Sage Journals: Discover world-class research

Abstract

Background: Colorectal cancer is a highly aggressive malignant tumor that primarily affects the digestive system. It is frequently diagnosed at an advanced stage. Cuproptosis is a copper-dependent form cell death mechanism, distinct from all other known pathways underlying cell death, tumor progression, prognosis, and immune response. Although the role of cuproptosis in colorectal cancer has been investigated over time, there is still an urgent need to explore new methods and insights to understand its potential function. Methods: The Gene Expression Omnibus and The Cancer Genome Atlas gene expression data were systematically explored to investigate the role of cuproptosis in colon adenocarcinoma. The weighted gene coexpression network analysis was used to construct a gene coexpression network and identify the critical module and cuproptosis-related genes correlated with colon adenocarcinoma prognosis. A cuproptosis-related genes prognostic signature for colon adenocarcinoma was identified and validated. To validate the identified gene signature, quantitative reverse transcription-polymerase chain reaction was performed. Cell proliferation assays were analyzed by CCK8 and cell cycle detection. In addition, reactive oxygen species assay was also analyzed. Results: Five hub cuproptosis-related genes (Dihydrolipoamide S-acetyltransferase, Cyclin-dependent kinase inhibitor 2A, ATOX1, VEGFA, and ULK1) were screened and a prognostic risk model for predicting overall survival was established based on these genes. The model was successfully tested in the validation cohort and the GEPIA database. Colon adenocarcinoma patients were categorized into high-risk and low-risk groups based on risk scores. The study revealed that patients with higher risk scores were more likely to have a poor prognosis. Moreover, Dihydrolipoamide S-acetyltransferase was a tumor suppressor gene that can induce cell death and affected the redox reactions in the colon cancer cell line. Conclusions: These findings suggest that the newly identified 5-gene signature may serve as a more reliable prognostic factor than clinical factors such as age and stage of disease. These findings offer a theoretical foundation for further investigation into potential cuproptosis-related biomarkers for predicting colon adenocarcinoma prognosis in the future.

Keywords

cuproptosis colon adenocarcinoma gene signature prognostic model dihydrolipoamide S-acetyltransferase

Introduction

Colorectal cancer (CRC) ranks the third most common cancer worldwide among males and females, accounting for approximately 12% of new cancer cases and 9% of cancer-related deaths in 2020.¹ The incidence has increased continuously and rapidly in recent decades.² The most frequent histological type of CRC is colon adenocarcinoma (COAD), which accounts for approximately 95% of all cases.³ Due to the absence of efficient diagnostic methods, most patients are diagnosed in advanced stages of the disease. In addition, it is necessary to note that nearly 50% of patients with early-stage disease will develop metastasis.² While radical surgery has traditionally been an effective therapy, its effectiveness has decreased significantly over time.⁴ In addition to surgical intervention, drug treatments such as gene-targeted therapy and immunotherapy have been found to be effective. However, despite these options, most patients with COAD still experience tumor recurrence due to drug resistance and lack of responsiveness to current immune checkpoint therapy.^5,6 As a significant global public health challenge, it is imperative to comprehend the molecular mechanism behind the recurrence and metastasis of COAD development.

Cuproptosis is a novel type of regulated cell death (RCD) linked to copper metabolism. Copper is an essential trace element that plays a crucial role in various biological processes (BPs), though the redox properties make it both beneficial and toxic to the cell.^7,8 Studies have shown that the levels of copper in cancer patients are significantly higher in both serum and tumor tissues compared to healthy counterparts, such as breast cancer,⁹ lung cancer,¹⁰ gastrointestinal,¹¹ and thyroid cancer.¹² Tumor cells also have a higher level of copper, which indicates the potential of cuproptosis inducers in new antitumor therapy. In their study, Peter et al demonstrated that copper could bind directly to lipid-acylated components of the tricarboxylic acid (TCA) cycle.¹³ This binding increases lipoylated protein aggregation and loss of iron-sulfur cluster proteins, leading to proteotoxic stress and ultimately resulting in cell death.¹³ Despite numerous studies on cuproptosis in various diseases, its potential function in COAD remains poorly understood as it is a newly discovered form of RCD.

Here, the gene datasets with sample size and prognostic information for COAD from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) datasets were adopted. We utilized the weighted gene coexpression network analysis (WGCNA) and least absolute shrinkage and selection operator (LASSO) regression analysis to identify key modules and hub cuproptosis-related genes (CRGs) correlated with COAD prognosis. A new 5-gene signature to highlight the regulatory functions of CRGs in COAD progression was created. Our findings provide insights into novel strategies for predicting prognosis and suggest potential therapeutic applications for individualized treatment of patients with COAD.

Materials and Methods

Data Collection and Processing

The reporting of this study conforms to TRIPOD guidelines.¹⁴ The dataset GSE17536, GSE39582, GSE29621 (platform file GPL570), and GSE28722 (platform file GPL13425) were downloaded from the GEO public database (https://www.ncbi.nlm.nih.gov/geo/, accessed August 26, 2022), according to the number of colon cancer patients and possessed complete clinical information. Above 4 files as the training cohort enrolled for our subsequent analysis, the expression data of 177, 585, 65, and 125 samples were obtained from GSE17536, GSE39582, GSE29621, and GSE28722, respectively. All of 952 cases were used for the following WGCNA and prognosis model construction. Then we merged all the files after gene reannotation and normalization, removing the batch effects from different datasets using the “sva” package.¹⁵ Furthermore, the RNA sequencing (Fragments per kilobase million, FMPK) and clinical data of colon cancer patients as the validation cohort analyzed the expression of genes, downloaded from TCGA Genomic Data Commons portal (https://cancer genome.nih.gov/, accessed August 18, 2022) database, including 451 cases. To reduce the number of errors caused by confounding factors, we excluded samples without survival information (n = 3). All the GEO data and TCGA data were noted and transferred to gene expression matrix with the R program (version 4.2.2) and Perl (version 5.38.2.2) for subsequent analysis. Forty-six CRGs were acquired from previous literature for subsequent analyses^7,13,16,17 with a false discovery rate < 0.05 as a filter condition listed in Table 1.

Table 1.

The Identification of Cuproptosis-Related Genes.

