Skip to main content
NCBI home page
Discover Oncology logoLink to Discover Oncology
. 2025 Mar 12;16:306. doi: 10.1007/s12672-024-01733-3

Analysis of expression characteristics of ferroptosis-related lncRNAs in gastrointestinal cancer patients in Asia

Ning Ding 1,2, Ying Kang 1, Xiaoxiao Tan 1, Yanbo Tang 2, Yingjie Zhang 1,✉, Yongheng He 2,✉
PMCID: PMC11904047  PMID: 40072763

Abstract

Background

Asian cancer patients have become the highest morbidity and mortality group, and gastrointestinal tumors account for the majority of them, so it is urgent to find effective targets. Therefore, ferroptosis-related lncRNAs models were established to predict the prognosis and clinical immune characteristics of GI cancer.

Methods

RNA sequencing and clinical data were collected from the TCGA database (LIHC, STAD, ESCA, PAAD, COAD, CHOL, and READ) of patients with gastrointestinal cancer in Asia. Download ferrodroptosis genes from FerrDb. Through R language, differential genes were identified, prognostic related LncRNAs were screened, and risk scores were obtained by risk formula to build models. Survival analysis, risk heat map, COX regression and ROC were used to evaluate the risk model. Establish Nomogram and clinically relevant heat maps. GSEA software was used to analyze gene enrichment and immune-related characteristics in high and low risk groups. LncRNA expression was validated through paired sample differential analysis and qRT-PCR, and the drug sensitivity of genes was also analyzed.

Results

The transcriptome data of 297 cases and clinical data of 322 cases were downloaded from TCGA, and the intersection of ferroptosis-related genes were obtaine. Cox analysis revealed 48 ferroptosis-related LncRNAs associated with prognosis. Through survival analysis, risk heatmap, COX regression and ROC, it was found that the risk model was highly accurate and efficient in predicting prognosis. KEGG-related GSEA enrichment analysis showed that 12 related pathways were significantly expressed in the low-risk group. Four immune-related functions were significantly higher in the high-risk group than in the low-risk group, and the expression of all immune checkpoints were significantly higher in the high-risk group than in the low-risk group. The three LncRNAs in the model exhibited varying expression levels across different tumors and obtained drug sensitivity data.

Conclusions

Our results reveal innovative and strong evidence that ferroptosis-related lncRNAs can be used as biomarkers for the treatment and prognosis of Asian GI cancer.

Keywords: Gastrointestinal cancer, Asian patients, Ferroptosis, LncRNAs, QRT-PCR, Immune

Introduction

According to the incidence statistics of new cancer patients in 2020, the incidence of cancer in Asia accounts for 50% of the total incidence, and the mortality rate accounts for 58.3% of the total mortality rate. Both the mortality rate and the morbidity rate are the first, and the mortality rate is higher than the morbidity rate. Combined with the morbidity and mortality of all tumors, digestive tract tumors, including esophagus, stomach, liver, gallbladder, pancreas, colorectal, etc. accounted for the highest proportion [1]. At present, the treatment of gastrointestinal tumors is mainly radiotherapy, chemotherapy and surgery, and the therapeutic effect is better than before but still poor [2]. The main reason for this result is that the incidence of many GI cancers is mostly in the middle and late stages, so finding new therapeutic targets and prognostic markers has become the current approach.

Programmed cell death (PCD) is closely related to the occurrence and development of tumors. Ferroptosis, a kind of PCD, refers to a special iron-dependent non-apoptosis-dependent cell death process caused by membrane damage and cystine depletion mediated by massive lipid peroxidation [3, 4]. Recent studies have found that ferroptosis in tumors is due to abnormal intracellular iron metabolism and iron-dependent lipid ROS metabolism [5]. Ferroptosis in tumors can inhibit tumor genesis, development and invasion, and affect tumor metabolic reprogramming and tumor microenvironment (TME) [6]. Recent studies have found that it can also reshape tumor ecological niche through metabolic pathways, thereby affecting tumor immune escape, and ultimately promoting tumor genesis and development [7].

Studies related to colorectal cancer have found that a variety of molecules can inhibit GPX4 in colorectal cancer cells, thereby inducing ferroptosis and inhibiting the occurrence and development of colorectal cancer (CRC) [8]. ACSL4 overexpression in colorectal cancer can promote ferroptosis of cancer cells and ultimately inhibit the progression of CRC [9]. CRC-related TME studies have shown that molecular studies related to ferroptosis have a significant role in improving the understanding of TME infiltration and predicting immune efficacy [10]. Ferroptosis plays a dual role in Hepatocellular Carcinoma (HCC), not only helping the pathogenesis of the liver and the development of cancer but also inhibiting the carcinogenesis of HCC. In addition, iron deficiency-related genes and ferroptosis-related non-coding RNAs can be used as diagnostic and prognostic markers of HCC [11]. Studies related to gastric cancer (GC) have found that ferroptosis-related genes can be used as molecular markers to effectively predict immune, clinical and prognostic factors of patients [12, 13]. At present, ferroptosis has become a popular research target for clinical diagnosis, immunological and prognostic analysis of various tumors. It has been confirmed in CRC, HCC, GC and other tumors. However, there are no overall studies of gastrointestinal cancer. Therefore, it is particularly important to further search for ferroptosis markers related to clinical factors, prognosis and immune status of GI cancer in Asian patients.

Long non-coding RNAs (LncRNAs) refer to RNAs with a non-coding protein length of 200-100000nt, which play a variety of physiological and pathological functions such as cell growth, invasion and migration, and anti-apoptosis in vivo [14]. In tumors, lncRNA mutations and disorders play an important role, which can not only inhibit cancer or oncogene, but also promote the occurrence and metastasis of tumors and affect the immune microenvironment of tumors, and are closely related to a variety of gastrointestinal tumors, including gastric cancer, pancreatic cancer, colorectal cancer and gallbladder cancer [15]. Through the establishment of the ferroptosis-related LncRNA model, the study found that it has important guiding value for the prognosis of gastric cancer, immune and chemotherapy response and other personalized treatments [16]. It has a high ability to predict the survival prognosis, key clinical features and immune function of HCC patients, which is conducive to the development of individualized treatment plans [17]. It is of value in predicting the clinical characteristics and prognostic factors of CRC patients [18]. Current research has demonstrated that LncRNA can serve as disease markers. Studies indicate that LINC01089 has tumor-suppressive effects in various cancers, but its role in hepatocellular carcinoma shows significant differences across different studies. In-depth research suggests that LINC01089 has the potential to be a future biomarker and therapeutic target [19]. In cancers including esophageal cancer, liver cancer, pancreatic cancer, gastric cancer, and colorectal cancer, LncRNA SNHG16 is expected to become a potential biomarker and therapeutic target for human digestive system cancers [20].

In this study, we used the TCGA database to construct the Asian GI cancer-related LncRNAs prognostic risk model. The diagnostic, prognostic, gene aggregation analysis (GSEA) and immune infiltration were evaluated. Three LncRNAs and their drug sensitivity were verified by experiments. It provides a new idea for the study and evaluation of GIC and ferroptosis.

Materials and methods

Data collection and sample collation

RNA sequencing data were downloaded from the TCGA [21] (https://portal.gdc.cancer.gov) database for Asian patients with gastrointestinal cancer. Perl script was used to process transcriptome data and transform it into a symbol to obtain the number of normal samples and tumor samples. Perl script was also used to distinguish mRNA and LncRNA. The TCGA database was also used to collect clinical data on patients with gastrointestinal cancer in Asia.

