Integrative analysis of S100A2 in pan-cancer and verification in glioma

doi:10.21203/rs.3.rs-2741027/v1

Download PDF

Research Article

Integrative analysis of S100A2 in pan-cancer and verification in glioma

https://doi.org/10.21203/rs.3.rs-2741027/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

S100A2 (S100 Calcium Binding Protein A2) is one of the 25 members of S100 protein family. S100A2 was involved with suppress tumor invasion and cancer progression. However, there was no pan-cancer analysis of S100A2 reported recently. This study sought to identify the effect of A100A2 on the immunological role and prognosis in pan-cancer. The expression of S100A2 was analyzed in 33 different cancers by using TCGA and GTEx database. The analysis of S100A2 gene expression was performed using the glioma samples from the cancer genome atlas (TCGA)-lower grade glioma and glioblastoma (GBMLGG) cohort. We assessed the association of S100A2 expression with molecular and immune subtypes in pan cancer. Then, we explored the S100A2-bingding proteins, the biological and molecular functions of S100A2, diagnosis and prognosis of S100A2 in pan cancer. Subsequently, based on the pan cancer analysis, we narrowed the focus to the relationship of S100A2 expression with clinical characteristics, overall survival (OS), disease-specific survival (DSS), and progression- free interval (PFI) in the GBMLGG. In addition, we analyzed the correlation of S100A2 and immune cells in pan cancer from the TIMER2 database. S100A2 was differently expressed in pan cancer, and it was associated with molecular subtypes and immune subtypes in various cancers. The infiltration levels of CD8⁺T cells, CD4⁺T cells were associated with S100A2 expression. Pan cancer cox regression analysis showed that S100A2 high expression can cause a poor prognosis. Further investigation showed that the higher expression S100A2 had a worse OS, DSS and PFI in some clinical subgroups of GBMLGG. Our results showed that S100A2 expression was linked to WHO grade, IDH status, gender, age, and histological type of GBMLGG. We also found that the first 50 important genes that positively correlated with S100A2 expression in GBMLGG. In conclusion, our research provides a molecular biomarker for the therapy in GBMLGG.

pan-cancer

S100A2

biomarker

prognosis

immune

Globally, the leading cause of death and morbidity is cancer (1). In both developing and developed countries, the burden of cancer is significant, as it has enormous social and economic consequences (2). Tumorigenesis is a dynamic and complex process with characteristics that are responsible for tumor growth and dissemination (3). Inactivation and mutations of tumor suppressor genes as well as oncogene activation account for the molecular pathogenesis of cancer (4). Therefore, the understanding of the molecular mechanisms controlling the malignant transformation of cancer cells provides us with new therapeutic strategies for fighting the cancer (5). Based on public databases like the Cancer Genome Atlas (TCGA), we could perform pan-cancer analysis to elucidate the role of different genes in cancers.

S100 calcium binding protein A2 (S100A2) belongs to the small calcium-binding protein family S100(6). There is some evidence that the expression of the s100 protein family members is associated with a wide variety of cancers, such as pancreatic, ovarian and other cancers (7). Previous study has reported that S100A2 in cell nuclei could inhibit the recurrence of oral cancer, and the loss of nuclear S100A2 may be a prognostic marker in early-stage oral cancer (8). S100A2 has also been found to be dysregulated in human prostate cancer, and to be associated with tumor suppressing functions (9). In cancers such as laryngeal squamous cell carcinoma and non-small cell lung cancer, S100A2 is known to play a role in controlling tumor progression and growth (10, 11). Recently, S100A2 has been proven to be over-expressed in endometrial carcinoma tissues (12). The role of S100A2 has been highlighted in various cancers. However, a limited study has been reported in the functions of S100A2 in pan cancer.

In our study, we conducted a pan cancer analysis to illustrate the roles of S100A2 in human cancers. We evaluated the potential diagnostic and prognostic value of S100A2 in pan cancer. There is a discover that S100A2 may have a critical role in the GBMLGG. Notably, we found that the higher expression S100A2 had a worse OS, DSS and PFI in some clinical subgroups of All in all, our study showed that S100A2 may be a prognostic predictor for the development of cancer treatment in GBMLGG.

Expression of S100A2 in pan cancer

The Tumor Immune Estimation Resource (TIMER2, v. 2) web page was used to analyze the S100A2 expression levels in different cancers. The analysis of differential S100A2 expression in 33 tumor types, the gene expression data of GTEx and TCGA were downloaded from the University of California Santa Cruz (UCSC) Xena database (https://xenabrowser.net/datapages/). R software v3.6.3 was used for data analysis, and data visualization were performed using ggplot2 package.