	Gene symbol
Cuproptosis-related genes	LIAS, LIPT1, LIPT2, CDKN2A, DLAT, DLST, DLD, FDX1, COA6, ATP7A, ATP7B, SLC31A1, SLC31A2, SLC25A3, PDHA1, PDHB, MAP2K1, MTF1, GLS, UBE2D1, H3C1, CD274, COX11, UBE2D2, PDK1, SCO1, COX17, UBE2D4, UBE2D3, DBT, SOD1, ULK2, CP, AOC3, DBH, LOXL2, ULK1, VEGFA, GCSH, CCS, MAP2K2, ATOX1, MTCO2P12, TYR, NFE2L2, NLRP3

Weighted Gene Coexpression Network Analysis Analysis of Training Cohort

Weighted gene coexpression network analysis was constructed with an R package named “WGCNA,” according to the method published in the literature.¹⁸ We used the goodSampleGenes method to dislodge nonexpressed genes and leveraged a standard deviation of >1 to identify expressed genes in colon cancer cases for cluster analysis. Firstly, the Person's correlation coefficient was calculated to construct a gene matrix for all pair-wise genes, investigating the relationships between gene networks and clinical traits. The power function was utilized to construct a scale-free coexpression network. After selecting appropriate R² and soft-threshold value β, the adjacency matrix was transformed into a topological overlap matrix (TOM). The TOM facilitated the measurement of a specific gene's network connectivity. Next, the gene modules were identified by hierarchical clustering, with the minimum size number of 50 genes per module for the gene dendrogram and the minimum size of 2 for deepSpilt, and the eigengene was also calculated. Furthermore, we merged all close modules into new modules with the height of 0.25 via cluster analysis. The correlations between the modules’ characteristic genes and the disease phenotypes of the training group were calculated using the eigenvectors of modules and gene expression of samples, hub genes were selected for further prognosis model building with R > 0.7.

Gene Functional and Pathway Enrichment Analysis

Enrichment analysis of the key module consisted of Gene Ontology (GO) analysis and Kyoto Encyclopedia of Genes (KEGG) pathway analysis, performed by “clusterProfiler” R package.¹⁹ Gene Ontology was performed 3 main parts, including BP, cellular component, and molecular function (MF). Via ggplot2 package in R software to plot the top 10 GeneRatio of BP, cell components, or MF of GO analysis and the top 10 GeneRatio of KEGG pathways. The “ggplot2” package was used to plot the figures results based on the enrichment analysis.²⁰

Construction of the LASSO-Cox Regression Model and Validation

Before establishing the model, taking the intersection of CRGs with candidate genes of the turquoise module screened from WGCNA, then the selected genes were used to build the eventual model. After all datasets was integrated the survival time, survival status, gene expression data, and common clinical characteristics (age, sex, clinical stages), the files from GEO database, containing gene expression and overall survival (OS) related clinical prognosis information, were as the training group to construct the LASSO-Cox regression model using the “glmnet” package,²¹ a machine learning algorithm applying to select the cuproptosis-related signature through the risk score. Meanwhile, we used the 10-fold cross-validation method and took the minimized λ as lambda to obtain an optimized model. The risk score for OS (RSO) was calculated via multiplying the expression level based on the coefficients, which affected the prognosis of COAD. The calculation of RSO formula for each patient as follows: RSO = coef 1 × Exp 1 + coef 2 × Exp 2 + …+ coef_i× Exp_i (coef_i represents the corresponding coefficient of each coproptosis-related signature, Exp_i identifies the expression of hub gene.)

The median RSO of training set was set as the cutoff value, all samples were categorized into high-risk and low-risk groups according to the cutoff. Next, the R package “ggplot2” was used to generate the gene expression heatmaps of RSO. In addition, the univariate survival analysis was used to construct a forestplot, containing the clinical information of COAD patients, hazard ratio (HR) and 95%CI. Moreover, a nomogram capable of predicting the prognosis was created using the “rms” package to illustrate the independent factors. The “Points” indicate the effect of each factor on the prognosis. Meanwhile, the principal component analysis (PCA) was performed to observe the distinction between the 2 risk clusters. We also compared the OS difference between the 2 clusters by K-M survival analysis. For further validation, TCGA cohort as the validation set was divided into high and low risk group based on the cutoff RSO value, K-M curves were constructed. We also analyzed the 5 target genes Survival Analysis in the Gene Expression Profiling Interactive Analysis (http://gepia.cancer-pku.cn/, GEPIA, accessed December 26, 2022) database individually.²² In addition, the expression of the 5 genes in normal tissues and tumor tissues of colon cancer was also analyzed through this dataset, the survival heatmap in pan cancers of these genes was also evaluated.

The Immune Cell Infiltration Estimation

To investigate the immune cell infiltration, the ssGSEA enrichment analysis performed by the R GSVA package was used to analyze the tumor-infiltrating immune cell subgroups and immune function between the high-risk and low-risk groups. The final visualization was presented via box plots. The open website of TIMER (https://cistrome.shinyapps.io/timer/) was used to study the correlation between the 5 hub CRGs and 6 tumor infiltrating immune cells.

Cell Culture and Plasmid Transfection

Human CRC cell line HT29, HCT116, and LoVo were purchased from the Cell Resources Center, Shanghai Academy of Life Sciences, Chinese Academy of Sciences. All cells were cultured in DMEM supplemented with 10% fetal bovine serum and penicillin (100 U/mL)/streptomycin (0.1 mg/mL) at 37 °C in a humidified atmosphere containing 5% CO2. Dihydrolipoamide S-acetyltransferase (DLAT) overexpression plasmid and the empty control were constructed using pcMV-HA-C-DLAT and vector, respectively. HT29 cell line was transfected with plasmid by LipoFiter reagents (Cat# HB-TRLF-1000) purchased from Hanbio according to the manufacturer's protocol.

RNA Isolation and Quantitative Reverse Transcription-Quantitative Polymerase Chain Reaction

After being treated with DLAT overexpression plasmid for 48 h, total RNA was extracted from the HT29 cell line using the RNA Quick Purification kit (Cat# RC112-01) according to the manufacturer's instruction purchased from Vazyme, and 500 ng of RNA was reverse transcribed into cDNA using the PrimeScript RT Reagent Kit (Cat# Q222-01) purchased from Vazyme according to the manufacturer's instructions. Gene expression levels were measured by reverse transcription-quantitative polymerase chain reaction using TB Green Premix Ex Taq (Cat# RR820Q) purchased from TaKaRa on a CFX96 Real-Time System (Bio-Rad). The PCR primers were synthesized via Hzykang and the sequence were as follows: 5′-CCGCCGCTATTACAGTCTTCC (sense) and 5′-CTCTGCAATTAGGTCACCTTCAT (antisense).

