Sage Journals: Discover world-class research

Abstract

BACKGROUND:

Intercellular adhesion molecules (ICAMs) in the tumor microenvironment are closely related to immunity and affect the prognosis of cancer patients.

OBJECTIVE:

The aim of our study is to explore the correlation between ICAM expression, mutation, methylation and immunity and their prognostic value in breast cancer (BC) is not clear.

METHODS:

Online databases and tools such as UALCAN, COSMIC, cBioPortal, MethSurv, PrognoScan, Kaplan-Meier Plotter, GSCA and TIMER were utilized in this study.

RESULTS:

We found that the mRNA and protein expression levels of ICAM1 were upregulated in triple-negative breast cancer (TNBC) compared with normal tissues, and TNBC patients with high expression of ICAM1 had better overall survival (OS) and recurrence-free survival (RFS). The main types of ICAM1 gene variants were missense mutation and amplification, and ICAM1 showed a lower level of methylation in TNBC cancer tissues than in normal tissues, which was contrary to the high expression levels of ICAM1 mRNA and protein. Next, the function of ICAM1 was mainly related to the activation of apoptosis, epithelial-mesenchymal transition (EMT) and inhibition of the androgen receptor (AR) and estrogen receptor (ER) pathways. Meanwhile, functional pathway enrichment results showed that ICAM1 was also involved in the immune regulation process of BC. Furthermore, the expression of ICAM1 was positively associated with 6 types of tumor-infiltrating immune cells (CD8 $+$ T cells, CD4 $+$ T cells, B cells, neutrophils, macrophages and dendritic cells) and was also positively related to the expression of programmed cell death-1 (PD-1), programmed cell death-ligand 1 (PD-L1) and cytotoxic T lymphocyte-associated antigen-4 (CTLA4).

CONCLUSIONS:

Our research indicated that ICAM1 was likely to be a potential therapeutic target in TNBC.

Keywords

Breast cancer ICAMs prognostic biomarkers immune checkpoint therapy tumor microenvironment

1. Introduction

Breast cancer (BC) is the most common malignant tumor among women worldwide [1]. BC is a heterogeneous disease with different treatment strategies according to different molecular subtypes [2]. Recently, immunotherapy for cancer has developed rapidly and has been proven successful in the treatment of lymphoma, non-small cell lung cancer, melanoma, renal cell carcinoma and other tumor types [3]. Traditionally, BC is considered to have weak immunogenicity, which varies among different molecular subtypes. Triple-negative breast cancer (TNBC) and human epidermal growth factor receptor 2-positive (HER2 $+$ ) tumors have higher immunogenicity than luminal A and luminal B subtypes [4]. Meanwhile, there is increasing evidence that immune infiltration plays a prognostic role in BC and can also predict the immune efficacy of TNBC and HER2 $+$ tumors [5].

In the past few years, growing evidence has shown that tumor-infiltrating immune cell infiltration exhibits a positive impact on the neoadjuvant chemotherapy response and prognosis of BC, and tumor-infiltrating immune cells can reflect the intensity of the tumor bed immune response [6]. Several immune-based therapies have been evaluated in BC, including cancer vaccines, immune checkpoint inhibitors, and cytokine adoptive T-cell therapy. The drugs or antibodies that block the immune checkpoint include lymphocyte activating 3 (LAG3) [7], cytotoxic T lymphocyte-associated antigen-4 (CTLA4) [8], programmed cell death-1 (PD-1) [9] and programmed cell death-ligand 1 (PD-L1) [10], which have been used in clinical trials for BC treatment. However, although immunotherapy for BC has been extensively studied, only a small number of patients showed a response, and little is known about the underlying mechanisms. Therefore, it is necessary to identify more potential biomarkers of antitumor immunity, which may help the development of individualized immunotherapy for BC patients.

Intercellular adhesion molecules (ICAMs) are type I transmembrane glycoproteins containing 2–9 immunoglobulin-like C2 domains and binding to leukocyte adhesion lymphocyte function-associated antigen-1 (LFA-1) proteins [11], with a total of 5 members, namely, ICAM1, ICAM2, ICAM3, ICAM4 and ICAM5. ICAMs are closely related to immunity, such as mediating the antigen-specific immune response, lymphocyte circulation, NK-cell-mediated clearance and other adhesion interactions [12]. To date, ICAM1/3 have been confirmed to play an important role in the development of BC. For example, ICAM1/3 expression was upregulated in endothelial cells of BC, especially in ER-negative tumors, and tumors with high ICAM1/3 expression were proven to have more invasive characteristics and worse prognosis [13]. ICAM1 was also demonstrated to be a prognostic molecular target of TNBC [14]. Some studies have proven that ICAM2 plays a key role in immunotherapy of pancreatic cancer [15] and in preventing the spread of gastric cancer [16]. Overexpression of ICAM2 induced sensitivity of oral squamous cell carcinoma cells to radiotherapy [17]. ICAM4 variants are associated with glioma [18], and ICAM-5 may play a role in tumorigenesis and perineural invasion, most likely through the P13K/Akt signaling pathway [19]. However, ICAM2/4/5 has not been reported in BC.

The immune response of BC was initiated by neoantigens expressed by tumor cells that encode altered genes, such as gene mutations and abnormal methylation [5]. In summary, the literature mainly reports the expression and prognosis of ICAM1/3 in BC and its involvement in EMT and inflammation-related functional pathways. Mutations and methylation patterns of ICAMs in BC and their relationship with immune response and immunotherapy have not been reported. Therefore, the purpose of this study was to investigate the expression, mutation and methylation levels of ICAM family members in breast cancer and the correlation of ICAMs expression with survival, tumor immune cell invasion levels, and PD-1/PD-L1/CTLA4 expression levels.

