Skip to main content
NCBI home page
IET Systems Biology logoLink to IET Systems Biology
. 2024 Feb 20;18(2):41–54. doi: 10.1049/syb2.12088

Anoikis‐related genes as potential prognostic biomarkers in gastric cancer: A multilevel integrative analysis and predictive therapeutic value

Yongjian Lin 1, Jinlu Liu 1,✉
PMCID: PMC10996445  PMID: 38377622

Abstract

Background

Gastric cancer (GC) is a frequent malignancy of the gastrointestinal tract. Exploring the potential anoikis mechanisms and pathways might facilitate GC research.

Purpose

The authors aim to determine the significance of anoikis‐related genes (ARGs) in GC prognosis and explore the regulatory mechanisms in epigenetics.

Methods

After describing the genetic and transcriptional alterations of ARGs, we searched differentially expressed genes (DEGs) from the cancer genome atlas and gene expression omnibus databases to identify major cancer marker pathways. The non‐negative matrix factorisation algorithm, Lasso, and Cox regression analysis were used to construct a risk model, and we validated and assessed the nomogram. Based on multiple levels and online platforms, this research evaluated the regulatory relationship of ARGs with GC.

Results

Overexpression of ARGs is associated with poor prognosis, which modulates immune signalling and promotes anti‐anoikis. The consistency of the DEGs clustering with weighted gene co‐expression network analysis results and the nomogram containing 10 variable genes improved the clinical applicability of ARGs. In anti‐anoikis mode, cytology, histology, and epigenetics could facilitate the analysis of immunophenotypes, tumour immune microenvironment (TIME), and treatment prognosis.

Conclusion

A novel anoikis‐related prognostic model for GC is constructed, and the significance of anoikis‐related prognostic genes in the TIME and the metabolic pathways of tumours is initially explored.

Keywords: cancer, genetics, patient treatment

In summary, an APM for gastric cancer based on ARGs to predict prognosis and assess TIME is generated. Anchorage‐dependent properties also mediated the crosstalk of signalling pathways, cellular metabolism, and chat. These evidences provide new insights into precise treatment strategies for GC patients guided by the anoikis pattern.

graphic file with name SYB2-18-41-g008.jpg

1. INTRODUCTION

Anoikis is an integral part of the secondary site colonisation process in primary tumour development, overcoming apoptosis for continued dissemination and infiltration of growing cells in the secondary environment [1]. In general, anchorage‐dependent cell death balances organismal development, endogenous organisation, disease development, and tumour metastasis. The anoikis‐related gene (ARG) family is not only overexpressed in specific tumour types or certain cell cycle processes but also fails to induce apoptosis in the corresponding stroma. Cumulative results have shown that overexpression or silencing of ARGs were both factors affecting the prognostic background of tumours [2]. Silencing CEACAM6 would impair the anoikis resistance and metastatic ability of pancreatic cancer cells in vivo, whereas overexpression of ARGs marked a worse prognosis in epithelial ovarian cancer, glioma, oesophageal squamous cell carcinoma, melanoma, and colorectal cancer [3]. Therefore, the distinctive anti‐anoikis transformation might offer targets for cancer therapy, which facilitates prognosis.

Gastric cancer (GC) continues to develop with limited treatment options and a poor prognosis due to heterogeneity, drug resistance, and insidiousness. Du et al. reported that anoikis‐resistant cells promoted angiogenesis and peritoneal metastasis through the mitogen‐activated protein kinase/ extracellular signal‐regulated kinase (ERK) signalling pathway [4]. PLAUR, TCF7L2, and other anoikis resistance genes could enhance GC distant metastasis and are candidate therapeutic targets [5]. In conclusion, the identification of prognostic targets is critical for diseases, and relevant advances in epigenetics and genomics would guide immunotherapy.

Anoikis relies on various contexts to promote tumour progression, including strengthening metastatic potential, expanding tumour stem cells, chemoresistance, and evasion of the immune surveillance system. In solid tumours, anti‐anoikis is considered the cornerstone of the cell metastasis signature, which involves circulating tumour cells negating natural killer (NK) cell and macrophage activity [6]. Lucía et al. explained that the metastatic status of colorectal cancer cells was associated with the ability of overexpressed mitochondrial Inhibitory Factor 1 to enhance the immune surveillance of NK cells [3]. The TIME is the surrounding microenvironment where tumour cells exist and where tumour‐infiltrating immune cells (TIICs) differentiate and destroy cancer cells, so we explored the necessity of TIME for immunotherapy and immunoscreening. Novel evidence demonstrated that anoikis resistance also helped tumour stem cells drive invasion in circulation, directly regulating epithelial‐mesenchymal transition (EMT) early in the metastatic cascade, as well as recruiting immunosuppressive cell populations and evading surveillance by downregulating tumourigenicity [7]. This cancer cell trait reflects the chemotherapy resistance and metastatic phenotype upon EMT reactivation. Smad interacting protein one induces EMT to regulate peritoneal carcinomatosis and promote intestinal GC [8]. While glycosaminoglycans exert an effect with trastuzumab in the control of anti‐anoikis [9]. In summary, stem cell events and extracellular matrix (ECM) might be factors in tumourigenic cell development and transformation.

For GC, potent drug targets such as cytotoxic T lymphocyte antigen 4 and programmed cell death protein 1 (PD‐1) are extremely helpful for precision therapy and targeted control of early metastasis. PD‐1 has given advanced cancer high‐tech therapies that have demonstrated favourable safety and tolerability. Caspase‐8, as a member of the cysteine protease family, activates during anoikis, augments cell death, and inhibits peritoneal dissemination of human GC cells [10]. Besides, immune checkpoint expression (including Tyro3, Gas6, MFGE8, Stab2, Tim‐4, CX3CL1, IDO1, Rac1, and PD‐L1) is directed against cancer cells to avoid immune attack, and they were associated with reducing the half‐maximal drug inhibitory concentration (IC50) for multiple anti‐cancer drugs [11]. Therefore, mining immunotherapeutic biomarkers and the role of epigenetics in gene regulation further assist GC patients in formulating precise immunotherapy regimens for durable efficacy.

