Skip to main content
Download PDF

Multiomic traits reveal that critical irinotecan-related core regulator FSTL3 promotes CRC progression and affects ferroptosis

Cancer Cell International volume 25, Article number: 115 (2025) Cite this article

Abstract

Background

Irinotecan is a widely used chemotherapy drug in colorectal cancer (CRC). The evolution and prognosis of CRC involve complex mechanisms and depend on the drug administered, especially for irinotecan. However, the specific mechanism and prognostic role of irinotecan-related regulators remain to be elucidated.

Methods

Data from public databases were used to explore the multiomic traits of irinotecan-related regulators through bioinformatics analysis. RT‒qPCR, western blotting, transmission electron microscopy and flow cytometry were used as experimental validations.

Results

Iriscore (irinotecan-related score) was constructed based on irinotecan-related regulators, and a high iriscore predicted a poor prognosis, poor therapeutic response and the MSS/MSI-L status. Single-cell analysis revealed that FSTL3 and TMEM98 were mainly expressed in CRC stem cells. Potential transcription factors (E2F1, STAT1, and TTF2) and therapeutic drugs (telatinib) that target irinotecan-related regulators were identified. FSTL3 was the core risk irinotecan-related regulator. Some ferroptosis regulators (GPX4, HSPB1 and RGS4) and related metabolic pathways (lipid oxidation and ROS metabolism) were correlated significantly with FSTL3. In vitro, irinotecan inhibited the expression of FSTL3 and ferroptotic defence proteins (GPX4 and SLC7A11), and induced lipid peroxidation and intracellular Fe (2+) ions concentration increased.

Conclusions

We confirmed that irinotecan-related regulators, especially FSTL3, have effective prognostic value in CRC and speculated that FSTL3 may promote CRC progression and affect ferroptosis, which is beneficial for identifying candidate targeted irinotecan-related regulators and accurate individualized treatment strategies for CRC.

Introduction

Colorectal cancer (CRC) is the third most common cancer and the second leading cause of cancer-related deaths globally. The number of CRC patients is increasing annually worldwide [1]. Despite the rapid development of immunotherapy and targeted therapy, traditional chemotherapy remains the first-line treatment for CRC, particularly metastatic CRC. Drug resistance is the main limiting factor in CRC treatment. The development of drug resistance is a complex process that involves many different mechanisms and signaling pathways and is essentially driven by changes in gene expression [2]. Hence, it is necessary to identify differentially expressed genes (DEGs) related to chemotherapy. Moreover, the analysis of the effects of DEGs on prognosis and drug sensitivity is of great significance for finding new targets for reversing drug resistance and enhancing prognosis.

The camptothecin-derived drug irinotecan is a classical anticancer drug with a broad spectrum of activities, characterized by a multistep and complex pharmacology. It mainly functions as a topoisomerase I (Topo1) inhibitor that triggers cell death by trapping the enzyme in DNA and generating cytotoxic protein-linked DNA breaks [3]. Moreover, studies have shown that irinotecan has an efficient radiosensitizing effect, which is triggered by ATM/CHK/CDC25C/CDC2 signaling, leading to G2/M phase arrest and apoptosis in colorectal cancer cells [4]. Currently, irinotecan is mainly used in combination with fluorouracils, such as FOLFIRI and FOLFIRINOX, to treat metastatic or advanced solid tumors such as colon, gastric, and pancreatic cancers [5,6,7]. Therefore, irinotecan plays a critical role in tumor treatment.

Ferroptosis is a unique cell death mechanism that has attracted great interest in the cancer research community because targeting ferroptosis may provide new therapeutic opportunities for the treatment of cancers that are refractory to conventional therapies [8]. Previous studies have shown that ferroptosis plays a critical role in CRC progression and tumor microenvironment rebalancing [9,10,11,12]. On one hand, ferroptosis is involved in the activities of multiple tumor suppressors, such as p53 and BAP1, which help to inhibit the occurrence and development of tumors [13, 14]. The inhibition of ferroptosis by oncogenes or oncogenic signaling contributes to tumor initiation, progression, metastasis, and therapeutic resistance [15]. On the other hand, several tumor-suppressor and oncogenic signaling pathways have been shown to promote or suppress ferroptosis, including limiting PUFA-PL synthesis and peroxidation, restricting labile iron availability, and upregulating cellular defense systems against ferroptosis [16,17,18]. SLC7A11-GSH-GPX4 is one of the most important ferroptosis defense systems, and its upregulation can promote cancer progression through inhibiting ferroptosis [19,20,21]. However, no study has reported an association between irinotecan and ferroptosis. Although SN38 (the active metabolite of irinotecan) [22] has been confirmed to trigger a reactive oxygen species (ROS) burst synergistically with an autophagy inhibitor, the exact molecular mechanism has not been elucidated [23].

In this study, irinotecan-related DEGs (IRDEGs) were selected by analyzing gene expression data from the Gene Expression Omnibus (GEO) dataset of CRC cell lines treated with irinotecan, and functional enrichment analysis was conducted. We then constructed an irinotecan-related score (iriscore) model based on the consistent cluster results by combining genes related to the prognosis of CRC in database TCGA with screened IRDEGs. Next, we evaluated the value of the iriscore in terms of prognosis, immunity, tumor mutation burden, and chemotherapeutic response. scRNA-seq was performed to identify the expression signatures of selected IRDEGs in CRC cells. We validated the 18 most important IRDEGs using a random forest analysis and explored their correlations with transcription factors (TFs) and immunity; among them, FSTL3 was the most important IRDEG. We then validated the prognostic value of FSTL3 and constructed prognostic models and corresponding nomograms for CRC based on FSTL3. Finally, we assessed the correlation between FSTL3 and ferroptosis and concluded that irinotecan may trigger ferroptosis by regulating FSTL3 expression. Our study will help to understand the action and resistance mechanisms of irinotecan in ferroptosis and develop an individualized treatment plan for patients with CRC.

Graphical abstracts

Multi-omic traits of irinotecan-related regulators and the prognostic and ferroptosis relevance of FSTL3.

Materials and methods

Downloaded irinotecan-related genes associated with CRC

Irinotecan-related gene expression data was downloaded from the GEO database of the GSE145356 cohort. The R package “limma” (version 4.1.3) was used to obtain DEGs between irinotecan-treated and untreated CRC cell lines with a threshold log2-fold change (FC) > 1 and an adjusted P-value < 0.05.

Functional enrichment analysis of IRDEGs

We conducted gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses using Metascape (http://metascape.org/), a public database that explores the potential functions of the irinotecan-related DEGs (IRDEGs). The 20 most important signaling pathways were selected when the data showed an adjusted P-value of < 0.05.

Identification of IRDEGs associated with prognosis in CRC

We collected transcriptome data and corresponding clinical information associated with CRC from the Cancer Genome Atlas (TCGA) website (https://portal.gdc.cancer.gov/), which includes 383 CRC and 51 normal tissue samples. DEGs between CRC and normal tissues were obtained using the “limma” R package (version 4.1.3), with the screening conditions log2-FC > 1 and P < 0.05. To identify DEGs related to prognosis, we used the Kaplan-Meier “survival” package (version 3.8.3) for univariate Cox regression analysis (P < 0.05).

Identification of molecular subgroups by consistent clustering

We conducted a consistent clustering analysis through the R package “ConsensusClusterPlus” (version 3.16). The TCGA CRC cohort was divided into k clusters (k = 2–9) using the Euclidean squared distance metric and K-means clustering algorithm. A consistent cumulative distribution function (CDF) graph and delta region graph were used to identify the optimal number of clusters [24].

Calculation of the irinotecan-related score (iriscore) based on consistent clustering

Principal component analysis (PCA) was conducted to extract the main components from the 51 feature genes as the signature score for the respective irinotecan-related gene signature. Subsequently, a method similar to the gene expression grade index was adopted to define an iriscore for each patient: iriscore = ∑PCA1A-∑PCA1B.

