ORIGINAL ARTICLE

Open Access

Construction and experimental validation of a novel ferroptosis-related gene signature for myelodysplastic syndromes

Yidong Zhu

orcid.org/0000-0002-8505-8257

Department of Traditional Chinese Medicine, Shanghai Tenth People's Hospital, Tongji University School of Medicine, Shanghai, China

Contribution: Conceptualization, Formal analysis, Writing - original draft

Search for more papers by this author

Jun He,

Jun He

Department of Hematology, Shanghai Tenth People's Hospital, Tongji University School of Medicine, Shanghai, China

Contribution: Validation

Search for more papers by this author

Rong Wei,

Corresponding Author

Rong Wei

[email protected]

Department of Hematology, Shanghai Tenth People's Hospital, Tongji University School of Medicine, Shanghai, China

Correspondence Rong Wei, Shanghai Tenth People's Hospital, Tongji University School of Medicine, Shanghai, 200072, China.

Email: [email protected]

Contribution: Writing - review & editing

Search for more papers by this author

Jun Liu,

Jun Liu

Department of Traditional Chinese Medicine, Shanghai Tenth People's Hospital, Tongji University School of Medicine, Shanghai, China

Contribution: Funding acquisition

Search for more papers by this author

Yidong Zhu,

Yidong Zhu

orcid.org/0000-0002-8505-8257

Department of Traditional Chinese Medicine, Shanghai Tenth People's Hospital, Tongji University School of Medicine, Shanghai, China

Contribution: Conceptualization, Formal analysis, Writing - original draft

Search for more papers by this author

Jun He,

Jun He

Department of Hematology, Shanghai Tenth People's Hospital, Tongji University School of Medicine, Shanghai, China

Contribution: Validation

Search for more papers by this author

Rong Wei,

Corresponding Author

Rong Wei

[email protected]

Department of Hematology, Shanghai Tenth People's Hospital, Tongji University School of Medicine, Shanghai, China

Correspondence Rong Wei, Shanghai Tenth People's Hospital, Tongji University School of Medicine, Shanghai, 200072, China.

Email: [email protected]

Contribution: Writing - review & editing

Search for more papers by this author

Jun Liu,

Jun Liu

Department of Traditional Chinese Medicine, Shanghai Tenth People's Hospital, Tongji University School of Medicine, Shanghai, China

Contribution: Funding acquisition

Search for more papers by this author

First published: 05 April 2024

https://doi.org/10.1002/iid3.1221

Yidong Zhu and Jun He contributed equally to this work.

Share a link

Email
Wechat
Bluesky

Abstract

Background

Myelodysplastic syndromes (MDS) are clonal hematopoietic disorders characterized by morphological abnormalities and peripheral blood cytopenias, carrying a risk of progression to acute myeloid leukemia. Although ferroptosis is a promising target for MDS treatment, the specific roles of ferroptosis-related genes (FRGs) in MDS diagnosis have not been elucidated.

Methods

MDS-related microarray data were obtained from the Gene Expression Omnibus database. A comprehensive analysis of FRG expression levels in patients with MDS and controls was conducted, followed by the use of multiple machine learning methods to establish prediction models. The predictive ability of the optimal model was evaluated using nomogram analysis and an external data set. Functional analysis was applied to explore the underlying mechanisms. The mRNA levels of the model genes were verified in MDS clinical samples by quantitative real-time polymerase chain reaction (qRT-PCR).

Results

The extreme gradient boosting model demonstrated the best performance, leading to the identification of a panel of six signature genes: SREBF1, PTPN6, PARP9, MAP3K11, MDM4, and EZH2. Receiver operating characteristic curves indicated that the model exhibited high accuracy in predicting MDS diagnosis, with area under the curve values of 0.989 and 0.962 for the training and validation cohorts, respectively. Functional analysis revealed significant associations between these genes and the infiltrating immune cells. The expression levels of these genes were successfully verified in MDS clinical samples.

Conclusion

Our study is the first to identify a novel model using FRGs to predict the risk of developing MDS. FRGs may be implicated in MDS pathogenesis through immune-related pathways. These findings highlight the intricate correlation between ferroptosis and MDS, offering insights that may aid in identifying potential therapeutic targets for this debilitating disorder.

1 INTRODUCTION

Myelodysplastic syndromes (MDS) are a group of clonal hematopoietic diseases characterized by abnormal development of bone marrow cells, ineffective hematopoiesis, peripheral cytopenia, and a risk of transformation to acute myeloid leukemia.¹ The clinical presentation is nonspecific and includes symptoms and signs of thrombocytopenia, anemia, and neutropenia.² Diagnosis of MDS primarily relies on the presence of cytopenia, morphological evidence of dysplasia in bone marrow aspirate and biopsy, and clonal cytogenetic abnormalities.^{1, 3, 4} However, MDS cases vary in clinical presentation, with some exhibiting atypical features, including inapparent or absent cytopenia and dysplasia, upon examination.^{5, 6} Diagnostic discrepancies may occur in 12% of patients at the time of initial presentation, affecting therapeutic decision-making.⁷ Timely and accurate diagnosis, followed by effective management, can significantly interfere with disease progression and improve overall prognosis.⁸ Therefore, the identification of novel and reliable biomarkers is imperative for improving diagnostic accuracy, which is essential for the effective treatment of MDS.