Identification of ferroptosis-related genes and LncRNAs and their differential expression

From the FerrDb database [22] (http://www.zhounan.org/ferrdb/)download ferroptosis-related genes, LncRNAs and genes linked to ferroptosis were obtained through the R software, as well as in patients with tumor and normal ferroptosis-related genes differentially expressed. The co-expression network of ferroptosis-related genes and LncRNA was constructed by Cytoscape [23] software, in which genes were represented by green polygons and LncRNA was represented by red diamonds. The differential gene lines Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) were then analyzed using the org.hs.eg. db, DOSE, clusterProfiler and enrichPlot packages in R.

Screening of prognostic LncRNAs and construction of risk models

The Survival package of R language was used to find prognostic LncRNA. 0.05 was used as the screening condition for prognostic correlation, and whether HR value was greater than 1 was used to classify high and low risks. A forest map was further used to display the risk classification. The LncRNAs prognostic model was constructed by the following formula: Riskscore = coef(lncRNA1) × expr(lncRNA1) + coef(lncRNA2) × expr(lncRNA2) + … + coef(lncRNAn) × expr(lncRNAn), and the risk score was obtained, in which different risk groups with high and low risk were obtained by comparing with the median risk score.

Validation of risk models

Survival analysis of high and low risk groups was presented by survival and survminer package, and risk rows were visualized by pheatmap package. Univariate and multivariate COX regression were performed using R language to evaluate the relationship between clinical factors and risk scores. Survival, survMiner and timeROC packages were used to obtain the area under the ROC curve. 0.7 Indicates that the model can predict patient survival with high accuracy.

Construction of Nomogram and clinical heatmap

Survival and regplot packages were used for the prediction of the nomogram, and comprehensive scores were obtained through various clinical scores to obtain the 1-year, 3-year and 5-year survival probability of the patients. Limma and pheatmap packages were used to construct clinically relevant heat maps, which were used to intuitively understand the expression of LncRNA in high and low risk groups. High risk was shown in red and low risk was shown in blue, while high expression was represented in red and low expression was represented in green. At the same time, it can also understand the differential expression of clinical traits in the high and low risk groups. when pvalue < 0.001 is "***", when pvalue < 0.01 is "**" when pvalue < 0.05 is "*".

Gene set enrichment analysis

To understand the enrichment of different pathways between high and low risk groups, GSEA software [24] was used for analysis, and the filtering condition was FDR < 0.05, selecting the top 100 pathways with significant enrichment.

Immunization analysis

TIMER2.0 (http://timer.cistrome.org) [25] was used to download the results of immune cell infiltration, and R was used to analyze the correlation of immune cells, so as to obtain the expression of immune cells predicted by different software in high and low risk groups. Limma, GSVA, GSEABase, ggpubr and reshape2 were used to analyze the difference of immune function. 0.05 Differences in expression were determined in high and low risk groups. Using Limma, reshape2, ggplot2 and ggpubr to analyze the difference of immune checkpoint genes, the expression of immune checkpoint related genes in the high and low risk groups can be obtained, and the differential expression icon can be obtained according to the P-value, when P < 0.001 is indicated by "***", When P < 0.01 is "**", when P < 0.05 is "*".

Analysis of differences in paired samples

By using the RComplexHeatmap package to analyze the differential expression of genes in paired samples of LIHC, STAD, ESCA, PAAD, CRC, and CHOL.

Quantitative real‑time PCR (qRT‑PCR)

Total RNA of HIEC, Caco-2, Lovo, SW620, HT29, L-O2, HepG-2, CM-H048, and MKN-45 was extracted by Trizol Reagent (Vazyme, Nanjing, China). The primers are as follows: Human LINC01843, 5'-CTGGAAGAGGCAAGGCAGGATC-3' (Forward) and 5'-CACAGGCAGGGTGGCTTACAG-3' (Reverse); Human LINC01943, 5'-ACAGGAAGCGTGAGGACAGAATC-3' (Forward) and 5'-ATCCAACCAGACTGATGCCACAG-3' (Reverse); Human ZNF667-AS1, 5'-GGTGACCGTTGCGTAATTGTGAG-3' (Forward) and 5'- TCTGTGCTCTCCTGCGATCATTC-3' (Reverse); The relative expression levels of the three were calculated by 2-△△Ct.

Drug sensitivity analysis

On the basis of cancer drugs sensitivity genomics GDSC database (https://www.cancerrxgene.org/), The drug sensitivity of LINC01843, LINC01943 and ZNF667-AS1 against 251 antitumor agents were predicted by R packet "prorophetic.

Statistical analysis

All statistical analyses and visualizations were processed in R-4.1.0. Differences between groups were analyzed using the Student T-test or the Wilcoxon test. The standard of statistical significance was P < 0.05.

Results

Identification of genes associated with ferroptosis in Asians

This research flow chart was shown in Fig. 1. Transcriptome data were downloaded from 297 Asian patients with gastrointestinal cancer, including 158 LIHC, 74 STAD, 38 ESCA, 11 PAAD, 11 COAD, 3 CHOL, and 1 READ. According to the processing of transcript data by Perl script, the dataset included 14 normal samples, 298 tumor samples and 16,773 LncRNA samples. Clinical data of 322 Asian patients with gastrointestinal cancer were downloaded from TCGA, including 161 LIHC patients, 89 STAD patients, 46 ESCA patients, 11 PAAD patients and 11 COAD patients, 3 CHOL patients, and 1 READ patient.

Fig. 1.

Fig. 1

Open in a new tab

Study flowchart

Differentially expressed genes and LncRNAs and their enrichment analysis

123 ferroptosis-related markers, 109 ferroptosis-related suppressors and 150 ferroptosis-related drivers were downloaded from the FerrDb database. 248 ferroptosis-related genes and 2246 ferroptosis-related LncRNAs were obtained by the R package. Further, according to the corresponding R package, there were 53 differentially expressed ferroptosis-related genes in normal and tumor tissues (Table 1), and 409 differentially expressed ferroptosis-related LncRNAs. The relationship between lncRNAs and ferroptosis-related genes were shown in Fig. 2, where the red quadrilateral represented LncRNA and the green octagon represented ferroptosis-related genes. After GO and KEGG enrichment analysis of differentially expressed genes related to ferroptosis. 529 GO-related items were obtained, including 493 biological processes (BP), 3 cellular components (CC) and 33 molecular functions (MF). Barplot was used to display the top 10 items of GO enrichment respectively, as shown in Fig. 3A, and the top 30 items of 52 KEGG enrichment were visualized, as shown in Fig. 3B.

Table 1.

