Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2022 Nov 21:13:1056932.
doi: 10.3389/fimmu.2022.1056932. eCollection 2022.

Identification of cuproptosis-related subtypes, construction of a prognosis model, and tumor microenvironment landscape in gastric cancer

Jinyan Wang  1   2 , Dongmei Qin  3 , Zhonghua Tao  1   2 , Biyun Wang  1   2 , Yizhao Xie  1   2 , Ye Wang  1   2 , Bin Li  1   2 , Jianing Cao  1   2 , Xiaosu Qiao  1   2 , Shanliang Zhong  4 , Xichun Hu  1   2
Affiliations

Identification of cuproptosis-related subtypes, construction of a prognosis model, and tumor microenvironment landscape in gastric cancer

Jinyan Wang et al. Front Immunol. .

Abstract

Introduction: Cuproptosis is a novel identified regulated cell death (RCD), which is correlated with the development, treatment response and prognosis of cancer. However, the potential role of cuproptosis-related genes (CRGs) in the tumor microenvironment (TME) of gastric cancer (GC) remains unknown.

Methods: Transcriptome profiling, somatic mutation, somatic copy number alteration and clinical data of GC samples were downloaded from the Cancer Genome Atlas (TCGA) and the Gene Expression Omnibus (GEO) database to describe the alterations of CRGs from genetic and transcriptional fields. Differential, survival and univariate cox regression analyses of CRGs were carried out to investigate the role of CRGs in GC. Cuproptosis molecular subtypes were identified by using consensus unsupervised clustering analysis based on the expression profiles of CRGs, and further analyzed by GO and KEGG gene set variation analyses (GSVA). Genes in distinct molecular subtypes were also analyzed by GO and KEGG gene enrichment analyses (GSEA). Differentially expressed genes (DEGs) were screened out from distinct molecular subtypes and further analyzed by GO enrichment analysis and univariate cox regression analysis. Consensus clustering analysis of prognostic DEGs was performed to identify genomic subtypes. Next, patients were randomly categorized into the training and testing group at a ratio of 1:1. CRG Risk scoring system was constructed through logistic least absolute shrinkage and selection operator (LASSO) cox regression analysis, univariate and multivariate cox analyses in the training group and validated in the testing and combined groups. Real-time quantitative polymerase chain reaction (RT-qPCR) was used to evaluate the expression of key Risk scoring genes. Sensitivity and specificity of Risk scoring system were examined by using receiver operating characteristic (ROC) curves. pRRophetic package in R was used to investigate the therapeutic effects of drugs in high- and low- risk score group. Finally, the nomogram scoring system was developed to predict patients' survival through incorporating the clinicopathological features and CRG Risk score.

Results: Most CRGs were up-regulated in tumor tissues and showed a relatively high mutation frequency. Survival and univariate cox regression analysis revealed that LIAS and FDX1 were significantly associated with GC patients' survival. After consensus unsupervised clustering analysis, GC patients were classified into two cuproptosis molecular subtypes, which were significantly associated with clinical features (gender, age, grade and TNM stage), prognosis, metabolic related pathways and immune cell infiltration in TME of GC. GO enrichment analyses of 84 DEGs, obtained from distinct molecular subtypes, revealed that DEGs primarily enriched in the regulation of metabolism and intracellular/extracellular structure in GC. Univariate cox regression analysis of 84 DEGs further screened out 32 prognostic DEGs. According to the expression profiles of 32 prognostic DEGs, patients were re-classified into two gene subtypes, which were significantly associated with patients' age, grade, T and N stage, and survival of patients. Nest, the Risk score system was constructed with moderate sensitivity and specificity. A high CRG Risk score, characterized by decreased microsatellite instability-high (MSI-H), tumor mutation burden (TMB) and cancer stem cell (CSC) index, and high stromal and immune score in TME, indicated poor survival. Four of five key Risk scoring genes expression were dysregulated in tumor compared with normal samples. Moreover, CRG Risk score was greatly related with sensitivity of multiple drugs. Finally, we established a highly accurate nomogram for promoting the clinical applicability of the CRG Risk scoring system.

Discussion: Our comprehensive analysis of CRGs in GC demonstrated their potential roles in TME, clinicopathological features, and prognosis. These findings may improve our understanding of CRGs in GC and provide new perceptions for doctors to predict prognosis and develop more effective and personalized therapy strategies.