2. Materials and methods

2.1 UALCAN

UALCAN (http://ualcan.path.uab.edu) is a comprehensive, interactive online platform for analysing cancer omics data mainly from The Cancer Genome Atlas (TCGA), MET500 and Clinical Proteomics Tumor Analysis Consortium (CPTAC) [20, 21]. An internal PERL (Practical Extraction and Report Language) program of UALCAN was utilized to obtain transcripts per million (TPM) as a measure of gene expression values from the TCGA database. Interquartile range (IQR) of TPM values used to generate the boxplot that assessed differences in gene expression levels between groups. Protein values downloaded from the CPTAC data portal were log ${}_{2}$ normalized in each sample. Then, a $Z$ value for each sample for each protein was calculated as the standard deviation from the median across samples. Additionally, the TCGA-assembler pipeline was used to download TCGA DNA methylation data, which were generated using the Illumina Infinium HumanMethylation 450 BeadChip and further processed to calculate an average methylation (beta) value for each gene, considering CpG sites located in the promoter region of the gene (1500 bases upstream and 200 bases downstream of transcription start sites [TSS]) via the CalculateSingleValueMethylationData in TCGA-assembler. In this study, we utilized the UALCAN platform to analyse the mRNA, protein, and promoter methylation levels of ICAM family members in BC.

2.2 Catalog of somatic mutations in cancer (COSMIC)

COSMIC (https://cancer.sanger.ac.uk/cosmic/) is the largest and most comprehensive global study on the impact of somatic mutation on human cancer, which is merged with genome-wide annotations from 466 whole genome and large-scale systematic screen publications, as well as open-access data from The Cancer Genome Atlas (TCGA) and the International Cancer Genome Consortium (ICGC) [22]. COSMIC introduced a new FAthmm-MKL algorithm, which uses a hidden Markov model to predict the functional, molecular and phenotypic consequences of protein missense variation. For each type of mutation, they calculate the percentage as num_samples/total samples * 100. In this study, we utilized COSMIC to explore the ICAMs’ mutation patterns in BC.

2.3 cBioPortal

cBioPortal (http://cbioportal.org) is an open source and online data for the interactive exploration of multiple cancer genomes. Data such as DNA copy numbers, nonsynonymous mutations, DNA methylation and clinical information were collected on the website [23]. The TCGA data in cBioPortal were retrieved from the Broad Institute’s GDAC Firehose (http://gdac.broadinstitute.org), an analysis pipeline that automatically performs standard processing and analyses. cBioPortal utilizes the mutation calls provided by each publication without performing any additional filtering. The only processing is to standardize the annotation of the mutations using Genome Nexus (which utilizes VEP with the canonical UniProt transcript). In our study, TCGA pancancer data were analysed to determine the mutation rate of the ICAM family, the distribution of mutations in different molecular types, and the association between TP53 mutations and their mutations via cBioPortal.

2.4 MethSurv databases

MethSurv (https://biit.cs.ut.ee/methsurv/) is a user-friendly interactive website that provides survival analysis based on DNA methylated biomarkers [24]. MethSurv matched the downloaded methylation data (level 3 data, HM450K) with the clinical data using the Biotab utility of TCGA. DNA methylation status was represented as $\beta$ -values (ranging from 0 to 1), which were derived from methylated (M) and unmethylated (U) probe intensities, using the formula M/(M $+$ U $+$ 100). Additionally, MethSurv annotated TCGA methylation data (covering 486,428 CpG sites) into genome regions based on an annotated file (Human genome build 37) provided by Illumina, Briefly, which includes two major classifications: First, in relation to CGI including “shores” (0–2 kb from CGI), “shelves” (2–4 kb from CGI) and “open sea” (remaining regions); Second, in relation to gene subregions including TSS200 (region from the transcription start site [TSS] to $-$ 200 nucleotides upstream of TSS), TSS1500 (covering $-$ 200 to $-$ 1500 nucleotides upstream of TSS), first exon, 5’UTR, body and 3’UTR. In this study, MethSurv was used to further explore the correlation between the methylation patterns of the ICAM family and their clinical characteristics and is presented as heatmaps.

2.5 Kaplan-Meier Plotter

The Kaplan-Meier Plotter (http://kmplot.com/analysis/) is the most comprehensive prognostic analysis website, which mainly includes the Gene Expression Omnibus (GEO), European Genome-phenome Archive (EGA), and TCGA [25, 26]. Survival analysis was performed using Cox proportional hazards regression. The package “survival” version 3.3.1 was used to calculate Kaplan-Meier survival curves, and the number-at-risk is indicated below the main plot. The hazard ratio (and 95% confidence intervals) and log-rank $p$ are calculated and displayed. In this study, we used it to determine the prognostic values of the ICAM family mRNA expression in different molecular subtypes in BC. Then, all prognostic outcomes were visualized using “forestplot” package version 2.0.1 in R software (version 4.1.1).

2.6 PrognoScan databases

Figure 1.

The flow-process diagram of this study.

PrognoScan (http://www.prognoscan.org/) is a database used for meta-analysis of genetic prognostic value. Its data come from GEO, ArrayExpress and individual lab websites [27]. Survival analysis in PrognoScan employs the minimum P value approach to find the cutpoint in continuous gene expression measurement for grouping patients. First, patients are ordered by the expression value of a given gene. Next, patients were divided into two (high and low) expression groups at the most pronounced P value cut-off point, and the risk differences of the two groups were estimated by the log-rank test. We used it to determine the prognostic value of ICAM mRNA expression in BC. Then, all prognostic outcomes were visualized using “forestplot” package version 2.0.1 in R software (version 4.1.1).