A considerable percentage of patients with MDS experience iron overload, which is primarily attributed to intermittent blood transfusions and ineffective hematopoiesis.⁹ In severe cases, this overload may trigger an excessive accumulation of reactive oxygen species, thereby inducing ferroptosis, a unique form of iron-dependent cell death.¹⁰ Ferroptosis arises from a redox imbalance in the production of oxidants and antioxidants as a consequence of the abnormal expression and activity of several redox-active enzymes involved in the production or detoxification of lipid oxidation products and free radicals.^11-15 Increasing evidence indicates that ferroptosis plays a critical role in cancer progression.^{16, 17} For instance, Ubellacker et al. reported that lymphoid tissues protected tumor cells from ferroptosis and promoted melanoma metastasis.¹⁸ Nagpal et al. suggested that the induction of ferroptosis inhibited brain metastasis of tumors in a spontaneous mouse model of HER2-positive breast cancer.¹⁹ Additionally, ferroptosis injury could trigger inflammation-related immunosuppression in the tumor microenvironment, thereby favoring tumor growth.²⁰ Given its distinct features, ferroptosis has attracted substantial interest as a potential treatment target for hematological malignancies, such as MDS, leukemia, lymphoma, and multiple myeloma.²¹ A recent study showed that the antileukemic drug decitabine induces ferroptosis in MDS. Treatment of MDS cell lines with decitabine leads to increased levels of reactive oxygen species and reduced activity of glutathione and glutathione peroxidase 4, suggesting a significant association between ferroptosis and MDS.²² However, the role of ferroptosis-related genes (FRGs) in the pathogenesis of MDS remains unclear.

Studies have demonstrated that FRGs have the potential to predict immune responses and clinical outcomes in hematological malignancies, including lymphoma,^23-25 leukemia,^26-28 and multiple myeloma.^29-31 In this study, we aimed to examine and validate the accuracy of FRGs as biomarkers for MDS. This study was initiated by performing differential expression analysis between MDS and control samples, followed by various machine learning methods to build a predictive model with differentially expressed FRGs. The efficacy of the model was evaluated using nomogram analysis and an external data set. Functional analysis was conducted to explore potential mechanisms. In addition, the mRNA levels of the identified genes in clinical MDS samples were confirmed by quantitative real-time polymerase chain reaction (qRT-PCR). This study is the first to establish an FRG diagnostic signature in patients with MDS, which will enhance our understanding of the pathogenesis and progression of MDS and aid in the development of personalized therapy strategies in clinical practice.

2 MATERIALS AND METHODS

2.1 Data collection

Microarray data and clinical information from patients with MDS and control samples were downloaded from the Gene Expression Omnibus database (GEO, https://www.ncbi.nlm.nih.gov/geo/). All data sets were normalized to remove batch effects. The GSE19429 data set, consisting of 183 MDS and 17 normal samples, was used as the training cohort. The GSE58831 data set, containing 159 MDS and 17 normal samples, was chosen as the validation cohort. In total, 728 FRGs were analyzed based on the FerrDb database (Supplementary Table S1).

2.2 Screening of differentially expressed FRGs

In the training cohort, the expression of FRGs in both MDS and control samples was extracted. Differential expression analysis was conducted using the “limma” package. FRGs with p-values < 0.05 were retained.

2.3 Establishment of the optimal prediction model

Multiple machine learning methods were employed to develop a prediction model using the differentially expressed FRGs. Random forest (RF) is an ensemble of classification and regression trees that addresses overfitting and exhibits greater stability in high-dimensional parameter spaces and in the presence of outliers.^{32, 33} Support vector machine (SVM) is a powerful method for building classifiers and creating decision boundaries between 2 classes to predict labels from one or more feature vectors.^{34, 35} Generalized linear models (GLMs) are statistical models that enable the modeling of relations between a response variable and one or more predictor variables, extending linear regression models to handle non-normal distributions of the response variable, including binary, count, or continuous data with nonconstant variance.^{36, 37} Extreme gradient boosting (XGB) is a machine learning model that integrates multiple weak learners to achieve a stronger learning effect, exhibiting strong flexibility and scalability advantages.^{38, 39} The “Caret” package adjusts the parameters of these models through a grid search. These machine learning models were implemented with default parameters. Prediction models based on these algorithms were constructed accordingly. Subsequently, the residual distributions and feature importance of these models were analyzed. Receiver operating characteristic (ROC) curves were generated to evaluate the model accuracy and specificity. Based on the aforementioned performance, the optimal model was determined, and the top 6 feature variables in the model were selected as the optimal combination of gene signatures for MDS diagnosis. The results of these analyses were visualized using the “caret,” “DALEX,” “ggplot2,” “randomForest,” “kernlab,” “pROC,” and “xgboost” packages.

2.4 Nomogram construction

A nomogram was generated using the identified gene signature to predict the risk of MDS. We conducted nomogram analysis to evaluate the predictive ability of the model using the “rms” and “rmda” packages.

2.5 External data set validation

The expression levels of signature genes were verified in the validation cohort. Furthermore, ROC curve analysis was used to evaluate the diagnostic value of the gene signature using the “ggpubr” and “pROC” packages.

2.6 Functional analysis

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were performed to explore the underlying mechanisms on the differentially expressed FRGs. Gene set enrichment analysis (GSEA) was used to select significant pathways associated with each signature gene. Moreover, the CIBERSORT algorithm was used to calculate the relative abundance of 22 types of immune cells in each sample, and the differences between MDS and control samples were compared to explore the potential correlation between MDS and immunity. Spearman correlation analysis was conducted to evaluate the relation between signature genes and immune cells. The outcomes were visualized using the “clusterProfiler,” “enrichplot,” “DOSE,” “pheatmap,” “GSVA,” “GSEABase,” “reshape2,” “ggpubr,” “ggplot2,” and “tidyverse” packages.

