Skip to main content
NCBI home page
Scientific Reports logoLink to Scientific Reports
. 2025 Aug 22;15:30864. doi: 10.1038/s41598-025-16737-4

Functional characteristics of fresh antitumor immune interferer GDF-15 in multiple cancers

Min Fang 1,2, Ziyu Wang 1, Yuxin Zhou 1, Fengrui Yang 1, Zhu Wu 3, Jian Xiao 4,5,✉
PMCID: PMC12373771  PMID: 40846894

Abstract

We designed our study with the aim of conducting a comprehensive analysis of the role that GDF-15 plays in pan-cancer. Multiple databases were used to obtain GDF-15 expression. XIANTAO Academic and Sangerbox were utilized to conduct the diagnostic significance of GDF-15 expression across pan-cancer. The relationships between GDF-15 expression and clinical data, tumor stemness, genomic instability, and tumor prognosis in pan-cancer were evaluated. The interaction between molecules and RNA methylation phenotypes related to GDF-15 were explored. The correlations between GDF-15 expression and immune cell infiltration, immune-related genes, as well as immune checkpoints were also under investigation. Besides, immunohistochemistry (IHC)-based tissue microarray analysis of LUAD validated GDF-15 protein expression and its correlation with survival prognosis. GDF-15 exhibited differential expression across various cancers, being upregulated in BRCA, CHOL, COAD, HNSC, LIHC, LUAD, READ, STAD, THCA, and UCEC, but downregulated in KIRC and LUSC. Its expression correlated with clinical parameters and GDF-15 expression had prognostic value in some cancers. Notably, it had a remarkable diagnostic value in CHOL, COAD, ESCC, GBM, LAML, READ and THCA. GDF-15 was involved in influencing tumor stemness, genomic instability. Besides, GDF-15 interacted with molecules involved in iron ion homeostasis and acute-phase response, receptor activator, ligand activities and cellular senescence. It was intertwined with key tumor processes. Moreover, GDF-15 expression further influenced immune cell infiltration, related immune genes, and immune checkpoints. More importantly, IHC staining results of tissue microarrays confirmed its differential protein expression in LUAD tissues, where elevated levels were significantly associated with poorer overall survival, underscoring its clinical relevance. GDF-15 functioned as a prospective indicator for both the diagnosis and prognosis in various cancers.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-16737-4.

Keywords: GDF-15, Stemness, Diagnosis, Prognosis, Pan-cancer

Subject terms: Cancer, Oncology

Introduction

Tumors, like insidious demons lurking within the human body, bring devastating harm on human health1. Economically, the long and complicated treatment processes, including surgery, chemotherapy, radiotherapy and various targeted drug therapies, cost an astonishing amount of money and can be regarded as a bottomless black hole. According to an authoritative study, cancer claimed 9.7 million lives globally in 2022, with 20 million new cases diagnosed2. Against this grim backdrop, molecule-level pan-cancer research has emerged as an imperative3.Tumors exhibit profound complexity, with marked variations across individuals and organ systems, making molecular investigations indispensable as a precision scalpel to dissect their underlying pathological roots. By analyzing key molecules, such as PD-1, PD-L1 and VEGF4,5, researchers can uncover common pathological mechanisms, thereby developing universal and efficient diagnostic and therapeutic strategies. Such advancements would directly translate to earlier detection, overcome existing treatment limitations, and enhance the precision of tumor targeting. Ultimately, this research promises to improve patient survival rates and quality of life, bringing humanity closer to conquering tumors.

Growth differentiation factor 15 (GDF-15), also known as macrophage inhibitory cytokine-1 (MIC-1), is a member of the transforming growth factor-β (TGF-β) superfamily with diverse physiological and pathological functions6. Extensive research have revealed that GDF-15 exhibits a wide expression pattern in various tissues and organs7,8 and emerging experimental evidence further highlighted its crucial role as a multifaceted regulator in tumor biology across different cancer types. For instance, in epithelioid hemangioendothelioma, GDF-15 has been identified as a novel biomarker that reliably correlated with tumor invasiveness9. In breast cancer, knocking down GDF-15 successfully reduced cancer stem cell properties and enhanced chemosensitivity9. In glioblastoma, studies have established GDF-15 as both a diagnostic marker and prognostic factor10 and in metastatic colorectal cancer, GDF-15 was considered a potential serum biomarker11. The discoveries not only provide a robust scientific foundation for further mechanistic investigations but also offer potential clinical applications, paving the way for innovative approaches in cancer diagnosis and treatment.