2.7 GeneMANIA

GeneMANIA (http://www.genemania.org) is a user-friendly and flexible web interface for the construction of gene interaction networks [28]. GeneMANIA uses the GO biological process (BP) function categories to evaluate gene function prediction, measures pair similarity between gene expression profiles and constructs networks using Pearson’s correlation coefficient. In addition, the site can set the sources of network edges with a variety of bioinformatics methods, such as physical interaction, gene colocalization, gene coexpression and gene enrichment analysis. We visualized the gene-gene interaction network of ICAM coexpression via GeneMANIA in this study.

2.8 Genome set cancer analysis (GSCA)

GSCA (http://bioinfo.life.hust.edu.cn/GSCA/) is an integrated database of genomic and immunogenomic genomic sets for cancer analysis. In GSCA, a range of services are provided to perform genomic genome and immunogenomic (24 immune cells) analysis [29]. The expression module provides the Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) pathway enrichment for the input gene set. KEGG and GO analyses were performed by the R package clusterProfiler version 2.4.3. The p value was adjusted to obtain the false discovery rate (FDR) (Benjamini & Hochberg). Only terms with FDR $<$ 0.05 remained. In this study, “Single gene level analysis” and “Gene set enrichment analysis” of the “Expression” module in GSCA were used to carry out ICAM1-related pathway and functional enrichment analysis.

Figure 2.

The mRNA expression levels of the ICAM family in BC (UALCAN). (A) Differences in the mRNA expression levels of the ICAM family in breast cancer tissues and normal tissues. (B) Differences in the mRNA expression levels of the ICAM family in different molecular subtypes of BC. ${}^{*}p<$ 0.05, ${}^{**}p<$ 0.01, ${}^{***}p<$ 0.001 vs. Normal. A $p$ value $<$ 0.05 was considered statistically significant.

Figure 3.

The protein expression levels of the ICAM family in BC (UALCAN) (UALCAN). (A) Differences in the protein expression levels of the ICAM family in breast cancer tissues and normal tissues. (B) Differences in the protein expression levels of the ICAM family in different molecular subtypes of BC. ${}^{*}p<$ 0.05, ${}^{**}p<$ 0.01, ${}^{***}p<$ 0.001 vs. Normal. A $p$ value $<$ 0.05 was considered statistically significant.

2.9 Tumor IMmune Estimation Resource (TIMER)

TIMER (https://cistrome.shinyapps.io/timer/) is a reliable, intuitive tool that systematically evaluates 32 types of cancer in 10897 samples of diverse immune cells (B cells, CD4 $+$ T cells, CD8 $+$ T cells, neutrophils, macrophages and dendritic cells) and their clinical effect in TCGA [30]. TIMER used the RSEM-processed tumor gene expression profiles from TCGA and the Human Primary Cell Atlas (HPCA) as the reference dataset of gene expression profiles of sorted immune cell types. Partial correlations of immune cell abundance and gene expression of chemokines and receptors, somatic mutation counts, gene expression, and immunosuppressive molecule expression were calculated and corrected using the $q$ value and FDR thresholds were applied based on the abundance of signals in the data. The Pearson correlation was applied to calculate purity and gene expression, while for others, it used the Spearman correlation. Spearman’s rho values and statistical significance after purity correction are shown on the upper right of the scatterplots, which were used to present our results In this study, we analysed the relationship between ICAM1 expression and immune cells and PD-1/PD-L1 and CTLA4 expression in the gene module and correlation module of TIMER, respectively.

Finally, we provided a flowchart figure (Fig. 1), which shows our research step by step.

Figure 4.

The genetic alteration characteristics of the ICAM family in BC. (A–B) Pie chart showing the percentage of the different mutation types of ICAMs in BC (COSMIC). Mutation types mainly include missense mutations, nonsense mutations, and synonymous mutations. (B) Pie chart showing the percentage of the different base mutation types of ICAMs in BC (COSMIC). (C) ICAMs mutation level in BC (cBioPortal); (D) Distribution of ICAMs mutations in different molecular subtypes of BC (cBioPortal). (E) Relationship between ICAMs mutations and common mutations of BC (cBioPortal).

Figure 5.

The ICAM family promoter methylation profile based on sample types, major subclasses and TP53 mutation status in BC (UALCAN). (A) The ICAM family promoter methylation profile based on sample types. (B) The ICAM family promoter methylation profile based on major subclasses. (C) The ICAM family promoter methylation profile based on TP53 mutation status. The beta value indicates the level of DNA methylation ranging from 0 (unmethylated) to 1 (fully methylated). Different beta value cut-offs have been considered to indicate hypermethylation [beta value: 0.7–0.5] or hypomethylation [beta-value: 0.3–0.25]. ${}^{*}p<$ 0.05, ${}^{**}p<$ 0.01, ${}^{***}p<$ 0.001 vs. Normal. ${}^{\scriptscriptstyle\#\#\#}p<$ 0.001 vs. mutant. A $p$ value $<$ 0.05 was considered statistically significant.

3. Results

3.1 The ICAM family mRNA and protein expression in BC

First, we used the TCGA analysis and CPTAC analysis modules of the UALCAN platform to explore the differences in the mRNA and protein expression levels of ICAMs in normal breast tissue and breast cancer tissue, as well as in different molecular subtypes. The results of TCGA analysis showed that the mRNA expression levels of ICAM1 and ICAM3 in breast primary tumors were higher than those in breast normal tissues, whereas the expression patterns of ICAM2, ICAM4 and ICAM5 were reversed (Fig. 2A). From the perspective of different molecular subtypes of BC, the mRNA expression levels of ICAM1 in triple-negative tumors and ICAM3 in HER2 $+$ tumors were significantly higher than those of other molecular subtypes (Fig. 2B). Additionally, the results of CPTAC analysis showed that the protein expression levels of ICAM1, ICAM2 and ICAM3 in breast cancer tissues were downregulated compared with those in breast normal tissues (Fig. 3A); however, the protein expression data of ICAM4 and ICAM5 were not retrieved from the UALCAN database. Interestingly, from the perspective of different molecular subtypes, ICAM1 protein in triple-negative breast tumors was upregulated compared to normal breast tissues, while the expression of ICAM2 and ICAM3 proteins was still higher in normal breast tissues (Fig. 3B). It should be noted that we did not retrieve the protein data of ICAM4 and ICAM5 from the CPTAC database, so it was not analysed here.