Cell Viability Assay

Cell proliferation rate was detected with Cell Counting Kit-8 kit (Cat# C0038) according to the manufacturer's instruction purchased from Beyotime. Briefly, HT29 cells were plated in 96-well plats (2500 cells/well) and transfected with plasmid. A 10 μL CCK-8 reagent was added to each well with 100 μL medium at the indicated time points (0, 1, 2, 3, and 4 days after transfection) and then incubated for another 1 h. A 450 nm of the absorbance was measured by a microplate reader (BioTek).

Cell Cycle Analysis

Cell cycle analysis was performed after transfection with plasmid for 24 h. In detail, HT29 cells were collected by trypsinization, fixed in ice-cold 70% ethanol at 4 °C overnight, stained for Propidium (PI) and performed by CytoFLEX S flow cytometry (Beckman), based on the protocol provided by the Cell Cycle and Apoptosis Analysis Kit (Cat# C1052) purchased from Beyotime.

Reactive Oxygen Species Assay

Intracellular reactive oxygen species (ROS) levels were detected by the ROS kit (Cat# S0033S) purchased from Beyotime. After the transfecting plasmid, cells in 6-well plates were washed with PBS and incubated with 10 μM DCFH-DA for 30 min. Then washing the cells with serum-free solution and collecting, the mean fluorescence intensity of DCFH-DA was measured by flow cytometry (Beckman), which was representative of ROS level. Subsequently, FlowJo software V10 was used to analyze the results.

Statistic

The Student t test was used to compare differences between 2 groups. GraphPad Prism 9.0 software was used to conduct the statistical analyses. P < .05 was considered statistically significant.

Results

Identification of Key Modules in Colon Cancer

A detailed flowchart of this study is displayed in Figure 1. A total 952 samples and 23 412 genes in the gene expression obtained from the training sets were downloaded from GEO database. After normalization, the average RNA expression in every sample was basically the same. The WGCNA algorithm was used to build a network, including a group of genes frequently possessing high coexpression level and high topological overlap similarity, as shown in Figure 2. A soft threshold power of β = 4 was selected to establish that the network was scale-free, and the corresponding was R²= 0.9 and with the high average connectivity (Figure S1A and B). A gene hierarchy clustering dendrogram was constructed using the dynamic tree cut package, and genes in the new data expression profiles were allocated to eleven biologically similar modules (Figure 2A). Meanwhile, we analyzed the correlation between modules and phenotypes, and the turquoise module containing 1935 genes had the most significant clinically correlation in COAD, based on the correlation of module feature values (Figure 2B). The scatter plot (Figure 2C) shows a strong correlation between the gene significance and module membership analyzes in the turquoise (Cor = 0.6, P = 2.3e-59). The scatter plots of other modules are established in (Figure S2).

Figure 1.

The flowchart of the overall study.

Figure 2.

Identification of weighted gene coexpression network analysis (WGCNA) key module. (A) The cluster dendrogram of 11 colored modules based on a dissimilarity measure (1-TOM) in colon cancer patients. (B) The heatmap of correlation between the biologically significant modules and clinical traits (age, stage, vital status, and survival time). (C) Scatter plot between gene salience and module membership of turquoise module (cor = 0.6, P = 2.3e-59).

Functional Enrichment Analysis

The genes in the turquoise module were studied though GO and KEGG enrichment analysis, to explore the biological functions and potential pathways. The GO analysis results displayed that BPs were mainly concentrated in extracellular matrix organization, extracellular structure organization, epithelial cell proliferation, and ossification (Figure 3A). Cellular components were also analyzed, including collagen-containing extracellular matrix, apical part of cell, endoplasmic reticulum lumen, and many other parts (Figure 3B). Additionally, the genes of the turquoise module enriched in the part of molecular: extracellular matrix structure constituent, integrin binding, and growth factor binding (Figure 3C). The results of KEGG signaling pathway analysis indicated that the genes of turquoise module were most enriched in Protein digestion absorption, ECM-receptor interaction, and Focal adhesion pathway (Figure 3D).

Figure 3.

Functional enrichment analysis of Gene Ontology (GO) and Kyoto Encyclopedia of Genes (KEGG) pathway with the turquoise. (A) Biological processes of GO enrichment. (B). Cell components process of GO enrichment. (C) Molecular function of GO enrichment. (D) KEGG pathways enrichment.

Identification of Hub CRGs and Construction of a 5-Gene Signature for Predicting OS

To further explore the regulatory role of cuproptosis in the pathogenesis of COAD, we intersected the CRGs with turquoise module genes to obtain a total of 5 signature genes serving as candidates: DLAT, Cyclin-dependent kinase inhibitor 2A (CDKN2A), ATOX1, VEGFA, and ULK1 (Figure 4A). Subsequently, LASSO regression module was established based on the 5 genes to streamline the important characteristic variables (Figures 4B and C). The relative regression coefficients of the 5 genes were calculated by LASSO regression analysis in training group and the detailed information are in Table 2.

Figure 4.

Prognostic module constructed by least absolute shrinkage and selection operator (LASSO) regression analysis. (A) Venn diagram of the intersection of cuproptosis-related genes and turquoise module, obtaining 5 genes: Dihydrolipoamide S-acetyltransferase (DLAT), CDKN2A, ATOX1, VEGFA, and ULK1. (B, C) The relative regression coefficients of the 5 genes. (D) Distribution plot of the risk score for OS (RSO) scores and survival status. (E) principal component analysis (PCA) plot.

Table 2.

The Detailed Information of the 5 Modeling Genes and Their Correlation R Values in the Turquoise Module.

Ensembl ID	Gene symbol	Gene description	Regression coefficient
ENSG00000150768.15	DLAT	Dihydrolipoamide S-acetyltransferase	0.06423223
ENSG00000147889.1	CDKN2A	cyclin-dependent kinase inhibitor 2A	0.30363910
ENSG00000177556.11	ATOX1	antioxidant 1 copper chaperone	−0.198567004
ENSG00000112715.20	VEGFA	vascular endothelial growth factor A	0.22006697
ENSG00000177169.9	ULK1	unc-51 like autophagy activating kinase 1	0.25098770

We used the relative expression level and relative regression coefficients of the 5 genes to calculate the RSO of every sample. The cutoff values of RSO divided the patients with COAD into a high-risk group and a low-risk group and the regression equation was calculated as follows:

RSO = (0.06423223) × expression level of DLAT + (0.30363910) × expression level of CDKN2A + (−0.198567004) × expression level of ATOX1 + (0.22006697) × expression level of VEGFA + (0.25098770) × expression level of ULK1

The distribution plot of RSO and survival status was analyzed by ranking and risk scores. The results indicated that individuals with higher risk scores were more likely to have dead. Moreover, the different expression profiles of the cuproptosis-related 5 genes were listed by the heatmap between high-risk and low-risk group (Figure 4D). The PCA was also analyzed to prove that the 5-gene signature based on RSO could distinguish 2 separate subgroup of COAD patients (Figure 4E).