S100A2 Expression in Molecular Subtypes and Immune Subtypes of Cancers

We explored the correlations between S100A2 expression and molecular and immune subtypes by the TISIDB database (http://cis.hku.hk/TISIDB/in-dex.php). The TISIDB database is an online integrated repository portal for the analysis of tumors and the immune system from the TCGA database.

Protein-Protein Interaction (PPI) Network Analysis

The STRING (https://string-db.org/) database was used to analyze the protein-protein interactions (PPI) of S100A2. We obtained a total of 50 S100A2-bingding proteins from the STRING website. The PPI network was visualized using the following parameters: “medium confidence (0.400), active interaction sources “Experiments, Text mining, Databases” and “no more than 50 interactors”. The Cytoscape (v3.9.1) was applied for molecular network visualization. And Cytoscape was also used to identify key modules.

KEGG and GO Analysis

Gene Ontology (GO) function and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were carried to examine the biological and molecular functions of S100A2 in cancers by the Xiantao web (https://www.xiantao.love).

Diagnosis and Prognosis Analysis ROC

The diagnostic values of S100A2 gene in various cancer types were calculated by the receiver operating characteristic (ROC) curve. ROC curves and AUC were calculated using package “pROC” (v1.17.0.1) of R software and visualized by package “ggplot2” (13). The Area under Curve (AUC) and sensitivity, were also calculated. The closer the AUC is to 1, the better the diagnostic effect is. AUC in 0.5–0.7 has low accuracy, 0.7–0.9 AUC has good accuracy, and 0.9 or more AUC has a high accuracy.

We assessed the relationship between the S100A2 expression and the prognosis of patients, including overall survival (OS), disease-specific survival (DSS), and progression- free interval (PFI) in pan cancer. We used days as the survival time unit, cutoff high (50%) and cutoff low (50%) values to split the high-expression and low expression cohorts. The associations between S100A2 expression and prognosis in different clinical subgroups of GBMLGG were analyzed and the Kaplan-Meier (KM) curves were drawn through the survminer package.

Univariate and Multivariate Cox Regression Analysis in GBMLGG

We used univariate and multivariate cox regression analyses to determine the dependency of clinical characteristics and the association between S100A2 and OS, DSS and PFI. The survival package was for the effect of S100A2 on survival.

Differentially Expressed Genes Between S100A2 High Expression and Low Expression Groups in GBMLGG

Using the deseq2 package, we analyzed the differently expressed genes (DEGs) between different S100A2 expression in low expression group: 0–50% and high expression group:50–100% of GBMLGG. The volcano plot and heat map were plotted using the ggplot2 package. The results were filtered using |log2 fold-change (FC)| > 1.0 and adjusted p-value < 0.05. Furthermore, we conducted GO and KEGG enrichment analyses of DEGs using the ggplot2 package for visualization and the cluster Profiler package for statistical analysis.

Timer Database Analysis

The TIMER2 database was used to examine the correlations between immune cells and the levels of immune infiltration of S100A2 in pan cancer. We investigated the correlation of S100A2 and immune cells which included activated DC (aDC), B cells, CD8 T cells, Cytotoxic cells, DC, Eosinophils, immature DC (iDC), Macrophages, Mast cells, Neutrophils, NK CD56bright cells, NK CD56dim cells, NK cells, Plasmacytoid DC (pDC), T cells, T helper cells, T central memory (T cm), T effector memory (T em), T follicular helper (T fh), T gamma delta (Tgd), Th1 cells, Th17 cells, Th2 cells and Treg.

Expression analysis of S100A2

In our study, we analyze the expression of S100A2 in different cancers. As shown in Fig. 1A, the expression level of S100A2 was higher in the tumor tissues of CHOL, COAD, ESCA, HNSC, LUAD, LUSC, READ, STAD, THCA and UCEC than in the adjacent normal tissues. It was significantly lower expressed in the tumor tissues of KICH KIRC KIRP and PRAD. Then, in the TCGA tumors with the data of the GTEx database, the S100A2 gene showed a high expression in BLCA, CESC, CHOL, COAD, DLBC, ESCA, GBM, HNSC, LGG, LUAD, LUSC, OV, PAAD, STAD, THCA, THYM, UCEC and UCS (Fig. 1B). For TCGA tumors and adjacent normal tissues, S100A2 expression was upregulated in BRCA, CHOL, COAD, LUAD, LUSC, THCA and UCEC, while lower expressed in KICH, KIRC, KIRP and PRAD (Fig. 1C).