Figure 6.

The prognostic value of the ICAM family for OS and RFS in BC patients. (A–B) The OS and RFS of the ICAM family mRNA expression in BC (Kaplan-Meier Plotter). (C–D) The OS and RFS of the ICAM family mRNA expression in BC (PrognoScan). A $p$ value $<$ 0.05 was considered statistically significant.

3.2 Genomic alterations of the ICAM family in BC

COSMIC provides information on ICAM mutations in different cancers, resulting in a pie chart. Only ICAM2 was found to have a 3.7% nonsense mutation rate in BC, while missense mutations in ICAM1 (25%), ICAM2 (25.93%), ICAM3 (33.33%), ICAM4 (58.33%) and ICAM5 (53.85%) were found in BC. In addition, synonymous mutations were observed in ICAM1 (17.86%), ICAM2 (3.7%), ICAM3 (16.67%), ICAM4 (8.33%) and ICAM5 (15.38%) (Fig. 4A). C $>$ T mutation was most common in the ICAMs. Other types of base mutations occasionally occur in ICAMs (Fig. 4B).

Then, we used cBioPortal to determine the type and frequency of genetic alterations in the ICAMs. Alterations in ICAM1, ICAM2, ICAM3, ICAM4 and ICAM5 were 6%, 10%, 5%, 6% and 6% in 994 BC patients, respectively (Fig. 4C). Most of the alterations were amplifications. From the perspective of molecular subtypes of BC, the mutation rate of ICAMs in the basal-like subtype was the highest, followed by the mutation rate in the luminal B subtype (Fig. 4D). In addition, the mutation rate of ICAMs in the TP53 mutant group was significantly higher than that in the TP53 wild-type group, while the mutation rate of ICAMs in the phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha (PIK3CA) mutant group was lower than that in the wild-type group in BC (Fig. 4E).

Figure 7.

Functional pathway enrichment analysis of ICAM1 and its interacting genes (GSCA). (A) The biological processes involved in ICAM1. (B) ICAM1 is involved in cell composition. (C) Molecular functions in which ICAM1 was involved. (D) ICAM1 participates in the KEGG pathway. BP, biological process; CC, cell composition; MF, molecular function; KEGG, Kyoto Encyclopedia of Genes and Genomes; FDR, false discovery rate.

3.3 Methylation pattern characteristics of the ICAM family in BC

The methylation level of the oncogene promoter region is closely related to the occurrence and development of cancer, including BC. The UALCAN database was used to explore the levels of promoter region methylation in ICAMs in BC from multiple perspectives. In general, the promoter methylation level of the ICAM family in BC tissues was higher than that in breast normal tissues (Fig. 5A). From the perspective of the molecular subtype of BC, ICAM1 and ICAM5 showed similar methylation patterns: the methylation level of the luminal subtype was higher than that of normal tissues, while the methylation level of the HER2 $+$ and TNBC subtypes was lower than that of normal breast tissues. However, ICAM2, ICAM3 and ICAM4 showed similar methylation patterns: the methylation levels in luminal, HER2 $+$ or TNBC subtypes were all higher than those in breast normal tissues (Fig. 5B). From the perspective of TP53 mutation in BC, the methylation level of ICAM1 in the TP53 mutant group was lower than that in the wild-type group, while ICAM3 showed the opposite pattern. Additionally, the methylation levels of ICAM2/ICAM4/ICAM5 showed no difference between the TP53-mutated group and the TP53 wild-type group (Fig. 5C).

MethSurv is also a well-known DNA methylation research database, whose data are mainly derived from TCGA. We used it to further investigate CpG methylation sites in the ICAMs in BC. The results showed that ICAM1 methylation sites were concentrated in CpG island, N-shelf and S-Shore regions, ICAM2 was only located in the Open Sea, and ICAM3 had four concentrated methylation site areas, including CpG Island, S-shelf, S-Shore and Open Sea regions (Fig. S1A–C). Additionally, ICAM4 and ICAM5 have the same methylation site regions, including island and N-shore regions (Fig. S1D–E).

3.4 Prognostic value of the ICAM family in BC

According to Kaplan-Meier Plotter’s research results, ICAM1/2/3 were favorable prognostic factors for OS in BC patients. Based on specific molecular subtypes, it was found that ICAM1 and ICAM2 were favorable prognostic factors for OS in BC patients with HER2 $+$ and basal-like subtypes. ICAM1 and ICAM3 were favorable prognostic factors for OS in patients with the luminal B subtype (Fig. 6A). In addition, ICAM1/2/4/5 were favorable prognostic factors for all BC patients’ RFS, and BC patients with higher expression of ICAM1/2/5 had better RFS in all subtypes. ICAM4 showed a favorable prognostic effect on the RFS of all subtypes except the HER2 $+$ subtype. ICAM3 had no significant prognostic effect on RFS in BC patients (Fig. 6B). However, data from PrognoScan suggested that ICAM1/4/5 were favorable prognostic factors for BC patients, while ICAM3 showed adverse effects. Interestingly, Duke et al. reported that ICAM2 was an adverse prognostic factor for OS in BC patients, whereas Stockholm et al. showed the opposite finding (Fig. 6C). In addition, ICAM3/4 were considered to be adverse prognostic factors for RFS in BC patients, while ICAM2 showed an adverse effect (Fig. 6D). It was also interesting to note that research by GUYT et al. reported that ICAM1 was a favorable prognostic factor for RFS in BC patients (Fig. 6D).