Comprehensive evaluation of Iriscore from aspects of prognosis, tumor mutation burden (TMB), and immunology

The TCGA CRC (383 total) and the GSE39582 (CRC, 566 total) cohorts were used to evaluate the prognostic value of the iriscore. The median iriscore in each cohort was regarded as the cutoff value. Kaplan–Meier (K–M) curves were used to visualize the differences in overall survival (OS) in different iriscore molecular subgroups. ROC curves were used to assess the sensitivity and specificity of the iriscore for predicting OS. The immunophenoscore (IPS) information of patients from the TCGA CRC cohort was downloaded from http://tcia.at/home. IMvigor210 dataset was downloaded using R package “IMvigor210CoreBiologies” (http://research-pub.gene.com/IMvigor210 CoreBiologies/packageVersions/).

Prediction of chemotherapeutic response based on Iriscore

The chemotherapeutic response of CRC patients in high- and low-iriscore groups was assessed using the R package “pRRophetic” (http://genemed.uchicago.edu/ ~pgeeleher/pRRophetic). The results were identified based on the IC50 values for human CRC cell lines on the Genomics of Drug Sensitivity in Cancer (GDSC) website (https://www.cancerrxgene.org/).

Single-cell RNA-seq data analysis of irinotecan-related regulators

scRNA-seq data from a CRC cohort (GSE188711) of 27,927 single human CRC cells were analyzed to identify the expression signatures of critical irinotecan-related regulators in CRC cells. We used the R package “Seurat” for PCA and ran the t-Distributed Stochastic Neighbor Embedding (tSNE). The cell population was annotated based on the R package “Single R” (version 3.16) and feature markers. Next, the relative contents and locations of critical irinotecan-related regulators were visualized using feature and dot plots.

Identification of critical irinotecan-related regulators by random forest

The “RandomForestSRC” R package (version 3.3.3) was used to select critical irinotecan-related regulators with positive importance for the 51 regulators.

Semiflexible Docking of drugs and FSTL3

We conducted semiflexible docking using the AutoDock Vina software [25] to explore the interaction between telatinib and the most important risk factor, FSTL3. We used Open Babel [26] to transcode the 3D structure of telatinib from the PubChem database (https://pubchem.ncbi.nlm.nih.gov/); 3D structural information of the FSTL3 protein was obtained from the Protein Database (PDB; https://www.rcsb.org). PyMOL was used to visualize the model docking [27].

Exploration of potential drug- and irinotecan-related targets

We downloaded the RNA expression data for 28 predicted drugs targeting irinotecan from the CellMiner database (https://discover.nci.nih.gov/cellminer/home.do). Correlation analysis with the RNA expression of the 18 selected irinotecan-related regulators was conducted based on the mean z-score of activity (DTP NCI-60). We analyzed the target genes of telatinib using SwissTargetPrediction (http://www.swisstargetprediction.ch/) to identify potential targets [28].

Construction and validation of a predictive nomogram

The independence of the FSTL3 signature in predicting the OS of patients with CRC was identified using univariate and multivariate COX analyses. The sensitivity and specificity of the risk score were examined using time-dependent ROC curve analysis. We constructed the nomogram through the “rsm” R package (version 2.10.3) based on the FSTL3 expression and clinical characteristics with prognostic value to predict the survival risk of patients with CRC quantitatively. The predictive accuracies of the nomograms were evaluated using a calibration curve. The survival benefits based on the nomogram were evaluated using decision curve analysis (DCA).

Gene set variation analysis (GSVA)

Gene set variation analysis (GSVA) is a special type of gene set enrichment method for identifying specific gene set scores and performing differential analyses at the gene set (pathway) level [29]. Gene sets associated with ferroptosis were obtained from the Molecular Signatures Database (http://software.broad-institute.org/ gsea/msigdb). Next, the “GSVA” R package (version 3.16) was utilized to calculate the scores of each sample in each ferroptosis-related pathway as an evaluation of their biological functions.

Cell culture, cytotoxicity assay, and drug administration

The HCT116 and LS180 cell lines were purchased from the American Type Culture Collection (Manassas, VA, USA). The cells were cultured in a 37 °C humidified cell incubator with 5% CO2 using RPMI-1640 medium (Gibco, Thermo Fisher, USA) supplemented with 10% fetal bovine serum (Gibco, Thermo Fisher, USA) and 1% G5 supplement (Thermo Fisher, USA).

The HCT116 and LS180 cells were seeded in 96-well plates (5 × 103 cells/well) with 100 µl medium each well, respectively, and cultured for 24 h. Then, cells were treated with SN38 with different concentrations (0, 10, 20, 40, 80, 160, 320 and 640 nM) for 72 h. A Cell Counting Kit-8 assay (CCK8; MedChemExpress, USA) was used to evaluate the cytotoxicity of SN38 on the two cell lines. After one hour of incubation, absorbance values were determined at 450 nm by a multi-mode reader (Synergy H1, BioTek). The two cell lines were treated with SN38 (MedChemExpress, USA) at 10, 30, or 100 nM for 24 h.

Real-time quantitative polymerase chain reaction (RT-qPCR) of the core irinotecan-related regulators

TRIzol Reagent (Invitrogen, Carlsbad, USA) was used to extract mRNA from the cells, and the RNA concentration was determined using a spectrophotometer (Nanodorf, ThermoFisher, USA). The extracted mRNA was reverse transcribed using the PrimeScript RT Reagent Kit (Takara, Otsu, Japan). Quantitative real-time PCR was performed using Super SYBR Green qPCR Master Mix Kit (Esscience, China). The sequences of the specific primers are shown in the supplemental materials (Supplementary Table S1).

Flow cytometry

HCT116 and LS180 cells were harvested and resuspended in serum-free RPMI-1640 medium (Gibco, Thermo Fisher, USA). BODIPY 581/591 C11 (Thermo Fisher, USA) was added to the cell suspension and incubated for 30 min at 37 °C. The labelled cells were analyzed using an SH800S Cell Sorter (Sony Biotechnology, USA). ROS was detected using a ROS Assay Kit (Beyotime, China), and the procedure of cell processing was similar to the above. The data obtained were processed using FlowJo version 10.8 software (BD Biosciences).

Western blotting

Total protein from the two colorectal cell lines treated with SN38 was extracted using RIPA lysis buffer (Invitrogen, USA) containing PMSF (Bio-Rad, Shanghai, China), separated by SDS-PAGE, and transferred to a polyvinylidene fluoride membrane (PVDF) (Invitrogen, Carlsbad, USA). After blocking with 5% BSA for 1 h, the membrane was incubated with the corresponding primary antibody at 4 °C overnight. Next, the appropriate secondary antibody was added and incubated for 1 h at room temperature. Finally, the labeled membranes were analyzed using a ChemiDoc XRS + imaging system (Bio-Rad, USA). The antibodies used in this study are detailed in Supplementary Materials (Supplementary Table S2).

Intracellular ferrous ions (Fe2+) detection

The HCT116 and LS180 cells were seeded in 6-well plates (3 × 105 cells/well) with 2 mL medium each well, respectively, and cultured for 24 h. Then, cells were treated with SN38 (30 nM) for 24 h. Cell Ferrous Iron (Fe2+) Fluorometric Assay Kit (MA0647, Meilunbio, China) was used to detect the content of Fe2+ in the two cells on the fluorescence microplate reader (Synergy H1, BioTek, USA) (Ex/Em = 543/580 nm).

Transmission electron microscopy (TEM)

The ultrastructure of mitochondria was assessed using TEM. HCT116 and LS180 cells were digested with 0.25% trypsin and fixed in a 2.5% glutaraldehyde solution at room temperature in the dark for 12 h. After fixation, the samples were postfixed in 1% aqueous osmium tetroxide for 2 h, dehydrated in a series of graded ethanol solutions (from 50 to 100%), and then in 100% acetone, embedded in the 812 embedding agent (SPI, 90529-77-4), and incubated at 60 °C for 48 h. Samples were sliced with an ultrathin microtome (60–80 nm) and stained with 2% uranyl acetate and lead citrate. The ultrastructural images were captured using a transmission electron microscope (HT7700, HITACHI, Japan).