Correlations between S100A2 Expression and Molecular Subtypes and Immune Subtypes of Different Cancers

Violin plots demonstrated that S100A2 expression was statistically associated with different molecular and immune subtypes of different caner types including UCEC, HNSC, STAD, BRCA, PRAD, OV, LIHC, LUSC, COAD, ESCA, KIRP and GBM (Fig. 2A-L, Fig. 3A-L). For molecular subtypes, S100A2 was highly expressed in the Atypical subtype and expressed at a low level in the Basal subtype of HNSC. S100A2 was highly expressed in the immunoreactive subtype of OV, the basal subtype of LUSC, and the ESCC subtype of ESCA. Immune subtypes were classified into six types including C1 (wound healing), C2 (IFN-γ dominant), C3 (inflammation), C4(lymphocytopenia depleted), C5 (immune silencing) and C6 (TGF-β dominant). S100A2 was highly expressed in wound healing of MESO and BRAC. For CESC, PRAD, OV and PAAD, the highest expression of S100A2 was found in IFN-γ dominant. The lowest levels expression of S100A2 in LUAD, KIRC, OV and STAD was found in the inflammation subtype. S100A2 had lowest levels expression in the TGF-β dominant subtype of SARC.

PPI Network and GO and KEGG Enrichment Analyses

We identified 50 targeted proteins interacting with S100A2 by the STRING database in Fig. 4A. The Cytoscape was performed for the obtained network visualization in Fig. 4B. Then we explored the functional mechanism of S100A2 expression by GO and KEGG enrichment analysis in Fig. 4C, D. GO analysis suggested that the enriched terms in the biological process (BP) were skin development, epidermal cell differentiation, keratinocyte differ entiation, peptide cross-linking. The cellular component (CC) was mainly involved in cornified envelope, collagen-containing extracellular matrix, cell cortex and azurophil granule lumen. For the molecular function ontology, the terms were S100 protein binding calcium-dependent protein binding, structural constituent of epidermis and RAGE receptor binding. In addition, the KEGG analysis showed that S100A2 expression were enriched in cellular senescence, wnt signaling pathway, cell cycle, colorectal cancer and chronic myeloid leukemia.

Diagnostic Value of S100A6 in Pan-Cancer

We performed the ROC curve to evaluate the prognostic predictive ability of S100A2 in pan-cancer. ROC curves analysis revealed that S100A2 gene had a high diagnostic value (AUC > 0.7) in COAD (AUC = 0.983) (Fig. 5A), GBMLGG (AUC = 0.852) (Fig. 5B), KICH (AUC = 0.922) (Fig. 5C), KIRP (AUC = 0.901) (Fig. 5D), LGG (AUC = 0.818) (Fig. 5E), LIHC (AUC = 0.723) (Fig. 5F), LUAD (AUC = 0.855) (Fig. 5G), LUSC (AUC = 0.962) (Fig. 5H), OSCC (AUC = 0.748) (Fig. 5I), PAAD (AUC = 0.968) (Fig. 5J), STAD (0.823) (Fig. 5K), UCEC (AUC = 0.920) (Fig. 5L)

Kaplan-Meier analysis demonstrated a relationship between S100A2 expression and prognosis significance in the patients with GBMLGG, LGG, PAAD and OV. For GBMLGG, Cox regression results revealed that upregulated S100A2 expression was associated with low OS (HR = 3.57, 95% CI: 2.74–4.64, p < 0.001) DSS (HR = 3.66, 95% CI: 2.77–4.83, p < 0.001), AND PFI (HR = 2.84, 95% CI: 2.27–3.55, p < 0.001). For LGG, Cox regression results showed that upregulated S100A2 expression was associated with worse prognosis, including OS (HR = 2.05, 95% CI: 1.44–2.90, p < 0.001) (Fig. 6A), DSS (HR = 2.10, 95% CI: 1.46–3.04, p < 0.001) (Fig. 6B), AND PFI (HR = 1.79, 95% CI:1.35–2.36, p < 0.001) (Fig. 6C). For PAAD, the results showed that S100A2 was significantly associated with poor OS (HR = 1.62, 95% CI: 1.07–2.46, p = 0.023) (Fig. 6D), DSS (HR = 1.80, 95% CI: 1.12–2.89, p = 0.015) (Fig. 6E), AND PFI (HR = 1.81, 95% CI:1.22–2.69, p = 0.003) (Fig. 6F). For OV, the results showed that S100A2 was significantly associated with poor OS (HR = 1.36, 95% CI: 1.05–1.77, p = 0.002) (Fig. 6G), DSS (HR = 1.33, 95% CI: 1.00-1.76, p = 0.048) (Fig. 6H), AND PFI (HR = 1.28, 95% CI:1.01–1.63, p = 0.039) (Fig. 6I).