We constructed an anoikis‐related prognostic model (APM) for GC, which has contributions to clinical endpoints, epigenetics, and TIME. Firstly, the distribution, mutation, and prognosis were identified after comparing the inheritance and transcription of ARGs. Then, the biological functional annotation of significant gene activation was enriched to further characterise the pathways of ARG variants. Next, we localised hub gene expression and alterations by histology and described the relevance of target genes in the tumour microenvironment (TME). We also assessed TIME by TIICs and cellular communication, including ARG expression profiles, response signalling between L‐R pairs, and immune subpopulations. Finally, the regulatory process of ARGs guided immunoscreening and small‐molecule therapeutic development. Our findings point to the sensitivity of APM to clinical endpoints, biological pathways, and TIME, and anoikis also offers an orientation to GC regulatory mechanisms in epigenetics.

2. MATERIALS AND METHODS

2.1. Acquisition of raw data

The mRNA sequencing matrix and complete clinicopathology files of GC patients were derived from the cancer genome atlas (TCGA) and gene expression omnibus (GEO) libraries, and four duplicates were filtered to obtain 804 tumour samples and 32 normal samples. UCSC Xena covered somatic mutation counts and copy number variation (CNV). Anoikis‐related genes were obtained from the novel studies and reports. The “limma” R package was applied for integration, transformation of TPMs, and differential analysis. The high‐throughput sequencing expression profiles used for single‐cell analysis were selected from the GSE series. Moreover, tools such as cBioPortal for Cancer Genomics, Genomics of Drug Sensitivity in Cancer, and the Human Protein Atlas were applied for online analysis. The design of this study process (Additional File 1: Figure S1).

2.2. Anoikis‐related gene distribution and genetic variation

Descriptions of genes focused on the interaction network and distribution on chromosomes, and human gene annotation files were downloaded from Ensembl. Genetic variation was accessible from the somatic mutation data (including Kirsten ratsarcoma viral oncogene homolog and V‐Raf murine sarcoma viral oncogene homolog B1 mutations) and the CNV, including the waterfall plot, panorama, and lollipop chart, respectively. After calculating Log2 (TPM+1) for 804 patients, the differential expression of ARGs between tissues was analysed by the Wilcoxon test. Based on the above analysis, we used a forest map to visualise the univariate Cox regression results and plotted survival curves for expression levels of 16 ARGs.

2.3. Non‐negative matrix factorisation algorithm and DEG‐related pathways

The non‐negative matrix factorisation (NMF) R package extracts the biological correlation coefficients of the data in the gene expression matrix based on the NMF algorithm, which captures the internal structural features of the data by organising the genes and samples so as to classify GC patients into C1 and C2 cohorts. The “limma” package sets the corresponding threshold (log = 0.585) to screen for significant anoikis‐related DEGs (ARDEGs) in GC. Through correlation and survival analyses, we plotted the heatmap and Kaplan–Meier (K‐M) curve. Employing the single sample gene set enrichment analysis scoring algorithm, we estimated 16 immune cell subpopulations and associated pathways in GC. Gene ontology and the Kyoto Encyclopaedia of Genes and Genomes (KEGG) could describe the functions and enrichment pathways of ARGs. The gene set variation analysis (GSVA) specified the sample genes in each pathway and complemented the KEGG enrichment results with a heatmap.

2.4. Weighted gene co‐expression network analysis (WGCNA) and least absolute shrinkage and selection operator algorithms

WGCNA was utilised to identify gene‐clustering, stemness, and GC prognosis. We then conducted univariate Cox regression analysis and constructed the least absolute shrinkage and selection operator (LASSO) regression model for ARDEGs. The risk scores = ∑(exp(mRNAs) *β), where exp denotes the expression of anoikis‐related prognostic genes (APGs) and β represents its coefficient. Based on the median APM‐based risk score (APM_score), patients were equally categorised into the training and test groups. We complemented the prognostic correlation analyses of the two groups and then applied receiver operating characteristic (ROC) curves to assess predictive accuracy. Sample stratification and clinical characteristics were visualised by heatmap.

2.5. Construction and validation of a prognostic nomogram

The nomogram was presented according to the prognostic factors and validated by employing the concordance index and calibration curves. The use of this nomogram predicted 1‐, 3‐, and 5‐year overall survival (OS) in GC patients. Then, we compared clinicopathological factors between the two APM_score subgroups and displayed the results in a clinical correlation heatmap, applying independent t tests to evaluate the relationship between the risk score and different factors. Furthermore, the results of representative protein expression and gene changes in cancer and normal tissues were verified in the Human Protein Atlas database and the cBioPortal online database, respectively. We collected six detectable variable genes to compare differences between normal and pathological tissues, including AKR1C2, cystatin 2 (CST2), POLE2, pituitary tumour‐transforming gene 1 (PTTG1), SLC27A2, and SNGG. Subsequent analyses of 10 ARDEGs were conducted for genetic alterations, including missense mutation, truncating mutation, amplification, and deep deletion.

2.6. Immune surveillance and the tumour microenvironment

According to the box plot, we revealed differential expression of immune checkpoints, including CD‐274 and CTLA4. The Microenvironment Cell Populations‐counter (MCP‐counter) method was used to quantify the absolute abundance of TIICs. The ESTIMATE is able to evaluate the immune and tumour components in each patient. As the score rose, it indicated relatively high cell purity. We plotted marginal scatter diagrams of risk scores against TME, monocytes, T regulatory cells, and cancer stemness based on Spearman correlation analysis and then analysed the relationship between risk scores and clustering results.

2.7. Single‐cell genomics and communication

We summarised the relationship between 22 TIICs (including B cells, CD4+/CD8+ T cells, monocytes, and non‐specific cells) and ARDEGs. Single‐cell RNA‐sequencing and spatial transcriptomics allow the study of transcriptional activity at the single‐cell or spatial level. By subgroup classification and annotation, we visualised Single‐cell RNA‐sequencing data from the GSE184198 microarray. For individual datasets, we compared the distribution of each cell type and the expression density of ARGs. The “SCENIC” R package was introduced to construct CellChat between different cell populations, which required the input of normalised expression data. The GSE184198 dataset belonged to the standard 10X genomics platform, which displayed the CellChatDB database structure after creating the communication object, and then the ARGs expression data was projected onto the cell marker heatmap, hierarchy, and signalling patterns. Single‐cell RNA‐sequencing data and bulk gene profiles in GC were exploited to reveal anoikis activity, paired ligand‐receptors, and network events during development.

2.8. Immunotherapy predictors and drug sensitivity analysis

We analysed the relationship between APM_score and microsatellite instability (MSI), gene clustering, and tumour mutation burden (TMB). In order to assess the applicability of the relevant indicators, the semi‐inhibitory concentration (IC50) values of immune compounds used to treat GC were further analysed. Furthermore, the Genomics of Drug Sensitivity in Cancer, a database that matches drug sensitivity with molecular markers of drug response, selected the human gastric cancer cell lines from the Cancer Cell Line Encyclopaedia database for small molecular drug repositioning. With the help of pharmacogenomics, we offered a relatively sensitive predictive drug outcome and response information heatmap in the anoikis mode.