Statistical analysis

All data from the public database were analyzed and visualized using R software (version 4.1.3). Figures related to experimental verification were generated using GraphPad Prism 9.5 software (version 9.5; San Diego, CA, USA). A t-test was used for the measurement data. The log-rank test was used for K–M curves. The level of significance was set at P < 0.05.

Results

Identification of prognostic IRDEGs and functional analysis

We analyzed DEGs in the mRNA expression profiles of irinotecan-treated and untreated CRC cell lines using the GSE145356 cohort in the GEO database. A total of 548 IRDEGs were screened; of them, 233 were upregulated, and 315 were downregulated (Fig. 1A–B) under the threshold of FDR < 0.05 and|log2 FC| > 1. Pathway enrichment analysis was conducted to explore the functions of IRDEGs. The results showed that IRDEGs were the most enriched in terms related to tumor cell functions and tumor growth, such as cell adhesion, cellular component movement, tissue morphogenesis, and the VEGFA-VEGFR2 signaling pathway (Fig. 1C and Supplementary Figure S1). These results indicate that genes related to irinotecan treatment may play a pivotal role in the malignant occurrence and progression of CRC. Subsequently, to screen the IREDGs related to the prognosis of patients with CRC, univariate Cox regression analysis was performed to identify prognosis-related genes in the TCGA database. We then intersected the IREDGs with CRC prognosis-related genes (Fig. 1D), and 51 overlapping genes were selected for subsequent analyses (Fig. 1E–F). The details of gene expression, interaction, and prognostic roles among the IRDEGs are shown in the Supplementary Materials (Supplementary Figure S2).

Fig. 1
figure 1

Identification of expression, biological function, and prognostic value of IRDEGs in CRC. (A) The volcano plot showed irinotecan-related DEGs (IRDEGs) in the GSE145356 cohort. (B) Heatmap showing IRDEGs in the GSE145356 cohort. (C) Protein-protein interactions reveal the biological functions of IRDEGs. Proteins with the same function are displayed in the same color. (D) Venn diagram showing the intersection of 497 IRDEGs in the GSE145356 cohort and 2412 prognostic genes in the TCGA cohort. (E) 51 prognostic IRDEGs are shown in the forest map. (F) Expression of the 51 IRDEGs in irinotecan-treated (T) and irinotecan-untreated (C) CRC cell lines from the GSE145356 cohort

Full size image

Establishment of IRDEG-based subtypes and clinical validation

Based on the expression features of 51 prognostic IRDEGs, we established irinotecan-related CRC subtypes using unsupervised consensus clustering. According to the CDF and delta region graphs (Fig. 2A–C), we separated patients in the CRC cohort from TCGA into two clusters (k value = 2): Cluster A and Cluster B. Principal component analysis (PCA) was conducted to validate the dependability of the clustering (Fig. 2D). The survival curve showed that patients in Cluster A had a significant survival advantage compared to patients in Cluster B (P < 0.001) (Fig. 2E). Next, we investigated the relationship between IRDEG-based subtypes and clinical information (Fig. 2F). As shown in the heatmap, the N stage, M stage, TNM stage, and survival status were significantly correlated with the IRDEG-based subtype.

Fig. 2
figure 2

Establishment of IRDEG-based subtypes and clinical validation. (A) Consensus cumulative distribution function (CDF) curves with k from 2 to 9. (B) Delta area plot showing the relative change in area under the CDF curves between k and k-1 (k = 2–9). (C) The result of consensus clustering (k value = 2). (D) Validation of the result of consensus clustering using PCA analysis. (E) K–M curve showing the difference in survival probability between cluster A and cluster B. (F) The clinical characteristics differences between cluster A and cluster B

Full size image

Construction of the Iriscore model based on IRDEG-based subtypes and validation of its prognostic value

We further constructed an iriscore model using PCA based on IRDEG-based subtypes to explore the value of the signature genes in terms of prognosis. We then divided the CRC patients into a high iriscore group and a low iriscore group based on the median value of the iriscore. The low-iriscore patients had a longer OS than patients with a high iriscore (P < 0.001), which indicated that the iriscore had a significant prognostic value in CRC (Fig. 3A). The iriscore results for those who died were higher than those for those who survived (P < 0.001), and the high-iriscore group had a higher proportion of patients who died (Fig. 3B–C). The areas under the curves (AUCs) were 0.693, 0.640, 0.701, and 0.753 at 1, 3, 5, and 10 years, respectively (Fig. 3D).

To assess the prognostic value of the iriscore, we used a validation cohort (GSE39582) from the GEO database. Consistent with the above results, patients in the low-iriscore group presented a longer OS than those in the high-iriscore group (P < 0.001) (Fig. 3E). The incidence of death was also strongly correlated with the iriscore (P < 0.05) (Fig. 3F–G). The AUCs of the GSE39582 cohort for 3-, 5-, 10-, and 13-year OS rates were 0.544, 0.560, 0.573, and 0.661, respectively (Fig. 3H). Subgroup analysis was conducted according to T stage. Based on Kaplan–Meier curve analysis, patients with T1-2 in the low-iriscore group exhibited no significant differences, compared with the high-iriscore group in TCGA CRC (Supplementary Figure S3) and GSE39582 (Supplementary Figure S5) cohorts. In contrast, patients with T3-4 in the low-iriscore group had a better prognosis than patients in the high-iriscore group in both TCGA CRC (P < 0.001) (Supplementary Figure S4) and GSE39582 (P = 0.004) cohorts (Supplementary Figure S6). The similarity between the training and validation datasets indicates that the prognostic value of the iriscore is specific and sensitive.

Fig. 3
figure 3

Prognostic value of iriscore in TCGA CRC cohort and GSE39582 cohort. (A, E) K–M survival curve analysis of prognoses between the high- and low-iriscore groups. (B, F) Proportion of different survival statuses in the high- and low-iriscore groups. (C, G) Patients with different status had statistically different iriscore. (D, H) ROC curves for assessing the predictive ability of iriscore model

Full size image

Correlation of the TMB with iriscore

Recently, the TMB, which refers to the number of somatic mutations after the removal of germline mutations from tumor genomes, has been increasingly proven to have predictive value for tumor response to immunotherapy in patients with metastatic CRC [30]. Therefore, we confirmed the features of TMB in the high- and low-iriscore groups (Fig. 4A–B) and explored the relationship between TMB and iriscore (Fig. 4C). The TMB of the low iriscore group was higher compared with that of the high iriscore (P < 0.01) (Supplementary Figure S7), and TMB correlated negatively with the iriscore (R = -0.2, P < 0.001) (Supplementary Figure S8). Survival analysis revealed no statistically significant difference in OS between the high- and low-TMB groups (P > 0.05) (Fig. 4D). To further investigate whether the iriscore could be used as a marker to distinguish survival probability in patients with various levels of TMB, patients were separated into high-iriscore with high TMB, high-iriscore with low TMB, low-iriscore with high TMB, and low-iriscore with high TMB groups. The results indicated that patients with a low iriscore in the low TMB group had a significant survival advantage compared with those in the other groups; patients with a high iriscore in the high TMB group exhibited the worst prognosis (Fig. 4E). These results revealed that the constructed iriscore model is closely associated with TMB in patients with CRC.