Keywords: cuproptosis-related genes (CRGs); drugs susceptibility; gastric cancer (GC); prognosis model; tumor microenvironment (TME).

PubMed Disclaimer

Conflict of interest statement

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

Figures

Figure 1
Figure 1
Genetic and transcriptional alterations of CRGs in GC. (A) The expression levels of 19 CRGs between 375 GC samples and 32 normal samples. Wilcoxon test was used to compare two groups. (B) The maftool exhibited incidence of somatic mutations of CRGs in 431 GC patients from TCGA database. (C) The CNV frequency of CRGs in 440 GC samples from TCGA database. (D) Locations of CNV alterations on 23 chromosomes. P < 0.05 was considered as significant importance. * indicated P < 0.05, ** indicated P < 0.01, *** indicated P < 0.001.
Figure 2
Figure 2
The survival analyses of CRGs and a comprehensive landscape of cuproptosis network in GC. (A–I) The survival analyses of CRGs (ATP7A, DLAT, DLD, FDX1, LIAS, LIPT1, MTF1, NLRP3 and SLC31A1) in 732 GC patients. Kaplan–Meier plot and log-rank tests were conducted for survival analyses. (J) Mutual correlations among CRGs in 732 GC samples. Spearman correlation analyses were used. The line between two CRGs indicated their interaction, and the stronger the correlation, the thicker the line. Pink line represented positive correlation and blue line represented negative correlation. P < 0.05 was considered to be statistically significant.
Figure 3
Figure 3
CRG molecular subtypes and their clinicopathological features. (A) Identification of two molecular subtypes (k = 2) and their correlation area through consensus clustering analysis in 732 GC samples. (B) PCA presented a great difference in transcriptomes between different molecular subtypes. (C) Survival analysis showed a significant difference of survival between molecular subtype A and (B) Kaplan–Meier plot and log-rank tests were conducted for survival analyses. (D) The heat-map showed the CRGs expression profile in molecular subtype A and B, and the associations between clinicopathologic characteristics and different molecular subtypes. Chi-square test was used for the comparison. Red color indicated up-regulated expression level and blue color indicated down-regulated expression level. P < 0.05 was considered to be statistically significant. Molecular subtype A contained 339 GC samples and molecular subtype B contained 393 GC samples. * indicated P < 0.05, ** indicated P < 0.01.
Figure 4
Figure 4
Correlations between TME and CRG molecular subtypes. (A) GO GSVA enrichment analyses between molecular subtype A and (B) Red color indicated more enriched in pathways and blue color indicated less enriched in pathways. Adjusted p value <0.05 was considered to be statistically significant. (B) KEGG GSVA enrichment analyses between molecular subtype A and (B) Red color indicated more enriched in pathways and blue color indicated less enriched in pathways. Adjusted p value <0.05 was considered to be statistically significant. (C) GO GSEA enrichment analyses of genes between molecular subtype A and (B) NES>1, nominal p value<0.05, FDR<0.25 were considered to be statistically significant. (D) KEGG GSEA enrichment analyses of genes between molecular subtype A and (B) NES>1, nominal p value<0.05, FDR<0.25 were considered to be statistically significant. (E) ssGSEA indicated differences between the infiltration levels of TICs and distinct molecular subtypes. P value<0.05 was considered to be statistically significant. Molecular subtype A contained 339 GC samples and molecular subtype B contained 393 GC samples. * indicated P < 0.05, *** indicated P < 0.001.
Figure 5
Figure 5
Identification of CRG gene subtypes based on 84 DEGs derived from different molecular subtypes. (A, B) GO enrichment analyses of 84 DEGs from molecular subtype A and (B) Adjusted p value<0.05 was considered to be statistically significant. (C) Identification of two gene subtypes (k = 2) and their correlation area through consensus clustering analysis according to the expression of 32 prognosis-related DEGs. (D) The heat-map showed the gene profiles in gene subtypes A and B, and the associations between clinicopathologic characteristics and distinct gene subtypes. Chi-square test was used for the comparison. P < 0.05 was considered to be statistically significant. (E) Differential analysis of the expression of CRGs in different gene subtypes. P < 0.05 was considered to be statistically significant. (F) Survival analysis of two gene subtypes. Kaplan–Meier plot and log-rank tests were conducted for survival analyses. P < 0.05 was considered to be statistically significant. Molecular subtype A contained 339 GC samples and molecular subtype B contained 393 GC samples. Gene subtype A and B contained 329 and 403 GC samples, respectively. * indicated P < 0.05, ** indicated P < 0.01, *** indicated P < 0.001.