Figure 8.

Correlation between ICAM family expression and tumor-infiltrating immune cells in BC (TIMER). The horizontal axis is the infiltration level of tumor-infiltrating immune cells, and the vertical axis is the expression level of the ICAM family. A $p$ value $<$ 0.05 was considered statistically significant.

3.5 Functional and pathway enrichment analysis of ICAM1 in BC

Since ICAM1 showed favorable prognostic value in all the datasets analysed, it is recognized as the hub gene of BC among the ICAM family. Then, a gene interaction network was constructed via the GeneMANIA database. The results suggested that among the top 20 genes associated with ICAMs, integrin subunit alpha L (ITGAL) was the most significant (Fig. S2A). Next, we used GSCA to explore the activity of ICAM1 in 10 cancer-related pathways in BC. The results suggested that ICAM1 was potentially involved in the activation of apoptosis and epithelial-mesenchymal transition (EMT) pathways and the inhibition of the hormone androgen receptor (AR) and hormone ER pathways (Fig. S2B). Additionally, we utilized the GSCA website to conduct KEGG and GO analyses on the 20 genes that were most significantly associated with ICAM1 obtained from GeneMANIA. The results of GO analysis showed that ICAM1 was mainly involved in biological processes such as immune response regulation and cell adhesion (Fig. 7A). It was mainly involved in the formation of the integral components of the extracellular region or plasma membrane (Fig. 7B) and the function of binding integrin (Fig. 7C). The results of KEGG pathway analysis showed that ICAM1 was mainly enriched in the pathways regulating the actin cytoskeleton and cell adhesion molecules (Fig. 7D).

Figure 9.

The correlation of ICAM1 expression with PD-1, PD-L1 and CTLA4 expression in BC (TIMER). (A) The correlation of ICAM1 expression with PD-1, PD-L1 and CTLA4 expression in all BC. (B) The correlation of ICAM1 expression with PD-1, PD-L1 and CTLA4 expression in the luminal subtype. (C) The correlation of ICAM1 expression with PD-1, PD-L1 and CTLA4 expression in the HER2+ subtype. (D) The correlation of ICAM1 expression with PD-1, PD-L1 and CTLA4 expression in the basal-like subtype. A $p$ value $<$ 0.05 was considered statistically significant.

3.6 Relationships between ICAM1 expression and immune cell infiltration/PD-1/PD-L1/CTLA4 expression in BC

Tumor-infiltrating immune cells can affect the survival of patients, and the results of enrichment analysis of the above functional pathways supported the important role of ICAMs in immune regulation. Therefore, it is of great significance to explore the relationships between ICAM1 expression and immune infiltration/PD-1/PD-L1/CTLA4 expression in BC.

First, we used TIMER to calculate the coefficient of ICAM1 expression and immune cell invasion level in all BCs and different molecular subtypes to determine whether ICAM1 expression is related to the abundance of neoplastic infiltrating lymphocytes. The results showed that ICAM1 expression was negatively correlated with the purity of BC and was significantly positively associated with the infiltration level of B cells, CD8 $+$ T cells, CD4 $+$ T cells, macrophages, neutrophils, and dendritic cells in BC (Fig. 8A). Additionally, in terms of different molecular subtypes, the expression level of ICAM1 was positively correlated with the infiltration level of B cells, CD8 $+$ T cells, CD4 $+$ T cells, macrophages, neutrophils, and dendritic cells in the luminal subtype (Fig. 8B). In the basal-like subtype, the expression level of ICAM1 was also positively correlated with B cells, CD8 $+$ T cells, CD4 $+$ T cells, neutrophils, and dendritic cells but not with macrophages (Fig. 8C). In the HER2 $+$ subtype, the ICAM1 expression level was positively associated with the infiltration level of CD8 $+$ T cells, neutrophils and dendritic cells but not with the infiltration level of B cells, CD4 $+$ T cells, or macrophages (Fig. 8D).

Furthermore, we explored the correlation between the ICAM1 expression level in BC and the expression of PD-1, PD-L1, and CTLA4 via TIMER. The results suggested that the ICAM1 expression level in BC was significantly positively associated with the expression of PD-1, PD-L1 and CTLA4 (Fig. 9A). In different molecular subtypes, the expression level of ICAM1 was positively correlated with PD-1, PD-L1 and CTLA4 (Fig. 9B–D).

4. Discussion

This study was the first to comprehensively analyse the correlation of the expression, mutation and methylation of ICAM family members in BC via multiple online databases. It is noteworthy that ICAM1 mRNA was upregulated and ICAM1 protein was downregulated in breast cancer tissues. There are two explanations for this phenomenon. First, breast cancer is a highly heterogeneous tumor [31], so it is common for the same molecule to have different expression levels in different breast cancer subtypes. Second, the biological process from mRNA to protein involves much posttranscriptional regulation [32], which is also one of the reasons for the inconsistency of mRNA and protein expression levels. Additionally, our study found that both the mRNA and protein levels of ICAM1 were elevated in TNBC but opposite in luminal and HER2 $+$ tumors, suggesting that more complex regulatory mechanisms may be involved in these two subtypes. Previous studies proved that ICAM1/3 expression was upregulated in endothelial cells of BC, especially in ER-negative tumors, and tumors with high ICAM1/3 expression were proven to have more invasive characteristics and worse prognosis [33, 34], which was consistent with our findings.