Fig. 4
figure 4

Correlation of the tumor mutation burden (TMB) with iriscore. (A–B) Characteristics of TMB in the high- and low-iriscore groups from TCGA CRC cohort. (C) Sankey diagram showing the relationship between iriscore and TMB in patients from TCGA CRC cohort. (D) K–M survival curve analysis of patients with low and high TMB from TCGA CRC cohort. (E) K–M curve analysis of the different survival probability among high-iriscore with low TMB, high-iriscore with high TMB, low-iriscore with low TMB and low-iriscore with high TMB groups from TCGA CRC cohort

Full size image

Iriscore as an efficient predictive tool of responses in chemotherapeutic and targeted therapy

To distinguish the treatment response of various iriscore levels in patients with CRC, half the maximum inhibitory concentration (IC50) of anticancer drugs was downloaded from the Genomics of Drug Sensitivity in Cancer (GDSC) website to predict chemotherapeutic and targeted therapeutic responses. Compared with the low-iriscore group, the high-iriscore group had lower IC50 values for traditional chemotherapeutic drugs, including docetaxel, camptothecin, doxorubicin, and etoposide. (P < 0.05) (Fig. 5A–D). In addition, CRC patients with higher iriscores had lower IC50 values for targeted drugs, such as elesclomol, axitinib, midostaurin, lenalidomide, pazopanib, vorinostat, nilotinib, and temsirolimus (P < 0.05) (Fig. 5E–L). Based on these results, the iriscore can be regarded as an efficient predictor of treatment response in patients receiving chemotherapeutic and targeted therapy.

Fig. 5
figure 5

Iriscore as an efficient predictive tool of responses in chemotherapeutic and targeted therapy in CRC patients with high- and low-iriscores. (A–D) Box plot showing the predictive IC50 value of traditional chemotherapeutic drugs, including docetaxel, camptothecin, doxorubin, and etoposide, between the high- and low-iriscore groups of patients with CRC form the TCGA cohort. (E–L) Box plot showing the predictive IC50 value of targeted drugs, including elesclomol, axitinib, midostaurin, lenalidomide, pazopanib, vorinostat, nilotinib, and temsirolimus, in patients with CRC with high-iriscore compared with patients with CRC with low-iriscore from TCGA cohort

Full size image

Immune landscapes of low- and high-iriscore CRC patients

The state of the microsatellites is closely associated with the efficacy of immunotherapy. However, approximately 15% of CRC patients have microsatellite instability-high (MSI-H), which is highly sensitive to immunotherapy [31]. The remainder are microsatellite stable (MSS) or microsatellite instability-low (MSI-L), which are resistant to immunotherapy. Hence, the identification of patients with MSI-H CRC is of great importance. Here, we investigated the correlation between the iriscore and the state of the microsatellites. The proportion of patients in the low-iriscore group was 27%, which was much higher than that in the high-iriscore group (7%; Fig. 6A). The iriscore in patients with MSI-H CRC was significantly different from that in patients with MSS or MSI-L CRC (P < 0.01) (Fig. 6B), and the immune checkpoint TIM-3 was significantly different between the high- and low-iriscore group (P < 0.05) (Fig. 6C). Next, we investigated the relationship between irisin levels and immunotherapy. In the IMvigor210 cohort, the CR/PR (response) group had a higher iriscore than that of the SD/PD (non-response) group (P = 0.0009) (Fig. 6D). In addition, differences in the immunophenoscore (IPS) when using cytotoxic T lymphocyte-associated antigen-4 (CTLA-4)/programmed death-1 (PD-1) inhibitors between the low- and high-iriscore groups were assessed to further verify the value of the iriscore in immunotherapy (Fig. 6E–H). The IPS of the low-iriscore group treated with PD-1 inhibitors was higher than that of the high-iriscore group (P < 0.05) (Fig. 6F). The low-iriscore group also benefited from CTLA-4 inhibitors (P < 0,01) (Fig. 6H). Nonetheless, there was no benefit when PD-1 and CTLA-4 inhibitors were used simultaneously (P > 0.05) (Fig. 6G). Overall, the iriscore displayed a noticeable correlation with immune signaling pathways and immune cells (Fig. 6I-J). These results show that the iriscore is closely associated with the immunotherapy response and tumor immune microenvironment.

Fig. 6
figure 6

Landscape of immunotherapy, immune checkpoints, and immune cell infiltration in iriscore model. (A) Proportion of different microsatellite statuses in high- and low-iriscore groups in TCGA CRC cohort. (B) The differences of iriscore among patients with MSS, MSI-L and MSI-H in TCGA CRC cohort. (C) Box blot showing the expression of TIM-3 in patients with CRC with high-iriscore compared with patients with CRC with low-iriscore from the TCGA cohort. (D) The differences of iriscore between responders (CR/PR) and non-responders (SD/PD) in IMvigor210 cohort. (E–H) Violin plots showing the immunophenoscore (IPS) when using cytotoxic T lymphocyte-associated antigen-4 (CTLA-4)/programmed death-1 (PD-1) inhibitors between the low- and high-iriscore groups in the TCGA CRC cohort. (I–J) The correlation of the iriscore and immune cell infiltration (I) or immune-related pathways (J) in patients from TCGA CRC cohort

Full size image

Single-RNA sequencing of CRC cells revealed the expression signature of crucial irinotecan-related regulators

We investigated the expression profiles of 10 crucial irinotecan-related regulators in CRC cells using scRNA-seq. As depicted in Figs. 7A–B and 18 cell clusters were identified using unsupervised clustering. The annotation of these clusters revealed the expression signatures of the most important irinotecan regulators (Fig. 7C). FSTL3 and TMEM98 are highly expressed in tissue stem cells and fibroblasts in CRC and may promote cancer progression by regulating tumor cell proliferation. BCL10, TBC1D10A, CASZ1, CYFIP2, ECHDC1, and LUZP1 are mainly expressed in T cells, macrophages, and monocytes. Hence, these regulators may influence the effect of immunotherapy by regulating the innate and acquired tumor immune microenvironment. Furthermore, BCL10, CYFIP2, ECHDC1, and LUZP1 have been identified in B cells. ANKRD22 has only been identified in innate immune cells, such as macrophages, monocytes, neutrophils, and dendritic cells. PTP4A2 was highly expressed in all cell clusters.

Fig. 7
figure 7

Single-cell RNA sequencing revealing the expression signature of the crucial irinotecan-related regulators. (A)18 cell clusters were identified by unsupervised clustering from 6 CRC patients. (B) Annotation of these clusters according to known marker genes. (C) The 10 crucial irinotecan-related regulators were identified in the single-cell tumor atlas

Full size image

Identification of crucial irinotecan-related regulators and the transcriptional and immune microenvironment landscape

We used random forest analysis to further screen core irinotecan-related regulators, and the results revealed 18 crucial irinotecan-related regulators (Fig. 8A–B). The state of gene alterations, interactions, and prognostic features of the 18 irinotecan-related regulators are illustrated in Fig. 8C. FSTL3, ENO2, RIMKLB, PTPRU, and CREG2 were the five most important risk factors associated with CRC prognosis (Fig. 8C), with FSTL3 being the most important factor. Therefore, FSTL3 was selected for further analysis. The co-expression of irinotecan-related regulators in the CRC cohort from TCGA is shown (Fig. 8D). We found a positive correlation between FSTL3 and PTPRU expression (R = 0.52, P < 0.0001). Considering that irinotecan-related regulators may influence the expression of transcription involved in CRC progression, we used a Sankey diagram to investigate the specific relationship between TFs and irinotecan-related regulators (Fig. 8E). The functions of irinotecan-related regulators in CRC may also be reflected in the immune microenvironment. Immune cell infiltration (Fig. 8F) and checkpoints (Fig. 8G) were closely associated with these irinotecan-related regulators. Additionally, FSTL3 presented was strongly associated with TNFSF4 (R = 0.57, P < 0.001), HAVCR2 (R = 0.57, P < 0.001), and macrophages. M2 (R = 0.37, P < 0.001) (Supplementary Figures S9–S11). Overall, the multiple traits of irinotecan-related regulators indicate that targeting these regulators at the transcriptional and immune microenvironment levels may be a potential therapeutic approach for irinotecan resistance in CRC.