Differential ferroptosis-related genes in normal and adjacent tumor tissues

Gene logFC P value fdr
SLC7A11 4.67470575747032 8.76547591801238e−06 7.62215297218468e−05
AKR1C3 1.85554031779979 0.000273463693684769 0.00105178343724911
SQSTM1 1.49514164944303 0.000455170822207451 0.00162561007931233
NQO1 1.90720173010804 0.00106631170358776 0.00349610394618938
SLC3A2 1.2062569176576 1.803329844707e−05 0.000124367575497034
MT1G − 1.40354910331061 0.00824462966562774 0.019630070632447
FANCD2 2.89808588330887 3.78685612973638e−08 2.52457075315759e−06
HSPA5 1.33381357784008 9.1481266289265e−08 2.70230003074075e−06
TP53 1.00586298240954 0.023452151289741 0.046904302579482
HELLS 3.40110204205735 1.08091689856809e−07 2.70230003074075e−06
SCD 2.01018867860725 0.00208109513215446 0.00621222427508794
AIFM2 1.53085090136131 1.53734520527222e−06 2.19620743610317e−05
ZFP36 − 1.9476422750191 0.000988993253767431 0.00335251950429638
PTGS2 − 1.0376291756756 0.0216584443199141 0.044656586226627
DUSP1 − 1.48059793099784 3.29184138729035e−06 3.34058514202661e−05
TXNRD1 1.74812481603039 2.21068988602371e−05 0.000142625153937014
GPX2 2.38065038878832 0.000137839991981885 0.000610423545236417
ANGPTL7 − 3.76058317812983 0.0103471023837912 0.0232519154691937
DDIT4 1.38052800372618 0.00854456643976247 0.0198710847436337
DDIT3 1.47093741163971 4.41071485963138e−05 0.000245039714423966
SLC1A4 1.83485133072312 3.34058514202661e−06 3.34058514202661e−05
TXNIP − 1.59721378943353 0.00359490479899319 0.00958641279731517
TRIB3 1.918665015188 9.50253819740723e−05 0.000463538448654011
ZFP69B 1.60687264138479 0.000140397415404376 0.000610423545236417
ATP6V1G2 − 2.58446827235884 0.00925455227190748 0.0212748328089827
VEGFA 1.3868203897332 1.42820486830362e−05 0.00011425638946429
GDF15 1.55701277373916 0.00180330919147052 0.00546457330748642
CEBPG 1.2420962595383 3.38996569790015e−07 6.7799313958003e−06
RGS4 − 1.54307328283145 0.00942029279420206 0.0214097563504592
IL6 − 3.22610337790389 0.0118030029239128 0.0262288953864729
RPL8 1.31565891764299 1.03357749164921e−06 1.72262915274868e−05
TFRC 2.15024073798322 1.53001301133286e−05 0.000117050799233906
SLC2A6 1.00675409768555 0.00102147174396334 0.00340490581321113
HBA1 − 2.33236536746483 0.00355515901037857 0.00958641279731517
PLIN4 − 2.57960217189562 0.000118896770850059 0.000540439867500268
STMN1 2.03713316595408 1.0809200122963e−07 2.70230003074075e−06
RRM2 3.7138280786876 8.65631405096804e−09 8.65631405096804e−07
CAPG 1.08757032431033 0.000267096499128969 0.00104743725148615
AURKA 3.18735553255077 3.42615021244686e−09 6.85230042489372e−07
NOX1 4.9128106855572 7.9397878371894e−06 7.5310386476939e−05
DUOX1 3.35323458966808 0.000863688406222564 0.00297823588352608
G6PD 2.75193639838652 3.43634376627249e−05 0.000196362500929857
PGD 1.39781450109226 0.000106330754951375 0.000506336928339881
ACSL4 2.30164833147306 0.000213164739284053 0.000870060160343074
NRAS 1.22650101145892 2.10930986599247e−06 2.63663733249059e−05
HRAS 1.19756667332714 8.32562998548424e−07 1.51375090645168e−05
SLC1A5 2.72610829303973 0.00366529367694245 0.00964550967616434
ALOX15B 3.09974692602206 0.00448362428048605 0.0116457773519118
BID 1.5449436946537 1.02824519389379e−07 2.70230003074075e−06
CDKN2A 4.63957125582791 3.2307971982756e−07 6.7799313958003e−06
MYB 3.48319173664555 0.00230291059352839 0.00657974455293826
MIOX 4.06810024645654 0.000112326724294059 0.000522449880437484
TAZ 1.20028094255773 6.53496450456539e−08 2.70230003074075e−06
Open in a new tab

Fig. 2.

Fig. 2

Open in a new tab

Co-expression network of ferroptosis-related genes and LncRNAs

Fig. 3.

Fig. 3

Open in a new tab

GO and KEGG enrichment analysis of differential genes associated with ferroptosis. A GO enrichment analysis. B KEGG pathway enrichment analysis

Prognostic LncRNAs and risk models

According to the established screening criteria, 48 ferroptosis -related LncRNAs associated with prognosis were obtained by Cox regression analysis and presented in the form of forest map, as shown in Fig. 4, where red represents high risk LncRNAs and blue represents low risk LncRNAs. According to multivariate Cox regression, 11 LncRNAs were found to be the best prognostic genes (Table 2). The risk score for each Asian patient with GI cancer was calculated as follows: 0.0340959849827028 × EXPR (LINC01843) + COEF (LNC RNA2) × EXPR (AC092535.5) + (−1.68400825548866) × EXPR (AL355482.1)0.186822843824712 × expR (AP003469.2) + 0.186822843824712 × expR (AC008808.2) + 0.210239846807106 × expR (AC139100.2) + 0.210239846807106 × expR (AC145207.8) + 0.899006832361096 × expR (AC005479.2) + 0.0374673701777225 × expR (AL365181.3) + Expr (LINC01943) + expr(znf667-AS1) × expR (LINC01943) + expr(znf667-AS1). Patients were divided into high—and low-risk groups based on median risk scores.

Fig. 4.

Fig. 4

Open in a new tab

Univariate Cox regression analysis of independent prognostic factors of forest map

Table 2.

Ferroptosis LncRNA associated with significantly prognosis

Id Coef HR HR.95L HR.95H pvalue
LINC01843 0.0340959849827028 1.03468391607973 1.0172341615988 1.0524330057019 8.53013322235115e−05
AC092535.5 0.0291663349934633 1.02959583804983 1.00020341103757 1.0598520041337 0.0484127135699427
AL355482.1 − 1.68400825548866 0.185628436681279 0.0446181143486725 0.772285360054908 0.0206008388419795
AP003469.2 0.186822843824712 1.20541371964906 1.00111748785373 1.45140031329716 0.0486379808111571
AC008808.2 0.204321014705805 1.22669187638922 0.975985993257392 1.54179769996192 0.0798449286361249
AC139100.2 0.210239846807106 1.23397398918784 1.00410015825037 1.51647402251725 0.0456227718331664
AC145207.8 0.358246959531854 1.4308189312248 1.06250214913941 1.92681286866994 0.018313504885865
AC005479.2 0.899006832361096 2.45716152559498 0.928378197577884 6.50343015229819 0.0702469634622197
AL365181.3 0.0374673701777225 1.03817812096156 1.00079490720387 1.07695772938593 0.045238434906186
LINC01943 0.34990726953832 1.41893596390537 0.937118872422416 2.14847798813352 0.0983040741188328
ZNF667-AS1 0.310523299251742 1.36413878018838 0.958374504713199 1.94169878524755 0.0847213417444212
Open in a new tab