2.9. Statistical analysis

All data calculations were executed in R software (v.4.0.3) and the accompanying R packages. A Wilcoxon test was applied to compare two independent samples, and a t‐test was used for validation. Pearson correlation analysis is sensitive to continuous, normal, and linear data, while Spearman correlation analysis has a wide range, such as clinicopathological characteristics, checkpoints, TME, and drugs. Cox regression, survival forests, and K‐M analysis were the primary algorithms and components of the prognostic analysis. All results were screened for a statistically two‐tailed p‐value <0.05.

3. RESULTS

3.1. Genetic variations and expression of anoikis‐related genes

The 27 ARGs exhibited an interacting co‐expression network, with anoikis mainly bearing risk factors (Figure 1a). The incidence of somatic mutations was 30.25% in 433 samples, and PIK3CA had the highest frequency of mutation at 15% (Figure 1b). Although ARGs were not significantly modified in GC, the high mutation frequency of TCGA data may be associated with epigenetic alterations in tumours (Figure 1c). The circos diagram contributed to identifying the location of ARGs on human chromosomes and discovering multiple change loci on chromosome 9 (Figure 1d). There were significant alterations in ARGs between somatic cells, cancer tissues, and paracancerous tissues, such as NOTCH1, CEACAM6, and ITGB1 (Figure 1e), and genetic variation and overexpression were likely to be factors affecting the OS of patients. Anoikis‐related genes also participated in a wide range of epigenetic mechanisms to mediate the prognosis, including PDK4, TLE1, and ITGA5 (Figure 1f). The results indicated that ARDEGs could predict the prognosis of patients (Additional File 2: Figure S2).

FIGURE 1.

FIGURE 1

Open in a new tab

Genetic and transcriptional alterations of anoikis‐related genes (ARGs) in Gastric cancer (GC). (a) Interactions among ARGs. The line connecting the genes represents their interaction, with the line thickness indicating the strength of the association between genes. Green and violet represent favourable and risk correlations, respectively. (b) Mutation frequencies of 27 ARGs in 433 patients from the TCGA cohort. (c) Panorama of genomic mutations in the TCGA cohort. (d) Location of CNVs in ARGs on 23 chromosomes. (e) copy number variation (CNV) gain, loss, and ARG expression were different in normal and tumour tissues. (f) Forest plot of univariable regression analysis results for the 25 selected genes. CNV, copy number variant; TCGA, the cancer genome atlas. The “*” represents the statistically significant p value < 0.05, ** = p < 0.01, and *** = p < 0.001.

3.2. Anoikis‐related multiclass cancer pathway

The 804 GC patients were categorised into C1 and C2 groups, thus analysing the clinical and immunological characteristics of the two groups (Figure 2a). Clinicopathological characteristics of the C1/C2 clusters, such as T and N staging, may be the result of ARG regulation in epigenetics (Figure 2b). Kaplan–Meier analysis showed a disparity in prognosis between patients in the C1 and C2 clusters (p < 0.001, Figure 2c). Further exploring the TIME between the two clusters, we found that most immune cells (B cells, T cells, and macrophages) were substantially infiltrated in the C1 subpopulation, which might be prognostic markers (Figure 2d). In addition, ARDEGs activated biosynthetic and immunometabolic pathways. On the one hand, they were involved in mitosis and matrix structural constituent binding (Figure 2e). On the other hand, functional nodes were mainly found in proliferation, receptor‐ligand interaction, and immune signalling (Figure 2f). DNA, RNA, and nutrient metabolism pathways were more active in C2 in GSVA, where biosynthesis and DNA modification were the major enrichment pathways for ARDEGs (Figure 2g). We obtained key modules and 363 candidate biomarker genes by WGCNA (Figure 2h, i), and the above co‐expression network and modular genes provided clues for prognostic analyses and multi‐mechanism research.

FIGURE 2.

FIGURE 2

Open in a new tab

Identification of anoikis‐related gene (ARG) clustering and enrichment analysis. (a) non‐negative matrix factorisation (NMF) clustering based on ARGs divides the samples in the merged Gastric cancer (GC) dataset. The co‐correlation coefficient corresponding to the k value at 2 is given. (b) Differences in clinicopathologic features and expression levels of ARGs between the two clusters. (c) Kaplan–Meier (K‐M) survival curves of overall survival (OS) in C1 and C2 clusters. (d) Barplot of single sample gene set enrichment analysis (ssGSEA) quantifying the tumour‐infiltrating cell proportions. (e) The Gene ontology (GO) analysis for anoikis‐related differentially expressed genes (DEGs). (f) The Kyoto Encyclopaedia of Genes and Genomes (KEGG) analysis for anoikis‐related DEGs (ARDEGs). (g) Biological pathway between two variant isoforms. The red colour represents the activated pathway, while the blue colour represents the negative pathway. (h) WGCNA identified gene modules with highly synergistic changes. (i) The heatmap of module‐trait relationships. WGCNA, weighted gene co‐expression network analysis.

3.3. Construction and validation of the GC‐APM

The 363 candidate genes were divided into gene‐cluster A and B after univariate Cox regression and clustering, and we assessed the expression differences of ARGs between gene‐clusters, the prognosis of GC, and the tumour stemness (Figure 3a–c). Overall, patients with gene‐cluster B had elevated the expression of ARGs, which led to the activation of the anoikis function, decreased tumour stemness, and a better prognosis. The LASSO algorithm was employed to identify the prognostic gene set of GC and finally determine 10 APGs (Figure 3d, e). Ten APGs and corresponding coefficients constituted the APM (Figure 3f). The TCGA and GEO cohorts presented different subgroups of gene clusters, APM_score subtypes, and survival states (Figure 3g).

FIGURE 3.

FIGURE 3

Open in a new tab

Construction and validation of anoikis‐related prognostic model (APM). (a) Expression of anoikis‐related genes (ARGs) in the high and low gene‐clusters. (b) Kaplan–Meier (K‐M) curves of overall survival (OS) in A and B gene‐clusters. (c) Association between APM_score and CSC index. (d) Lasso coefficient profiles. (e) The partial likelihood deviance graph. (f) The APM_score for predicting survival is based on the model coefficients. (g) Alluvial diagram of clusters in groups with different subtypes and survival status in the TCGA and GSE cohorts. (h) The K‐M curves show that the high APM_score group had inferior OS in the test set and training set. (i) ROC curves of 1, 3, and 5‐year OS for the test set and training set. (j) Differences in clinicopathologic features and expression levels of anoikis‐related DEGs (ARDEGs) between clusters and gene‐clusters. ROC, receiver operating characteristic; TCGA, the cancer genome atlas.