Fig. 8
figure 8

Identification of crucial irinotecan-related regulators and the transcriptional and immune microenvironment landscape. (A–B) Random forest analysis identifying the 18 crucial irinotecan-related regulators according to their importance. (C) Protein–protein network revealing the state of gene alteration, interaction, and prognostic features of 18 crucial irinotecan-related regulators. (D) Co-expression of crucial irinotecan-related regulators in the CRC cohort. (E) Sankey diagram showing the specific relationship between TFs and irinotecan-related regulators. (F) The relationship between the 18 irinotecan-related regulators and immune cells. (G) The correlation between crucial irinotecan-related regulators and immune checkpoint proteins

Full size image

Validation of therapeutic agents targeting irinotecan-related regulators in CRC

We further explored potential therapeutic drugs (DTP NCI-60) for irinotecan-related regulators and predicted 28 potential irinotecan-related agents, including telatinib (Fig. 9A). Telatinib is a selective small-molecule inhibitor of VEGFR2 VEGFR3 and PDGFR-β tyrosine kinases [32] and is widely used in solid tumor treatment. The three-dimensional structure of telatinib and the molecular docking sites of telatinib and FSTL3 proteins are shown in Fig. 9B and C, respectively. FSTL3 forms a non-covalent bond with telatinib via asparagine at position 83 and aspartic acid at position 132 (Fig. 9C). The minimum binding energy obtained from the docking of FSTL3 with telatinib was calculated to be -7.95 kcal/mol with Autodock. As telatinib can bind to the FSTL3 protein, it could serve as a target drug in patients with a high level of FSTL3 and might have a potential synergistic effect with irinotecan. Next, the potential targets of telatinib were confirmed using Swiss Target Prediction. We performed the GO and KEGG enrichment analyses of these targets from the database and the results indicated that these genes were mostly associated with cell proliferation, differentiation, and metabolism, particularly “regulation of MAPK cascade,” “regulation of kinase activity,” and “PI3K-Akt signaling pathway” (Fig. 9D, E). Overall, therapeutic drugs targeting irinotecan-related regulators, especially telatinib, might act with irinotecan in CRC, as they target proto-oncogenes such as FSTL3.

Fig. 9
figure 9

Prediction of therapeutic agents targeting FSTL3. (A) Correlation between 28 predicted irinotecan-related regulators targeted therapeutic drug activities and the expression of 18 crucial irinotecan-related regulators. Red represents a positive correlation, and blue represents a negative correlation. *P-value < 0.05, **P-value < 0.01. (B) The molecular structure of telatinib. (C) 3D diagram of molecular docking of telatinib and FSTL3. (D–E) GO and KEGG analysis of targeted genes of telatinib predicted by the Swiss Targeted Prediction

Full size image

FSTL3 is a core irinotecan-related regulator associated with prognostic and clinical features

The K–M survival curves showed that patients with higher FSTL3 expression had poorer survival (Fig. 10A). The ROC curves showed that the AUCs at 1-, 3-, and 5 years were 0.718, 0.682, and 0.674, respectively, indicating the superior predictive ability of FSTL3 expression for CRC prognosis (Fig. 10B). The distribution of the relative expression levels of FSTL3 in patients with CRC in the cohort is shown in Fig. 9C. The expression level of FSTL3 was significantly higher in CRC patients who did not survive than in those who did (Fig. 10C). In addition, the expression level of FSTL3 was closely associated with TNM stage, survival status, and T stage (P < 0.05) but not with age or sex (P > 0.05) in CRC (Fig. 10D–H). In general, high expression levels of FSTL3 often predicted poor survival in patients with T stages III-IV; however, in patients with T stages I–II, there was no significance of FSTL3 expression with CRC prognosis (Fig. 10L). Regardless of age or sex, patients with high FSTL3 expression had a poorer prognosis compared with those in the low expression group (Fig. 10I–J). Notably, there was no significant difference between the high- and low-FSTL3 groups with respect to TNM stages I–II or III–IV (Fig. 10K).

Fig. 10
figure 10

Identification of prognostic and clinical features of FSTL3. (A) K–M survival curves analysis of patients between the high-FSTL3 and low-FSTL3 groups. (B) ROC curves for verifying the predictive ability of FSTL3. (C) Risk curve and survival status plot revealing the distribution of risk scores of patients with different gene expression levels in the TCGA cohort. (D) Differences in clinical characteristics (survival status, age, sex, T stage and TNM stage) between the high- and low-FSTL3 groups in the TCGA cohort. (E–H) Correlations of FSTL3 expression and clinical characteristics (age, sex, T stage and TNM stage) in the TCGA cohort. (I–L) K–M survival curve analysis of patients in the high- and low-FSTL3 groups in different clinical subgroups

Full size image

Construction and validation of a nomogram integrating FSTL3 and other independent prognostic factors

To determine whether FSTL3 could serve as an independent prognostic predictor, we screened the clinical characteristics (including age, sex, and T, N, M, and TNM stages) of patients in the CRC cohort using univariate and multivariate Cox regression analyses. The results revealed that Age (P < 0.05, HR = 2.986), T stage (P < 0.05, HR = 2.113), M stage (P < 0.05, HR = 2.902), and FSTL3 expression (P < 0.05, HR = 1.365) were independent predictive factors in CRC patients (Fig. 11A). Based on these results, we constructed a nomogram using these independent predictive factors to predict the survival probability of patients with CRC at 1-, 3-, and 5 years (Fig. 11B). A calibration curve was used to verify the reliability of the nomogram for predicting the OS probability at 1-, 3-, and 5 years of CRC patients with CRC. The slope of the calibration curve was approximately 45°, indicating the reliability of the established nomogram (Fig. 11C–E). Next, we used DCA to evaluate the predictive performance of the clinical model and nomogram for OS at 1-, 3-, and 5 years. DCA showed a net benefit of FSTL3 expression in predicting survival possibilities, compared with the clinical model, including independent prognostic factors in CRC (Fig. 11F–H). Hence, the FSTL3 expression level could be considered an independent prognostic factor, and the nomogram we established has superior performance in predicting OS.

Fig. 11
figure 11

Construction and validation of a nomogram integrating FSTL3 and other independent prognostic factors. (A) A forest plot showing the univariate and multivariate Cox regression analyses with OS in the TCGA database. (B) A nomogram predicting the survival probability of patients with CRC at 1-, 3-, and 5 years. (C–E) The calibration curve of the constructed nomogram at 1-, 3-, and 5 years. (F-H) DCA curve at 1-, 3-, and 5 years for assessing the advantage of nomogram compared with the clinical model in CRC patients

Full size image

FSTL3 might contribute to cancer progression by inhibiting ferroptosis

Ferroptosis is a unique cell death process that is mechanistically and morphologically different from other forms of cancer and plays an indispensable role in colorectal cancer progression. To study the mechanism by which FSTL3 promotes cancer progression at the cell death level, we analyzed the correlations between FSTL3 and crucial ferroptosis-related regulators and metabolic pathways. The ferroptosis score of each sample in the CRC cohort from TCGA was calculated using the GSVA algorithm as described above. The ferroptosis score of the low-FSTL3 group was higher than that of the high-FSTL3 group (P < 0.05), and there was a significant correlation between FSTL3 and ferroptosis (R = -0.13, P < 0.05) (Fig. 12A–B). Next, we confirmed the close correlation between FSTL3 and ferroptosis-related hub genes; the hub genes GPX4, HSPB1, PTGS2, FTH1, TFRC, SLC40A1, RGS4, ACSL4, TF, and DHODH were significantly different between the low- and high-FSTL3 groups (P < 0.05) (Fig. 12C). We further found that FSTL3 expression was closely associated with ferroptosis markers such as RGS4 (R = 0.59, P < 0.001) and HSPB1 (R = 0.31, P < 0.001) (Fig. 12D–G). FSTL3 also showed a significant correlation with ferroptosis-related metabolic pathways (Fig. 12H). Importantly, some crucial ferroptosis-related metabolic pathways, including lipid oxidation, regulation of lipid metabolism, reactive oxygen species (ROS) metabolism, phosphatidic acid metabolism, triglyceride acid metabolism, fatty acid metabolism, glutathione metabolism, and glutathione peroxidase activity, were significantly different between the high- and low-FSTL3 groups (P < 0.05).