2.7 Establishment of a competitive endogenous RNA (ceRNA) network

We constructed a ceRNA network by integrating the interactions among mRNAs, microRNAs (miRNAs), and long noncoding RNAs (lncRNAs). The interaction pairs between mRNAs and miRNAs were predicted by intersecting the miRDB, TargetScan, and miRanda databases using signature genes. To identify possible interactions between lncRNAs and miRNAs, we searched the SpongeScan database. The network was visualized using Cytoscape version 3.8.2.

2.8 qRT-PCR

Bone marrow samples were collected from 4 patients with MDS and 4 healthy volunteers. This study received approval from the Ethics Committee of Shanghai Tenth People's Hospital (24K14), and informed consent was obtained from all participants. Peripheral blood mononuclear cells were isolated to validate the expression patterns of the identified genes. Total RNA was extracted using TRIzol reagent (Thermo Fisher Scientific, USA). Subsequently, cDNA was reverse-transcribed using a PrimeScript^TM RT Reagent Kit (Takara, Japan). Quantitative real-time PCR was carried out with the SYBR Green PCR Master Mix (KAPA, Japan) and the Applied Biosystems 7900HT Fast Real-Time PCR System (Thermo Fisher Scientific, USA). The relative expression levels of genes were detected using the 2^−ΔΔCt method. The primer sequences are listed in Supplementary Table S2. GAPDH served as an internal control.

2.9 Statistical analysis

Data were analyzed using R 4.1.3 and GraphPad Prism 8.0.1. Student's t-test was used to compare normally distributed measurement data, whereas Wilcoxon test was used for skewed data. A two-sided p-value < 0.05 was considered significant.

3 RESULTS

3.1 Identification of differentially expressed FRGs

The flowchart is presented in Figure 1. After analyzing the training set, we identified 117 FRGs as differentially expressed genes between the MDS and control samples. Of these, 61 were upregulated and 56 were downregulated in patients with MDS, meeting the established criteria (Supplementary Table S3).

Details are in the caption following the image — **Figure 1**
Open in figure viewer PowerPoint

Flowchart of this study. MDS, Myelodysplastic syndromes; qRT-PCR, Quantitative real-time polymerase chain reaction; ceRNA, competitive endogenous RNA; GEO, Gene Expression Omnibus.

3.2 Establishment of the optimal prediction model

Based on the differentially expressed FRGs, several machine learning models were developed, among which the XGB-based prediction model exhibited the lowest residuals (Figure 2A and B). In each model, the top 10 feature variables were ranked based on the root mean square error (Figure 2C). The ROC curves demonstrated that the XGB-based prediction model had the highest area under the curve (AUC) value of 0.989 (Figure 2D). These findings suggested that the XGB algorithm was superior to the other algorithms in diagnosing MDS. Consequently, the top 6 feature variables (SREBF1, PTPN6, PARP9, MAP3K11, MDM4, and EZH2) in the XGB model were considered as the optimal combination of signature genes. In comparison to normal controls, the expression levels of SREBF1, MAP3K11, PARP9, and PTPN6 were upregulated, whereas EZH2 and MDM4 were downregulated in the MDS samples (Figure 3A–F). Moreover, these signature genes showed strong synergistic or antagonistic effects, and their interaction relations were visualized (Figure 3G and H).

3.3 Nomogram construction

A nomogram was developed to predict the risk of developing MDS (Figure 4A). The calibration curve demonstrated a high degree of concordance between the observed and predicted probabilities (Figure 4B). Furthermore, the decision curve illustrated that the nomogram provided substantial benefits for clinical decision-making (Figure 4C). These findings indicated that the nomogram model exhibited excellent performance in predicting MDS diagnosis.

3.4 External data set validation

In the independent validation cohort, the signature genes showed gene expression patterns consistent with those in the training cohort (Figure 5A–F). The ROC curve also revealed that the gene signature exhibited a high diagnostic value for MDS (AUC = 0.962) (Figure 5G).

3.5 Functional analysis

GO analysis revealed a significant enrichment of differentially expressed FRGs in peptidyl-serine phosphorylation, modification, and cellular responses to chemical stress (Figure 6A). KEGG analysis revealed that these FRGs were significantly involved in immune-related diseases, such as hepatitis B, Epstein-Barr virus infection, Kaposi's sarcoma-associated herpes virus infection, and human cytomegalovirus infection (Figure 6B). Furthermore, GSEA demonstrated that the top 6 pathways enriched by each signature gene were mainly involved in immune-related pathways, such as the NOD-like receptor signaling pathway, primary immunodeficiency, and the RIG-I-like receptor signaling pathway (Figure 6C–H). In addition, immune cell infiltration analysis showed that memory B cells and activated dendritic cells were significantly lower in MDS samples than in control samples, whereas plasma cells, follicular helper T cells, activated CD8⁺ T cells, activated natural killer cells, and both M1 and M2 macrophages were significantly higher in MDS samples (Figure 6I). Moreover, correlation analysis revealed significant correlations between the signature genes and immune cells (Figure 6J).

3.6 Establishment of a ceRNA network

To investigate the regulatory mechanisms of the signature genes, a ceRNA network was constructed (Figure 7). This network encompassed 425 nodes, comprising 6 mRNAs, 206 miRNAs, and 213 lncRNAs interconnected by 508 edges (Supplementary Table S4).