Clinical Characteristics S100A2 GBMLGG

Baseline characteristics of S100A2 patients from TVGA are shown in Table 1. Results showed that S100A2 expression in patients was related to WHO grade, IDH status, gender, age, and histological type of GBMLGG. In additional, S100A2 was expressed higher in patients with age > 60 (Fig. 7A), WHO grade 3(Fig. 7B), 1p/19q codeletion non-codel (Fig. 7C) and male (Fig. 7D), while it was lower in patients with Glioblastoma (Fig. 7E) and IDH status Mut (Fig. 7F).

Table 1

Clinical characteristics of GBMLGG patients
Characteristic	Low expression of S100A2		High expression of S100A2	p
n	348		348
WHO grade, n (%)				< 0.001
G2		165 (26%)	59 (9.3%)
G3		129 (20.3%)	114 (18%)
G4		19 (3%)	149 (23.5%)
IDH status, n (%)				< 0.001
WT		45 (6.6%)	201 (29.3%)
Mut		301 (43.9%)	139 (20.3%)
1p/19q codeletion, n (%)				< 0.001
codel		143 (20.8%)	28 (4.1%)
non-codel		204 (29.6%)	314 (45.6%)
Gender, n (%)				0.007
Female		167 (24%)	131 (18.8%)
Male		181 (26%)	217 (31.2%)
Age, n (%)				< 0.001
<=60		306 (44%)	247 (35.5%)
> 60		42 (6%)	101 (14.5%)
Histological type, n (%)				< 0.001
Astrocytoma		92 (13.2%)	103 (14.8%)
Glioblastoma		19 (2.7%)	149 (21.4%)
Oligoastrocytoma		84 (12.1%)	50 (7.2%)
Oligodendroglioma		153 (22%)	46 (6.6%)
Age, median (IQR)		39.5 (32, 51)	52.5 (38, 62)	< 0.001

Prognostic Value of S100A2 in Pan-Cancer

We explored the correlations between S100A2 and prognosis value for OS, DSS, and PFI in different subgroups of MGBLGG. According to Kaplan-Meier, we found that the higher expression S100A2 had a worse OS in some clinical subgroups, subgroup of age < = 60 (Fig. 8A), subgroup of white (Fig. 8B), subgroup of female (Fig. 8C), subgroup of WHO grade G2 (Fig. 8D) and subgroup of 1p/19q codeletion non-codel (Fig. 8E). For DSS, it was found that the high S100A2 gene expression correlated with worse DSS in subgroup of age < = 60 (Fig. 9A), subgroup of white (Fig. 9B), subgroup of female (Fig. 9C), subgroup of WHO grade G2 (Fig. 9D) and subgroup of 1p/19q codeletion non-codel (Fig. 9E). For PFI, it was found that the high S100A2 gene expression correlated with worse PFI in subgroup of age < = 60 (Fig. 10A), subgroup of white (Fig. 10B), subgroup of female (Fig. 10C), subgroup of WHO grade G2 (Fig. 10D), subgroup of 1p/19q codeletion non-codel (Fig. 10E), and subgroup of IDH status Mut (Fig. 10F).

Univariate and Multivariate Cox Regression Analyses in GBMLGG

In our study, we performed univariate and multivariate cox regression analyses for the S100A2 and clinical characteristics in GBMLGG. In the univariate and multivariate Cox regression analyses, WHO grade, 1p/19q codeletion, primary therapy outcome, IDH status and age were associated with OS, DSS and PFI (Table 2–4).