While GDF-15 has been established as a critical mediator in tumor progression, its functional mechanisms remain poorly defined in several cancer types, representing a significant research gap. Notably, as a fresh antitumor immune interferer, GDF-15 neutralization has demonstrated efficacy in overcoming resistance to PD-1/PD-L1 inhibitors in solid tumors, highlighting its promising implications for immunotherapy12. Inspired by this evidence12, we conducted a systematic pan-cancer analyses to elucidate the functional landscape of GDF-15, aiming to address the aforementioned knowledge gaps.​ To this end, we first investigated GDF-15 abundance across diverse cancer types and matched normal tissues using multi-omics databases. We then systematically analyzed the relationships between GDF-15 expression and oncological features, including clinical features, prognostic and diagnostic value, tumor stemness, genomic instability parameters, RNA methylation modification markers, the interactions with potential molecules, immune-related molecules, immune checkpoint molecules and immune cell infiltration. Importantly, we employed clinical tissue microarrays to verify and explore the impact of GDF-15 expression on the survival prognosis of patients with LUAD. This comprehensive study aims to establish a theoretical foundation for GDF-15’s diagnosis, treatment and research in multiple tumors.

Methods

Analysis of the expression difference of GDF-15

To explore the expression disparity of GDF-15, we integrated its transcriptomic and proteomic profiles across human tissues from multiple datasets, including the Human Protein Atlas (HPA), Genotype-Tissue Expression (GETx), and FANTOM5. Protein expression data were exclusively obtained from HPA. We systematically compared GDF-15 expression between diverse tumor types and their matched normal cells and tissues using data from HPA, The Cancer Genome Atlas Program (TCGA), GTEx, and the University of California Santa Cruz (UCSC) Xena databases, with analyses performed through XIANTAO Academic13 under default parameters. Furthermore, we obtained GDF-15 expression patterns in human tissues, single cells, and cell lines from HPA while paired comparison between pan-cancer samples and matched adjacent normal tissues sourced from the TCGA database were conducted using XIANTAO Academic with standardized settings for both the software and the HPA database.

Immunohistochemistry (IHC) staining of samples from HPA database

IHC images illustrating the protein expression situation of GDF-15 protein in a variety of normal tissues and their corresponding cancerous counterparts were obtained from HPA database14.

Analysis of the diagnostic and prognostic value of GDF-15

To evaluate the diagnostic and prognostic potential of GDF-15 in cancer, we acquired its expression variations from TCGA and GTEx and performed statistical analyses using XIANTAO Academic. This platform enabled the construction of receiver operating characteristic (ROC) curves and calculation of the area under the curve (AUC) to assess GDF-15’s diagnostic performance across multiple cancer types. The detailed process of data analysis of ROC was placed in the supplementary materials (Supplementary material 1) and the cutoff values of GDF-15 expression for ROC in different cancer type were shown in Table S1 in supplementary material 2. The diagnostic value was categorized into three tiers with cutoff 0.7 and 0.9: high accuracy (AUC > 0.9), moderate accuracy (AUC < 0.9 and > 0.7), and low accuracy (AUC 0.5–0.7). For prognostic evaluation, we employed the Sangerbox15 gene expression and prognosis analysis module to examine overall survival (OS), disease-specific survival (DSS), disease-free interval (DFI), and progression-free interval (PFI), thereby elucidating GDF-15’s clinical relevance in multiple cancers.

Analysis of clinical data and GDF-15 expression

The Sangerbox module for clinical staging and gene expression was employed to study the link between GDF-15 expression and clinical traits including tumor (T) stage, node (N) stage, metastasis (M) stage, overall clinical stage, tumor grade (G), and gender. T stage was grouped into T1 - T4, where a larger number after T indicates a bigger tumor size and/or more extensive invasion into adjacent tissues. N stage had classifications of N0-N3, with N1-N3 showing increased spread to nearby lymph nodes in terms of size, location, and/or quantity. M stage was split into M0 (no detected distant metastasis) and M1 (tumor spread to remote organs/tissues). Overall clinical stage ranged from Ⅰ to Ⅳ, with a higher number signifying more severe tumor spread. Tumor grades were typically G1-G4, where G1 meant low-grade malignancy, G2 cannot be generalized, G3 was highly malignant, and G4 was even more so. Gender was simply male or female.

Analysis of tumor stemness and GDF-15 expression

To explore the relationship between tumor stemness and GDF-15 expression, we employed the “tumor stemness and gene expression” module available in Sangerbox, in conjunction with Spearman analysis. Specifically, we focused on examining the correlation between GDF-15 expression and various tumor stemness characteristics including the DNA stemness score (DNAss), the epigenetic regulation of DNA methylation (EREG-METHss), the differential methylation probe (DMPss), the enhancer element/DNA methylation (ENHss), the RNA stemness score (RNAss), as well as the epigenetic regulation of RNA expression (EREG.EXPss). The detailed process of analyzing correlation between GDF-15 expression and tumor stemness was placed in the supplementary materials (Supplementary material 1).