Next, this study further explored the variation and methylation status of the ICAM family. ICAMs in BC had a relatively high frequency of missense mutations, among which the most common base replacement was C $>$ T. Amplification mutation was the most common genetic mutation of ICAMs, and missense mutation mainly appeared in ICAM1 in BC. In addition, in the ICAMs mutation group, patients with TNBC accounted for the largest proportion, accompanied by a higher TP53 mutation frequency. Previous studies have shown that TP53 mutation is the most common genetic alteration and is associated with poor overall survival in BC [35]. Only 13% of TP53 mutations in luminal A and 71%, 86%, 75% of those in luminal B, HER2 $+$ , TNBC subtypes were discovered, respectively [35]. Next, the results of our study showed that the methylation level of ICAMs in BC tissues was higher than that in breast normal tissues. Interestingly, from the perspective of molecular subtypes, only the methylation level of ICAM1 in TNBC was lower than that in the normal group. Collectively, we found that ICAM1 expression and methylation levels in TNBC were opposite, which was consistent with the previous conclusion that DNA methylation could inhibit gene expression [36]. In addition, the missense mutation and TP53 mutation rates of ICAMs in TNBC were higher than those of other subtypes.

Furthermore, based on the analysis results of the Kaplan-Meier Plotter database, only ICAM1 and ICAM2 were found to be favorable prognostic factors for OS and RFS in patients with BLBC in this study. However, previous literature has reported that ICAM1 overexpression in breast cancer cells (especially TNBC cells) and downregulation of ICAM1 expression can inhibit the invasion and metastasis of tumor cells [33, 37, 38], indicating that ICAM1 is a potential therapeutic target for TNBC. Next, in the process of exploring its functional pathway, ICAM1-related genes were mainly involved in activating apoptosis and inhibiting the hormone AR and ER pathways. Current studies have shown that the hormone AR and ER pathways are involved in the occurrence, invasion and metastasis of TNBC, and their inhibition can lead to a better prognosis for these patients [39]. Additionally, our study found that ICAM1 was highly expressed in TNBC, and its high expression was positively correlated with the immune cell infiltration level. Collectively, we found that ICAM1 was highly expressed in TNBC but was correlated with better OS and RFS, and the most likely reason is that ICAM1 expression was positively correlated with the level of immune cell infiltration. Therefore, we believe that ICAM1 is closely related to the occurrence, development and immunotherapy effect of TNBC.

A large number of studies on the precision treatment of immune-related cancers are ongoing. Many studies have found that the methylation levels of oncogenes and TP53 mutations are closely related to the immune microenvironment and immune response of tumors. Tumor-specific antigens, or neoantigens, produced by somatic mutations can influence patients’ response to immunotherapy and promote tumor shrinkage [40]. Wild-type TP53 genes are involved in cell cycle regulation, DNA damage repair and damage-induced apoptosis. Previous studies have shown that TP53 mutation is closely related to the immune microenvironment and immunotherapy response of liver cancer and head-neck cancer [41, 42]. In addition, TP53 mutation has also been found to be related to the immunotherapy response of anti-PD-1 and anti-PD-L1 in ovarian cancer [43]. Our study found that in the altered group of ICAMs, the proportion of patients with TNBC or TP53 mutation rates was the highest. Therefore, from the perspective of mutation, ICAMs may play a regulatory role in the immune regulation and immunotherapy of BC, especially in TNBC patients.

It is well known that the first epigenetic change described in human cancer is the loss of DNA methylation. In BC, general hypomethylation of cancer genes can lead to gene activation, oncogene upregulation and chromosome instability, or local abnormal hypermethylation can lead to DNA repair gene silencing, further leading to gene suppression and genetic instability [44]. In addition, some studies have shown that DNA methylation plays an important role in the immune response of tumors or inflammation [45, 46, 47], and evidence has shown that the function of DNA methylation can be used to evaluate the immune response of BC and can improve the diagnosis and prognosis of BC. Overall, high or low levels of DNA methylation reflect, to some extent, a better predictor of the host’s response to immunotherapy. Our results indicated that the methylation level of ICAM1 in TNBC was lower than that in the normal group, the methylation level of ICAM1 in BC was negatively correlated with the immune cell infiltration level, and high expression of ICAM1 was associated with better OS and RFS, which was consistent with the research results of Jeschke J and Zhou et al. [48, 49]. In summary, from the perspective of epigenetics, ICAM1 may become the latest predictive biomarker of immunotherapy in TNBC.

In BC, the immune biomarkers of the TNBC subtype and the expression levels of PD-1, PD-L1 and CTLA4 are relatively high [50]. Through TCGA RNA sequencing, Mittendorf et al. found that the expression levels of PD-L1 in the TNBC group ( $n=$ 120) were higher than those in the non-TNBC group ( $n=$ 716), so anti-PD-L1 immunotherapy may be more effective for patients with TNBC [51]. Moreover, some studies have shown that TNBC and HER2 $+$ BC have a higher degree of immune cell infiltration. In the TNBC subgroup, lymphocyte-dominated breast cancer (LPBC) has a significantly better prognosis than non-LPBC, suggesting that tumor-infiltrating immune cells can be used as an indicator to predict the prognosis of TNBC [52]. In addition, ICAM1-specific CAR T cells have high therapeutic potential for ICAM1-positive TNBC tumors [53]. The results of this study found that the ICAM1 expression level in BC was not only positively associated with the expression of PD-1, PD-L1 and CTLA4 but also positively associated with the infiltration level of tumor-infiltrating immune cells. Therefore, it is reasonable to predict that ICAM1 plays a critical role in the direction of immune checkpoint treatment of BC, which needs further verification through in vitro and in vivo experiments.

Our study has some limitations. We analysed the expression, mutation and methylation of the ICAM family in BC mainly at the transcriptional level and further explored their correlation with immune infiltration and immunotherapy. However, we did not further investigate these relationships at the protein level, mainly because there are few publicly available databases on proteins. In addition, in vivo or in vitro basic studies are needed to confirm our results.