Table 2

Univariate and multivariate Cox regression analyses of clinical characteristics associated with OS of MGBLGG
Characteristics	Total(N)	Univariate analysis		Multivariate analysis
Characteristics	Total(N)	Hazard ratio (95% CI)	P value	Hazard ratio (95% CI)	P value
WHO grade(G3&G4 VS. G2)	634	0.177 (0.123–0.255)	< 0.001	0.496 (0.314–0.783)	0.003
1p/19q codeletion (codel VS. non-codel )	688	4.428 (2.885–6.799)	< 0.001	1.596 (0.928–2.747)	0.091
Primary therapy outcome (PD VS. SD&PR&CR)	461	0.282 (0.196–0.407)	< 0.001	0.314 (0.203–0.486)	< 0.001
IDH status (WT VS. Mut)	685	0.117 (0.090–0.152)	< 0.001	0.535 (0.322–0.887)	0.015
Age ( < = 60 VS. >60)	695	4.668 (3.598–6.056)	< 0.001	5.049 (3.085–8.266)	< 0.001

Table 3

Univariate and multivariate Cox regression analyses of clinical characteristics associated with DSS of MGBLGG
Characteristics	Total(N)	Univariate analysis		Multivariate analysis
Characteristics	Total(N)	Hazard ratio (95% CI)	P value	Hazard ratio (95% CI)	P value
WHO grade (G3&G4 vs. G2 )	614	0.172 (0.117–0.253)	< 0.001	0.510 (0.314–0.828)	0.006
1p/19q codeletion (codel vs. non-codel)	668	4.987 (3.117–7.978)	< 0.001	1.759 (0.979–3.160)	0.059
Primary therapy outcome (PD vs. SD&PR&CR)	457	0.236 (0.160–0.347)	< 0.001	0.261 (0.164–0.417)	< 0.001
IDH status (WT vs. Mut)	664	0.110 (0.083–0.146)	< 0.001	0.526 (0.309–0.894)	0.018
Age ( < = 60vs. >60)	674	4.500 (3.409–5.940)	< 0.001	4.522 (2.688–7.606)	< 0.001

Table 4

Univariate and multivariate Cox regression analyses of clinical characteristics associated with PFI of MGBLGG
Characteristics	Total(N)	Univariate analysis		Multivariate analysis
Characteristics	Total(N)	Hazard ratio (95% CI)	P value	Hazard ratio (95% CI)	P value
WHO grade (G3&G4 vs. G2 )	634	0.363 (0.279–0.473)	< 0.001	0.843 (0.600-1.184)	0.324
1p/19q codeletion(codel vs non-codel)	688	3.373 (2.438–4.666)	< 0.001	1.476 (0.988–2.207)	0.057
Primary therapy outcome (PD vs. SD&PR&CR)	461	0.212 (0.160–0.282)	< 0.001	0.241 (0.175–0.332)	< 0.001
IDH status (WT vs. Mut)	685	0.151 (0.119–0.191)	< 0.001	0.399 (0.266-0.600)	< 0.001
Age ( < = 60 vs. >60 )	695	2.873 (2.268–3.640)	< 0.001	2.214 (1.477–3.320)	< 0.001
ACC Adrenocortical carcinoma
APOBEC3B Apolipoprotein B mRNA editing enzyme catalytic subunit 3B
BLCA Bladder urothelial carcinomas
BP Biological process
CC Cellular component
CEP55 Centrosomal protein
CESE Cervical squamous cell carcinomas and endocervical adenocarcinomas
CHOL Cholangiocarcinomas
CNA Copy number alteration
COAD Colon adenocarcinomas
CPTAC Clinical proteomic tumor analysis consortium
DBLC Lymphoid neoplasm diffuse large B-cell lymphomas
DFS Disease-free survival
ESCA Esophageal carcinomas
GBM Glioblastoma multiforme
GEPIA Gene expression profiling interactive analysis
GO Gene ontology
GTE Genotype-tissue expression
HNSC Head and neck squamous carcinomas
HPV Human papilloma virus
KEGG Kyoto encyclopedia of genes and genomes
KICH Kidney chromophobes
KIRC Kidney renal clear cell carcinomas
KIRP Kidney renal papillary cell carcinomas
KPNA2 Karyopherin subunit alpha 2
LAML Acute myeloid leukemia
LGG Brain lower grade gliomas
LIHC Liver hepatocellular carcinomas
LUAD Lung adenocarcinomas
LUSC Lung squamous cell carcinomas
MCM2 Minichromosome maintenance complex component 2
MELK Maternal embryonic leucine zipper kinase
MESO Mesotheliomas
MF Molecular function
MSI Microsatellite instability
NCAPH Non-SMC condensin I complex subunit H
OV Ovarian cancers
PAAD Pancreatic adenocarcinomas
PCPG Pheochromocytomas and paragangliomas
PFS Progression-free survival
PRAD Prostate adenocarcinomas
READ Rectum adenocarcinoma
SKCM Skin cutaneous melanomas
STAD Stomach adenocarcinomas
TCGA The cancer genome atlas
TGCT Testicular germ cell tumors
THCA Thyroid carcinomas
THYM Thymomas
TIMER Tumor immune estimation resource
TK1 Thymidine kinase 1
TMB Tumor mutational burden. TPM: transcripts per million reads
UCEC Uterine corpus endometrial carcinomas
UCS Uterine carcinosarcomas
UVM Uveal melanomas