Evaluation results of multiple risk models

According to the survival analysis, the P < 0.001 of the high-low risk group was found, as shown in Fig. 5A, indicating that there was a difference in survival between the high-low risk groups. The establishment of the model could distinguish patients in the high-low risk group. With the increase of risk score, it was known that the risk score in the high-risk group was higher according to the risk score graph (Fig. 5B), and the number of patients dying was increased according to the survival status graph (Fig. 5C), the expression of LncRNAs in the high-risk group was increased according to the risk heat map, and the expression of LncRNAs in low-risk group was decreased according to the risk map (Fig. 5D), Overall high risk in red and low risk in green. Univariate and multivariate COX regression analyses were performed, as shown in Fig. 7. In univariate Cox (Fig. 5E), the Hazard ratio (HR) was 1.215(95% CI 1.150–1.284) and in multivariate Cox regression was 1.255(95% CI 1.177–1.336) (Fig. 5F), with a Pvalue < 0.001 for both. It was concluded that 11 ferroptosis-related LncRNAs could independently predict the prognosis of patients independently of the clinical prognostic factors of GI cancer: age, gender, grade and stage. According to the ROC curve, the area under the curve (AUC) in predicting patients 1 and 2 were 0.736, 0.766 and 0.772 respectively (Fig. 5G), indicating that the model had high accuracy in predicting patient survival. By comparing the model with other clinical traits, the maximum AUC of risk was 0.736(Fig. 5H), indicating that the survival of patients predicted by the model was optimal.

Fig. 5.

Fig. 5

Open in a new tab

Predictive efficacy of Asian GI cancer prognostic models for ferroptosis-related LncRNAs. A Survival analysis. B Risk score distribution. C Scatterplot of the survival state of each sample. D Heat map of expression of 11 ferroptosis-related LncRNAs. E Univariate Cox regression analysis of clinical indicators. F Multivariate Cox regression analysis of clinical indicators. G The 1-, 3-, and 5-years ROC curve for risk model. H ROC curves for clinical indicators and risk scores. GI gastrointestina, ROC receiver operating characteristic

Fig. 7.

Fig. 7

Open in a new tab

KEGG pathway enrichment analysis of ferroptosis-related LncRNAs predicted by GSEA. A–L Different enrichment pathways

Clinical nomogram and heatmap

According to the score of clinical traits and ferroptosis-related LncRNA model corresponding to the nomogram, as shown in Fig. 6A, the 1-year, 3-year and 5-year survival probability of patients could be known, which were convenient for guiding clinical prognosis. According to the clinically relevant heatmap, the clinical traits of T, N and Stage were significantly different between the high and low risk groups (p < 0.001), as shown in Fig. 6B.

Fig. 6.

Fig. 6

Open in a new tab

Clinical features and risk-related survivals. A Clinical nomogram predict 1 -, 3 -, and 5-year survival at different risks. B Heat map of clinical features and ferroptosis-related LncRNAs (***p < 0.001; **p < 0.01; *p < 0.05)

GSEA enrichment

KEGG-related GSEA enrichment analysis was performed with FDR less than 0.25 and Nominal P-value less than 0.001 as screening conditions. No pathway meet the conditions was found in the high-risk group, while 12 related pathways were actively expressed in the low-risk group, as shown in Fig. 7. They are Butanoate metabolism, Retinol metabolism, Tryptophan metabolism, Drug metabolism-cytochrome P450, PPAR signaling pathway, Valine, leucine and isoleucine degradation, Arginine and proline metabolism, Propanoate metabolism, Peroxisome, Fatty acid metabolism, Glycine, serine and threonine metabolism, and Steroid hormone biosynthesis.

Immune status in the high-low GI risk group

In order to understand the immune status of patients in the high and low risk groups. The immune cell response heatmaps of the high and low risk groups were obtained based on the analysis of TIMER, CIBERSORT, CIBERSORT-ABS, QUANTISED, MCPCOUNTER, XCELL and EPIC software, as shown in Fig. 8A. By analyzing differences in immune function, results showed that the immune-related functions of APC_co_inhibition, CCR, MHC_class_I and Parainflammation differed between the high and low risk groups, where the P-value of APC_co_inhibition was less than 0.01. P-values of the remaining three immune functions were less than 0.001, and all four immune-related functions were actively expressed in the high-risk group, as shown in Fig. 8B. The results of immune checkpoint difference analysis showed that, as shown in Fig. 8C, all immune checkpoint genes were different between the high and low risk groups, where the P-values of IDO2 and KIR3DL1 ranged from 0.01 to 0.05, and the pvalues of TNFRSF14 and TMIGD2 ranged from 0.001 to 0.01. The P-values of the remaining immune checkpoints were all less than 0.01, indicating that the immune checkpoints were significantly different between the high and low risk groups.

Fig. 8.

Fig. 8

Open in a new tab

Immune status of Asian GI cancer patients in different risk groups. A Immune heat maps of high and low risk groups under different algorithms. B Box plot of 13 immune-related functions. C Map of 42 immune checkpoints in patients at high and low risk

The expression level of LncRNAs and drug sensitivity analysis

The paired sample analysis results indicate that in LIHC, compared to the normal group, LINC01843 and LINC01943 are highly expressed in tumors, while ZNF667-AS1 expression is reduced, as shown in Fig. 9A. In STAD, compared to the normal group, LINC01843 and LINC01943 are highly expressed in tumors, while ZNF667-AS1 expression is reduced, as shown in Fig. 9B. In ESCA, compared to the normal group, LINC01843 and LINC01943 are highly expressed in tumors, while ZNF667-AS1 expression is reduced, as shown in Fig. 9C. In PAAD, compared to the normal group, LINC01843 is highly expressed in tumors, while LINC01943 and ZNF667-AS1 expression is reduced, as shown in Fig. 9D. In CRC, compared to the normal group, the expression of LINC01843, LINC01943, and ZNF667-AS1 is increased in tumors, as shown in Fig. 9E. In CHOL, compared to the normal group, the expression of LINC01843, LINC01943, and ZNF667-AS1 is increased in tumors, as shown in Fig. 9F.

Fig. 9.

Fig. 9

Open in a new tab

Pairing differential expression of LINC01843, LINC01943 and ZNF667-AS1 and mRNA levels in multiple cells in different GI cancers. A Paired differential expression in LIHCs. B Paired differential expression in STADs. C Paired differential expression in ESCAs. D Paired differential expression in PAADs. E Paired differential expression in CRCs. F Paired differential expression in CHOLs. G mRNA expression levels of 3 lncrnas in colorectal cells; H mRNA expression levels of 3 LncRNAs in hepatocellular carcinoma; I mRNA expression levels of 3 LncRNAs in gastric cancer

The PCR results show that compared to normal intestinal epithelial cells (HIEC), the mRNA expression levels of three LncRNAs are significantly elevated in colorectal cancer cells (such as Caco-2, Lovo, SW620, and HT29), as shown in Fig. 9G. Compared to normal liver cells (L-O2), liver cancer cells (HepG-2) exhibit high expression of LINC01843 and LINC01943, while the expression of ZNF667-AS1 is reduced, as shown in Fig. 9H. Compared to human gastric mucosal epithelial cells (CM-H048), gastric cancer cells (MKN-45) show high expression of LINC01843 and LINC01943, with reduced expression of ZNF667-AS1, as shown in Fig. 9I.