Fig. 12
figure 12

Close correlation of ferroptosis and FSTL3. (A) Ferroptosis score in high- and low-FSTL3 groups. (B) The relationship of FSTL3 expression with ferroptosis score. (C) Box plots revealing the differences in ferroptosis markers between high- and low-FSTL3 groups. (D) The difference of RGS4 expression between high- and low-FSTL3 groups. (E) The relationship of FSTL3 expression with RGS4. (F) The difference of HSPB1 expression between high- and low-FSTL3 groups. (G) The correlation of FSTL3 expression with HSPB1. (H) Boxing plots showing the differences in ferroptosis related metabolic pathway between high- and low-FSTL3 groups

Full size image

Inhibition of FSTL3 and crucial ferroptosis defense proteins in CRC cells treated with SN38 and ferroptosis induction by Irinotecan

At the transcriptional level, we determined the relative mRNA expression of the 10 most important regulators in the two CRC cell lines treated with SN38 (Fig. 13A–B). The cytotoxic assays of SN38 in two CRC HCT116 and LS180 cell lines were evaluated (Supplementary Figure S12-S13). The level of mRNA expression of FSTL3 in both LS180 and HCT116 cells was significantly inhibited after SN38 treatment (P < 0.05). To confirm the effect of ferroptosis induction by SN38, intracellular ROS was detected by flow cytometry analysis in HCT116 and LS180 cells treated with SN38 after 6, 12, 24, and 48 h, and the ROS levels increased with processing time in both cell lines (Fig. 13C, E). Moreover, SN38 induced lipid peroxidation in HCT116 and LS180 cells (Fig. 13D and F). Transmission electron microscopy was used to observe the mitochondrial ultrastructure to confirm the occurrence of ferroptosis induced by SN38 in the two cells (Fig. 13G). The results revealed that HCT116 and LS180 cells treated with SN38 after 24 h exhibited shrinking mitochondria, increased mitochondrial membrane density, decreased mitochondrial cristae, and mitochondrial outer membrane rupture. Moreover, intracellular ferrous ions in HCT116 and LS180 cells treated with SN38 were enhanced significantly in cells treated with SN38 compared with the control (P < 0.001) (Fig. 13H). Western blotting indicated that crucial ferroptosis defense proteins, including SLC7A11 and GPX4, were inhibited in both cell lines treated with SN38, and an increasing inhibitory effect was observed with increasing concentrations of SN38 (Fig. 13I). The results above demonstrate that irinotecan significantly inhibits FSTL3 and can effectively induce ferroptosis in CRC cells.

Fig. 13
figure 13

SN38 inhibited the expression of FSTL3 and induced ferroptosis in CRC cells. (A-B) The mRNA relative expression of FSTL3 in HCT116 and LS180 CRC cells treated with SN38. (C, E) ROS increasing in HCT116 and LS180 CRC cells treated with SN38 after 6, 12, 24 and 48 h. (D, F) Flow cytometry showing the levels of lipid peroxidation in HCT116 and LS180 cells using C11 BODIPY probe. (G) Transmission electron microscopy images of mitochondrial ultrastructure in HCT116 and LS180 cells. (H) Intracellular ferrous ions (Fe2+) detection in HCT116 and LS180 cells. (I) Western blotting showing GPX4 and SLC7A11 proteins expression in HCT116 and LS180 CRC cells treated with SN38 at the concentration of 10-, 30-, 100 nM

Full size image

Discussion

Colorectal cancer (CRC) is the second most common cancer in women and the third most common cancer in men and is the fourth leading cause of cancer-related deaths, accounting for 9.2% of deaths worldwide [33]. Despite the rapid advancement of immunotherapy and targeted therapy, the first-line treatment regimen for CRC still consists of combined classical chemotherapy. Irinotecan is one of the most fundamental chemotherapeutic drugs for CRC and plays a pivotal role in its progression and prognosis. Previously, we mainly focused on studying the molecular mechanisms of the chemotherapeutic drugs 5-FU and oxaliplatin instead of irinotecan in CRC. Therefore, it is necessary to identify potential irinotecan-targeting prognostic regulators.

Irinotecan, a camptothecin-derived drug, was first approved for cancer treatment in 1994, and this natural product derivative remains a major anticancer drug worldwide. Irinotecan mainly targets topoisomerase I (Topo1), a crucial nuclear enzyme involved in the maintenance of proper DNA topology during replication and transcription, and induces cell death by blocking DNA replication [34, 35]. However, recent studies have indicated that irinotecan may not act solely by inhibiting Topo1. The active metabolite of irinotecan, SN38, induces apoptosis via a TP53-dependent pathway. Specifically, the expression and phosphorylation of tumor antigen P53 (TP53) induced by SN38 can trigger the upregulation of downstream apoptosis-inducing proteins such as apoptosis regulator (BAX), caspase-3, and caspase-9 [36, 37]. Moreover, tumor necrosis factor receptor superfamily member 6 (FAS)-mediated and mitogen-activated serine/threonine protein kinases (MAPKs) are involved in apoptosis induced by irinotecan [37, 38]. The evidence suggests that various mechanisms and signaling pathways may participate in the mechanism of action of irinotecan in a simultaneous or complementary manner. In order to investigate the correct target of irinotecan at the genetic level, we comprehensively analyzed the potential mechanism and prognostic role of 51 irinotecan-related regulators in CRC; based on these prognostic irinotecan-related regulators, CRC was divided into two different groups, high- and low-iriscore groups. The overall survival (OS) of the high-iriscore group was notably poorer than that of the low-iriscore group. Moreover, we confirmed that the low-iriscore group may benefit from immunotherapy and some specific chemotherapy drugs, compared with the high-iriscore group, which may guide clinical decision-making. Next, scRNA-seq indicated that FSTL3 and TMEM98 are mainly expressed in CRC stem cells and contribute to multiple tumor malignant behaviors, such as recurrence, metastasis, heterogeneity, multidrug resistance, and radiation resistance [39,40,41]. In particular, we found that FSTL3 is a critical independent prognostic irinotecan-related regulator in CRC and that high FSTL3 expression predicted a worse survival rate. In addition, we identified drugs, particularly telatinib, that target FSTL3 and may have synergistic effects with irinotecan for the treatment of CRC.

FSTL3 (follistatin-like 3) is an extracellular matrix glycoprotein involved in a series of pathophysiological processes, such as cell proliferation, migration, differentiation, and embryonic development, and is mainly released by the placenta. Recently, significant overexpression of FSTL3 was observed in some malignant tumors [42,43,44]. FSLT3 expression has been shown to correlate with cancer progression in CRC, thyroid cancer, gastric cancer, and non-small-cell lung cancer cells [43,44,45]. We validated its negative association with ferroptosis to explore the underlying mechanism of FSTL3 expression in CRC progression at the level of cell death. Although the correlation with ferroptosis was not strong, FSTL3 showed a close correlation with ferroptosis-related marker genes, such as GPX4, and hub metabolic pathways, such as ROS metabolism and lipid oxidation. Importantly, FSTL3 showed the strongest correlation with the ferroptosis-related gene RGS4, which upregulates SLC7A11(xCT) [46]. SLC7A11-GSH-GPX4 is believed to constitute a major cellular defense mechanism against ferroptosis, and FSTL3 may inhibit ferroptosis by promoting RGS4 expression and boosting the SLC7A11-GSH-GPX4 pathway. Notably, a significant difference was found between the high- and low-FSTL3 expression groups in GPX4 expression, but not in SLC7A11 expression. It is believed that SLC7A11 expression is influenced by other mechanisms in the presence of irinotecan.

Heat shock protein beta-1 (HSPB1) is a negative regulator of ferroptosis [47]. In our study, a significant correlation with FSTL3 was found, revealing that FSTL3 may inhibit ferroptosis by targeting HSPB1. We further confirmed that irinotecan promoted ferroptosis by preventing the expression of ferroptosis defense proteins (SLC7A11 and GPX4) and accelerating lipid peroxidation. Based on the above analysis, irinotecan appears to promote ferroptosis by targeting FSTL3 and indirectly inhibiting cancer evolution and progression. This study reveals that irinotecan inhibits cancer progression at the ferroptosis level, which provides a valuable direction for exploring irinotecan and ferroptosis.