Figure 6
Figure 6
Construction of CRG Risk scoring system in the training group. (A) Alluvial diagram of subtype distributions in groups with different molecular subtypes, gene subtypes, Risk scores and survival outcomes. (B) Differential analysis of CRG Risk score in 339 molecular subtype A and 393 molecular subtype (B, C) Differential analysis of CRG Risk score in 329 gene subtype A and 403 gene subtype (B, D) RT-qPCR indicated the expression of five CRG risk score gene in 5 tumor and normal samples. * indicated P < 0.05. (E) Heat-map of five scoring genes expression profile in different risk sets of the training group. (F, G) Ranked dot and scatter plots of CRG Risk score distribution and patient survival in the training group. (H) Survival analysis in high- and low- CRG Risk score groups in the training set. Kaplan–Meier plot and log-rank tests were conducted for survival analyses. (I) ROC curve predicted the sensitivity and specificity of 1-, 3-, and 5-year survival according to CRG Risk score in the training group. The training group contained 364 GC samples. P < 0.05 was considered to be statistically significant.
Figure 7
Figure 7
Associations of TME and CRG Risk score. (A) Differential analyses of CRGs expression in the high- and low-risk score groups. (B–J) Correlation analyses between CRG Risk score and TICs. (K) Differential analyses between CRG Risk score and immune/stromal/estimate scores. (L) Correlation analyses between the abundance of TICs and five key Risk scoring genes in the proposed model. High-risk score group contained 352 GC samples and low-risk score group contained 376 GC samples. P < 0.05 was considered to be statistically significant. * indicated P < 0.05, ** indicated P < 0.01, *** indicated P < 0.001.
Figure 8
Figure 8
Associations of CRG Risk score with MSI, TMB and CSC. (A) The distribution of MSI in different Risk score groups. (B) Differential analyses between CRG Risk score and MSI. (C) Correlation analysis between CRG Risk score and CSC index. (D) Differential analysis of TMB in distinct CRG Risk score groups. (E) Correlation analysis of CRG Risk score and TMB. (F, G) The waterfall plot of somatic mutation characteristics in high- and low- CRG Risk score groups. High-risk score group contained 352 GC samples and low-risk score group contained 376 GC samples. P < 0.05 was considered to be statistically significant.
Figure 9
Figure 9
Construction and validation of a nomogram in 728 GC samples. (A) Nomogram for predicting the 1-, 3-, and 5-year OS of GC patients. (B) Calibration curves of the nomogram. 728 GC samples contained 352 high-risk score and 376 low-risk score.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Smyth EC, Nilsson M, Grabsch HI, Grieken van NC, Lordick F. Gastric cancer. Lancet (2020) 396(10251):635–48. doi: 10.1016/S0140-6736(20)31288-5 - DOI - PubMed
    1. Joshi SS, Badgwell BD. Current treatment and recent progress in gastric cancer. CA Cancer J Clin (2021) 71(3):264–79. doi: 10.3322/caac.21657 - DOI - PMC - PubMed
    1. Slagter AE, Vollebergh MA, Caspers IA, van Sandick JW, Sikorska K, Lind P, et al. . Prognostic value of tumor markers and ctDNA in patients with resectable gastric cancer receiving perioperative treatment: results from the CRITICS trial. Gastric Cancer (2022) 25(2):401–10. doi: 10.1007/s10120-021-01258-6 - DOI - PMC - PubMed
    1. Tsang T, Davis CI, Brady DC. Copper biology. Curr Biol (2021) 31(9):R421–r427. doi: 10.1016/j.cub.2021.03.054 - DOI - PubMed
    1. Ge EJ, Bush AI, Casini A, Cobine PA, Cross JR, DeNicola GM, et al. . Connecting copper and cancer: From transition metal signalling to metalloplasia. Nat Rev Cancer (2022) 22(2):102–13. doi: 10.1038/s41568-021-00417-2 - DOI - PMC - PubMed

Publication types