3.7 qRT-PCR

Subsequently, qRT-PCR was employed to validate the expression of the identified genes in both MDS patients and healthy volunteers. As anticipated, compared to healthy volunteers, the expression levels of SREBF1, MAP3K11, PARP9, and PTPN6 were found to be upregulated, while the expression levels of EZH2 and MDM4 were downregulated in the MDS samples (Figure 8).

4 DISCUSSION

Given the critical role of prompt and accurate diagnosis coupled with effective MDS management in significantly influencing disease progression and improving overall prognosis, it is essential to investigate the potential effect of FRGs on predicting MDS diagnosis. The key achievement of our study was the pioneering success in constructing and validating a novel signature for patients with MDS based on FRGs. Further analyses revealed a significant correlation between the signature genes and immune-related pathways. These findings provide a theoretical molecular framework that enhances our understanding of MDS and may inform the development of future diagnostic strategies.

The present study identified 117 differentially expressed FRGs between MDS and control groups, highlighting the crucial role of FRGs in MDS development. Using these differentially expressed FRGs, we constructed prediction models using various machine learning algorithms. Our analysis indicated that the XGB model delivered the most accurate diagnostic predictions for MDS compared with the other tested algorithms. With the highest AUC values and lowest residuals, the XGB model demonstrated superior performance among all the tested models. Consequently, we selected 6 key variables identified using the XGB model to form the optimal signature gene panel: SREBF1, PTPN6, PARP9, MAP3K11, MDM4, and EZH2. In MDS samples, the expression of SREBF1, MAP3K11, PARP9, and PTPN6 was significantly elevated, while EZH2 and MDM4 was notably reduced compared with that in controls. These gene expression patterns were consistent between the training and validation cohorts. It was also worth noting that the gene signature-based model had AUCs of 0.989 and 0.962 for the 2 cohorts, indicating its robust performance in predicting MDS diagnosis. Among the signature genes, PTPN6 hypermethylation has been associated with poor prognosis in patients with high-risk MDS.^40-42 MDM4 has been implicated as a common mechanism for the transition from preleukemia to acute myeloid leukemia in several genetic disease subtypes.⁴³ EZH2 loss-of-function mutations are common in MDS and contribute significantly to its pathogenesis.^44-46 In addition, as a key transcription factor, SREBF1 regulates genes involved in lipid homeostasis and cholesterol biosynthesis.^47-49 PARP9 is involved in immune response and DNA damage repair.^50-52 MAP3K11 activates the JUN N-terminal pathway and affects microtubule organization during the cell cycle.^53-55 Notably, the expression of these model genes was verified in clinical MDS samples, strengthening the validity of the findings. The exact roles of SREBF1, PARP9, and MAP3K11 in MDS remain unclear, and our results provide new insights into the pathogenesis of MDS. Furthermore, the distinct functions of signature genes in MDS warrant further investigation. To the best of our knowledge, there have been some studies on FRG-based prediction models for human cancers. For example, Shao et al. identified 12 FRGs and generated a prognostic model in acute myeloid leukemia.⁵⁶ Wu et al. established a prognostic prediction model for triple-negative breast cancer based on 15 FRGs.⁵⁷ Both studies selected gene signatures using LASSO regression. In contrast, we constructed a prediction model using various machine learning algorithms, which increased the reliability of the model.

Functional analysis was performed to investigate the mechanism of action of the signature genes in MDS. KEGG analysis revealed that the differentially expressed FRGs were significantly associated with different types of infection, highlighting their close relation with immunity. GSEA further confirmed this by showing enrichment of signature genes in immune-related pathways, including the NOD-like receptor signaling pathway, primary immunodeficiency, and the RIG-I-like receptor signaling pathway. Among these pathways, the NOD-like receptors are intracellular proteins that play a central role in both innate and adaptive immunity.^{58, 59} Similarly, the RIG-I-like receptors have a significant impact on sensing RNA viral infections and initiating and modulating antiviral immunity.^{60, 61} To determine the differences in immune cell infiltration between the 2 groups, the CIBERSORT algorithm was applied and revealed significant differences in the number of infiltrating immune cells between the 2 groups, supporting the notion that immunity plays a key role in MDS progression. The existing literature has increasingly recognized the involvement of different types of immune cells and their downstream pathways in the hematopoietic niche of MDS,^62-65 which aligns with our research. Further analysis revealed a significant correlation between the identified genes and immune cells, indicating that these genes may participate in MDS development through immune-related pathways. Understanding the regulation of these genes by the ceRNA network offers valuable insights into the molecular mechanisms underlying MDS, thereby paving the way for future studies.

Although our prediction model demonstrated satisfactory predictive power and was validated using an external data set, some limitations must be considered. First, our model was established based on available public data. Although qRT-PCR confirmed the differential gene expression identified in this study, further prospective studies were required to verify its predictive power. Second, the present model did not consider alternative splice forms, which were significantly altered in patients with MDS. Third, the underlying mechanisms linking ferroptosis to MDS progression and the relation between the immune microenvironment and MDS remained unclear. The molecular mechanisms underlying these new biomarkers require further clarification and experimental validation. These elements will be the focus of future research.

In conclusion, our study pioneered the development of a novel 6-FRG predictive model for patients with MDS. This prediction model demonstrated high diagnostic predictivity with an AUC > 0.9, suggesting that it has a robust potential for effective integration into clinical practice. Furthermore, functional analysis highlighted meaningful associations between the signature genes and immune-associated pathways. The expression patterns of these genes were consistent in MDS clinical samples, indicating their potential for effective clinical application.