Drug sensitivity results showed that the drug sensitivity of the above three lncrnas ranked the top four. In LINC01843, the first four digits are IGF1R_3801 (p = 2.2e−0.9), JQ1 (p = 4.2e−14), PF-4708671 (p = 6.7e−13), and Selumetinib (p = 5.6e−08), respectively, as shown in Fig. 10A. In LINC01943, the first four places were AZ960 (p = 2.2e−0.9), Dasatinib (p = 4.2e−14), Entospletinib (p = 6.7e−13), XAV939 (p = 5.6e−08), as shown in Fig. 10B. In ZNF667-AS1, the first four digits are BMS-754807 (p = 2.2e−9.0), Crizotinib (p = 4.2e−14), JQ1 (p = 6.7e−13), and MK-1775 (p = 5.6e−08), respectively, as shown in Fig. 10C.

Fig. 10.

Fig. 10

Open in a new tab

Drug sensitivity of LINC01843, LINC01943 and ZNF667-AS1. A Drug sensitivity of LINC01843. B Drug sensitivity of LINC01943. C Drug sensitivity of ZNF667-AS1

Discussion

Asians rank first in cancer incidence, and gastrointestinal cancer is the most common malignant tumor in the Asian population [1]. TME also plays an important role in the occurrence and development of GI cancer [26]. Therefore, it is very necessary to find a valuable prognostic model to accurately predict prognosis, clinical features and immune status. Ferroptosis is different from other PCD death modes such as autophagy and necrotizing apoptosis. Cancer cells with RAS mutation are highly sensitive to Iron death induced by Erastin or RSL3, resulting in cancer disease and iron death accumulated in iron-dependent ROS [27]. Ferroptosis can cause the disorder of anti-tumor immune function and the polarization of pretumor immune cells, and produce a microenvironment conducive to tumor growth [28]. Studies have shown that lncRNAs affect the occurrence and development of malignant tumors by modifying the process of normal cell division, growth and differentiation or by being influenced by metabolites of oncogenes and tumor suppressor genes [29]. In GI cancer, LncRNA inhibits or promotes tumor occurrence by affecting chromatin modification, translation, regulating transcription and acting as a sponge-mirRNA. In addition, LncRNA's open reading frame can encode small peptides in GI cancer to regulate its occurrence and development [30]. Abnormal expression is associated with the occurrence, progression, invasion and overall survival of GI tumors [31]. In the process of tumor occurrence and metastasis, LncRNAs also play an important role in influencing immune function and then TME [32].

Recent studies have found that Long Non-coding RNA-ZFAS1 exhibits both pro-cancer and anti-cancer dual functions in various human malignancies. It can play a key role in influencing tumor progression, metastasis, invasion, apoptosis, cell cycle regulation, and drug resistance, and due to its long half-life, it demonstrates exceptional stability, making it a potential biomarker [33]. It has also been discovered that the abnormal expression of LncRNA ROR1-AS1 can affect cell growth, proliferation, invasion, and metastasis, increasing tumor occurrence and spread, and is associated with liver cancer, colon cancer, osteosarcoma, glioma, cervical cancer, bladder cancer, lung adenocarcinoma, and mantle cell lymphoma, serving as a marker in tumors or a potential therapeutic target for various cancers [34]. In previous studies, prognostic risk models based on ferroptosis-associated LncRNAs have mainly focused on individual types of cancers, such as colorectal cancer (CRC), hepatocellular carcinoma (HCC), and gastric cancer (GC). These studies have provided valuable insights into disease diagnosis, treatment, immune-related factors, and prognostic assessments. However, this study is the first to introduce an innovative approach by considering gastrointestinal cancers (GIC) as a unified entity, exploring the potential roles of ferroptosis-associated LncRNAs within this group. This research not only addresses the interrelationships between different gastrointestinal tumor types but also incorporates the specific genetic background and environmental factors of Asian populations, revealing the clinical prognostic value of ferroptosis-associated LncRNAs across the entire gastrointestinal cancer spectrum. This strategy offers a novel perspective for precision medicine in GICs and lays the foundation for further investigation into the mechanisms of ferroptosis in tumorigenesis and progression.

In this study, 48 prognostic LncRNAs of ferroptosis were identified by COX analysis. Further analysis obtained 11 optimal risk models for ferroptosis -related LncRNAs, LINC01843, AC092535.5, AL355482.1, AP003469.2, AC008808.2, AC139100.2, AC145207.8, AC005479.2, AL365181.3, LINC01943, and ZNF667 − AS1 respectively. The down-regulation of ZNF667 − AS1 expression in esophageal carcinoma cells and esophageal squamous cell carcinoma (ESCC) tissues may be related to the influence of hypermethylation of the CpG site in the proximal promoter on its transcription and expression. Overexpression of ZNF667-AS1 reduced the growth, invasion, and migration of esophageal cancer cells, and the possible mechanism of action was that ZNF667-AS1 inhibited the expression of mir-1290-mediated PRUNE2 to achieve cancer inhibition [35]. ZNF667 − AS1 affects the survival of ESCC patients [36]. ZNF667 − AS1 is down-regulated in CRC tissues, and overexpression of ZNF667-AS1 in VOLO cells inhibits the proliferation, invasion and migration of cancer cells by promoting ANK2 and inhibiting the expression of JAK2. In addition, ZNF667-AS1 is associated with the occurrence, development and prognosis of CRC [37]. In addition, ZNF667-AS1 is closely associated with the diagnosis and survival of glioma, chronic lymphocytic leukemia and other tumors [38, 39]. This study also confirmed the high expression of ZNF667-AS1 in multiple colorectal cancer cells from the mRNA level. At present, the remaining 10 ferroptosis-related LncRNAs (LINC01843, AC092535.5, AL355482.1, AP003469.2, AC008808.2, AC139100.2, AC145207.8, AC005479.2, AL365181.3, LINC01943,) have not been reported in tumors, suggesting the need for further exploration in our study.

In this study, we observed significant differential expression of ferroptosis-associated long non-coding RNAs (LncRNAs) across various gastrointestinal and liver cancers. Specifically, LINC01843 and LINC01943 were consistently upregulated, while ZNF667-AS1 was downregulated in hepatocellular carcinoma, gastric cancer, esophageal cancer, and pancreatic cancer compared to normal tissues. These findings suggest that the dysregulation of these LncRNAs may play a critical role in the tumorigenesis of these cancers. Furthermore, in colorectal cancer and cholangiocarcinoma, the expression levels of all three LncRNAs were notably elevated in tumor tissues, highlighting their potential as biomarkers for these malignancies. PCR validation in cell lines further confirmed the upregulation of LINC01843 and LINC01943, along with the downregulation of ZNF667-AS1, in colorectal, liver, and gastric cancer cells compared to normal cell types. These results reinforce the potential of these LncRNAs as therapeutic targets and prognostic indicators. Given their widespread dysregulation across different cancer types, these LncRNAs may offer novel insights into the molecular mechanisms underlying tumor progression and the role of ferroptosis in cancer biology. Moreover, their potential as biomarkers could lead to improved diagnostic and therapeutic strategies for gastrointestinal and liver cancers, thus contributing to personalized treatment approaches in the clinic. Moreover, the three LncRNAs were significantly sensitive to multiple drugs. KM survival curve analysis was performed on the constructed prognostic model, which indicated that the model predicted better survival. The ROC curve indicated that the prognostic model had high accuracy in predicting the 1, 2, and 3-year survival rate and was superior to other clinical factors. COX regression results of the prognostic analysis suggested that the model could be used as an independent prognostic factor.