The effects of GDF-15 expression on genomic instability

To investigate the connection between GDF-15 expression and genomic instability in the cancers, the online data analysis tool Sangerbox was employed. Several key genomic instability indicators included tumor mutation burden (TMB), mutant-allele tumor heterogeneity (MATH), microsatellite instability (MSI), neoantigen (NEO), purity, ploidy, homologous recombination deficiency (HRD), and loss of heterozygosity (LOH) were explored. The detailed process of analyzing correlation between GDF-15 expression and genomic instability was placed in the supplementary materials (Supplementary material 1).

The influences of GDF-15 expression on the markers of RNA methylation modification

Using the Sangerbox database, we analyzed the relationship between GDF-15 expression on markers of RNA methylation modifications like N1-methyladenosine (m1A), 5 methylcytidine (m5C), and N6-methyladenosine (m6A). There are three types of genes related to these modifications: writers, readers, and erasers. We then looked at how GDF-15 expression relates to genes involved in changes of RNA methylation markers in different tumor types. The detailed process of analyzing correlation between GDF-15 expression and RNA methylation modifications was placed in the supplementary materials (Supplementary material 1).

Analysis of molecules interacting with GDF-15 and analysis of functions of GDF-15

To probe molecules interacting with GDF-15, we utilized three databases in below: The European Molecular Biology Laboratory’s European Bioinformatics Institute (EMBL-EBI), gene function using a multiple association network integration algorithm (GENEMANIA) and The Search Tool for the Retrieval of Interacting Genes/Proteins (STRING). Focusing on the STRING-derived interactors, we performed functional enrichment analysis using Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG)16,17 through XIANTAO Academic. Additionally, we investigated the link between GDF-15 and 14 cancer-related functions including angiogenesis, apoptosis, cell cycle, differentiation, DNA damage, DNA repair, EMT, hypoxia, inflammation, invasion, metastasis, proliferation, quiescence and stemness using the CancerSEA database.

Analysis of immune-related genes/immune checkpoints/immune cell infiltration and GDF-15 expression

In Sangerbox, the immune regulatory gene analysis module was harnessed to probe the relationship between GDF-15 expression and the marker genes of five crucial immune pathways, including chemokine, receptor, MHC, immunoinhibitor, and immunostimulator. Simultaneously, the immune checkpoint genetic analysis module within Sangerbox was applied to scrutinize the correlations between GDF-15 expression and immune checkpoints (inhibitory or stimulatory). Moreover, the xCELL and CIBERSORT modules dedicated to immune cell analysis in Sangerbox were utilized to dissect the associations between GDF-15 expression and the infiltration patterns of diverse immune cells. The detailed process of data processing was placed in the supplementary materials (Supplementary material 1).

Tissue microarray

This study utilized tissue microarrays from Shanghai Outdo Biochip Company (China). A total of three LUAD tissue microarrays were employed (HLugA150CS04, HLugA180Su09, and HLugA180Su12). It is noteworthy that the HLugA180Su09 and HLugA180Su12 microarrays contained comprehensive overall survival follow-up data. All tissue microarrays applied in this study underwent ethical review and obtained approval. Shanghai Outdo Biochip Company had previously completed the ethical approval process with the approval numbers YB M-05-01 and YB M-05-02, thus ensuring compliance with ethical norms and regulations.

IHC assay of tissue microarrays

For the tissue microarray, the dewaxing and hydration processes were first performed, followed by antigen retrieval. The GDF-15 antibody 1:100 (F1086, Selleck, USA) was then added and incubated at 4 °C overnight. The next day, sections were incubated with HRP-labeled anti-rabbit IgG (AWI0629, abiowell, China) at room temperature for 45 min, and 3,3-diaminobenzidine (DAB) was used for staining.​.

GDF-15 protein expression assessment

To thoroughly assess GDF-15 expression, the staining index (SI) was calculated by averaging the scores assigned by two independent investigators, following a previously reported method18. The proportion of positive tumor cells was scored as: 0 for no positive cells; 1 for < 10% positive cells; 2 for 10–35% positive cells; 3 for 35–70% positive cells; and 4 for > 70% positive cells. Staining intensity was categorized as: 0 (no staining), 1 (weak staining), 2 (moderate staining), or 3 (strong staining). The SI was determined by multiplying the intensity score by the proportion score, yielding possible values of 0, 1, 2, 3, 4, 6, 8, 9, or 12. The median SI across all samples was 4. Based on this, samples were stratified into different groups: SI ≤ 4 was classified as low GDF-15 expression; and SI > 4 was considered high GDF-15 expression.​.

Statistical analysis

The Kaplan-Meier approach was utilized to investigate the correlations between GDF-15 protein expression and prognosis of LUAD. Following that, the log-rank test was applied to make comparisons among these curves. A P-value smaller than 0.05 was defined as the criterion for statistical significance.