Prognostic Value of the CRGs 5-Gene Signature

The survival curves for the patients with high-risk and low-risk were performed by the Kaplan-Meier analysis in training and validation groups. The results of both 2 cohorts suggested that the patients with a higher RSO had a statistically lower probability of survival. Because of the different platforms and normalization methods for each group, the P value was .002 in the training group (Figure 5A) and .032 in the validation group (Figure 5B). Furthermore, to prove the prognostic value of the 5 genes and integrated signature in COAD, the GEPIA database was chosen to analyze. The high and low expression was set 70% and 30% median cutoff values. The 5-gene signature with a low expression group showed a better prognosis (P = .019, HR = 2.1, Figure 5C). The DLAT high expression group also showed a better prognosis with the P = .049, HR = 0.52 (Figure 5D). The CDKN2A low expression group had a better prognosis with the P = 0.029 and HR = 2 (Figure 5E). Despite the ATOX1, VEGFA, and ULK1 had similar results to the gene of CDKN2A, there were no statistically significant differences (P = .54, HR = 1.2; P = .21, HR = 1.5; P = .098, HR = 1.7, Figure S3A). The expression levels of 5 genes in COAD were compared between cancer and normal tissues using this database. A |Log2FC| cutoff value of 0.5 and a P-value cutoff of .05 were set. The results indicated that the expression levels of DLAT (P < .05, Figure 5D), CDKN2A (P < .05, Figure 5E), ATOX1 (P < .05, Figure S3B), and VEGFA (P < .05, Figure S3B) in cancer tissues were significantly higher compared to those in normal tissues, the gene of ULK1 had the contrary tendency (P < .05, Figure S3B). Moreover, the expression of the 5 genes in pan cancers also analyzed, the heatmap suggested the expression of each gene has a slight variation in different cancer patients (Figure 5F).

Figure 5.

Prognostic value of the cuproptosis-related genes (CRGs) 5-gene signature. (A) Univariate survival analysis of the high-risk group and low-risk group in train cohort. (B) Univariate survival analysis of the high-risk group and low-risk group in validation cohort. (C) K-M survival curve of the low and high 5-gene signature groups in the GEPIA database. (D) Through the GEPIA database to analyze the 2 Dihydrolipoamide S-acetyltransferase (DLAT) groups’ K-M survival curve and the expression difference between cancer and normal tissues. (E) Through the GEPIA database to analyze the 2 CDKN2A groups’ K-M survival curve and the expression difference between cancer and normal tissues. (F) The difference expression of the 5 genes in pan cancers. The asterisks represented the statistical P value (*P < .05; **P < .01; ***P < .001).

Analysis of Immune Cell Infiltration and Function

The relationship of cuproptosis-related 5-gene signature with immune infiltrates was investigated. The results showed that B cells, CD8⁺ T cells, dendritic cells (DCs), Mast cells, Neutrophils, NK cells, Th2 cells, TIL, and Treg cells had obvious difference between high-risk and low-risk group (all P < .001). Besides T helper cells, T follicular helper cells (both P < .01) and immune DCs (P < .05) were distinct in 2 groups too (Figure 6A). In addition, immune-related functions analysis also found that nearly all had a different expression in the 2 groups, such as APC coinhibition, Cytolytic activity, T-cell coinhibition (Figure 6B). Immune inﬁltration through ssGSEA algorithm showed that the immune status in each group was totally different, which indicated that 5-gene signature may be further elucidated to develop tumor immunotherapy in COAD. The TIMER database was also used to investigate the immune infiltration associated with 5 genes in colon cancer. The results revealed that the expression level of the 5 CRGs had correlation with different tumor immune cell, and P < .05 was statistically significant (Figure 6C).

Figure 6.

The immune status difference in the 5-gene signature and each gene expression levels. (A) Comparison of immune cell abundance between the high-risk and low-risk groups of colon adenocarcinoma (COAD) patients. (B) Comparison of immune-related functions between the high-risk and low-risk groups of COAD patients. (C) Expression of the target 5 genes relating to immune infiltration cells in COAD. The asterisks represented the statistical P value (*P < .05; **P < .01; ***P < .001).

The CRGs 5-Gene Signature as a Better Independent Prognostic Factor for COAD

The univariate cox regression analysis was performed to investigate prognosis-related factors of colon cancer samples in the training set. The RSO calculated from the 5 genes and 2 common essential clinical characteristic factors (age and stage) were included in the cox regression analysis. The result demonstrated that RSO (P < .0001), age (P < .0001), and stage (P < .0001) might be as independent risk factors for COAD influencing OS (Figure 7A). Moreover, the RSO had a higher HR value of 2.745 (95%CI: 1.653-4.558) compared with other clinical factors, which indicated that the risk of death in the high RSO group was 2.745 times than the low RSO group. Additionally, a nomogram was constructed to establish a method for quantitatively capable of predicting the survival probabilities of 1-, 3-, and 5-year OS in colon cancer patients based on the high- and low-risk score signature (Figure 7B). The result showed that RSO had a wider range of points compared with age or stage, as a stronger factor for predicting the survival rates. Thus, it can be concluded that the predictive value of RSO is higher than that of the 2 important clinical factors (age and stages, respectively) in the training set.

Figure 7.

The 5-gene signature as a better prognostic factor than clinical factors. (A) Univariate analysis of the 5-gene signature and clinic pathological factors (age and stage) for overall survival in colon adenocarcinoma (COAD) patients. (B) Nomogram with the risk score for OS (RSO) and clinicopathological features for predicting the 1-, 3-, and 5-year survival.

Upregulation of DLAT Accelerates Cell Death of CRC Cells In Vitro

To ascertain the biological significance and the correlation of DLAT with the malignancy of CRC, we performed functional studies using overexpression plasmid to upregulation DLAT in HT29, HCT116 and LoVo cell lines in vitro. The map of plasmid showed in Figure S4. We transfected DLAT overexpression plasmids and empty vectors into CRC cells and the efficiency of interference were confirmed by mRNA levels. Meanwhile, we found that the transfection group's protein expression was obviously elevated (Figure 8A). Subsequently, through CCK-8 assays, the cell growth decreased remarkably upon DLAT upregulation, as shown in Figure 8B. In addition, the impact on the cell cycle was assessed by flow cytometry analysis, and the results indicated an increased tendency in the proportion of G0/G1 phases in the HT29 and HCT116 cells with DLAT overexpression group compared with the vector group, and the results showed an increase G2/M phase in the LoVo cells with DLAT overexpression (Figure 8C).