However, this study has a few limitations that warrant further improvement. First, we obtained only RNA-seq data and corresponding clinical information from public databases; more cohorts are needed for external validation, and it is essential to conduct RNA-seq validation to confirm the findings of the bioinformatics analysis in the future. Second, the prognostic roles of the iriscore and FSTL3 in CRC require validation in a clinical cohort. We will conduct related research to confirm their prognostic roles in the future. Third, further molecular biology and animal experiments are necessary to explore the potential regulatory mechanisms of FSTL3 and irinotecan at the ferroptotic level in CRC. Lastly, the biological and clinical relevance of using irinotecan-related regulators to build the model would need a further justification to validate the specificity and generalizability of the iriscore model across diverse tumor types and treatment contexts.

Conclusions

In summary, iriscore had a close relationship with prognosis, TMB, immunization, and chemotherapeutic response. The most important irinotecan-related risk regulator, FSTL3, had a strong prognostic value in patients with CRC. scRNA-seq revealed that FSTL3 was mainly expressed in cancer stem cells and may serve as a critical target in cancer therapy. Finally, we validated the correlation between FSTL3 and ferroptosis and speculated that irinotecan may promote ferroptosis by targeting FSTL3.

Data availability

No datasets were generated or analysed during the current study.

References

  1. Siegel RL, Miller KD, Wagle NS, Jemal A. Cancer statistics, 2023. CA Cancer J Clin. 2023;73:17–48.

    Article  PubMed  Google Scholar 

  2. Lyu N, Pedersen B, Shklovskaya E, Rizos H, Molloy MP, Wang Y. SERS characterization of colorectal cancer cell surface markers upon anti-EGFR treatment. Explor (Beijing). 2022;2:20210176.

    CAS  Google Scholar 

  3. Bailly C. Irinotecan: 25 years of cancer treatment. Pharmacol Res. 2019;148:104398.

    Article  CAS  PubMed  Google Scholar 

  4. Wang Y, Yang L, Zhang J, Zhou M, Shen L, Deng W, et al. Radiosensitization by Irinotecan is attributed to G2/M phase arrest, followed by enhanced apoptosis, probably through the ATM/Chk/Cdc25C/Cdc2 pathway in p53-mutant colorectal cancer cells. Int J Oncol. 2018;53:1667–80.

    CAS  PubMed  Google Scholar 

  5. Zhu J, Li X, Shen Y, Guan Y, Gu W, Lian P, et al. Genotype-driven phase I study of weekly Irinotecan in combination with capecitabine-based neoadjuvant chemoradiation for locally advanced rectal cancer. Radiotherapy Oncology: J Eur Soc Therapeutic Radiol Oncol. 2018;129:143–8.

    Article  CAS  Google Scholar 

  6. Zhu J, Liu A, Sun X, Liu L, Zhu Y, Zhang T, et al. Multicenter, randomized, phase III trial of neoadjuvant chemoradiation with capecitabine and Irinotecan guided by UGT1A1 status in patients with locally advanced rectal cancer. J Clin Oncology: Official J Am Soc Clin Oncol. 2020;38:4231–9.

    Article  CAS  Google Scholar 

  7. Wang H, Jia H, Gao Y, Zhang H, Fan J, Zhang L, et al. Serum metabolic traits reveal therapeutic toxicities and responses of neoadjuvant chemoradiotherapy in patients with rectal cancer. Nat Commun. 2022;13:7802.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  8. Zhang X, Tang B, Luo J, Yang Y, Weng Q, Fang S, et al. Cuproptosis, ferroptosis and PANoptosis in tumor immune microenvironment remodeling and immunotherapy: culprits or new hope. Mol Cancer. 2024;23:255.

    Article  PubMed  PubMed Central  Google Scholar 

  9. Tang B, Zhu J, Li J, Fan K, Gao Y, Cheng S, et al. The ferroptosis and iron-metabolism signature robustly predicts clinical diagnosis, prognosis and immune microenvironment for hepatocellular carcinoma. Cell Commun Signal. 2020;18:174.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  10. Tang B, Yan R, Zhu J, Cheng S, Kong C, Chen W, et al. Integrative analysis of the molecular mechanisms, immunological features and immunotherapy response of ferroptosis regulators across 33 cancer types. Int J Biol Sci. 2022;18:180–98.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  11. Tang B, Zhu J, Fang S, Wang Y, Vinothkumar R, Li M, et al. Pharmacological Inhibition of MELK restricts ferroptosis and the inflammatory response in colitis and colitis-propelled carcinogenesis. Free Radic Biol Med. 2021;172:312–29.

    Article  CAS  PubMed  Google Scholar 

  12. Chen Q, Zheng W, Guan J, Liu H, Dan Y, Zhu L, et al. SOCS2-enhanced ubiquitination of SLC7A11 promotes ferroptosis and radiosensitization in hepatocellular carcinoma. Cell Death Differ. 2023;30:137–51.

    Article  CAS  PubMed  Google Scholar 

  13. Jiang L, Kon N, Li T, Wang S-J, Su T, Hibshoosh H, et al. Ferroptosis as a p53-mediated activity during tumour suppression. Nature. 2015;520:57–62.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  14. Zhang Y, Shi J, Liu X, Feng L, Gong Z, Koppula P, et al. BAP1 links metabolic regulation of ferroptosis to tumour suppression. Nat Cell Biol. 2018;20:1181–92.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  15. Yi J, Zhu J, Wu J, Thompson CB, Jiang X. Oncogenic activation of PI3K-AKT-mTOR signaling suppresses ferroptosis via SREBP-mediated lipogenesis. Proc Natl Acad Sci USA. 2020;117:31189–97.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  16. Tan SK, Mahmud I, Fontanesi F, Puchowicz M, Neumann CKA, Griswold AJ, et al. Obesity-Dependent adipokine chemerin suppresses fatty acid oxidation to confer ferroptosis resistance. Cancer Discov. 2021;11:2072–93.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  17. Alvarez SW, Sviderskiy VO, Terzi EM, Papagiannakopoulos T, Moreira AL, Adams S, et al. NFS1 undergoes positive selection in lung tumours and protects cells from ferroptosis. Nature. 2017;551:639–43.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  18. Dixon SJ, Lemberg KM, Lamprecht MR, Skouta R, Zaitsev EM, Gleason CE, et al. Ferroptosis: an iron-dependent form of nonapoptotic cell death. Cell. 2012;149:1060–72.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  19. Koppula P, Zhuang L, Gan B. Cystine transporter SLC7A11/xCT in cancer: ferroptosis, nutrient dependency, and cancer therapy. Protein Cell. 2021;12:599–620.

    Article  CAS  PubMed  Google Scholar 

  20. Lei G, Zhuang L, Gan B. Targeting ferroptosis as a vulnerability in cancer. Nat Rev Cancer. 2022;22:381–96.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  21. Jin Y, Cai S, Zhou Y, Guo D, Zeng Y, Xu W, et al. Targeting SLC7A11/xCT improves radiofrequency ablation efficacy of HCC by dendritic cells mediated anti-tumor immune response. Imeta. 2024;3:e248.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  22. Kawato Y, Aonuma M, Hirota Y, Kuga H, Sato K. Intracellular roles of SN-38, a metabolite of the camptothecin derivative CPT-11, in the antitumor effect of CPT-11. Cancer Res. 1991;51:4187–91.

    CAS  PubMed  Google Scholar 

  23. Chen P, Luo X, Nie P, Wu B, Xu W, Shi X, et al. CQ synergistically sensitizes human colorectal cancer cells to SN-38/CPT-11 through lysosomal and mitochondrial apoptotic pathway via p53-ROS cross-talk. Free Radic Biol Med. 2017;104:280–97.

    Article  CAS  PubMed  Google Scholar 

  24. Wilkerson MD, Hayes DN. ConsensusClusterPlus: a class discovery tool with confidence assessments and item tracking. Bioinformatics. 2010;26:1572–3.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  25. Trott O, Olson AJ. AutoDock Vina: improving the speed and accuracy of Docking with a new scoring function, efficient optimization, and multithreading. J Comput Chem. 2010;31:455–61.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  26. O’Boyle NM, Banck M, James CA, Morley C, Vandermeersch T, Hutchison GR. Open babel: an open chemical toolbox. J Cheminform. 2011;3:33.