AUTHOR CONTRIBUTIONS

Yidong Zhu: Conceptualization (lead); writing—original draft (lead); formal analysis (lead). Jun He: Experimental validation (lead). Rong Wei: Writing—review and editing (equal). Jun Liu: Writing—review and editing (equal). All read and authors gave their approval for publication of the final version of the manuscript.

ACKNOWLEDGEMENTS

We thank the contributions of the GEO project. This work was sponsored by the National Facility for Translational Medicine (Shanghai) (TMSK-2021-413).

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

ETHICS APPROVAL AND CONSENT TO PARTICIPATE

This study was approved by the Ethics Committee of Shanghai Tenth People's Hospital (24K14). Informed consent was obtained from all subjects. All methods were performed in accordance with the Declaration of Helsinki.

CONSENT FOR PUBLICATION

Not applicable.

Open Research

DATA AVAILABILITY STATEMENT

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Supporting Information

REFERENCES

1Cazzola M. Myelodysplastic syndromes. N Engl J Med. 2020; 383(14): 1358-1374.
10.1056/NEJMra1904794
CAS PubMed Web of Science® Google Scholar
2Sekeres MA, Taylor J. Diagnosis and treatment of myelodysplastic syndromes: A review. JAMA. 2022; 328(9): 872-880.
10.1001/jama.2022.14578
PubMed Web of Science® Google Scholar
3Li H, Hu F, Gale RP, Sekeres MA, Liang Y. Myelodysplastic syndromes. Nat Rev Dis Primers. 2022; 8(1): 74.
10.1038/s41572-022-00402-5
PubMed Google Scholar
4Greenberg PL, Stone RM, Al-Kali A, et al. NCCN guidelines® insights: myelodysplastic syndromes, version 3.2022. Journal of the National Comprehensive Cancer Network. 2022; 20(2): 106-117.
10.6004/jnccn.2022.0009
PubMed Web of Science® Google Scholar
5Garcia-Manero G. Myelodysplastic syndromes: 2023 update on diagnosis, risk-stratification, and management. Am J Hematol. 2023; 98(8): 1307-1325.
10.1002/ajh.26984
CAS PubMed Web of Science® Google Scholar
6Valent P, Orazi A, Steensma DP, et al. Proposed minimal diagnostic criteria for myelodysplastic syndromes (MDS) and potential pre-MDS conditions. Oncotarget. 2017; 8(43): 73483-73500.
10.18632/oncotarget.19008
PubMed Google Scholar
7Naqvi K, Jabbour E, Bueso-Ramos C, et al. Implications of discrepancy in morphologic diagnosis of myelodysplastic syndrome between referral and tertiary care centers. Blood. 2011; 118(17): 4690-4693.
10.1182/blood-2011-03-342642
CAS PubMed Web of Science® Google Scholar
8Platzbecker U, Kubasch AS, Homer-Bouthiette C, Prebet T. Current challenges and unmet medical needs in myelodysplastic syndromes. Leukemia. 2021; 35(8): 2182-2198.
10.1038/s41375-021-01265-7
PubMed Web of Science® Google Scholar
9Shenoy N, Vallumsetla N, Rachmilewitz E, Verma A, Ginzburg Y. Impact of iron overload and potential benefit from iron chelation in low-risk myelodysplastic syndrome. Blood. 2014; 124(6): 873-881.
10.1182/blood-2014-03-563221
CAS PubMed Web of Science® Google Scholar
10Weber S, Parmon A, Kurrle N, Schnütgen F, Serve H. The clinical significance of iron overload and iron metabolism in myelodysplastic syndrome and acute myeloid leukemia. Front Immunol. 2020; 11:627662.
10.3389/fimmu.2020.627662
CAS PubMed Web of Science® Google Scholar
11Tang D, Chen X, Kang R, Kroemer G. Ferroptosis: molecular mechanisms and health implications. Cell Res. 2021; 31(2): 107-125.
10.1038/s41422-020-00441-1
CAS PubMed Web of Science® Google Scholar
12Jiang X, Stockwell BR, Conrad M. Ferroptosis: mechanisms, biology and role in disease. Nat Rev Mol Cell Biol. 2021; 22(4): 266-282.
10.1038/s41580-020-00324-8
PubMed Web of Science® Google Scholar
13Chen X, Li J, Kang R, Klionsky DJ, Tang D. Ferroptosis: machinery and regulation. Autophagy. 2021; 17(9): 2054-2081.
10.1080/15548627.2020.1810918
CAS PubMed Web of Science® Google Scholar
14Meng H, Yu Y, Xie E, et al. Hepatic HDAC3 regulates systemic iron homeostasis and ferroptosis via the hippo signaling pathway. Research. 2023; 6: 0281.
10.34133/research.0281
CAS PubMed Google Scholar
15Liu Y, Jiao C, Zhang T, et al. Early-Life gut microbiota governs susceptibility to colitis via Microbial-Derived ether lipids. Research. 2023; 6: 0037.
10.34133/research.0037
CAS PubMed Google Scholar
16Zou Y, Zheng S, Xie X, et al. N6-methyladenosine regulated FGFR4 attenuates ferroptotic cell death in recalcitrant HER2-positive breast cancer. Nat Commun. 2022; 13(1): 2672.
10.1038/s41467-022-30217-7
CAS PubMed Web of Science® Google Scholar
17Liu Y, Liu Y, Ye S, Feng H, Ma L. A new ferroptosis-related signature model including messenger RNAs and long non-coding RNAs predicts the prognosis of gastric cancer patients. J Transl Int Med. 2023; 11(2): 145-155.
10.2478/jtim-2023-0089
CAS PubMed Web of Science® Google Scholar
18Ubellacker JM, Tasdogan A, Ramesh V, et al. Lymph protects metastasizing melanoma cells from ferroptosis. Nature. 2020; 585(7823): 113-118.
10.1038/s41586-020-2623-z
CAS PubMed Web of Science® Google Scholar
19Nagpal A, Redvers RP, Ling X, et al. Neoadjuvant neratinib promotes ferroptosis and inhibits brain metastasis in a novel syngeneic model of spontaneous HER2(+ve) breast cancer metastasis. Breast Cancer Res. 2019; 21(1): 94.
10.1186/s13058-019-1177-1
PubMed Web of Science® Google Scholar
20Chen X, Kang R, Kroemer G, Tang D. Broadening horizons: the role of ferroptosis in cancer. Nat Rev Clin Oncol. 2021; 18(5): 280-296.
10.1038/s41571-020-00462-0