Figure 8.

Upregulation of Dihydrolipoamide S-acetyltransferase (DLAT) inhibits cell function of colorectal cancer cells in vitro. (A) Relative expression level of DLAT gene mRNA after transfecting DLAT overexpression plasmids (n = 3). (B) Cell viability analysis by CCK8 showing the effect of DLAT on the proliferation of HT29, HCT116 and LoVo cell lines at 24, 48, 72, and 96 h (n = 3). (C) Flow cytometry showing the DLAT influencing the cell cycle of HT29, HCT116, and LoVo cell lines (n = 3). (**P < .01; ***P < .001; ****P < .0001).

Dihydrolipoamide S-acetyltransferase, as a catalytic subunit of the human the pyruvate dehydrogenase complex, considered to facilitate TCA cycle metabolism.²³ To investigate whether DLAT affected the redox reactions of COAD, ROS analysis was conducted. And the results showed that ROS accumulation was higher in DLAT overexpression CRC cell lines than the negative control group (Figure 9A and B).

Figure 9.

The effect of Dihydrolipoamide S-acetyltransferase (DLAT) on the intracellular reactive oxygen species (ROS) levels of colorectal cancer cells in vitro. (A and B) DLAT overexpression increased the intracellular ROS levels in HT29, HCT116, and LoVo cells measured by DFCH-DA fluorescence (**P < .01; ***P < .001; ****P < .0001).

Discussion

Herein, we explored the role of CRGs in colon cancer and utilized various bioinformatics tools and statistical approaches to identify a new cuproptosis-related 5-gene signature biomarker for the prediction of prognostic. In addition, we further analyzed the functions of the 5-gene signature involved in the model construction and other biological functions.

Colon adenocarcinoma is a heterogenous disease and dominates a major driver of morbidity and mortality in recent years. There are many molecular alterations that contribute to the dysregulation of different signaling pathways, ultimately influencing tumor onset, progression, and invasiveness.²⁴ Despite significant advances in clinical treatment, the exact pathogenesis of COAD remains a pressing problem. Over the past 2 decades, research on the genetic and epigenetic regulations has shown promising potential in improving this disease's diagnosis, therapy, and survival prediction.²⁵ Furthermore, numerous studies have shown that tumorigenesis and progression involve various molecular pathways and biomarkers, such as phosphatidylinositol3 kinase/AKT signaling pathway,²⁶ mitogen-activated protein kinase pathway,²⁷ and KRAS mutation.²⁸ Active research on novel genes is essential for improving the prognosis of COAD.

In 2022, Tsvetkov et al proposed cuproptosis as a nonapoptotic mode of cell death characterized by an abundance of intracellular copper.¹³ This induces cell death through various subroutines, regulates mitochondrial respiration, and leads to glycolysis, insulin resistance, and lipid metabolism changes. Recently, a growing body of research indicates that cuproptosis is a significant factor in the development of various diseases.^10,29–32 That means there may be potential therapeutic options for targeting tumors with a specific metabolic profile using copper ions, and reliable biomarkers for detecting these tumors in humans.

This study presents the first implementation of WGCNA and LASSO algorithm to investigate the evidence the value of CRGs in COAD. The study provides a full view of the differential expression of obvious hallmark gene signatures. Weighted gene coexpression network analysis is commonly used to screen key module information from chip data and conduct significant association analyses with phenotype. In the present study, we analyzed 4 microarray datasets from GEO as the training group and identified a key module that is significantly correlated with clinical prognosis. Functional enrichment analysis (GO and KEGG) based on the target module showed that it was enriched in multiple crucial pathways consistent with previous studies, such as epithelial cell proliferation²⁷ and ECM-receptor interaction.³³ These findings suggest that these functions may play important roles in the pathogenesis and tumor suppression of COAD. And we identified 5 pivotal genes (DLAT, CDKN2A, ATOX1, VEGFA, and ULK1) that play a key regulatory role in the development of a prognosis signature.

Cyclin-dependent kinase inhibitor 2A is usually a tumor suppressor gene located on chromosome 9p21.3. It encodes 2 proteins, p14 and p16, which primarily function as cell cycle regulators.³⁴ Various studies have indicated that CDKN2A deregulation expression is linked to a negative prognosis in HPV-negative head neck squamous cell carcinomas.³⁵ Additionally, it has been associated with a poor prognosis in soft tissue sarcoma.³⁶ On the contrary, we find that CDKN2A high expression is associated with a poor prognosis, which is downregulated in COAD patients compared with the healthy controls, due to the heterogeneity between samples possibly.

ATOX1 serves as both a copper chaperone and an antioxidant, playing a crucial role in the copper-trafficking pathway that helps maintain intracellular copper homeostasis. It binds and transports cytosolic copper to ATPase proteins in the trans-Golgi network, which is subsequently incorporated into ceruloplasmin.³⁷ Furthermore, many activities characterize important cytology processes, such as cell proliferation, migration, autophagy, DNA damage repair, and programmed death.^38,39 Recently, studies have found that ATOX1 plays a role in tumorigenesis in various types of cancer.⁸ In human breast cancer, ATOX1 was upregulated and localized at the lamellipodia edges of aggressive breast cancer cells.⁴⁰ Our study also found higher expression of ATOX1 in colon cancer tumor tissue. High expression of ATOX1 has been linked to poor patient outcomes,⁴¹ and we observed a consistent trend in COAD via GEPIA analysis, although statistical significance was not reached.

ULK1 is a fundamental human autophagy-related gene that encodes a serine-threonine kinase and is the mammalian counterpart of the yeast ATG1 gene. It is situated on chromosome 12q24.3.⁴² Autophagy is a recycling system in organisms regulated by autophagy-related proteins and their partners. Heavy metals, such as copper, can be cytotoxic due to the ROS generated and ROS-induced activation can trigger the autophagy pathway of cell death.⁴³ Furthermore, recent studies have highlighted that copper induces autophagy in normal and cancer cells, which can serve as a cellular defense against copper-mediated toxicity.^44–46 Thus, ULK1 gene was also act the CRGs involved in our study. Recent studies have indicated that ULK1 has a distinct and specific role in mitophagy, particularly in hypoxic conditions.⁴⁷ And emerging studies demonstrate that the ULK1 expression protein is downregulated in breast cancer patients,⁴⁸ which was same as our results of low expression in colon cancer. In addition, we also found a correlation between the high expression and worsened prognosis in patients through public online analysis, but the P > .5. There need more studies to further prove the role of ULK1 in COAD.