Co-expression Gene Analysis of S100A2 in GBMLGG

To find out the role of S100A2 in GBMLGG, we analyzed the first 50 important genes that positively correlated with S100A2 expression in GBMLGG as shown in the heat map (Fig. 11A). We selected the top 10 genes, including S100A6 (r = 0.752) (Fig. 11B), VAMP5 (r = 0.697) (Fig. 11C), RHOC (r = 0.667) (Fig. 11D), TRIP6 (r = 0.663) (Fig. 11E), GNG5 (r = 0.662) (Fig. 11F), HSPB1 (r = 0.659) (Fig. 11G), S100A16 (r = 0.656) (Fig. 11H), BCL2L12 (r = 0.667) (Fig. 11I), LAMTOR2 (r = 0.667) (Fig. 11J) and AGTRAP(r = 0.667) (Fig. 11K).

Differentially Expressed Genes Between S100A2 High Expression and Low Expression Groups in GBMLGG

There were 2750 DEGs obtained with the threshold values of |log2 fold-change (FC)| > 1.0 and adjusted p-value < 0.05, including 1741 upregulated genes and 1009 downregulated genes in Fig. 12A. Among them, 392 DEGs were obtained with the threshold values of |log2 fold-change (FC)| > 2.0 and adjusted p-value < 0.05, including 312 upregulated genes and 80 downregulated genes. Then we conducted GO and KEGG functional enrichment analyses of DEGs in Fig. 12B, C. It was found that the primary BP was involved in regulation of ion transmembrane transport, regulation of membrane potential and regulation of trans-synaptic signaling. CC analysis showed that S100A2 may be associated with synaptic membrane, transmembrane transporter complex and ion channel complex. MF analysis revealed that S100A2-realted genes were enriched in passive transmembrane transporter activity, channel activity and substrate-specific channel activity. The KEGG analysis confirmed the enrichment of neuroactive ligand-receptor interaction, systemic lupus erythematosus and staphylococcus aureus infection.

Immune Cell Infiltration Analyses

The TIMER2 was used to assess the relationship between S100A2 expression and the infiltration of different immune cells. The results showed that S100A2 expression was positively correlated with CD8⁺ T cells, CD4⁺ T cells and Tregs cells in most cancer types (Fig. 13A, B, C). We found that macrophages, eosinophils, neutrophils and aDC were positively correlated with S100A2, and tcm, tgd, pDC and TFH were negatively correlated with S100A2 (Fig. 13D).

S100 proteins are relatively small calcium-binding proteins which expression in vertebrates (14).There are various malignancies that express S100 family proteins in varying amounts (15).There are cancer-related members of the S100 family (S100A2, S100A4, S100B) that interact with p53 or HDM2, but their modes of interaction differ, and the outcomes vary from cancer-suppressive to cancer-promoting (16, 17, 18). Most cancers, including breast cancer, melanoma, gastric cancer, and others, can be suppressed by S100A2, but lung cancer may have the opposite effect (19). It has been found that patients with high S100A2 expression are related to fewer CD8⁺ T cells infiltration in the pancreatic cancer microenviroment (20). In oral cancer patients, cytoplasmic S100A2 overexpression is associated with a poor outcome (21). However, there have been no studies describing the expression pattern of S100A2 in multiple cancer types. Our study showed that the association between S100A2 expression level and different tumor progression.