Results

Abnormal expression of GDF-15 in pan-cancer

Analysis of data from multiple databases, including HPA, GTEx, and FANTOM5, revealed that GDF-15 is widely expressed in normal tissues (Fig. S1A–E in supplementary material 2), suggesting its potential involvement in essential physiological functions. Besides, we observed that GDF-15 was widely expressed in various tumor cell lines and numerous tumor types (Fig. 1A,B). Moreover, in 33 tumor types, GDF-15 was abnormally upregulated in BRCA, CHOL, COAD, GBM, HNSC, LIHC, LUAD, PARD, READ, STAD, THCA and UCEC, while it was downregulated in KICH, KIRC and LUSC (Fig. 1C). A similar trend was observed in matched tumor-normal tissue pairs from 23 kinds of tumors, where GDF-15 expression was elevated in BRCA, CHOL, COAD, HNSC, LIHC, LUAD, PRAD, READ, STAD, THCA and UCEC, but declined in KICH, KIRC and LUSC (Fig. 1D). Subsequently, analyses of TCGA and GETx databases revealed that GDF-15 expression was evidently enhanced in GBM, GBMLGG, LGG, UCEC, BRCA, LUAD, ESCA, STES, COAD, COADREAD, PRAD, STAD, HNSC, LIHC, SKCM, THCA, READ, OV, TGCT, UCS, ACC and CHOL, while it was observably decreased in KIPAN, KIRC, LUSC, WT and ALL (Fig. 1E). Further validation using IHC data from the HPA database revealed apparently elevated GDF-15 protein expression in melanoma, prostate cancer and renal cancer (Fig. S2 in supplementary material 2). Taken together, GDF-15 was abnormally expressed in various tumors, suggesting a potential critical role in tumor biology.

Fig. 1.

Fig. 1

Open in a new tab

Abnormal expression of GDF-15 in pan-cancer. (A) Expression profile of GDF-15 in various tumor cells from HPA database. (B) Expression profile of GDF-15 in different cancer cell lines from HPA database. (C, D) Expression profile of GDF-15 in unpaired and paired samples across pan-cancer using XIANTAO Academic. (E) Expression profile of GDF-15 in samples across pan-cancer from TCGA and GETx databases.

Clinical significance and prognostic efficacy of GDF-15 in cancer patients

We further measured the relationship between GDF-15 expression and various clinical parameters in cancer cohorts. In 26 kinds of tumors, GDF-15 expression was only shown to affect T stage in LUAD, BRCA, ESCA, STES, KIRP, PRAD and THCA (Fig. 2A). In addition, among 25 cancers, GDF-15 expression was correlative with N stage in BRCA, ESCA, STES, KIRP, KIPAN, THCA and TGCT (Fig. 2B). Besides, GDF-15 expression influenced M stage in KIPAN and TGCT (Fig. 2C). Moreover, across 30 cancers, GDF-15 expression was significantly associated with overall clinical stage in BRCA, STES, KIRP, HNSC, THCA and TGCT (Fig. 2D). In 14 kinds of tumors, GDF-15 expression correlated with tumor grade in GBMLGG, LGG, ESCA, STES, STAD and HNSC (Fig. 2E). Notably, GDF-15 expression exhibited a strong gender-specific association in BRCA, STES, KIRP, KIPAN, HNSC, and KIRC across 31 tumors (Fig. 2F). Collectively, these results underscored the involvement of dysregulated GDF-15 expression in tumor progression across diverse cancer types.

Fig. 2.

Fig. 2

Open in a new tab

The relationships of GDF-15 expression and clinical data in various cancers. (A) The relationships between GDF-15 expression and T stage in pan-cancer using Sangerbox. (B) Correlation between GDF-15 expression and N stage in pan-cancer using Sangerbox. (C) Correlation between GDF-15 expression and M stage in pan-cancer using Sangerbox. (D) Correlation between GDF-15 expression and overall clinical stage in pan-cancer using Sangerbox. (E) Correlation between GDF-15 expression and tumor grade in pan-cancer using Sangerbox. (F) Correlation between GDF-15 expression and gender in pan-cancer using Sangerbox.

As shown in Fig. 3A, elevated GDF-15 expression correlated with worse OS in GBMLGG, LGG, ESCA, KIRP, MESO, UVM and LAML while paradoxically associating with improved OS in LUAD, KIPAN, COAD and COADREAD. For DSS outcomes analyzed across 38 tumors, high GDF-15 expression predicted unfavorable prognosis in GBMLGG, LGG, KIRP, MESO and UVM, but showed the reverse results in LUAD, KIPAN, COAD, COADREAD, STAD, KIRC and BLCA (Fig. 3B). Furthermore, DFI analysis (32 tumors) demonstrated that increased GDF-15 expression adversely impacted in LGG, GBMLGG and KICH while benefitting STES (Fig. 3C). Notably, across 38 kinds of tumors, there was a positive relationship between higher GDF-15 expression and poor PFI in GBMLGG, LGG and GBM but a negative relationship in LUAD, STES, KIPAN, STAD, KIRC and BLCA (Fig. 3D). The Kaplan-Meier survival curves of the most significant prognostic associations for OS, DSS, DFI and PFI were shown in Fig. S3 in supplementary material 2. Therefore, GDF-15 exerted tumor-type-dependent prognostic effects.