VEGFA is a protein that belongs to the VEGF family and is closely associated with angiogenesis and development. It promotes the formation of new blood vessels and is often linked to various human diseases.⁴⁹ VEGFA binds with VEGFR-2, also known as KDR or Flk-1, the main receptor expressed on endothelial cells found in the tumor vasculature. This binding leads to receptor dimerization and trans-autophosphorylation of multiple tyrosine residues. It is important to note that the increased expression of VEGFA in tumors and the tumor microenvironment leads to increased tumor microvessel density, invasiveness, metastasis, and worsened patient prognosis.⁵⁰ VEGFA plays a crucial role in the formation of structurally abnormal and leaky tumor vasculature, while also promoting the growth of endothelial cells. This process leads to an improvement in antitumor immunity. Moreover, antiangiogenetic drugs targeting the VEGFA pathway are used in various cancer therapy, especially in nonsmall cell lung cancer (NSCLC).⁵¹ Our study has little attention on the function of VEGFA in colon cancer, more experimental studies could be performed to verify the value in the future.

Dihydrolipoamide S-acetyltransferase, a mitochondrial protein involved in the multienzyme pyruvate dehydrogenase complex and as the key molecules of cuprotosis, may play a crucial role in tumorigenesis.¹³ Assuming the Warburg effect, it is intuitive to expect that DLAT would be downregulated in cancer.⁵² One study suggested that gastric cancer patients had a higher expression of DLAT than usual. And knockdown in human gastric cancer cells substantially inhibited cell proliferation significantly, revealing a role for this gene in cancer development.²³ Moreover, we also found that the expression of DLAT was upregulated in COAD. Another study demonstrated that DLAT was a novel oncogene and higher DLAT expression was related to poorer outcome in NSCLC patients.⁵³ But the DLAT possessed a role of cancer inhibition in COAD by GEPIA analysis and HT29 cell experiments. We constructed the overexpression plasmid of DLAT and performed the cell cycle analysis, ROS analysis, and cell proliferation analysis to investigate the role in colon cancer. Apparently, DLAT is a tumor suppressor gene that can induce cell death in the colon cancer cell line. Nevertheless, the detailed tumorigenesis mechanism research of DLAT in COAD is unknown, thus it is urgent to conduct more experimental studies.

In the present study, we used the above 5 CRGs characterized by copper metabolism participation to construct a signature associated with prognosis. Recently, several studies have applied signature analysis to predict characteristics or prognosis for colon cancer. For instance, a 7 Ferroptosis-related lncRNAs (LINC01503, AC004687.1, AC010973.2, AP001189.3, ARRDC1-AS1, OIP5-AS1, and NCK1-DT) signature was developed as a biomarker to predict clinical outcomes and therapeutic responses in colon cancer patients.⁵⁴ Although the prognostic model was validated, this study did not present cell experiments or patient data from actual hospitals. Moreover, the gene number in this study was more than our model. Another 8-gene signature was reported to dichotomize patients with different OS significantly, serve as independent predictor consisting with our study.⁵⁵ Our study has shown that the prognostic influence of our 5-gene signature is superior to that of clinical factors such as age and stages. Additionally, our signature was constructed using fewer genes than some other similar reports, making it potentially easier to apply in subsequent clinical translational research or for the development of a detection kit to promote clinical applications.

Although the prognostic value of the RSO related to cuproptosis has been validated, some limitations need to be acknowledged. Firstly, our analysis was conducted using retrospective public databases (GEO and TCGA databases), and we did not divide the samples into independent training and test sets, though used the cross-validation method to evaluate the model's predictive performance. This may lead to some bias in the error estimation of the model and may not fully reflect the model's generalization ability on new data. Therefore, we will use stratified sampling or stratified random splitting methods to train and validate the model on larger data sets, to improve the model's stability and robustness in future studies. Secondly, while the DLAT gene was validated through cell function tests, it is crucial to conduct additional experiments in vitro and in vivo to explore the mechanism of the CRGs in COAD, what we should do in the near future.

In the present study, we used the above 5 CRGs characterized by copper metabolism participation to construct a signature associated with prognosis. Our comprehensive analysis highlights the crucial clinical applications of CRGs-RSO and might be potentially applied in subsequent clinical translational research.

Conclusion

In conclusion, we utilized bioinformatics analysis to identify 5 hub genes closely associated with cuproptosis in COAD. We systematically generated and evaluated a risk score signature based on these genes as a potential prognostic factor for colon cancer patients. These findings could enhance our comprehension of invasion, provide a theoretical foundation for exploring potential regulatory biomarkers for prognosis prediction, and even contribute to developing more precise target therapy strategies.

Supplemental Material

sj-docx-1-tct-10.1177_15330338241250285 - Supplemental material for Identification of a 5-Gene Cuproptosis Signature Predicting the Prognosis for Colon Adenocarcinoma Based on WGCNA

Supplemental material, sj-docx-1-tct-10.1177_15330338241250285 for Identification of a 5-Gene Cuproptosis Signature Predicting the Prognosis for Colon Adenocarcinoma Based on WGCNA by Dongxue Wang, Funing Yang, Guiping Han, Jifeng Zhang, Hongjia Wang, Zunyu Xiao, Weiyu Chen and Ping Li in Technology in Cancer Research & Treatment

Footnotes

Acknowledgments

The authors express our gratitude to the TCGA and GEO databases for providing platforms and collaborating with us in uploading crucial datasets.

Authors’ Note

Availability of Data and Materials: The datasets collected and analyzed in the study are available in the TCGA repository ( https://gdc.cancer.gov/ ) and GEO portal (

). The original contributions made in this study are included in the article/Supplemental Material. For any further inquiries,please contact the corresponding author.

Author Contributions

All authors in this study have made substantial contributions to the study conception and design. GH and PL contributed to the study design. FY and JZ contributed to data collection. WC and ZX performed statistical analysis and interpretation. HW and DW drafted the manuscript. All authors read and approved the final manuscript.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research,authorship,and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research,authorship,and/or publication of this article: This research was funded by the major subject of science and technology of Jinhua,China (grant number 2022-3-039).