CAS PubMed Web of Science® Google Scholar
21Li J, Zhang W. From iron chelation to overload as a therapeutic strategy to induce ferroptosis in hematologic malignancies. Hematology. 2022; 27(1): 1163-1170.
10.1080/16078454.2022.2132362
PubMed Web of Science® Google Scholar
22Lv Q, Niu H, Yue L, et al. Abnormal ferroptosis in myelodysplastic syndrome. Front Oncol. 2020; 10: 1656.
10.3389/fonc.2020.01656
PubMed Web of Science® Google Scholar
23Xiong D, Li M, Zeng C. Construction and validation of a risk scoring model for diffuse large b-cell lymphoma based on ferroptosis-related genes and its association with immune infiltration. Transl Oncol. 2022; 16:101314.
10.1016/j.tranon.2021.101314
CAS PubMed Web of Science® Google Scholar
24Weng J, Chen L, Liu H, Yang XP, Huang L. Ferroptosis markers predict the survival, immune infiltration, and ibrutinib resistance of diffuse large B cell lymphoma. Inflammation. 2022; 45(3): 1146-1161.
10.1007/s10753-021-01609-6
CAS PubMed Web of Science® Google Scholar
25Chen H, He Y, Pan T, et al. Ferroptosis-Related gene signature: A new method for personalized risk assessment in patients with diffuse large B-Cell lymphoma. Pharmgenomics Pers Med. 2021; 14: 609-619.
PubMed Web of Science® Google Scholar
26Song Y, Tian S, Zhang P, Zhang N, Shen Y, Deng J. Construction and validation of a novel Ferroptosis-Related prognostic model for acute myeloid leukemia. Front Genet. 2021; 12:708699.
10.3389/fgene.2021.708699
CAS PubMed Web of Science® Google Scholar
27Wang J, Zhuo Z, Wang Y, et al. Identification and validation of a prognostic Risk-Scoring model based on Ferroptosis-Associated cluster in acute myeloid leukemia. Front Cell Dev Biol. 2021; 9:800267.
10.3389/fcell.2021.800267
PubMed Web of Science® Google Scholar
28Cui Z, Fu Y, Yang Z, et al. Comprehensive analysis of a ferroptosis pattern and associated prognostic signature in acute myeloid leukemia. Front Pharmacol. 2022; 13:866325.
10.3389/fphar.2022.866325
CAS PubMed Web of Science® Google Scholar
29Su Q, Li Q, Zhang W, Li B, Zhuang W. Integrative analysis of enrichment and prognostic value of ferroptosis-related genes and pathways in multiple myeloma. Carcinogenesis. 2022; 43(11): 1050-1058.
10.1093/carcin/bgac080
CAS PubMed Web of Science® Google Scholar
30Gao D, Liu R, Lv Y, et al. A novel ferroptosis-related gene signature for predicting prognosis in multiple myeloma. Front Oncol. 2023; 13:999688.
10.3389/fonc.2023.999688
CAS PubMed Google Scholar
31Fu B, Shao R, Wang H, Chen G, Bai S, Wang H. Integrated assessment of the clinical and biological value of ferroptosis-related genes in multiple myeloma. Cancer Cell Int. 2022; 22(1): 326.
10.1186/s12935-022-02742-4
CAS PubMed Web of Science® Google Scholar
32Rigatti SJ. Random forest. J Insur Med. 2017; 47(1): 31-39.
10.17849/insm-47-01-31-39.1
PubMed Google Scholar
33Sarica A, Cerasa A, Quattrone A. Random forest algorithm for the classification of neuroimaging data in alzheimer's disease: A systematic review. Front Aging Neurosci. 2017; 9: 329.
10.3389/fnagi.2017.00329
PubMed Web of Science® Google Scholar
34Noble WS. What is a support vector machine? Nature Biotechnol. 2006; 24(12): 1565-1567.
10.1038/nbt1206-1565
CAS PubMed Web of Science® Google Scholar
35Huang S, Cai N, Pacheco PP, Narrandes S, Wang Y, Xu W. Applications of support vector machine (SVM) learning in cancer genomics. Cancer Genomics Proteomics. 2018; 15(1): 41-51.
CAS PubMed Web of Science® Google Scholar
36Lindsey JK. A review of some extensions to generalized linear models. Stat Med. 1999; 18(17-18): 2223-2236.
10.1002/(SICI)1097-0258(19990915/30)18:17/18<2223::AID-SIM251>3.0.CO;2-6
CAS PubMed Web of Science® Google Scholar
37Zareie B, Poorolajal J, Roshani A, Karami M. Outbreak detection algorithms based on generalized linear model: a review with new practical examples. BMC Med Res Methodol. 2023; 23(1): 235.
10.1186/s12874-023-02050-z
PubMed Web of Science® Google Scholar
38Yu B, Qiu W, Chen C, et al. SubMito-XGBoost: predicting protein submitochondrial localization by fusing multiple feature information and extreme gradient boosting. Bioinformatics. 2020; 36(4): 1074-1081.
10.1093/bioinformatics/btz734
CAS PubMed Web of Science® Google Scholar
39Zhao Z, Yang W, Zhai Y, Liang Y, Zhao Y. Identify DNA-Binding proteins through the extreme gradient boosting algorithm. Front Genet. 2021; 12:821996.
10.3389/fgene.2021.821996
CAS PubMed Web of Science® Google Scholar
40Han Y, Zhang J, Pang Y, Wang Y, Zhang X, Zhang H. The role of src homology region 2 domain-containing phosphatase-1 hypermethylation in the classification of patients with myelodysplastic syndromes and its association with signal transducer and activator of transcription 3 phosphorylation in skm-1 cells. J Int Med Res. 2021; 49(3):030006052199955.
10.1177/0300060521999550
Web of Science® Google Scholar
41Zhang Y, Zhao D, Zhao H, et al. Hypermethylation of SHP-1 promoter in patient with high-risk myelodysplastic syndrome and it predicts poor prognosis. Med Oncol. 2012; 29(4): 2359-2363.
10.1007/s12032-012-0163-6
CAS PubMed Web of Science® Google Scholar
42Mena-Duran AV, Togo SH, Bazhenova L, et al. SHP1 expression in bone marrow biopsies of myelodysplastic syndrome patients: a new prognostic factor. Br J Haematol. 2005; 129(6): 791-794.
10.1111/j.1365-2141.2005.05516.x
CAS PubMed Web of Science® Google Scholar