Training and test sets had similar trends in survival status and time distribution (Additional File 3: Figure S3a, b). Age, N staging, and APM_score were independent prognostic factors for GC (Additional File 3: Figure S3c–f). The survival curves suggested that the OS rate was higher in the low‐risk group than in the high‐risk group (Figure 3h). In general, the predictive model had the performance of detecting prognosis, and the test set had a more accurate 5‐year AUC‐ROC value (Figure 3i). Therefore, ARGs had predictive capabilities for stratified analysis and grouping (Figure 3j).

3.4. Clinicopathological features and nomogram

Combining prognostic factors and APM_score, we obtained a nomogram for clinical applicability (Figure 4a). This nomogram has higher sensitivity and specificity for prognosis (Figure 4b). The calibration curve compares the agreement between predicted and actual probabilities with good accuracy (Figure 4c). Also, we utilised stratification analysis to test the correlation of APM_score in TN staging. The results indicated that patients with GC maintained pronounced predictive capability in different subgroups (Figure 4d). According to the neoplasm histologic grade in patients with GC [12], grade 3 patients had an incremental APM_score compared to grade 2 (p = 0.013). Strong consistency was also observed for tumor node metastasis staging, for example, patients with stages II–IV both had higher risk scores than stage I (p = 0.0071, p = 0.039, and p = 0.00,035). With the progression of T staging from T1 to T2‐4 (p = 0.0062, p = 0.0019, p = 0.019) and N staging from N0 to N3 (p = 0.017), the APM_score demonstrated a clear upward trend, respectively (Figure 4e). Further analysing the expression of model genes through the human protein atlas database, structural disorders and irregular glandular morphology were predominantly formed in the tissues with high expression of ARGs, and aberrant expression increased the heterogeneity of tumour tissues (Figure 4f). The 10 variable genes were genetically altered with predominant amplification, which probably induced replication catastrophe and apoptosis (Figure 4g).

FIGURE 4.

FIGURE 4

Open in a new tab

Construction and validation of a nomogram. (a) Nomogram for both clinicopathological factors and APM_score. (b) The ROC curves for 1‐, 3‐, and 5‐year nomograms. (c) Calibration curve of the nomogram for predicting 1‐, 3‐, and 5‐year overall survival (OS). (d) Stratified analysis for risk and clinical factors. (e) Association of risk scores with clinicopathological characteristics of patients, p. values for stratified analyses based on the Wilcoxon test. (f) Representatively expressed proteins of six genes in Gastric cancer (GC) and normal tissues from the Human Protein Atlas (http://www.proteinatlas.org). (g) Altered expression profiles of 10 genes in the RNA‐seq dataset of TCGA GC. ROC, receiver operating characteristic; TCGA, the cancer genome atlas.

3.5. Immune microenvironment and correlation in Gastric cancer

We observed a dramatic association between APGs and 18 immune checkpoints, with a positive trend in the expression of 13 checkpoints, which reflected that patients with a high APM_score had more immune loci and protein expression (Figure 5a). Simultaneous co‐expression analysis revealed that the risk score was positively correlated with the aggregation of TIICs, especially with B lineage, myeloid dendritic cells, endothelial cells, and fibroblasts, which would predict an enriched immune microenvironment (Figure 5b). Figures 5c, d displayed that the immune and tumour components of the samples were relatively elevated as the APM_score rose. It was reasonable to believe that the risk scores between the C1 and C2 clusters reflected variations in TIICs, with clustering somewhat revealing the link between immunisation and the risk score (Figure 5e). Based on Spearman correlation coefficients, the APM_score was positively linked to monocytes (R = 0.2, p < 0.001) and regulatory T cells (R = 0.17, p < 0.001), while tumour stem cells (R = −0.414, p < 0.001) were reversed with the response to anoikis in TME, which might be the mechanism of stemness resistance to anoikis in GC (Figure 5f). Therefore, anoikis patterns were probably favourable prognostic and therapeutic mediators in the GC immune microenvironment.

FIGURE 5.

FIGURE 5

Open in a new tab

Immunity analysis for risk score correlation. (a) Expression of immune checkpoints in the high and low APM_score groups. (b) Correlation plots between risk score and 10 immune cell types. The colour and depth stand for correlation and degree, respectively. (c) Correlations between APM_score and both immune and stromal scores. (d) tumour microenvironment (TME) scores in the anoikis‐related prognostic model (APM) of stomach adenocarcinoma,STAD. (e) Differential risk scoring analysis of clusters. (f) Evaluation of three immune cell infiltration levels in Gastric cancer (GC) by the CIBERSORT algorithm. STAD, stomach adenocarcinoma.

3.6. Single‐cell transcript landscape

The seven cell types were sequenced from patients with primary GC, and anoikis held great relevance not only in malignant cells but also in many kinds of immune cells. Anoikis‐related genes were mainly expressed on T cells, B cells, NK cells, neutrophils, and monocytes, especially in the number and intensity of T cell subpopulations. The heatmap depicted the extent to which six T cell subtypes, two B cell subtypes, and 12 non‐specific immune cells were meaningfully associated with ARGs (Figure 6a). The uniform manifold approximation and projection (UMAP) algorithm visualised the distribution and number of cells and showed the regions of expression of ARGs in T cells, B cells, neutrophils, NK cells, and monocytes (Figure 6b). 26 ARGs were depicted by high‐resolution UMAP analysis of gene expression in Supplementary Figure S4, where MCL1 stood for the anoikis signature preferentially expressed by T cells. On the one hand, CellChat presented tumour interactions with peripheral blood vessels, immune cells, fibroblasts, signalling molecules, and ECM through circle and scatter plots (Figure 6c, d). The results revealed the strength and number of interactions among cancer cells, stromal cells, and immune cells (Additional File 5: Figure S5). On the other hand, the cell marker heatmap included the maximum Z‐score values for tissue stem cells, the measures for T cells, and ARDEGs (e.g., MCL1, AKT1, ITGA5, and ITGB1) possessing biomarker potential in different immune, stromal, and functional cells (Figure 6e). Our collection of ligand receptor information uncovered that L‐R pairs of dendritic cells and epithelial cell subpopulations displayed strong immunomodulatory properties, which could impede this cellular phenotypic switch (Figure 6f). Figure 6g complements the outgoing and incoming signalling patterns to provide potential therapeutic targets for GC.