    Article  PubMed  PubMed Central  Google Scholar 

  27. Seeliger D, de Groot BL. Ligand Docking and binding site analysis with PyMOL and Autodock/Vina. J Comput Aided Mol Des. 2010;24:417–22.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  28. Daina A, Michielin O, Zoete V. SwissTargetPrediction: updated data and new features for efficient prediction of protein targets of small molecules. Nucleic Acids Res. 2019;47:W357–64.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  29. Hänzelmann S, Castelo R, Guinney J. GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinformatics. 2013;14:7.

    Article  PubMed  PubMed Central  Google Scholar 

  30. Schrock AB, Ouyang C, Sandhu J, Sokol E, Jin D, Ross JS, et al. Tumor mutational burden is predictive of response to immune checkpoint inhibitors in MSI-high metastatic colorectal cancer. Ann Oncol. 2019;30:1096–103.

    Article  CAS  PubMed  Google Scholar 

  31. André T, Shiu K-K, Kim TW, Jensen BV, Jensen LH, Punt C, et al. Pembrolizumab in Microsatellite-Instability-High advanced colorectal cancer. N Engl J Med. 2020;383:2207–18.

    Article  PubMed  Google Scholar 

  32. Strumberg D, Schultheis B, Adamietz IA, Christensen O, Buechert M, Kraetzschmar J, et al. Phase I dose escalation study of telatinib (BAY 57-9352) in patients with advanced solid tumours. Br J Cancer. 2008;99:1579–85.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  33. Li J, Ma X, Chakravarti D, Shalapour S, DePinho RA. Genetic and biological hallmarks of colorectal cancer. Genes Dev. 2021;35:787–820.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  34. Champoux JJ. DNA topoisomerases: structure, function, and mechanism. Annu Rev Biochem. 2001;70:369–413.

    Article  CAS  PubMed  Google Scholar 

  35. Nitiss JL. DNA topoisomerase II and its growing repertoire of biological functions. Nat Rev Cancer. 2009;9:327–37.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  36. Lee B, Min JA, Nashed A, Lee S-O, Yoo JC, Chi S-W, et al. A novel mechanism of Irinotecan targeting MDM2 and Bcl-xL. Biochem Biophys Res Commun. 2019;514:518–23.

    Article  CAS  PubMed  Google Scholar 

  37. Takeba Y, Kumai T, Matsumoto N, Nakaya S, Tsuzuki Y, Yanagida Y, et al. Irinotecan activates p53 with its active metabolite, resulting in human hepatocellular carcinoma apoptosis. J Pharmacol Sci. 2007;104:232–42.

    Article  CAS  PubMed  Google Scholar 

  38. Rudolf E, Kralova V, Rudolf K, John S. The role of p38 in irinotecan-induced DNA damage and apoptosis of colon cancer cells. Mutat Res. 2013;741–742:27–34.

    Article  PubMed  Google Scholar 

  39. Huang T, Song X, Xu D, Tiek D, Goenka A, Wu B, et al. Stem cell programs in cancer initiation, progression, and therapy resistance. Theranostics. 2020;10:8721–43.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  40. Garcia-Mayea Y, Mir C, Masson F, Paciucci R, Lleonart ME. Insights into new mechanisms and models of cancer stem cell multidrug resistance. Semin Cancer Biol. 2020;60:166–80.

    Article  CAS  PubMed  Google Scholar 

  41. Yang L, Shi P, Zhao G, Xu J, Peng W, Zhang J, et al. Targeting cancer stem cell pathways for cancer therapy. Signal Transduct Target Ther. 2020;5:8.

    Article  PubMed  PubMed Central  Google Scholar 

  42. Du J, Yan X, Mi S, Li Y, Ji H, Hou K, et al. Identification of prognostic model and biomarkers for cancer stem cell characteristics in glioblastoma by network analysis of Multi-Omics data and stemness indices. Front Cell Dev Biol. 2020;8:558961.

    Article  PubMed  PubMed Central  Google Scholar 

  43. Gao L, Chen X, Wang Y, Zhang J. Up-Regulation of FSTL3, regulated by LncRNA DSCAM-AS1/miR-122-5p axis, promotes proliferation and migration of Non-Small cell lung cancer cells. Onco Targets Ther. 2020;13:2725–38.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  44. Panagiotou G, Ghaly W, Upadhyay J, Pazaitou-Panayiotou K, Mantzoros CS. Serum follistatin is increased in thyroid cancer and is associated with adverse tumor characteristics in humans. J Clin Endocrinol Metab. 2021;106:e2137–50.

    Article  PubMed  Google Scholar 

  45. Liu Y-J, Li J-P, Zhang Y, Nie M-J, Zhang Y-H, Liu S-L, et al. Is a prognostic biomarker in gastric cancer and is correlated with M2 macrophage infiltration. Onco Targets Ther. 2021;14:4099–117.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  46. Huang MW, Lin YJ, Chang CW, Lei FJ, Ho EP, Liu RS, et al. RGS4 deficit in prefrontal cortex contributes to the behaviors related to schizophrenia via system x(c)(-)-mediated glutamatergic dysfunction in mice. Theranostics. 2018;8:4781–94.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  47. Sun X, Ou Z, Xie M, Kang R, Fan Y, Niu X, et al. HSPB1 as a novel regulator of ferroptotic cancer cell death. Oncogene. 2015;34:5617–25.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

We would like to thank Editage (www.editage.cn) for English language editing.

Funding

This study was supported by the Key Research Foundation of Zhejiang (2022C03015).

Author information

Author notes

  1. Chengyi Huang and Bufu Tang contributed equally to this work.

Authors and Affiliations

  1. Department of Radiation Oncology, Changhai Hospital Affiliated to Naval Medical University, Shanghai, 200433, China

    Chengyi Huang, Wenjuan Chen & Huojun Zhang

  2. Department of Radiation Oncology, Zhejiang Cancer Hospital, Hangzhou, Zhejiang, 310022, China

    Chengyi Huang, Jinggang Chen & Minghua Bai

  3. Zhejiang Key Laboratory of Radiation Oncology, Hangzhou, Zhejiang, 310022, China

    Chengyi Huang, Jinggang Chen & Minghua Bai

  4. Department of Radiation Oncology, Zhongshan Hospital, Fudan University, Shanghai, 200032, China

    Bufu Tang

Authors
  1. Chengyi Huang
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Bufu Tang
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Wenjuan Chen
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Jinggang Chen
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Huojun Zhang
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Minghua Bai
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

C.H. made substantial contributions to the conception of the study, performed the experiments, analysed the data and made figures for this manuscript, and wrote the original draft. B.T. made substantial contributions to the conception of the study, analysed the data and substantively revised the manuscript. W.C. and J.C. made substantial contributions to performing the experiments. H.Z. made substantial contributions to the conception of the study and the review of the paper. M.B. made substantial contributions to the conception of the study and the review of the paper. The final manuscript has been read and revised by all authors.

Corresponding authors

Correspondence to Huojun Zhang or Minghua Bai.

Ethics declarations

Ethics approval

We obtained the expression data and clinical information from the free TCGA and GEO databases. This study strictly followed the publication guidelines and access policies of these databases. TCGA and GEO belong to public databases. The patients involved in the database have obtained ethical approval. Users can download relevant data for free for research and publish relevant articles. Our study is based on open-source data, so there are no ethical issues or other conflicts of interest.

Consent to participate

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note

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

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Huang, C., Tang, B., Chen, W. et al. Multiomic traits reveal that critical irinotecan-related core regulator FSTL3 promotes CRC progression and affects ferroptosis. Cancer Cell Int 25, 115 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12935-025-03753-7

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12935-025-03753-7

Keywords

Cancer Cell International

ISSN: 1475-2867

Contact us