43Ueda K, Kumari R, Schwenger E, et al. MDMX acts as a pervasive preleukemic-to-acute myeloid leukemia transition mechanism. Cancer Cell. 2021; 39(4): 529-547.
10.1016/j.ccell.2021.02.006
CAS PubMed Web of Science® Google Scholar
44Sashida G, Harada H, Matsui H, et al. Ezh2 loss promotes development of myelodysplastic syndrome but attenuates its predisposition to leukaemic transformation. Nat Commun. 2014; 5: 4177.
10.1038/ncomms5177
CAS PubMed Web of Science® Google Scholar
45Bejar R, Stevenson K, Abdel-Wahab O, et al. Clinical effect of point mutations in myelodysplastic syndromes. N Engl J Med. 2011; 364(26): 2496-2506.
10.1056/NEJMoa1013343
CAS PubMed Web of Science® Google Scholar
46Nikoloski G, Langemeijer SMC, Kuiper RP, et al. Somatic mutations of the histone methyltransferase gene EZH2 in myelodysplastic syndromes. Nature Genet. 2010; 42(8): 665-667.
10.1038/ng.620
CAS PubMed Web of Science® Google Scholar
47Xu D, Wang Z, Xia Y, et al. The gluconeogenic enzyme PCK1 phosphorylates INSIG1/2 for lipogenesis. Nature. 2020; 580(7804): 530-535.
10.1038/s41586-020-2183-2
CAS PubMed Web of Science® Google Scholar
48Amemiya-Kudo M, Shimano H, Hasty AH, et al. Transcriptional activities of nuclear SREBP-1a, -1c, and -2 to different target promoters of lipogenic and cholesterogenic genes. J Lipid Res. 2002; 43(8): 1220-1235.
10.1194/jlr.M100417-JLR200
CAS PubMed Web of Science® Google Scholar
49Yokoyama C, Wang X, Briggs MR, et al. SREBP-1, a basic-helix-loop-helix-leucine zipper protein that controls transcription of the low density lipoprotein receptor gene. Cell. 1993; 75(1): 187-197.
10.1016/S0092-8674(05)80095-9
CAS PubMed Web of Science® Google Scholar
50Iwata H, Goettsch C, Sharma A, et al. PARP9 and PARP14 cross-regulate macrophage activation via STAT1 ADP-ribosylation. Nat Commun. 2016; 7:12849.
10.1038/ncomms12849
CAS PubMed Web of Science® Google Scholar
51Zhang Y, Mao D, Roswit WT, et al. PARP9-DTX3L ubiquitin ligase targets host histone H2BJ and viral 3C protease to enhance interferon signaling and control viral infection. Nature Immunol. 2015; 16(12): 1215-1227.
10.1038/ni.3279
CAS PubMed Web of Science® Google Scholar
52Yan Q, Xu R, Zhu L, et al. BAL1 and its partner E3 ligase, BBAP, link Poly(ADP-ribose) activation, ubiquitylation, and double-strand DNA repair independent of ATM, MDC1, and RNF8. Mol Cell Biol. 2013; 33(4): 845-857.
10.1128/MCB.00990-12
CAS PubMed Web of Science® Google Scholar
53Chadee DN, Kyriakis JM. MLK3 is required for mitogen activation of B-Raf, ERK and cell proliferation. Nature Cell Biol. 2004; 6(8): 770-776.
10.1038/ncb1152
CAS PubMed Web of Science® Google Scholar
54Swenson KI, Winkler KE, Means AR. A new identity for MLK3 as an NIMA-related, cell cycle-regulated kinase that is localized near centrosomes and influences microtubule organization. Mol Biol Cell. 2003; 14(1): 156-172.
10.1091/mbc.e02-02-0115
CAS PubMed Web of Science® Google Scholar
55Tibbles LA, Ing YL, Kiefer F, et al. MLK-3 activates the SAPK/JNK and p38/RK pathways via SEK1 and MKK3/6. EMBO J. 1996; 15(24): 7026-7035.
10.1002/j.1460-2075.1996.tb01094.x
CAS PubMed Web of Science® Google Scholar
56Shao R, Wang H, Liu W, et al. Establishment of a prognostic ferroptosis-related gene profile in acute myeloid leukaemia. J Cell Mol Med. 2021; 25(23): 10950-10960.
10.1111/jcmm.17013
CAS PubMed Web of Science® Google Scholar
57Wu S, Pan R, Lu J, et al. Development and verification of a prognostic Ferroptosis-Related gene model in Triple-Negative breast cancer. Front Oncol. 2022; 12:896927.
10.3389/fonc.2022.896927
CAS PubMed Web of Science® Google Scholar
58Zhou Y, Yu S, Zhang W. NOD-like receptor signaling pathway in gastrointestinal inflammatory diseases and cancers. Int J Mol Sci. 2023; 24(19):14511.
10.3390/ijms241914511
CAS PubMed Web of Science® Google Scholar
59Platnich JM, Muruve DA. NOD-like receptors and inflammasomes: A review of their canonical and non-canonical signaling pathways. Arch Biochem Biophys. 2019; 670: 4-14.
10.1016/j.abb.2019.02.008
CAS PubMed Web of Science® Google Scholar
60Loo YM, Gale, Jr. M. Immune signaling by RIG-I-like receptors. Immunity. 2011; 34(5): 680-692.
10.1016/j.immuni.2011.05.003
CAS PubMed Web of Science® Google Scholar
61Onomoto K, Onoguchi K, Yoneyama M. Regulation of RIG-I-like receptor-mediated signaling: interaction between host and viral factors. Cell Mol Immunol. 2021; 18(3): 539-555.
10.1038/s41423-020-00602-7
CAS PubMed Web of Science® Google Scholar
62Wang C, Yang Y, Gao S, et al. Immune dysregulation in myelodysplastic syndrome: clinical features, pathogenesis and therapeutic strategies. Crit Rev Oncol Hematol. 2018; 122: 123-132.
10.1016/j.critrevonc.2017.12.013
PubMed Web of Science® Google Scholar
63Trowbridge J, Starczynowski D. Innate immune pathways and inflammation in hematopoietic aging, clonal hematopoiesis, and MDS. the. J Exp Med. 2021; 218(7):e20201544.
10.1084/jem.20201544
CAS PubMed Web of Science® Google Scholar
64Sallman DA, List A. The central role of inflammatory signaling in the pathogenesis of myelodysplastic syndromes. Blood. 2019; 133(10): 1039-1048.
10.1182/blood-2018-10-844654
CAS PubMed Web of Science® Google Scholar
65Peng X, Zhu X, Di T, et al. The yin-yang of immunity: immune dysregulation in myelodysplastic syndrome with different risk stratification. Front Immunol. 2022; 13:994053.
10.3389/fimmu.2022.994053
CAS PubMed Web of Science® Google Scholar