FIGURE 6.

FIGURE 6

Open in a new tab

Single‐cell sequencing of GSE184198 in Gastric cancer (GC). (a) Correlations between the abundance of immune cells and the 10 genes in this model. (b) UMAP projection showing the landscape of immune cells and the cellular distribution of anoikis‐related target function scores. (c) Overview of CellChat in the GC tumour microenvironment (TME). (d) Main sending sources and receiving targets for visualisation in two dimensions. (e) Heatmap of the canonical and curated marker genes for major cell lineages. (f) Bubble plot showing the selected ligand‐receptor interactions between two cell types. (g) Heatmap to visualise the probability of outgoing or incoming communication between two cell types. UMAP, uniform manifold approximation and projection.

3.7. Immunoscreening and small‐molecule drugs

The MSI analysis sensitises the effectiveness of immunotherapeutic drugs and screening. High‐risk scores were significantly associated with the microsatellite stability of GC compared to the MSI (Figure 7a). The TMB is a way for somatic cells to increase antigenic species through mutations, whereas the anoikis‐resistant phenotype diminishes mutations to elicit immune responses (Figure 7b). As the risk related to anoikis increased, this added antigenicity gradually weakened ( R = −0.2, p = 0.00,012, Figure 7c). In high‐ and low‐risk groups, the TMB advantage group benefited GC patients against cancer progression (Figure 7d). This finding suggested that patients in the low‐risk group were sensitive to immunotherapy and also had the potential to screen. Additionally, we believed that the risk was linked to the IC50 value of common chemotherapeutics (Figure 7e). The summary heatmap of drug‐targeted metabolism and signalling pathways contributed to understanding the linkage of small‐molecule compounds, genes, and pathway sensitivities, resulting in network proximity estimates of drug‐targeted effects as measured by Z‐score. The small‐molecule agents included docetaxel as a microtubule stabiliser (Z = −1.15,254), elesclomol synergistically with HSP90 (Z = −2.2138), and the antimetabolite pathways related to gemcitabine (Z = −1.36,671) and methotrexate ( Z = 0.678,627, Figure 7c). The above combined analysis offered direction for clinical application and further validated the potential mechanisms and epigenetics.

FIGURE 7.

FIGURE 7

Open in a new tab

Comprehensive analysis of immunity and treatment. (a) Relationship between APM_score and microsatellite instability (MSI). (b) Spearman correlation analysis of the APM_score and tumour mutation burden (TMB). (c) Survival probabilities between the anoikis‐related prognostic gene (APG) subgroup and both high and low TMB groups. (d) Relationship between the APM_score and chemotherapy sensitivity. (e) Z scores of the drugs with the signalling pathway enrichment in the summary heatmap.

4. DISCUSSION

Anoikis, apoptosis following cell detachment from the correct ECM, is critical for colonisation during malignant metastasis [13]. When anchorage‐dependent growth and EMT occurred, dislodged cancer cells grew via anti‐anoikis adhering to an inappropriate stroma or epithelial cells settled elsewhere. Zhu et al. investigated TGF‐β‐driven mesenchymal transition from the expression of the enzyme PDK4, but the mechanism by which PDK4 further induces anoikis remains unknown [14]. Several studies have proven that cancer cells utilise multiple mechanisms to stay alive, which depend on extrinsic immune escape and intrinsic tumour metabolic pathways. For example, Snai2 promotes the epigenetic regulation of transcriptional programmes in tumour metastasis, and the inhibition of ITGA5 CpG island promoter methylation attenuates cancer cell‐matrix deadhesion [15, 16]. Anoikis‐related genes in epigenetics led to favourable environments for cancer cells, including hypoxia, oxidative stress, tumour cell dormancy, and EMT. Thus, anti‐anoikis mechanisms in disease might represent biomarkers and promote tumour invasion and migration, developing heterogeneity and drug resistance. However, few studies have explored single‐cell and clinicopathological features in conjunction with anoikis‐related epigenetic mechanisms in GC.

This study was the first to construct a prognostic model of GC associated with ARGs based on the NMF algorithm and to perform a comprehensive analysis of its underlying anoikis pattern. After NMF downscaling, WGCNA, and LASSO regression analysis for patient categorisation, we obtained preliminary clusters, gene‐clusters of DEGs, and the two APM_score subgroups. In the groups with a better prognosis, patients had a universal overexpression of ARGs and DEGs, mainly in intracellular pathways corresponding to cycle, structural components, drug metabolism, and trigger signals, as well as extrinsic pathways including ECM, cGMP‐PKG signalling, receptor‐ligand interaction, and immune factor binding. In contrast to conventional algorithms and single‐cell analyses, we further discussed the significance of risk scores in terms of clinicopathological characterisation, cellular communication, and the TIME [17].

Differing from previous markers, we extracted regulatory genes in epigenetics from ARDEGs. ITGB1 is involved in histone modifications leading to defects in enterocyte maturation and differentiation [18]. TLE1 and CAV1 serve as prognostic methylation biomarkers that regulate CpG island methylation‐associated inactivation and progression in lymphomas and pancreatic adenocarcinomas, respectively [19, 20]. Model genes focused on genes that were overexpressed in solid tumours. The PTTG1 was associated with lymph node metastasis in GC and regulated the cancer cell cycle to affect prognosis [21]. Dual specificity phosphatase 1 was the drug resistance gene that exerted significant effects on both metabolic signalling and apoptosis induction [22]. In general, ARGs played a pivotal role in the epigenetic regulation of GC, and MAGE‐A3 promoter methylation expression contributed to cancer cell proliferation and drug sensitivity [23]. Also, the enrichment pathways of ARDEGs provided a direction for understanding tumour aggressiveness and prognosis, which illustrated the innovative nature of the study. Lysyl oxidase‐like 4 served as a factor for independent prognosis and targeted drug classification in GC patients, and its overexpression promoted the malignant behaviour of cancer cells via the FAK/Src pathway [24]. The similar prognostic gene CST2 controlled the typical EMT and TGF‐β1 pathways, thus enhancing the preceding behaviours [25]. Among the various model genes, GAD1 emerged as a minority gastroprotective factor, which increased expression in modified Suanzaoren decoction treatment to inhibit cancer cell metabolism [26]. In contrast to normal gastric mucosal epithelial cells, SNCG was overexpressed in GC, and SNGG siRNA‐transfected cells downregulated the phosphorylation of Protein Kinase B (AKT) and ERK [27]. One study showed that the nullification of AKR1C2 reduced pre‐tumour transformation and played an essential function in tumour escape signalling [28]. The subsequent microenvironment section clarified that the anoikis phenotype and other early signals still enhanced cellular metabolism and cross‐talked immune surveillance.