5. Conclusions

In summary, ICAM1 was closely related to the occurrence, development and immunotherapy effect of TNBC. Additionally, we proposed that ICAM1 can be further explored from the perspective of mutation and methylation in immunoregulation and immunotherapy of TNBC. Therefore, our study may provide new insights into the selection of immune checkpoint therapeutic targets and prognostic biomarkers for BC.

Abbreviations

BC, breast cancer; CTLA4, cytotoxic T-lymphocyte associated protein 4; COSMIC, Catalog of Somatic Mutations in Cancer; ER, estrogen receptor; EMT, epithelial-mesenchymal transition; EGA, European Genome-phenome Archive; GEO, Gene Expression Omnibus; GSCA, Gene Set Cancer Analysis; HER2, human epidermal growth factor receptor 2; ITGAL, integrin subunit alpha L; ICAMs, intercellular adhesion molecules; LPBC, lymphocyte-dominated breast cancer; TNBC, triple negative breast cancer; PR, progesterone receptor; ICGC, International Cancer Genome Consortium; LAG3, lymphocyte activating 3; LncMAP, LncRNA Modulator Atlas in Pancancer; PD-1, programmed cell death 1; PD-L1, programmed cell death-ligand 1; PIK3CA, phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha; TP53, tumor protein P53; TNBC, triple negative breast cancer; TNF, tumor necrosis factor; TF, transcription factor; TCGA, The Cancer Genome Atlas; TIMER, Tumor IMmune Estimation Resource; TISDB, Tumor-Immune System Interaction Database; VCAM-1, vascular cell adhesion molecule 1.

Author contributions

Conception: Heyan Chen, Huimin Zhang and Jianjun He.

Interpretation or analysis of data: Heyan Chen, Huimin Zhang, Shengyu Pu and Nan Mei.

Preparation of the manuscript: Heyan Chen, Huimin Zhang.

Revision for important intellectual content: Heyan Chen and Xiaoxu Liu.

Supervision: Heyan Chen and Huimin Zhang.

Data availability statement

The data used to support the findings of this study are available from publicly available online databases. The access sites for each of the open databases used in this study were specifically described in the methodology.

Supplementary data

The supplementary files are available to download from http://dx.doi.org/10.3233/CBM-220073.

sj-pdf-1-cbm-10.3233_CBM-220073.pdf - Supplemental material

Supplemental material, sj-pdf-1-cbm-10.3233_CBM-220073.pdf

Footnotes

Acknowledgments

We thank The Cancer Genome Atlas, Gene Expression Omnibus and European Genome-phenome Archive for sharing the large amount of data. The authors also thank the Institutional Foundation of The First Affiliated Hospital of Xi’an Jiaotong University (No. 2022YQPY08) and Shaanxi Provincial Key Research and Development Project (No. 2020SF-294) for funding to do this work.

Conflict of interest

The authors declare no conflicts of interest.

References

Bray

et al., Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries, CA Cancer J Clin 68 (2018), 394–424.

Perou

C.M.

et al., Molecular portraits of human breast tumours, Nature 406 (2000), 747–52.

Veronesi

et al., Breast cancer, Lancet 365 (2005), 1727–41.

Farkona

Diamandis

E.P.

and Blasutig

I.M.

, Cancer immunotherapy: the beginning of the end of cancer? BMC Med 14 (2016), 73.

Nagarajan

and McArdle

S.E.B.

, Immune Landscape of Breast Cancers, Biomedicines 6 (2018).

Solinas

et al., Tumor-infiltrating lymphocytes in breast cancer according to tumor subtype: Current state of the art, Breast 35 (2017), 142–150.

Ruffell

et al., Leukocyte composition of human breast cancer, Proc Natl Acad Sci USA 109 (2012), 2796–801.

Dieci

M.V.

et al., International Immuno-Oncology Biomarker Working Group on Breast, Update on tumor-infiltrating lymphocytes (TILs) in breast cancer, including recommendations to assess TILs in residual disease after neoadjuvant therapy and in carcinoma in situ: A report of the International Immuno-Oncology Biomarker Working Group on Breast Cancer, Semin Cancer Biol 52 (2018), 16–25.

Savas

et al., Clinical relevance of host immunity in breast cancer: from TILs to the clinic, Nat Rev Clin Oncol 13 (2016), 228–41.

10.

Brignone

et al., First-line chemoimmunotherapy in metastatic breast carcinoma: combination of paclitaxel and IMP321 (LAG-3Ig) enhances immune responses and antitumor activity, J Transl Med 8 (2010), 71.

11.

Amaria

R.N.

et al., Neoadjuvant immune checkpoint blockade in high-risk resectable melanoma, Nat Med 24 (2018), 1649–1654.

12.

Nanda

et al., Pembrolizumab in Patients With Advanced Triple-Negative Breast Cancer: Phase Ib KEYNOTE-012 Study, J Clin Oncol 34 (2016), 2460–7.

13.

Rugo

H.S.

et al., Safety and Antitumor Activity of Pembrolizumab in Patients with Estrogen Receptor-Positive/Human Epidermal Growth Factor Receptor 2-Negative Advanced Breast Cancer, Clin Cancer Res 24 (2018), 2804–2811.

14.

Simmons

C.E.

et al., Positive progress: current and evolving role of immune checkpoint inhibitors in metastatic triple-negative breast cancer, Ther Adv Med Oncol 12 (2020), 1758835920909091.

15.

Liu

Guo

and Lopez

R.D.

, Expression of intercellular adhesion molecule (ICAM)-1 or ICAM-2 is critical in determining sensitivity of pancreatic cancer cells to cytolysis by human gammadelta-T cells: implications in the design of gammadelta-T-cell-based immunotherapies for pancreatic cancer, J Gastroenterol Hepatol 24 (2009), 900–11.

16.