Firstly, we evaluated S100A2 expression level in pan-cancer through the TIMER2 GTEx database and TCGA database. In the GTEx database, the results revealed that the S100A2 was upregulated in the BLCA, CESC, CHOL, COAD, DLBC, ESCA, GBM, HNSC, LGG, LUAD, LUSC, OV, PAAD, STAD, THCA, THYM, UCEC and UCS. In other database, the results also suggested S100A2 had pivotal roles in promoting cancer development. We also found that S100A2 expression was statistically associated with different molecular and immune subtypes of different caner types including UCEC, HNSC, STAD, BRCA, PRAD, OV, LIHC, LUSC, COAD, ESCA, KIRP and GBM. S100A2 was observed to express the highest in the Atypical subtype and express at a low level in the Basal subtype of HNSC. In the inflammation subtype, the expression of S100A2 was lowest in LUAD, KIRC, OV and STAD. We examined the biological function and pathways of S100A2 by the GO and KEGG analyses. The KEGG showed that S100A2 expression significantly correlated with cellular senescence, wnt signaling pathway, cell cycle, colorectal cancer and chronic myeloid leukemia. Then we explored the prognostic value of S100A2 in several tumor types. We performed the ROC curve analysis to validate the prognostic predictive ability of S100A2 in tumor types. The results showed that S100A2 had high diagnostic accuracy in the COAD, KICH, KIRP, LUSC, PAAD and UCEC (AUC: 1.0-0.9). We also observed that the upregulated expression of S100A2 correlated with worse OS, DSS and PFI in GBMLGG, LGG, PAAD and OV.

We further studied the associations between S100A2 expression and different clinical characteristics in GBMLGG in detail. We found that the high expression of S100A2 had caused a worse of the OS, DSS and PFI in clinical subgroups of Age > 60, Race: White, Gender: Female, WHO grade: G2 and 1p/19q codeletion: non-codel. Previous studies have reported that the expression of S100A2 was increased with glioma grade (22). In the univariate and multivariate Cox regression analyses, WHO grade, 1p/19q codeletion, primary therapy outcome, IDH status and age were linked to OS, DSS and PFI in GBMLGG patients. We selected the top 10 genes that positively correlated with S100A2 expression in GBMLGG. In this study, we ranked DEGs identified between S100A2 high-expression and low-expression groups. Tumor immune cytokines are important factors for the prognosis of GBMLGG (23). In our study, we found a strong correlation between S100A2 and the level of immune cell nitration through the TIMER2.

In summary, our study was the first to conduct analysis of S100A2 in pan-cancer and GBMLGG. Further studies consisting of in vivo and in vitro experiments are highly required to validate our results. In GBMLGG patients, S100A2 may serve as a specific prognostic biomarker.

Ethics approval and consent to participate N/A

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Author contributions

XL and QZ were involved in the study conception and design. XL and QZ performed statistical analysis. YY and C-EH critically revised original study design and the manuscript. QZ, XL, C-EH, NH, and YY contributed to the data acquisition, analysis, and interpretation. All authors have read and approved the final manuscript.

Funding

This work was supported by grants from the National Natural Science Foundation of China (Grant No. 82104303).