Anoikis‐related genes have pro‐carcinogenic potential in GC patients with high fibroblast infiltration [29]. We found that the expression levels of ARGs were highly correlated with immune components in TIME, such as T cells, NK cells, macrophages, dendritic cells, and B cells, and several key ARDEGs were particularly prominent in T cell infiltration. Extracellular matrix proteins mediate cell adhesion and density through the T cell protein tyrosine phosphatase and induce anoikis in isolated metastatic cancer cells to exert IFN‐γ [30]. This study related TIICs to the characterisation of anoikis, and T cells might be key biomarkers. The transient expression of genes in T cells could evoke apoptosis, which was particularly vital for the survival of circulating cancer cells during metastasis [31]. Besides, our interaction network described a portion of relevant mechanisms. For example, galectin acts as a counterreceptor to regulate tumour‐effector T cell growth via the ganglioside GM1, interferon (IFN) and STAT1 dephosphorylation systems work together to overcome immune escape, colony stimulating factor (CSF) receptor ligand co‐expression correlates with GC progression, and interleukins favour tumour metastasis [32, 33]. These insights further explained the impact of ARGs on the single‐cell landscape in GC.

Evidence now exists that most ARGs antagonise escape signals in the immune surveillance system, and anoikis resistance allows GC cells to remain with metastatic potential despite non‐specific immunity [34]. Although the anoikis‐associated cell network included complex NK cell communication, recent studies have shown that NK cells involved in immune surveillance could not completely respond to anoikis‐resistant mechanically stressed cancer cells and were prone to form metastatic lesions under enhanced cell motility [35]. On the basis of endothelial cells inducing EMT in epithelial cells and conferring tissues with carcinoma stemness phenotype, vascular endothelial cells promoted tumour resistance to anoikis through a series of extracellular secretory factors that confer dissemination and stemness to cancer cells [36]. Other investigations discovered that the overexpressed anoikis gene CEACAM6 was not only oncogenic, but its monoclonal antibodies and fragments acted as immune checkpoint inhibitors in haematological malignancies [37]. And the anti‐PD‐L1 antibody had anoikis sensitising activity to suppress EMT in human lung cancer cells [38]. Due to the anti‐apoptotic phenomenon and significant immune gene expression differences, the mechanism of cancer metastasis away from the anchorage‐dependent cell cycle was probably linked to the escape of TIICs and immune checkpoints. We proposed the novel idea that anoikis resistance may be compatible with T cell activation, endothelial cell transformation, and depletion of epithelial cells.

Finally, we proposed the clinical significance of APM in small‐molecule compounds. As a novel lead quinazoline‐based compound, DZ‐50 exerted its anoikis to impair tumour growth and metastasis by targeting intercellular functional structures, linkage regions, binding proteins, and adhesion pathways [39]. Future research would aim to improve patient prognosis with anti‐metastatic inhibitors targeting ARGs. Docetaxel has phenotypic changes in prostate cancer, and docetaxel‐resistance variants were established and characterised in terms of anoikis and functional relevance [40]. Elesclomol is an effective inducer of oxidative stress that promotes the immunoscreening interface, stem‐like features, and anoikis‐related processes [41]. Apart from this, it was identified that there is an intimately interrelated role between the determinants of gemcitabine efficacy and carcinoembryonic antigen (CEA)‐related cell adhesion molecule 6 in pancreatic cancer [42]. As well as specific inhibitors, for example, AKT inhibitors, methyl ethyl ketone small‐molecule inhibitors, farnesyltransferase, and Rho kinase inhibitors were likely to have additive effects and anoikis‐related events. Regarding the design of drugs directly on CEA expression levels, there was still no methotrexate significantly sensitive against anoikis cells [43]. Therefore, the accumulated small compounds point to the development of ARG in immunisation and precision medicine.

However, several limitations of our study remain. First, this is a preliminary study based on bioinformatics tools. Incorporating our own clinical patients to supplement normal samples and validation sets would allow design rigour. Second, the anoikis process is not a specific mechanism of GC, and the ARDEGs that mediate the TIME are still stuck in our correlation and algorithms. Third, owing to technical limitations, we were lacking animal models and molecular experiments. Finally, further validation would be conducted between ARGs and small‐molecule compounds.

5. CONCLUSION

In GC, we generated an APM based on ARGs for predicting prognosis and assessing the TME. Anoikis‐related DEGs mediate biological signalling pathways, regulate epigenetics, and cross‐talk immune cell communication. These evidences provide new insights into precise treatment strategies for GC patients guided by the anoikis pattern.

AUTHOR CONTRIBUTIONS

Conceptualisation: Yongjian Lin. Data curation: Yongjian Lin. Formal analysis: Yongjian Lin. Funding acquisition: Jinlu Liu. Investigation: Jinlu Liu. Methodology: Yongjian Lin. Project administration: Jinlu Liu. Resources: Jinlu Liu. Software: Yongjian Lin. Supervision: Jinlu Liu. Validation: Yongjian Lin. Visualisation: Jinlu Liu. Writing – original draft: Yongjian Lin. Writing – review & editing: Yongjian Lin.

CONFLICT OF INTEREST STATEMENT

All authors have no conflicts of interest to declare.

Supporting information

Supplementary Information S1

SYB2-18-41-s001.docx (6.7MB, docx)

ACKNOWLEDGEMENT

The authors thank the First Affiliated Hospital of Guangxi Medical University for the assistance offered with data collection. The authors did not receive specific funding.

Lin, Y. , Liu, J. : Anoikis‐related genes as potential prognostic biomarkers in gastric cancer: a multilevel integrative analysis and predictive therapeutic value. IET Syst. Biol. 18(2), 41–54 (2024). 10.1049/syb2.12088

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available in https://www.ncbi.nlm.nih.gov/gds/, reference number: GSE184198. These data were derived from the following resources available in the public domain: the Cancer Genome Atlas at https://www.cancer.gov/ccg/research/genome‐sequencing/tcga, UCSC Xena at https://xenabrowser.net/datapages/, cBioPortal for Cancer Genomics at https://www.cbioportal.org/, Genomics of Drug Sensitivity in Cancer at https://www.cancerrxgene.org/, and the Human Protein Atlas at https://www.proteinatlas.org/. Data on the microarrays and microarray matrix used in this study are available from the corresponding author.