Tanaka

et al., ICAM-2 gene therapy for peritoneal dissemination of scirrhous gastric carcinoma, Clin Cancer Res 10 (2004), 4885–92.

17.

Ishigami

et al., Inhibition of ICAM2 induces radiosensitization in oral squamous cell carcinoma cells, Br J Cancer 98 (2008), 1357–65.

18.

Rajaraman

et al., Common variation in genes related to innate immunity and risk of adult glioma, Cancer Epidemiol Biomarkers Prev 18 (2009), 1651–8.

19.

Maruya

S.I.

et al., ICAM-5 (telencephalin) gene expression in head and neck squamous carcinoma tumorigenesis and perineural invasion! Oral Oncol 41 (2005), 580–8.

20.

Chandrashekar

D.S.

et al., UALCAN: A Portal for Facilitating Tumor Subgroup Gene Expression and Survival Analyses, Neoplasia 19 (2017), 649–658.

21.

Chandrashekar

D.S.

et al., UALCAN: An update to the integrated cancer data analysis platform, Neoplasia 25 (2022), 18–27.

22.

Forbes

S.A.

et al., COSMIC: exploring the world’s knowledge of somatic mutations in human cancer, Nucleic Acids Res 43 (2015), D805–11.

23.

Cerami

et al., The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data, Cancer Discov 2 (2012), 401–4.

24.

Modhukur

et al., MethSurv: a web tool to perform multivariable survival analysis using DNA methylation data, Epigenomics 10 (2018), 277–288.

25.

Gyorffy

, Survival analysis across the entire transcriptome identifies biomarkers with the highest prognostic power in breast cancer, Comput Struct Biotechnol J 19 (2021), 4101–4109.

26.

Gyorffy

et al., An online survival analysis tool to rapidly assess the effect of 22,277 genes on breast cancer prognosis using microarray data of 1,809 patients, Breast Cancer Res Treat 123 (2010), 725–31.

27.

Mizuno

et al., PrognoScan: a new database for meta-analysis of the prognostic value of genes, BMC Med Genomics 2 (2009), 18.

28.

Warde-Farley

et al., The GeneMANIA prediction server: biological network integration for gene prioritization and predicting gene function, Nucleic Acids Res 38 (2010), W214–20.

29.

Liu

C.J.

et al., GSCALite: a web server for gene set cancer analysis, Bioinformatics 34 (2018), 3771–3772.

30.

et al., TIMER: A Web Server for Comprehensive Analysis of Tumor-Infiltrating Immune Cells, Cancer Res 77 (2017), e108–e110.

31.

Yeo

S.K.

and Guan

J.L.

, Breast Cancer: Multiple Subtypes within a Tumor?, Trends Cancer 3 (2017), 753–760.

32.

Cruz

and Lemos

, Post-transcriptional diversity in riboproteins and RNAs in aging and cancer, Semin Cancer Biol 76 (2021), 292–300.

33.

Guo

et al., ICAM-1 as a molecular target for triple negative breast cancer, Proc Natl Acad Sci USA 111 (2014), 14710–5.

34.

Schroder

et al., Prognostic value of intercellular adhesion molecule (ICAM)-1 expression in breast cancer, J Cancer Res Clin Oncol 137 (2011), 1193–201.

35.

Zhang

et al., Somatic Trp53 mutations differentially drive breast cancer and evolution of metastases, Nat Commun 9 (2018), 3953.

36.

Zhang

and Chang

, Epigenetics in Health and Disease, Adv Exp Med Biol 1253 (2020), 3–55.

37.

et al., Downregulation of human intercellular adhesion molecule-1 attenuates the metastatic ability in human breast cancer cell lines, Oncol Rep 35 (2016), 1541–8.

38.

Rosette

et al., Role of ICAM1 in invasion of human breast cancer cells, Carcinogenesis 26 (2005), 943–50.

39.

Khan

M.A.

et al., PI3K/AKT/mTOR pathway inhibitors in triple-negative breast cancer: a review on drug discovery and future challenges, Drug Discov Today 24 (2019), 2181–2191.

40.

Lehmann

B.D.

et al., Identification of human triple-negative breast cancer subtypes and preclinical models for selection of targeted therapies, J Clin Invest 121 (2011), 2750–67.

41.

Desrichard

Snyder

and Chan

T.A.

, Cancer Neoantigens and Applications for Immunotherapy, Clin Cancer Res 22 (2016), 807–12.

42.

Long

et al., Development and validation of a TP53-associated immune prognostic model for hepatocellular carcinoma, EBioMedicine 42 (2019), 363–374.

43.

Wieser

et al., BRCA1/2 and TP53 mutation status associates with PD-1 and PD-L1 expression in ovarian cancer, Oncotarget 9 (2018), 17501–17511.

44.

Harbeck

et al., Breast cancer, Nat Rev Dis Primers 5 (2019), 66.

45.

Ciechomska

Roszkowski

and Maslinski

, DNA Methylation as a Future Therapeutic and Diagnostic Target in Rheumatoid Arthritis, Cells textbf8(2019).

46.

Deleris

Halter

and Navarro

, DNA Methylation and Demethylation in Plant Immunity, Annu Rev Phytopathol 54 (2016), 579–603.

47.

Wichnieski

et al., DNA methylation profiles of immune response-related genes in apical periodontitis, Int Endod J 52 (2019), 5–12.

48.

Jeschke

et al., DNA methylation-based immune response signature improves patient diagnosis in multiple cancers, J Clin Invest 127 (2017), 3090–3102.

49.

Zhou

et al., DNA methylation loss in late-replicating domains is linked to mitotic cell division, Nat Genet 50 (2018), 591–602.

50.

Hendrickx

et al., Identification of genetic determinants of breast cancer immune phenotypes by integrative genome-scale analysis, Oncoimmunology 6 (2017), e1253654.

51.

Kok

Winer