, the innate immune cells like myeloid-derived suppressor cells, innate lymphoid cells, macrophages, and neutrophils, together with adaptive immune cells including T cells and B cells, may be associated with tumor development within TIME⁵⁰. Infiltration of immune cells is important for anti-tumor immunotherapy. Based on TISCH2 and TIMER2.0 database, we found that CEP55 was highly observed in T cell proliferation, which showed a positive relation to immune infiltration levels of memory CD4⁺ T (Tmem) cells, MDSCs and Th2 cells, but negative relation to central memory CD4⁺ T (Tcm) cells and effector memory CD4⁺ T (Tem) cells within various cancers. Tumor immune surveillance refers to the tumor cell detection and clearance capacity of immune system⁵¹. T cells are crucial for tumor immune monitoring⁵². Based on current evidence, type 2 immune responses mediated by Th2 cells also promote anti-cancer immunity and promote tumor growth and metastasis⁵³, while Tem and Tcm cells contribute to antitumor responses⁵⁴. In addition, MDSCs increased the tumor resistant to immunotherapy⁵⁵. CEP55 expression was highly related to immune infiltration, indicating that CEP55 affects the survival of patients in an immune-dependent manner. Immunomodulatory genes, such as PD-L1, CTLA-4, and IDO1, regulate immune responses within tumors by modulating the activity and infiltration of various immune cells, including T cells, natural killer (NK) cells, and macrophages. Patients with tumors showing high PD-L1 expression and abundant CD8 + T cell infiltration generally respond better to immune checkpoint inhibitors (ICIs)⁵⁶. Recent findings emphasize that manipulating the expression of these genes, either by targeted therapies or ICIs, can remodel the immune landscape within tumors, enhancing the recruitment of effector immune cells while reducing suppressive cell populations⁵⁷. The correlation analysis between CEP55 and the pan-cancer immunomodulatory factors suggested that the CEP55 expression was highly correlated to the expression of specific immunomodulatory genes, especially in KIRH, KIRC, LGG, THCA, and THYM. This dynamic interaction underscores the potential of CEP55 not only as biomarkers for immunotherapy efficacy but also as therapeutic targets to optimize immune responses in the tumor milieu.

Several limitations should be noted in this work. To begin with, the TCGA database was used for these analyses, but the internal heterogeneity of the same tumor was inevitably ignored. In addition, we only verified partial features of CEP55 in ccRCC patients. Further research should pay attention to using in vivo and in vitro biological experiments for verifying the predicted findings and promoting treatment utility in more malignant tumors.

Taken together, CEP55 showed high expression within many cancers and had oncogenic characteristics predicting poor prognosis. Furthermore, our study also identified that CEP55 overexpression may act on tumor proliferation by participating in cell cycle activity. CEP55 had potential roles on regulating TIME and correlated markedly with the curative effect of anticancer drugs in pan-cancers. We verified the expression, location and clinical features of CEP55 in patients with renal cell carcinoma. More prospective studies regarding CEP55 level and tumor immune milieu will provide a direction for CEP55 targeted therapy for pan-cancer.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

L.Z. H.L. and M.Y. contributed to the study conception and design. L.Z. and Y.Z. wrote the main manuscript text and performed the bioinformatics. L.Z. conducted tissue experiments. All authors reviewed the manuscript and approved the final manuscript.

Data availability

The datasets analyzed during the current study are available in the TCGA database (https://portal.gdc.cancer.gov). All additional information generated during and/or analyzed during the current study is available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.