REFERENCES

  • 1. Liu, W. , et al.: Microenvironmental influences on metastasis suppressor expression and function during a metastatic cell's journey. Cancer Microenviron. 7(3), 117–1131 (2014). 10.1007/s12307-014-0148-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2. Lee, H.Y. , et al.: The role of noncoding RNAs in the regulation of anoikis and anchorage‐independent growth in cancer. Int. J. Mol. Sci. 22(2), 627 (2021). 10.3390/ijms22020627 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3. Gonzalez‐Llorente, L. , et al.: Overexpression of mitochondrial IF1 prevents metastatic disease of colorectal cancer by enhancing anoikis and tumor infiltration of NK cells. Cancers 12(1), 22 (2020). 10.3390/cancers12010022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. Du, S. , et al.: Anoikis resistant gastric cancer cells promote angiogenesis and peritoneal metastasis through C/EBP beta‐mediated PDGFB autocrine and paracrine signaling. Oncogene 40(38), 5764–5779 (2021). 10.1038/s41388-021-01988-y [DOI] [PubMed] [Google Scholar]
  • 5. Zhang, T. , et al.: TCF7L2 promotes anoikis resistance and metastasis of gastric cancer by transcriptionally activating PLAUR. Int. J. Biol. Sci. 18(11), 4560–4577 (2022). 10.7150/ijbs.69933 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6. Majidpoor, J. , Mortezaee, K. : Steps in metastasis: an updated review. Med. Oncol. 38(1), 3 (2021). 10.1007/s12032-020-01447-w [DOI] [PubMed] [Google Scholar]
  • 7. Jhunjhunwala, S. , Hammer, C. , Delamarre, L. : Antigen presentation in cancer: insights into tumour immunogenicity and immune evasion. Nat. Rev. Cancer 21(5), 298–312 (2021). 10.1038/s41568-021-00339-z [DOI] [PubMed] [Google Scholar]
  • 8. Okugawa, Y. , et al.: Smad interacting protein 1 (SIP1) is associated with peritoneal carcinomatosis in intestinal type gastric cancer. Clin. Exp. Metastasis 30(4), 417–429 (2013). 10.1007/s10585-012-9547-4 [DOI] [PubMed] [Google Scholar]
  • 9. Sousa Onyeisi, J.O. , et al.: Heparan sulfate proteoglycans as trastuzumab targets in anoikis‐resistant endothelial cells. J. Cell. Biochem. 120(8), 13826–13840 (2019). 10.1002/jcb.28656 [DOI] [PubMed] [Google Scholar]
  • 10. Nishimura, S. , et al.: Adenovirus‐mediated transfection of caspase‐8 augments anoikis and inhibits peritoneal dissemination of human gastric carcinoma cells. Cancer Res. 61(19), 7009–7014 (2001) [PubMed] [Google Scholar]
  • 11. Cheng, L. , et al.: The expression and significance of efferocytosis and immune checkpoint related molecules in pancancer samples and the correlation of their expression with anticancer drug sensitivity. Front. Pharmacol. 13 (2022). 10.3389/fphar.2022.977025 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Ajani, J.A. , et al.: Gastric cancer, version 2.2022. J. Natl. Compr. Cancer Netw. 20(2), 167–192 (2022). 10.6004/jnccn.2022.0008 [DOI] [PubMed] [Google Scholar]
  • 13. Paoli, P. , Giannoni, E. , Chiarugi, P. : Anoikis molecular pathways and its role in cancer progression. Biochim. Biophys. Acta Mol. Cell Res. 1833(12), 3481–3498 (2013). 10.1016/j.bbamcr.2013.06.026 [DOI] [PubMed] [Google Scholar]
  • 14. Zhu, X. , et al.: Acetate controls endothelial‐to‐mesenchymal transition. Cell Metabol. 35(7), 1163–1178. e10 (2023). 10.1016/j.cmet.2023.05.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Zhou, W. , Gross, K.M. , Kuperwasser, C. : Molecular regulation of Snai2 in development and disease. J. Cell Sci. 132(23) (2019). 10.1242/jcs.235127 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Behera, J. , Govindan, S. , Ramasamy, M.S. : Nitric oxide promotes cell‐matrix adhesion of endothelial progenitor cells under hypoxia condition via ITGA5 CpG promoter demethylation. Biochem. Biophys. Res. Commun. 644, 162–170 (2023). 10.1016/j.bbrc.2023.01.008 [DOI] [PubMed] [Google Scholar]
  • 17. Cao, J. , et al.: Development of anoikis‐related genes signature to predict the prognosis in gastric cancer patients. Front. Oncol. 12 (2023). 10.3389/fonc.2022.1096608 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18. Mohan, M. , et al.: Focused examination of the intestinal epithelium reveals transcriptional signatures consistent with disturbances in enterocyte maturation and differentiation during the course of SIV infection. PLoS One 8(4), e60122 (2013). 10.1371/journal.pone.0060122 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19. Fraga, M.F. , et al.: Epigenetic inactivation of the Groucho homologue gene TLE1 in hematologic malignancies. Cancer Res. 68(11), 4116–4122 (2008). 10.1158/0008-5472.can-08-0085 [DOI] [PubMed] [Google Scholar]
  • 20. Guo, Y.‐l. , et al.: Aberrant CpG island shore region methylation of CAV1 is associated with tumor progression and poor prognosis in gastric cardia adenocarcinoma. Arch. Med. Res. 47(6), 460–470 (2016). 10.1016/j.arcmed.2016.10.005 [DOI] [PubMed] [Google Scholar]
  • 21. Yao, N.H. , et al.: Expression of PTTG1 in gastric adenocarcinoma tissues and its relationship with prognosis. Zhonghua Zhong Liu Za Zhi 43(3), 306–311 (2021) [DOI] [PubMed] [Google Scholar]
  • 22. Teng, F. , et al.: DUSP1 induces apatinib resistance by activating the MAPK pathway in gastric cancer. Oncol. Rep. 40(3), 1203–1222 (2018). 10.3892/or.2018.6520 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23. Xie, C. , et al.: Melanoma associated antigen (MAGE)‐A3 promotes cell proliferation and chemotherapeutic drug resistance in gastric cancer. Cell. Oncol. 39(2), 175–186 (2016). 10.1007/s13402-015-0261-5 [DOI] [PubMed] [Google Scholar]
  • 24. Li, R.‐k. , et al.: Lysyl oxidase‐like 4 (LOXL4) promotes proliferation and metastasis of gastric cancer via FAK/Src pathway. J. Cancer Res. Clin. Oncol. 141(2), 269–281 (2015). 10.1007/s00432-014-1823-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25. Zhang, W.P. , et al.: Cystatin 2 leads to a worse prognosis in patients with gastric cancer. J. Biol. Regul. Homeost. Agents 34(6), 2059–2067 (2020) [DOI] [PubMed] [Google Scholar]