Supplemental Material

Supplemental material for this article is available online.

ORCID iD

Dongxue Wang

References

Siegel

Miller

Wagle

Jemal

. Cancer statistics, 2023. CA Cancer J Clin. 2023;73(1):17-48. doi:10.3322/caac.21763

Ciardiello

Martini

Napolitano

Tabernero

Cervantes

. Clinical management of metastatic colorectal cancer in the era of precision medicine. CA Cancer J Clin. 2022;72(4):372-401. doi:10.3322/caac.21728

Siegel

Wagle

Cercek

Smith

Jemal

. Colorectal cancer statistics, 2023. CA Cancer J Clin. 2023;73(3):233-254. doi:10.3322/caac.21772

Dekker

Tanis

Vleugels

JLA

Kasi

Wallace

. Colorectal cancer. Lancet. 2019;394(10207):1467-1480. doi:10.1016/S0140-6736(19)32319-0

Topalian

Taube

Pardoll

. Neoadjuvant checkpoint blockade for cancer immunotherapy. Science. 2020;367(6477):eaax0182. doi:10.1126/science.aax0182

Dahan

Sadok

Formento

Seitz

Kovacic

. Modulation of cellular redox state underlies antagonism between oxaliplatin and cetuximab in human colorectal cancer cell lines: cetuximab/oxaliplatin antagonism in vitro. Br J Pharmacol. 2009;158(2):610-620. doi:10.1111/j.1476-5381.2009.00341.x

Bush

Casini

, et al. Connecting copper and cancer: from transition metal signalling to metalloplasia. Nat Rev Cancer. 2022;22(2):102-113. doi:10.1038/s41568-021-00417-2

Blockhuys

Celauro

Hildesjö

, et al. Defining the human copper proteome and analysis of its expression variation in cancers. Metallomics. 2017;9(2):112-123. doi:10.1039/c6mt00202a

Sha

, et al. Prognostic analysis of cuproptosis-related gene in triple-negative breast cancer. Front Immunol. 2022;13:922780. doi:10.3389/fimmu.2022.922780

10.

Wang

Zhang

Xue

. Cuproptosis predicts the risk and clinical outcomes of lung adenocarcinoma. Front Oncol. 2022;12:922332. doi:10.3389/fonc.2022.922332

11.

Liu

Zeng

, et al. Nucleophosmin 1 is a prognostic marker of gastrointestinal cancer and is associated with m6A and cuproptosis. Front Pharmacol. 2022;13:1010879. doi:10.3389/fphar.2022.1010879

12.

Wang

Yao

Yang

Tian

. Construction of a novel cuproptosis-related gene signature for predicting prognosis and estimating tumor immune microenvironment status in papillary thyroid carcinoma. BMC Cancer. 2022;22(1):1131. doi:10.1186/s12885-022-10175-5

13.

Tsvetkov

Coy

Petrova

, et al. Copper induces cell death by targeting lipoylated TCA cycle proteins. Science. 2022;375(6586):1254-1261. doi:10.1126/science.abf0529

14.

Collins

Reitsma

Altman

Moons

KGM

. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): the TRIPOD statement. Br Med J. 2014;350(jan07 4):g7594-g7594. doi:10.1136/bmj.g7594

15.

Parker

Leek

Favorov

, et al. Preserving biological heterogeneity with a permuted surrogate variable analysis for genomics batch correction. Bioinformatics. 2014;30(19):2757-2763. doi:10.1093/bioinformatics/btu375

16.

Lukanović

Herzog

Kobal

Černe

. The contribution of copper efflux transporters ATP7A and ATP7B to chemoresistance and personalized medicine in ovarian cancer. Biomed Pharmacother. 2020;129:110401. doi:10.1016/j.biopha.2020.110401

17.

Zhou

Zhao

, et al. Blockage of SLC31A1-dependent copper absorption increases pancreatic cancer cell autophagy to resist cell death. Cell Prolif. 2019;52(2):e12568. doi:10.1111/cpr.12568

18.

Langfelder

Horvath

. WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics. 2008;9:559. doi:10.1186/1471-2105-9-559

19.

Wang

Han

. Clusterprofiler: an R package for comparing biological themes among gene clusters. OMICS. 2012;16(5):284-287. doi:10.1089/omi.2011.0118

20.

Ito

Murphy

. Application of ggplot2 to pharmacometric graphics. CPT Pharmacometrics Syst Pharmacol. 2013;2(10):e79. doi:10.1038/psp.2013.56

21.

Engebretsen

Bohlin

. Statistical predictions with glmnet. Clin Epigenetics. 2019;11(1):123. doi:10.1186/s13148-019-0730-1

22.

Tang

Kang

Gao

Zhang

. GEPIA: a web server for cancer and normal gene expression profiling and interactive analyses. Nucleic Acids Res. 2017;45(W1):W98-W102. doi:10.1093/nar/gkx247

23.

Goh

WQJ

Kuznetsov

Chong

Lim

. DLAT subunit of the pyruvate dehydrogenase complex is upregulated in gastric cancer-implications in cancer therapy. Am J Transl Res. 2015;7(6):1140-1151.

24.

Vogelstein

Fearon

Hamilton

, et al. Genetic alterations during colorectal-tumor development. N Engl J Med. 1988;319(9):525-532. doi:10.1056/NEJM198809013190901

25.

Buikhuisen

Torang

Medema

. Exploring and modelling colon cancer inter-tumour heterogeneity: opportunities and challenges. Oncogenesis. 2020;9(7):66. doi:10.1038/s41389-020-00250-6

26.

Slattery

Mullany

Sakoda

, et al. The PI3K/AKT signaling pathway: associations of miRNAs with dysregulated gene expression in colorectal cancer. Mol Carcinog. 2018;57(2):243-261. doi:10.1002/mc.22752

27.

Song

Zhang

, et al. TIMP1 is a prognostic marker for the progression and metastasis of colon cancer through FAK-PI3K/AKT and MAPK pathway. J Exp Clin Cancer Res. 2016;35(1):148. doi:10.1186/s13046-016-0427-7

28.

Simanshu

Nissley

McCormick

. RAS proteins and their regulators in human disease. Cell. 2017;170(1):17-33. doi:10.1016/j.cell.2017.06.009

29.

Liu

Chang

, et al. Cuproptosis-associated lncRNA establishes new prognostic profile and predicts immunotherapy response in clear cell renal cell carcinoma. Front Genet. 2022;13:938259. doi:10.3389/fgene.2022.938259

30.