KEGG-related pathways were analyzed by GSEA to obtain pathways and functional mechanisms related to the difference between high and low risk prognosis. The related metabolic pathways include Butanoate, Retinol, Tryptophan, Cytochrome P450 drug, Arginine, Proline, Propanoate, Fatty acid, Glycine, Serine and Threonine metabolism were abundant in low-risk group. And PPAR signaling pathway, Valine, Leucine and isoleucine degradation, Peroxisome and Steroid hormone biosynthesis. Several metabolic pathways are associated with tumor and ferroptosis, such as increased butyric acid metabolism in CRC detected by metagenomic sequencing in colorectal cancer [40]. Retinol can regulate cell growth, differentiation and apoptosis, and single nucleotide polymorphisms (SNPs) in its metabolic pathway genes are associated with tumor genesis including pancreatic cancer [41]. Tryptophan metabolic pathway and its gene pair correlation coefficients were biomarkers for diagnosis and prognosis of gastrointestinal tumors, such as BLCA, COAD, ESCA, LIHC and STAD [42]. Abnormal arginine metabolism is a characteristic of tumor cell metabolism. Arginine is not only necessary for tumor cells, but also participates in tumor metabolisms such as nucleotide, nitric oxide, glutamate, proline and polyamine. Enzymes and transport molecules produced in the metabolism process are also involved in tumor progression [43]. Proline metabolism affects the production of ATP, redox homeostasis and protein synthesis of cancer cells, which further affects the proliferation, invasion and metastasis of cancer cells. Moreover, because proline is mainly stored in collagen, it affects the plasticity of cancer cells [44]. Fatty acid metabolism pathway provides energy for the growth of tumor cells. Fatty acid metabolism is mediated by pI3K-AKT-mTOR and other carcinogenic signaling pathways as well as molecular heterogeneity and can reshape the tumor microenvironment to affect its occurrence, development and metastasis. Because it can be blocked by drugs to inhibit tumor progression, it can be used as a tumor treatment target [45, 46]. Ferroptosis is sensitive to biological factors including fatty acid metabolism, amino acids, iron processing, mitochondrial respiration and glutathione processes [47, 48]. Abnormal serine/glycine metabolism promotes tumor cell survival and proliferation [49]. The peroxisome is a cellular redox balance and lipid metabolism organelle, which plays an important role in bile acid synthesis, fatty acid oxidation, ether phospholipid synthesis and reactive oxygen balance. When peroxisome changes abnormally, it leads to a variety of metabolic diseases including tumors [50]. When the peroxisome gene is deleted or overexpressed, the occurrence of ferroptosis is positively correlated with its change, and its mechanism is related to peroxisome mediated plasminogen production [51].

Because both ferroptosis and LncRNA are associated with the immune-associated tumor microenvironment, immunotherapy has attracted much attention in tumor-related therapy. Here we analyze the immune infiltration, immune-related functions and immune checkpoints in the risk-prognosis model. Literature study found that NK Cells, DCs, Macrophages, T Lymphocytes subsets including CD8+ T Cells, CD4+ helper T Cells, B lymphocytes, Natural killer T Cells (NKTs), And regulatory T cells are associated with gastrointestinal tumors [52]. It was consistent with our results of immune infiltration, and the results of immune infiltration suggested that the high risk group had a greater degree of immune cell infiltration and stronger tumor immune activity. Immune checkpoint-related inhibitors (ICIs) are one of the most significant methods in current tumor immunotherapy, so finding tumor immune checkpoints have become the primary condition for its treatment. Current studies have found that immunosuppressant related therapies have made significant breakthroughs in advanced gastric cancer, esophageal cancer and defect mismatch repair (dMMR)CRC [53]. It is feasible and valuable to find immune checkpoints in GI cancer. Our immune checkpoint results suggested that except for IDO2 and ADORA2A, the expression levels of immune checkpoints in other high-risk groups were lower than those in low-risk groups, suggesting that high-risk groups have a better effect on immunotherapy in our model.

In this study, we integrated and validated the results of immune infiltration analysis with the KEGG and GSEA pathway enrichment findings to explore the complex biological mechanisms underlying prognosis in gastrointestinal cancers. KEGG and GSEA analyses revealed that several key metabolic pathways—such as Butanoate, Retinol, Tryptophan, and Fatty acid metabolism—were enriched in the low-risk group, highlighting their roles in regulating cell differentiation, apoptosis, and energy production. Dysregulation of these pathways is strongly linked to cancer progression, with specific pathways like tryptophan metabolism and butyrate metabolism emerging as potential biomarkers for diagnosis and prognosis in gastrointestinal tumors. Concurrently, immune infiltration analysis showed that the high-risk group had a greater degree of immune cell infiltration, including NK cells, macrophages, T lymphocyte subsets, and regulatory T cells, aligning with existing literature on the active immune response in high-risk tumors. This suggests that the high-risk group may exhibit stronger tumor immune activity, potentially making them more responsive to immunotherapeutic strategies. Notably, we found that immune checkpoint expression was generally lower in the high-risk group, except for IDO2 and ADORA2A, suggesting a better therapeutic response to immune checkpoint inhibitors (ICIs) in these patients. By combining immune and metabolic pathway data, our study offers a deeper understanding of the tumor microenvironment and the interplay between immune responses and metabolic reprogramming in gastrointestinal cancers. These findings underscore the potential for personalized therapies that target both immune checkpoints and metabolic pathways, particularly in high-risk gastrointestinal cancer patients, paving the way for more effective treatment strategies.

While previous research has identified ferroptosis-associated LncRNAs and their prognostic value in specific cancer types such as colorectal, liver, and gastric cancers, our study extends this analysis to encompass a broader spectrum of gastrointestinal cancers. By analyzing the ferroptosis-related LncRNAs, we reveal shared molecular pathways and mechanisms that govern tumor progression in these malignancies. Specifically, we found common dysregulation of key metabolic pathways such as fatty acid, arginine, and tryptophan metabolism, which are closely associated with ferroptosis. These metabolic alterations influence tumorigenesis, immune responses, and the tumor microenvironment across various gastrointestinal cancers. Additionally, the integration of immune infiltration analysis further highlights the complex interplay between ferroptosis and the immune system in shaping the prognosis and therapeutic response in gastrointestinal tumors. By considering the collective mechanisms of ferroptosis in the context of all gastrointestinal cancers, our study offers new insights into the shared biological processes underlying these malignancies and their potential for targeted therapeutic interventions.

However, our study still has its limitations. Our data is from the TCGA database of Asian population information, but the sample size is limited, so it is necessary to further verify our research results in a large sample. The prognosis and clinical analysis of 11 ferroptosis LncRNAS-related genes obtained is a retrospective study prone to biased results, which need to be further verified by prospective studies. In addition, the obtained data need line experiments and relevant functional studies to verify the correctness of the results and explore the unknown mechanism.

Conclusion

In summary, we constructed prognostic detection models for 11 ferroptosis-related LncRNAs in GI cancer and divided them into a high-risk group and low-risk group. This model can independently predict the clinical characteristics and prognosis of GI cancer patients and can reflect the characteristics of immune-related functions. Therefore, our study has guiding significance for clarifying biomarkers, prognostic features and the design of immunotherapy regimens in Asian GI cancer.

Acknowledgements

We thank all the authors for their contributions to this study.