  • 26. Zhan, L.‐H. , et al.: Soporific effect of modified suanzaoren decoction and its effects on the expression of CCK‐8 and Orexin‐A. Evid. base Compl. Alternative Med., 2020 (2020) [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27. Fan, C. , et al.: siRNA targeting of the SNCG gene inhibits the growth of gastric carcinoma SGC7901 cells in vitro and in vivo by downregulating the phosphorylation of AKT/ERK. Cytogenet. Genome Res. 154(4), 209–216 (2018). 10.1159/000488571 [DOI] [PubMed] [Google Scholar]
  • 28. Martin, N. , et al.: Identification of a gene signature of a pre‐transformation process by senescence evasion in normal human epidermal keratinocytes. Mol. Cancer 13(1), 151 (2014). 10.1186/1476-4598-13-151 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Yang, X. , et al.: Comprehensive analysis of anoikis‐related genes in prognosis and immune infiltration of gastric cancer based on bulk and single‐cell RNA sequencing data. J. Cancer Res. Clin. Oncol. 149(14), 13163–13173 (2023). 10.1007/s00432-023-05157-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. Chen, Z. , et al.: Negative regulation of interferon‐gamma/STAT1 signaling through cell adhesion and cell density‐dependent STAT1 dephosphorylation. Cell. Signal. 23(8), 1404–1412 (2011). 10.1016/j.cellsig.2011.04.003 [DOI] [PubMed] [Google Scholar]
  • 31. Furui, T. , et al.: Overexpression of edg‐2/vzg‐1 induces apoptosis and anoikis in ovarian cancer cells in a lysophosphatidic acid‐independent manner. Clin. Cancer Res. 5(12), 4308–4318 (1999) [PubMed] [Google Scholar]
  • 32. Ledeen, R.W. , et al.: Beyond glycoproteins as galectin counterreceptors: tumor‐effector T cell growth control via ganglioside GM1. In: Glycobiology of the Immune Response, pp. 206–221 (2012) [DOI] [PubMed] [Google Scholar]
  • 33. Tsai, Y.‐F. , et al.: Interleukin 17A promotes cell migration, enhances anoikis resistance, and creates a microenvironment suitable for triple negative breast cancer tumor metastasis. Cancer Immunol. Immunother. 70(8), 2339–2351 (2021). 10.1007/s00262-021-02867-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34. Teijeira, A. , et al.: Metabolic consequences of T‐cell costimulation in anticancer immunity. Cancer Immunol. Res. 7(10), 1564–1569 (2019). 10.1158/2326-6066.cir-19-0115 [DOI] [PubMed] [Google Scholar]
  • 35. Fanfone, D. , et al.: Confined migration promotes cancer metastasis through resistance to anoikis and increased invasiveness. Elife 11, 11 (2022). 10.7554/elife.73150 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36. Zhang, Z. , et al.: Endothelial cell‐secreted EGF induces epithelial to mesenchymal transition and Endows head and neck cancer cells with stem‐like phenotype. Cancer Res. 74(10), 2869–2881 (2014). 10.1158/0008-5472.can-13-2032 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37. Johnson, B. , Mahadevan, D. : Emerging role and targeting of carcinoembryonic antigen‐related cell adhesion molecule 6 (CEACAM6) in human malignancies. Clin. Cancer Drugs 2(2), 100–111 (2015). 10.2174/2212697x02666150602215823 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38. Phetphoung, T. , et al.: Expression of plant‐produced anti‐PD‐L1 antibody with anoikis sensitizing activity in human lung cancer cells via., suppression on epithelial‐mesenchymal transition. PLoS One 17(11), e0274737 (2022). 10.1371/journal.pone.0274737 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39. Hensley, P.J. , et al.: Novel pharmacologic targeting of tight junctions and focal adhesions in prostate cancer cells. PLoS One 9(1), e86238 (2014). 10.1371/journal.pone.0086238 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40. Corcoran, C. , et al.: Docetaxel‐resistance in prostate cancer: evaluating associated phenotypic changes and potential for resistance transfer via exosomes. PLoS One 7(12), e50999 (2012). 10.1371/journal.pone.0050999 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41. Buccarelli, M. , et al.: Elesclomol‐induced increase of mitochondrial reactive oxygen species impairs glioblastoma stem‐like cell survival and tumor growth. J. Exp. Clin. Cancer Res. 40(1), 228 (2021). 10.1186/s13046-021-02031-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42. Duxbury, M.S. , et al.: A novel role for carcinoembryonic antigen‐related cell adhesion molecule 6 as a determinant of gemcitabine chemoresistance in pancreatic adenocarcinoma cells. Cancer Res. 64(11), 3987–3993 (2004). 10.1158/0008-5472.can-04-0424 [DOI] [PubMed] [Google Scholar]
  • 43. Eftekhar, E. , Naghibalhossaini, F. : Carcinoembryonic antigen expression level as a predictive factor for response to 5‐fluorouracil in colorectal cancer. Mol. Biol. Rep. 41(1), 459–466 (2014). 10.1007/s11033-013-2880-0 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information S1

SYB2-18-41-s001.docx (6.7MB, docx)

Data Availability Statement

The data that support the findings of this study are available in https://www.ncbi.nlm.nih.gov/gds/, reference number: GSE184198. These data were derived from the following resources available in the public domain: the Cancer Genome Atlas at https://www.cancer.gov/ccg/research/genome‐sequencing/tcga, UCSC Xena at https://xenabrowser.net/datapages/, cBioPortal for Cancer Genomics at https://www.cbioportal.org/, Genomics of Drug Sensitivity in Cancer at https://www.cancerrxgene.org/, and the Human Protein Atlas at https://www.proteinatlas.org/. Data on the microarrays and microarray matrix used in this study are available from the corresponding author.

Articles from IET Systems Biology are provided here courtesy of Wiley

RESOURCES