Fig. 3.

Fig. 3

Open in a new tab

Prognostic analysis of GDF-15 in various tumors. (A) The relationships between GDF-15 expression and OS in pan-cancer using Sangerbox. (B) The relationships between GDF-15 expression and DSS in pan-cancer using Sangerbox. (C) The relationships between GDF-15 expression and DFI in pan-cancer using Sangerbox. (D) The relationships between GDF-15 expression and PFI in pan-cancer using Sangerbox.

Diagnostic value of GDF-15 in 34 kinds of cancers

GDF-15 exhibited abnormal expression and influenced on clinical features and prognosis in a variety of tumors, suggesting its potential as a diagnostic biomarker for tumors. Here, we analyzed data from the TCGA and GTEx databases and generated ROC curves for GDF-15 in 34 tumors. The results demonstrated that GDF-15 had high diagnostic value in CHOL (AUC:0.902), COAD (AUC: 0.968), ESCC (AUC: 0.915), GBM (AUC: 0.966), LAML (AUC: 1.000), READ (AUC: 0.966), THCA (AUC: 0.910) (Fig. 4). Overall, these findings highlighted GDF-15 as a promising pan-cancer diagnostic biomarker.

Fig. 4.

Fig. 4

Open in a new tab

Diagnostic value of GDF-15 in 34 kinds of cancers. ROC curve analysis of GDF15 in 34 tumors using TCGA and GTEx data via XIANTAO Academic demonstrated its diagnostic potential in pan-cancer.

The relationship between GDF-15 expression and stemness score in various tumors

Tumor stemness affects tumor metastasis and drug resistance19. Given this connection, we investigated whether aberrant GDF-15 expression influences tumor stemness by analyzing its relationship with multiple stem-related indicators including DNAss, EREG-METHss, DMPss, ENHss, RNAss and EREG.EXPss across 37 kinds of tumors.

First, GDF-15 expression had the positive relationship with DNAss in GBMLGG (N = 558), LGG (N = 507), HNSC (N = 512), THYM (N = 119), THCA (N = 499), PCPG (N = 176), UVM (N = 79) and ACC (N = 76) while a negative relationship was observed in CESC (N = 301), LUAD (N = 451), BRCA (N = 774), STES (N = 548), KIPAN (N = 642), UCEC (N = 173), KIRC (N = 309), LIHC (N = 366), TGCT (N = 147) and KICH (N = 65) (Fig. 5A). Second, in GBMLGG (N = 558), LGG (N = 507), HNSC (N = 512), THYM (N = 119), THCA (N = 499), PCPG (N = 176), UVM (N = 79) and ACC (N = 76), there was a positive relationship between GDF-15 expression and EREG-METHss. Inversely, in CESC (N = 301), BRCA (N = 774), STES (N = 548), KIPAN (N = 642), UCEC (N = 173), KIRC (N = 309), LIHC (N = 366), TGCT (N = 147) and KICH (N = 65), there was a negative relationship between the two (Fig. 5B). Third, regarding DMPss, GDF-15 expression presented a positive correlation in GBMLGG (N = 558), LGG (N = 507), HNSC (N = 512), THYM (N = 119), PCPG (N = 176), UVM (N = 79) and ACC (N = 76), but a negative correlation in CESC (N = 301), LUAD (N = 451), KIPAN (N = 642), UCEC (N = 173), KIRC (N = 309), LIHC (N = 366), TGCT (N = 147) and KICH (N = 65) (Fig. 5C). Fourth, ENHss was positively associated with GDF-15 expression in GBMLGG (N = 558), LGG (N = 507), HNSC (N = 512), THYM (N = 119), PCPG(N = 176), UVM (N = 79) and ACC(N = 76), whereas an inverse relationship was observed in CESC (N = 301), KIPAN (N = 642), UCEC (N = 173), KIRC (N = 309), LIHC (N = 366), TGCT (N = 147), and KICH (N = 65) (Fig. 5D). Fifth, RNAss demonstrated a positive correlation with GDF-15 expression in ESCA (N = 179), KIRP (N = 283), KIPAN (N = 860), STAD (N = 399) and PAAD (N = 156), but a negative correlation with GBM (N = 152), GBMLGG (N = 659), LGG (N = 507), CESC (N = 301), LUAD (N = 507), LAML (N = 167), BRCA (N = 1080), SARC (N = 253), LUSC (N = 483), THYM (N = 119), TGCT (N = 147), UCS (N = 57), and BLCA (N = 403) (Fig. 5E). Finally, EREG.EXPss showed a positive association with GDF-15 expression in CESC (N = 301), ESCA (N = 179), STES (N = 578), KIRP (N = 283), STAD (N = 399), OV (N = 298), TGCT (N = 147) and PCPG (N = 176), while a negative association was observed in LAML (N = 167) and BRCA (N = 1080) (Fig. 5F).