Volume12, Issue4

April 2024

e1221

Filename	Description
iid31221-sup-0001-Supplementary_Table_S1.doc48.5 KB	Supplementary Table S1. Ferroptosis-related genes.
iid31221-sup-0002-Supplementary_Table_S2.doc34.5 KB	Supplementary Table S2. Primers used for quantitative real-time polymerase chain reaction.
iid31221-sup-0003-Supplementary_Table_S3.doc189 KB	Supplementary Table S3. Identification of differentially expressed ferroptosis-related genes.
iid31221-sup-0004-Supplementary_Table_S4.doc430 KB	Supplementary Table S4. Specific details of the competitive endogenous RNA network.

Construction and experimental validation of a novel ferroptosis-related gene signature for myelodysplastic syndromes

Abstract

Background

Methods

Results

Conclusion

1 INTRODUCTION

2 MATERIALS AND METHODS

2.1 Data collection

2.2 Screening of differentially expressed FRGs

2.3 Establishment of the optimal prediction model

2.4 Nomogram construction

2.5 External data set validation

2.6 Functional analysis

2.7 Establishment of a competitive endogenous RNA (ceRNA) network

2.8 qRT-PCR

2.9 Statistical analysis

3 RESULTS

3.1 Identification of differentially expressed FRGs

3.2 Establishment of the optimal prediction model

3.3 Nomogram construction

3.4 External data set validation

3.5 Functional analysis

3.6 Establishment of a ceRNA network

3.7 qRT-PCR

4 DISCUSSION

AUTHOR CONTRIBUTIONS

ACKNOWLEDGEMENTS

CONFLICT OF INTEREST STATEMENT

ETHICS APPROVAL AND CONSENT TO PARTICIPATE

CONSENT FOR PUBLICATION

Open Research

DATA AVAILABILITY STATEMENT

Supporting Information

REFERENCES

Figures

References

Related

Information