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Borchmann S,Selenz C,Lohmann,et al.Tripartite antigen-agnostic combination immunotherapy cures established poorly immunogenic tumors.Journal for immunotherapy of cancer.2022;10 (10).
Taritla S,Kumari M,Kamat S,et al.Optimization of PhysicoChemical Parameters for Production of Cytotoxic Secondary Metabolites and Apoptosis Induction Activities in the Culture Extract of a Marine Algal-Derived Endophytic Fungus Aspergillus sp.Frontiers in pharmacology.2021;12:542891.
Muresanu C,Somasundaram SG,Vissarionov SV,et al.Updated Understanding of Cancer as a Metabolic and Telomere-Driven Disease, and Proposal for Complex Personalized Treatment, a Hypothesis.International journal of molecular sciences.2020;21 (18).
Wang S,Yu ZH,Chai KQ.Identification of CFTR as a novel key gene in chromophobe renal cell carcinoma through bioinformatics analysis.Oncology Letters.2019;18 (2):1767–1774.
Wassef M,Michaud A,Margueron R.Association between EZH2 expression, silencing of tumor suppressors and disease outcome in solid tumors.CELL CYCLE.2016;15 (17):2256–62.
Zhang X,Ong C,Su G,et al.Characterization and engineering of S100A12-heparan sulfate interactions.GLYCOBIOLOGY.2020;30 (7):463–473.
Bresnick A R,Weber D J,Zimmer DB.S100 proteins in cancer.NATURE REVIEWS CANCER.2015;15 (2):96–109.
Tsai ST,Jin YT,Tsai WC,et al.S100A2, a potential marker for early recurrence in early-stage oral cancer.ORAL ONCOLOGY.2005;41 (4):349–57.
Hibi K,Fujitake S,Takase T,et al.Identification of S100A2 as a target of the DeltaNp63 oncogenic pathway.CLINICAL CANCER RESEARCH.2003;9 (11):4282–5.
Almadori G,Bussu F,Galli J,et al.Diminished expression of S100A2, a putative tumour suppressor, is an independent predictive factor of neck node relapse in laryngeal squamous cell carcinoma.Journal of otolaryngology - head & neck surgery = Le Journal d'oto-rhino-laryngologie et de chirurgie cervico-faciale.2009;38 (1):16–22.
Bulk E,Sargin B,Krug U,et al.S100A2 induces metastasis in non-small cell lung cancer.CLINICAL CANCER RESEARCH.2009;15 (1):22–9.
Zhang Q,Xia T,Qi C,et al.High expression of S100A2 predicts poor prognosis in patients with endometrial carcinoma.BMC cancer.2022;22 (1):77.
Vivian J,Rao AA,Nothaft FA,et al.Toil enables reproducible, open source, big biomedical data analyses.NATURE BIOTECHNOLOGY.2017;35 (4):314–316.
Guo Q,Wang J,Cao Z,et al.Interaction of S100A1 with LATS1 promotes cell growth through regulation of the Hippo pathway in hepatocellular carcinoma.INTERNATIONAL JOURNAL OF ONCOLOGY.2018;53 (2):592–602.
Destek S,Gul VO.S100A4 May Be a Good Prognostic Marker and a Therapeutic Target for Colon Cancer.Journal of Oncology.2018;2018:1828791.
Grigorian M,Andresen S,Tulchinsky E,et al.Tumor suppressor p53 protein is a new target for the metastasis-associated Mts1/S100A4 protein: functional consequences of their interaction.JOURNAL OF BIOLOGICAL CHEMISTRY.2001;276 (25):22699–708.
Mueller A,Schäfer BW,Ferrari S,et al.The calcium-binding protein S100A2 interacts with p53 and modulates its transcriptional activity.JOURNAL OF BIOLOGICAL CHEMISTRY.2005;280 (32):29186–93.
Lin J,Blake M,Tang C,et al.Inhibition of p53 transcriptional activity by the S100B calcium-binding protein.JOURNAL OF BIOLOGICAL CHEMISTRY.2001;276 (37):35037–41.
Hountis P,Matthaios D,Froudarakis M,et al.S100A2 protein and non-small cell lung cancer. The dual role concept.TUMOR BIOLOGY.2014;35 (8):7327–33.
Chen Y,Wang Ch,Song J,et al.S100A2 Is a Prognostic Biomarker Involved in Immune Infiltration and Predict Immunotherapy Response in Pancreatic Cancer.Frontiers in immunology.2021;12:758004.
Kumar M,Srivastava G,Kaur J,et al.Prognostic significance of cytoplasmic S100A2 overexpression in oral cancer patients.Journal of translational medicine.2015;13:8.
Yin A,Shang Z,Etcheverry A,et al.Integrative analysis identifies an immune-relevant epigenetic signature for prognostication of non-G-CIMP glioblastomas.Oncoimmunology.2021;10 (1):1902071.
Cloughesy TF,Mochizuki AY,Orpilla JR,et al.Neoadjuvant anti-PD-1 immunotherapy promotes a survival benefit with intratumoral and systemic immune responses in recurrent glioblastoma.NATURE MEDICINE.2019;25 (3):477–486.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Integrative analysis of S100A2 in pan-cancer and verification in glioma

Status:

Version 1

Abstract

Figures

Introduction

Materials And Methods

Expression of S100A2 in pan cancer

S100A2 Expression in Molecular Subtypes and Immune Subtypes of Cancers

Protein-Protein Interaction (PPI) Network Analysis

KEGG and GO Analysis

Diagnosis and Prognosis Analysis ROC

Univariate and Multivariate Cox Regression Analysis in GBMLGG

Differentially Expressed Genes Between S100A2 High Expression and Low Expression Groups in GBMLGG

Timer Database Analysis

Results

Expression analysis of S100A2

Correlations between S100A2 Expression and Molecular Subtypes and Immune Subtypes of Different Cancers

PPI Network and GO and KEGG Enrichment Analyses

Diagnostic Value of S100A6 in Pan-Cancer

Clinical Characteristics S100A2 GBMLGG

Prognostic Value of S100A2 in Pan-Cancer

Univariate and Multivariate Cox Regression Analyses in GBMLGG

Co-expression Gene Analysis of S100A2 in GBMLGG

Differentially Expressed Genes Between S100A2 High Expression and Low Expression Groups in GBMLGG

Immune Cell Infiltration Analyses

Discussion

Declarations

References

Additional Declarations

Status:

Version 1