Fig. 5.

Fig. 5

Open in a new tab

The relationship between GDF-15 expression and stemness score in various tumors. (A–F) The relationships between GDF-15 expression and tumor stemness based on DNAss/EREG-METHss/DMPss/ENHss/RNAss/EREG.EXPss across 37 cancers was analyzed by means of Sangerbox.

In summary, GDF-15 expression probably played a regulatory role in tumor stemness across multiple cancer types.

Genomic-instability parameters affected by GDF-15 expression in various tumors

Here, we comprehensively analyzed the association between GDF-15 expression and key genomic-instability parameters including TMB, MATH, MSI, NEO, purity, ploidy, HRD and LOH. All of them conducted analysis in 37 kinds of cancer except for NEO which performed in 32 cancer types.

Firstly, GDF-15 showed positive associations with TMB in STES (N = 589), STAD (N = 409) and UCEC (N = 175) but a negative relationship in LUAD (N = 509) and READ (N = 90) (Fig. 6A). Similarly, its expression positively correlated with MATH in SARC (N = 234), KIPAN (N = 679), THYM (N = 118) and LIHC (N = 356), while displaying inverse relationships in GBMLGG (N = 649), LGG (N = 500), COADREAD (N = 372), BRCA (N = 980), STES (N = 589) and UCEC (N = 175) (Fig. 6B). For MSI, GDF-15 its expression positively in STES (N = 592), KIPAN (N = 688), STAD (N = 412), UCEC (N = 180) and TGCT (N = 148) but negative regulation in GBMLGG (N = 657) (Fig. 6C). Notably, NEO exhibited limited associations, showing positive associations only in UCEC (N = 166) and KIRC (N = 337) (Fig. 6D). Tumor purity analysis revealed widespread positive correlations (COAD (N = 282), COADREAD (N = 372), BRCA (N = 1043), KIPAN (N = 844), STAD (N = 402), HNSC (N = 505), KIRC (N = 495), THYM (N = 102), TGCT (N = 147), BLCA (N = 397)) except in in KIRP (N = 283), UCS (N = 56) and DLBC (N = 46) where negative associations were observed (Fig. 6E). Ploidy measurements showed negative relationships in GBMLGG (N = 646), KIRP (N = 283), KIPAN (N = 844), PRAD (N = 470), UCEC (N = 178), HNSC (N = 505), MESO (N = 81), PAAD (N = 158), and TGCT (N = 147) (Fig. 6F). HRD analysis demonstrated complex regulation, with positive correlations in SARC (N = 241), HNSC (N = 505), LIHC (N = 356), MESO (N = 81), PAAD (N = 158) and PCPG (N = 160) but negative relationships in LUAD (N = 500), COAD (N = 283), COADREAD (N = 373), BRCA (N = 1044), STES (N = 563), PRAD (N = 470), UCEC (N = 178), KIRC (N = 495), THCA (N = 462), BLCA (N = 397) and KICH (N = 66) (Fig. 6G). Finally, there was a positive connection between GDF-15 expression and LOH in GBMLGG (N = 644),LGG (N = 501), SARC (N = 241), KIRP (N = 278), KIPAN (N = 821), HNSC (N = 500), LIHC (N = 353), TGCT (N = 147), PCPG (N = 159), UVM (N = 79) and CHOL(N = 36) and a negative connection in LUAD (N = 495), COADREAD (N = 362), BRCA (N = 1035), STES (N = 558), STAD (N = 400), UCEC (N = 175), THCA (N = 461), UCS (N = 56), BLCA (N = 397) and KICH (N = 65) (Fig. 6H).

Fig. 6.

Fig. 6

Open in a new tab

Genomic-instability parameters affected by GDF-15 expression in various tumors. (A) The relationships between GDF-15 expression and TMB across 37 cancers using Sangerbox. (B) The relationships between GDF-15 expression and MATH across 37 cancers using Sangerbox. (C) The relationships between GDF-15 expression and MSI across 37 cancers using Sangerbox. (D) The relationships between GDF-15 expression and NEO across 32 cancers using Sangerbox. (E) The relationships between GDF-15 expression and purity across 37 cancers using Sangerbox. (F) The relationships between GDF-15 expression and ploidy across 37 cancers using Sangerbox. (G) The relationships between GDF-15 expression and HRD across 37 cancers using Sangerbox. (H) The relationships between GDF-15 expression and LOH across 37 cancers using Sangerbox.