Bao

Duan

, et al. Identification of a novel cuproptosis-related gene signature and integrative analyses in patients with lower-grade gliomas. Front Immunol. 2022;13:933973. doi:10.3389/fimmu.2022.933973

31.

Zhang

Sun

Zhang

. A novel cuproptosis-related LncRNA signature to predict prognosis in hepatocellular carcinoma. Sci Rep. 2022;12(1):11325. doi:10.1038/s41598-022-15251-1

32.

Song

Zhou

Shu

. Cuproptosis scoring system to predict the clinical outcome and immune response in bladder cancer. Front Immunol. 2022;13:958368. doi:10.3389/fimmu.2022.958368

33.

Nersisyan

Novosad

Engibaryan

Ushkaryov

Nikulin

Tonevitsky

. ECM–receptor regulatory network and its prognostic role in colorectal cancer. Front Genet. 2021;12:782699. doi:10.3389/fgene.2021.782699

34.

Sievers

Hielscher

Schrimpf

, et al. CDKN2A/B homozygous deletion is associated with early recurrence in meningiomas. Acta Neuropathol. 2020;140(3):409-413. doi:10.1007/s00401-020-02188-w

35.

Deneka

Baca

Serebriiskii

, et al. Association of TP53 and CDKN2A mutation profile with tumor mutation burden in head and neck cancer. Clin Cancer Res. 2022;28(9):1925-1937. doi:10.1158/1078-0432.CCR-21-4316

36.

Sole

Grossetête

Heintzé

, et al. Unraveling Ewing sarcoma tumorigenesis originating from patient-derived mesenchymal stem cells. Cancer Res. 2021;81(19):4994-5006. doi:10.1158/0008-5472.CAN-20-3837

37.

Hatori

Yan

Schmidt

, et al. Neuronal differentiation is associated with a redox-regulated increase of copper flow to the secretory pathway. Nat Commun. 2016;7:10640. doi:10.1038/ncomms10640

38.

Reinhardt

Yaffe

. Phospho-Ser/Thr-binding domains: navigating the cell cycle and DNA damage response. Nat Rev Mol Cell Biol. 2013;14(9):563-580. doi:10.1038/nrm3640

39.

Chen

Zhang

, et al. APEX2-based proximity labeling of Atox1 identifies CRIP2 as a nuclear copper-binding protein that regulates autophagy activation. Angew Chem Int Ed. 2021;60(48):25346-25355. doi:10.1002/anie.202108961

40.

Blockhuys

Zhang

Wittung-Stafshede

. Single-cell tracking demonstrates copper chaperone Atox1 to be required for breast cancer cell migration. Proc Natl Acad Sci U S A. 2020;117(4):2014-2019. doi:10.1073/pnas.1910722117

41.

Blockhuys

Hildesjö

Olsson

Vahdat

Wittung-Stafshede

. Evaluation of ATOX1 as a potential predictive biomarker for tetrathiomolybdate treatment of breast cancer patients with high risk of recurrence. Biomedicines. 2021;9(12):1887. doi:10.3390/biomedicines9121887

42.

Kuroyanagi

Yan

Seki

, et al. Human ULK1, a novel serine/threonine kinase related to UNC-51 kinase of caenorhabditis elegans: CDNA cloning, expression, and chromosomal assignment. Genomics. 1998;51(1):76-85. doi:10.1006/geno.1998.5340

43.

Levine

Kroemer

. Biological functions of autophagy genes: a disease perspective. Cell. 2019;176(1-2):11-42. doi:10.1016/j.cell.2018.09.048

44.

Luo

Song

Kang

, et al. mtROS-mediated akt/AMPK/mTOR pathway was involved in copper-induced autophagy and it attenuates copper-induced apoptosis in RAW264.7 mouse monocytes. Redox Biol. 2021;41:101912. doi:10.1016/j.redox.2021.101912

45.

Xue

Wang

, et al. Suppressing autophagy enhances disulfiram/copper-induced apoptosis in non-small cell lung cancer. Eur J Pharmacol. 2018;827:1-12. doi:10.1016/j.ejphar.2018.02.039

46.

Guo

Ouyang

Yin

, et al. Induction of autophagy via the ROS-dependent AMPK-mTOR pathway protects copper-induced spermatogenesis disorder. Redox Biol. 2022;49:102227. doi:10.1016/j.redox.2021.102227

47.

Tian

, et al. ULK1 translocates to mitochondria and phosphorylates FUNDC1 to regulate mitophagy. EMBO Rep. 2014;15(5):566-575. doi:10.1002/embr.201438501

48.

Deng

Zhang

Huang

, et al. MAPK1/3 kinase-dependent ULK1 degradation attenuates mitophagy and promotes breast cancer bone metastasis. Autophagy. 2021;17(10):3011-3029. doi:10.1080/15548627.2020.1850609

49.

Claesson-Welsh

Welsh

. VEGFA and tumour angiogenesis. J Intern Med. 2013;273(2):114-127. doi:10.1111/joim.12019

50.

Chandrasekhar

Ding

Hugo

. The roles of TGF-β and VEGF pathways in the suppression of antitumor immunity in melanoma and other solid tumors. Pharmacol Ther. 2022;240:108211. doi:10.1016/j.pharmthera.2022.108211

51.

Chen

Liu

Wang

, et al. LINC00173.v1 promotes angiogenesis and progression of lung squamous cell carcinoma by sponging miR-511-5p to regulate VEGFA expression. Mol Cancer. 2020;19(1):98. doi:10.1186/s12943-020-01217-2

52.

Heiden MG

Cantley

Thompson

. Understanding the Warburg effect: the metabolic requirements of cell proliferation. Science. 2009;324(5930):1029-1033. doi:10.1126/science.1160809

53.

Chen

Wang

Yang

, et al. PM2.5 promotes NSCLC carcinogenesis through translationally and transcriptionally activating DLAT-mediated glycolysis reprograming. J Exp Clin Cancer Res. 2022;41(1):229. doi:10.1186/s13046-022-02437-8

54.

Cai

Zhuang

, et al. Development and validation of a ferroptosis-related lncRNAs prognosis signature in colon cancer. Bosn J Basic Med Sci. 2021;21(5):569-576. doi:10.17305/bjbms.2020.5617

55.

Dai

Yuan

Xiao

Ding

. A prognostic model for colon cancer patients based on eight signature autophagy genes. Front Cell Dev Biol. 2020;8:602174. doi:10.3389/fcell.2020.602174

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

18.30 MB