Abbreviations

GI

Gastrointestinal

TCGA

The Cancer Genome Atlas

LIHC

Liver Hepatocellular Carcinoma

STAD

Stomach Adenocarcinoma

ESCA

Esophageal Carcinoma

PAAD

Pancreatic Adenocarcinoma

COAD

Colon Adenocarcinoma

CHOL

Cholangiocarcinoma

READ

Rectum Adenocarcinoma

CRC

Colorectal Cancer

HCC

Hepatocellular Carcinoma

GC

Gastric Cancer

ESCC

Esophageal Squamous Cell Carcinoma

TME

Tumor Microenvironment

PCD

Programmed Cell Death

FerrDb

Ferroptosis Database

DEGs

Differentially Expressed Genes

qRT-PCR

Quantitative Reverse Transcription Polymerase Chain Reaction

ROS

Reactive Oxygen Species

LncRNA

Long Non-Coding RNA

COX

Cox Proportional Hazards

SNP

Single Nucleotide Polymorphism

KM

Kaplan–Meier

ROC

Receiver Operating Characteristic Curve

AUC

Area Under the Curve

GSEA

Gene Set Enrichment Analysis

KEGG

Kyoto Encyclopedia of Genes and Genomes

NK

Natural Killer Cells

DC

Dendritic Cells

T cells

T Lymphocytes

B cells

B Lymphocytes

NKT

Natural Killer T Cells

ICI

Immune Checkpoint Inhibitors

dMMR

Defective Mismatch Repair

HR

Hazard Ratio

FDR

False Discovery Rate

TIMER

Tumor Immune Estimation Resource

CIBERSORT

Cell-type Identification By Estimating Relative Subsets of RNA Transcripts

GSVA

Gene Set Variation Analysis

GSEABase

Gene Set Enrichment Analysis Base

IGF1R

Insulin-like Growth Factor 1 Receptor

JQ1

A small-molecule inhibitor targeting BET bromodomains

PF-4708671

A specific inhibitor of Akt1

Selumetinib

A selective MEK inhibitor

AZ960

A specific inhibitor of Aurora A kinase

Dasatinib

A drug used to treat cancer, particularly leukemia

Entospletinib

A potent inhibitor of SYK

XAV939

A selective tankyrase inhibitor

BMS-754807

A small-molecule insulin-like growth factor receptor (IGF-IR) inhibitor

Crizotinib

An inhibitor of ALK and ROS1 kinases

MK-1775

A specific inhibitor of WEE1 kinase

Author contributions

Ning Ding, Yingjie Zhang, and Yongheng He conceived and designed the study; Yingjie Zhang, Yongheng He, and Ning Ding provided administrative support; Ning Ding, Ying Kang, and Xiaoxiao Tan supplied study materials or patients; Ying Kang and Yanbo Tang collected and assembled data; Ying Kang and Xiaoxiao Tan performed data analysis and interpretation; all authors wrote the manuscript.

Funding

This work was supported in part by the National Natural Science Foundation of China (NO.82374462), and the National Funded Postdoctoral Researcher Program (NO.GZC20230773).

Data availability

The datasets analyzed during the current study are available from the Cancer Genome Atlas (TCGA) database (https://portal.gdc.cancer.gov/proje cts/TCGA-LIHC, STAD, ESCA, PAAD, COAD, CHOL, and READ), the FerrDb database (http://www.zhounan.org/ferrdb/) and the drugs sensitivity genomics GDSC database (https://www.cancerrxgene.org/).

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Contributor Information

Yingjie Zhang, Email: yingjiezhang@hnu.edu.cn.

Yongheng He, Email: 2320990685@qq.com.