Collectively, these comprehensive findings established GDF-15 as a significant modulator of genomic heterogeneity across diverse malignancies, with its effects exhibiting remarkable tissue-specific variation.

The effects of GDF-15 on RNA methylation modification markers

Here, we investigated the effects of GDF-15 on m1A, m5C and m6A related enzymes (writer, reader and eraser) in 40 kinds of tumors. The results exhibited that, in THYM, OV, HNSC, GBMLGG, TGCT, NB, THCA, KIPAN, LUAD and BRCA, GDF-15 expression was closely related to m6A related enzymes (Fig. 7). The specific details regarding sample size were shown in Table S2 in supplementary material 2. Overall, these findings suggested that GDF-15 may play a significant role in regulating RNA methylation, particularly m6A modification, across multiple cancer types.

Fig. 7.

Fig. 7

Open in a new tab

The effects of GDF-15 on RNA methylation modification markers. The relationship between GDF15 and m1A, m5C, m6A-related genes was analyzed across 40 types of tumors using Sangerbox.

GDF-15 interacted with various molecules and participated in biological functions

Subsequently, we identified GDF-15-interacting molecules using three databases (EMBL-EBI, GENEMANIA and String), with results visualized in Fig. 8A–C. Functional and pathway analyses through GO and KEGG of the String database interactions revealed GDF-15’s involvement in key biological processes. The results revealed that GDF-15 may be associated with iron ion homeostasis, cellular iron ion homeostasis and acute-phase response in the biological process (BP), signaling receptor activator activity and receptor ligand activity in molecular function (MF), and cellular senescence in KEGG (Fig. 8D). CancerSEA database demonstrated significant correlations between GDF-15 expression and 14 cancer functional states across 11 tumor types, showing positive associations with angiogenesis, apoptosis, differentiation, hypoxia, inflammation, metastasis stemness, while exhibiting negative correlations with cell cycle, DNA repair, EMT, invasion, proliferation and quiescence (Fig. 8E).

Fig. 8.

Fig. 8

Open in a new tab

GDF-15 could interact with various molecules and participate in biological functions. (A–C) The potential interactions of GDF-15 with related molecules were predicted using EMBL-EBI, GENEMANIA and String databases. (D) GO/KEGG analyses were performed on the GDF-15-interacting molecules retrieved from the STRING database. (E) Biological functions mediated by GDF-15.

To sum up, GDF-15 was involved in multiple biological processes and disease progression through complex molecular interaction networks.

Correlation between GDF-15 and immune-related genes, immune checkpoint molecules and immune cell infiltration in various tumors

The immune microenvironment is closely related to tumor progression. Thus, we investigated the relationship between GDF-15 expression and immune-related genes, immune checkpoint molecules and immune cell infiltration across multiple tumor types.

As exhibited in Fig. S4 in supplementary material 2, GDF-15 expression influenced on the expression of some chemokine, receptor, MHC, immunoinhibitor and immunostimulator, with particularly strong influences observed in TGCT, COAD, COADREAD, KIPAN, BRCA, BLCA, THCA, OV, PCPG and UVM. The specific details regarding sample size were shown in Table S3 in supplementary material 2.

In addition, GDF-15 expression showed robust correlations with some immune checkpoints in BLCA, BRCA, KIPAN, GBMLGG, LGG, THYM, PCPG, UVM, OV and WT (Fig. 9). The specific details regarding sample size were shown in Table S4 in supplementary material 2.

Fig. 9.

Fig. 9

Open in a new tab

Correlation between GDF-15 and immune checkpoint molecules in various tumors.The relationship between GDF15 and immune checkpoint molecules was analyzed across 40 types of tumors using Sangerbox.

For assessing immune cell infiltration, xCELL analysis revealed association between GDF-15 expression and immune cell infiltration in GBM (N = 152), GBMLGG (N = 656), LGG (N = 504), UCEC (N = 178), LAML (N = 142), BRCA (N = 1077), CESC (N = 291), LUAD (N = 500), ESCA (N = 181), STES (N = 569), SARC (N = 258), KIRP (N = 285), KIPAN (N = 878), COAD (N = 282), COADREAD (N = 373), PRAD (N = 495), STAD (N = 388), HNSC (N = 517), KIRC (N = 528), LUSC (N = 491), THYM (N = 118), LIHC (N = 363), WT (N = 80), SKCM-P (N = 101), SKCM (N = 452), BLCA (N = 405), SKCM-M (N = 351), THCA (N = 503), NB (N = 153), MESO (N = 85), READ (N = 91), OV (N = 417), UVM (N = 79), PAAD (N = 177), TGCT (N = 132), UCS (N = 56), LAML (N = 214), ALL (N = 86), PCPG (N = 177), ACC (N = 77), ALL-R (N = 99), DLBC (N = 46), KICH (N = 65), CHOL (N = 36) (Fig. 10A). Similarly, CIBERSORT analysis demonstrated that GDF-15 expression correlated with varying degrees of immune cell infiltration across different tumor types (Fig. 10B).