References

  • 1.Nierengarten MB. Global cancer statistics 2022: the report offers a view on disparities in the incidence and mortality of cancer by sex and region worldwide and on the areas needing attention. Cancer. 2024;130:2568. [DOI] [PubMed] [Google Scholar]
  • 2.Krause A, Stocker G, Gockel I, et al. Guideline adherence and implementation of tumor board therapy recommendations for patients with gastrointestinal cancer. J Cancer Res Clin Oncol 2022. [DOI] [PMC free article] [PubMed]
  • 3.Tang D, Chen X, Kang R, et al. Ferroptosis: molecular mechanisms and health implications. Cell Res. 2021;31:107–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Pei Z, Qin Y, Fu X, et al. Inhibition of ferroptosis and iron accumulation alleviates pulmonary fibrosis in a bleomycin model. Redox Biol. 2022;57:102509. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Wang H, Lin D, Yu Q, et al. A Promising future of ferroptosis in tumor therapy. Front Cell Dev Biol. 2021;9: 629150. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Wang H, Cheng Y, Mao C, et al. Emerging mechanisms and targeted therapy of ferroptosis in cancer. Mol Ther. 2021;29:2185–208. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Zeng Q, Ma X, Song Y, Chen Q, Jiao Q. Targeting regulated cell death in tumor nanomedicines. Theranostics. 2022;12:817–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Liu L, Yao H, Zhou X, et al. MiR-15a-3p regulates ferroptosis via targeting glutathione peroxidase GPX4 in colorectal cancer. Mol Carcinog. 2021;61:301–10. [DOI] [PubMed] [Google Scholar]
  • 9.Tian X, Li S, Ge G. Apatinib promotes ferroptosis in colorectal cancer cells by targeting ELOVL6/ACSL4 signaling. Cancer Manag Res. 2021;13:1333–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Luo WQ, Dai WX, Li QG, et al. Ferroptosis-associated molecular classification characterized by distinct tumor microenvironment profiles in colorectal cancer. Int J Biol Sci. 2022;18:1773–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Bekric D, Ocker M, Mayr C, et al. Ferroptosis in hepatocellular carcinoma: mechanisms, drug targets and approaches to clinical translation. Cancers (Basel). 2022;14:1826. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Shao YF, Jia HT, Li SC, et al. Comprehensive analysis of ferroptosis-related markers for the clinical and biological value in gastric cancer. Oxid Med Cell Longev. 2021;2021:7007933. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Ma JF, Hu X, Yao YX, et al. Characterization of two ferroptosis subtypes with distinct immune infiltration and gender difference in gastric cancer. Front Nutr. 2021;8: 756193. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Igor U, David PB. lincRNAs: genomics, evolution, and mechanisms. Cell. 2013;154:26–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Bhan A, Soleimani M, Mandal SS. Long noncoding RNA and cancer: a new paradigm. Cancer Res. 2017;77:3965–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Lai DL, Tan L, Zuo XJ, et al. Prognostic ferroptosis-related lncRNA signatures associated with immunotherapy and chemotherapy responses in patients with stomach cancer. Front Genet. 2021;12: 798612. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Huang AM, Li T, Xie XT, et al. Computational identification of immune- and ferroptosis-related LncRNA signature for prognosis of hepatocellular carcinoma. Front Mol Biosci. 2021;8: 759173. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Zhang WQ, Fang DQ, Li SH, et al. Construction and validation of a novel ferroptosis-related lncRNA signature to predict prognosis in colorectal cancer patients. Front Genet. 2021;12: 709329. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Yi Q, Zhu GF, Ouyang XT, et al. LINC01089 in cancer: multifunctional roles and therapeutic implications. J Transl Med. 2024;22:858. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Zhao LJ, Kan YL, Wang L, et al. Roles of long non-coding RNA SNHG16 in human digestive system cancer (Review). Oncol Rep. 2024;52:106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Tomczak K, Czerwińska P, Wiznerowicz M. The cancer genome atlas (TCGA): an immeasurable source of knowledge. Contemp Oncol (Pozn). 2015;19:A68-77. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Zhou N, Bao JK. FerrDb: a manually curated resource for regulators and markers of ferroptosis and ferroptosis-disease associations. Database (Oxford). 2020. 10.1093/database/baaa021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Otasek D, Morris JH, Bouças J, et al. Cytoscape automation: empowering workflow-based network analysis. Genome Biol. 2019;20:185. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Damian D, Gorfine M. Statistical concerns about the GSEA procedure. Nat Genet. 2004;36:663. [DOI] [PubMed] [Google Scholar]
  • 25.Li TW, Fu JX, Zeng ZX, et al. TIMER2.0 for analysis of tumor-infiltrating immune cells. Nucleic Acids Res. 2020;48:W509–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Li YT, Tan XY, Huang LN, et al. Research progress in immunosuppressive tumor microenvironment of gastrointestinal cancer. Sichuan Da Xue Xue Bao Yi Xue Ban. 2022;53:7–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Yu HT, Guo PY, Xie XZ, et al. Ferroptosis, a new form of cell death, and its relationships with tumourous diseases. J Cell Mol Med. 2017;21:648–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Wang Y, Cao XR, Yang CBX, et al. Ferroptosis and immunosenescence in colorectal cancer. Semin Cancer Biol. 2024;106–107:156–65. [DOI] [PubMed] [Google Scholar]
  • 29.Pandey GK, Kanduri C. Long non-coding RNAs: tools for understanding and targeting cancer pathways. Cancers (Basel). 2022;14:4760. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Chen Y, Long WL, Yang LQ, et al. Functional peptides encoded by long non-coding RNAs in gastrointestinal cancer. Front Oncol. 2021;11:777374. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Khajehdehi M, Khalaj-Kondori M, Ghasemi T, et al. Long noncoding RNAs in gastrointestinal cancer: tumor suppression versus tumor promotion. Dig Dis Sci. 2021;2:381–97. [DOI] [PubMed] [Google Scholar]
  • 32.Xu SP, Wang Q, Kang YJ, et al. Long noncoding RNAs control the modulation of immune checkpoint molecules in cancer. Cancer Immunol Res. 2020;8:937–51. [DOI] [PubMed] [Google Scholar]
  • 33.Liu X, Ma Z, Zhang XX, et al. Research progress of long non-coding RNA-ZFAS1 in malignant tumors. Cell Biochem Biophys. 2024. 10.1007/s12013-024-01441-3. [DOI] [PubMed] [Google Scholar]
  • 34.Fan H, Zhou YX, Zhang ZY, et al. ROR1-AS1: a meaningful long noncoding RNA in oncogenesis. Mini Rev Med Chem. 2024;24:1884–93. [DOI] [PubMed] [Google Scholar]
  • 35.Zheng YJ, Liang TS, Wang J, et al. Long non-coding RNA ZNF667-AS1 retards the development of esophageal squamous cell carcinoma via modulation of microRNA-1290-mediated PRUNE2. Transl Oncol. 2022;21: 101371. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Dong ZM, Li SM, Wu X, et al. Aberrant hypermethylation-mediated downregulation of antisense lncRNA ZNF667-AS1 and its sense gene ZNF667 correlate with progression and prognosis of esophageal squamous cell carcinoma. Cell Death Dis. 2019;10:930. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Zhuang L, Ding WB, Ding W, et al. lncRNA ZNF667-AS1 (NR_0365211) inhibits the progression of colorectal cancer via regulating ANK2/JAK2 expression. J Cell Physiol. 2021;236:2178–93. [DOI] [PubMed] [Google Scholar]
  • 38.Esmaiel MA, Mohamed MM, et al. Upregulation of long noncoding RNA Lnc-IRF2–3 and Lnc-ZNF667-AS1 is associated with poor survival in B-chronic lymphocytic leukemia. Int J Lab Hematol. 2020;42:284–91. [DOI] [PubMed] [Google Scholar]
  • 39.Yuan Q, Gao C, Lai XD, et al. Analysis of long noncoding RNA ZNF667-AS1 as a potential biomarker for diagnosis and prognosis of glioma patients. Dis Markers. 2020;2020:8895968. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Liu NN, Jiao N, Tan JC, et al. Multi-kingdom microbiota analyses identify bacterial-fungal interactions and biomarkers of colorectal cancer across cohorts. Nat Microbiol. 2022;7:238–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Cao DL, Meng YX, Li SW, et al. Association study between genetic variants in retinol metabolism pathway genes and prostate cancer risk. Cancer Med. 2020;9:9462–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Sakharkar MK, Dhillon SK, Rajamanickam K, et al. Alteration in gene pair correlations in tryptophan metabolism as a hallmark in cancer diagnosis. Int J Tryptophan Res. 2020;13:1178646920977013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Du T, Han JY. Arginine metabolism and its potential in treatment of colorectal cancer. Front Cell Dev Biol. 2021;9: 658861. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Geng PY, Qin WS, Xu GW. Proline metabolism in cancer. Amino Acids. 2021;53:1769–77. [DOI] [PubMed] [Google Scholar]
  • 45.Jin ZN, Chai YD, Hu S. Fatty acid metabolism and cancer. Adv Exp Med Biol. 2021;1280:231–41. [DOI] [PubMed] [Google Scholar]
  • 46.Luo YT, Wang HB, Liu BR, et al. Fatty acid metabolism and cancer immunotherapy. Curr Oncol Rep. 2022;24:659–70. [DOI] [PubMed] [Google Scholar]
  • 47.Zheng JS, Conrad M. The metabolic underpinnings of ferroptosis. Cell Metab. 2020;32:920–37. [DOI] [PubMed] [Google Scholar]
  • 48.Yang K, Wang XK, Song CH, et al. The role of lipid metabolic reprogramming in tumor microenvironment. Theranostics. 2023;13:1774–808. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Pan SJ, Fan M, Liu ZN, et al. Serine, glycine and one-carbon metabolism in cancer (Review). Int J Oncol. 2021;58:158–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Kim JA. Peroxisome metabolism in cancer. Cells. 2020;9(7):1692. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Tang D, Kroemer G. Peroxisome: the new player in ferroptosis. Signal Transduct Target Ther. 2020;5(1):273. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Abdul-Latif M, Townsend K, Dearman C, et al. Immunotherapy in gastrointestinal cancer: the current scenario and future perspectives. Cancer Treat Rev. 2020;88: 102030. [DOI] [PubMed] [Google Scholar]
  • 53.Zhuo N, Liu C, Zhang Q, et al. Characteristics and prognosis of acquired resistance to immune checkpoint inhibitors in gastrointestinal cancer. JAMA Netw Open. 2022;5(3): e224637. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

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

Data Availability Statement

The datasets analyzed during the current study are available from the Cancer Genome Atlas (TCGA) database (https://portal.gdc.cancer.gov/proje cts/TCGA-LIHC, STAD, ESCA, PAAD, COAD, CHOL, and READ), the FerrDb database (http://www.zhounan.org/ferrdb/) and the drugs sensitivity genomics GDSC database (https://www.cancerrxgene.org/).

Articles from Discover Oncology are provided here courtesy of Springer

RESOURCES