Fig. 10.

Fig. 10

Open in a new tab

Correlation between GDF-15 and immune cell infiltration in various tumors. (A) Correlation between GDF15 expression and immune cell populations estimated using xCELL algorithm. (B) Correlation between GDF15 expression and immune cell populations estimated using CIBERSORT algorithm.

Taken together, GDF-15 expression played a key role in immune regulation within the tumor microenvironment.

High GDF-15 protein expression was associated with unfavorable survival outcomes in LUAD patients

To investigate GDF-15 protein expression in LUAD tissues, we utilized LUAD tissue microarrays comprising samples from 228 LUAD patients and conducted IHC staining on these specimens. Figure 11A displayed the IHC staining results of all the samples. Figures 11B,C representatively illustrated the negative, weakly positive, positive, and strongly positive expression of GDF-15 protein in LUAD samples. Additionally, the distribution of the SIs for the 228 samples we studied was presented in Fig. 12A. Among them, 36.41% of patients with 228 LUAD samples showed high expression of GDF-15 (SI > 4). More importantly, for the 160 LUAD patients contained comprehensive OS follow-up data, Kaplan-Meier analysis revealed that high GDF-15 protein expression had an adverse effect on OS of LUAD patients (Fig. 12B). Subgroup analyses were further conducted, stratifying LUAD patients by gender (male/female) and age (< 60 years/≥60 years). The results showed that high GDF-15 protein expression was significantly associated with shortened OS in three aforementioned subgroups except for the female subgroup (P = 0.08) (Fig. 12C–F).

Fig. 11.

Fig. 11

Open in a new tab

The GDF-15 protein expression using IHC in the tissue microarray of LUAD. (A) LUAD tissue chips. (B, C) Representative images of GDF-15 protein.

Fig. 12.

Fig. 12

Open in a new tab

GDF-15 protein expression status and the relationship between its expression and survival prognosis in LUAD patients. (A) Statistical graph of GDF-15 expression in LUAD patients from the tissue microarray. (B) The relationship between GDF-15 expression and OS in 160 LUAD patients (92 cases with low expression, 68 cases with high expression). (C) The relationship between GDF-15 expression and OS in 87 male LUAD patients (47 cases with low expression, 40 cases with high expression). (D) The relationship between GDF-15 expression and OS in 73 female LUAD patients (45 cases with low expression, 28 cases with high expression). (E) The relationship between GDF-15 expression and overall survival in 65 LUAD patients under the age of 60 (33 cases with low expression, 32 cases with high expression). (F) The relationship between GDF-15 expression and overall survival in 95 LUAD patients aged 60 or above (59 cases with low expression, 36 cases with high expression).

Discussion

Pan-cancer analysis aims to identify genes, pathways and biomarkers across multiple cancers, uncovering both shared and distinct molecular features to advance cancer prevention, diagnosis and treatment20,21. Key genes such as PD-L1, GPC2, ALKBH5 and NCAPG have been comprehensively analyzed in pan-cancer, offering insights for the development of targeted therapies22–25. Recently, Melero I et al. found that neutralizing GDF-15 reduced resistance to PD-1 or PD-L1 inhibitors in solid tumors12, highlighting its critical role in tumor progression. In this study, we systematically analyzed GDF-15 expression patterns in normal, tumor cells and tissues by multiple databases. Our findings demonstrated that GDF-15 expression is widely expressed in healthy tissues but dysregulated in various tumors, suggesting its potential involvement in tumorigenesis. Moreover, we further investigated GDF-15’s clinical significance, prognostic efficacy diagnostic value, its association with tumor stemness and genomic instability. We also explored its interacting molecules, functional roles and immunomodulatory effects across various tumors. Finally, based on IHC staining results of tissue microarrays, we confirmed that high GDF-15 protein expression was significantly associated with poorer overall survival in LUAD patients, reinforcing its clinical relevance in specific malignancies.

GDF-15 expression has been extensively documented in both normal and tumor tissues8,26. Here, our comparative analysis further revealed that GDF-15 expression was significantly elevated in multiple cancer types (BRCA, CHOL, COAD, HNSC, LIHC, LUAD, READ, STAD, THCA and UCEC), while being markedly reduced in KIRC and LUSC. Importantly, these findings aligned with published experimental evidence. Specifically, similar trends have been observed in triple-negative breast cancer cells, perihilar cholangiocarcinoma, hepatocellular carcinoma and non-small cell lung cancer27–30, thereby validating the consistency between our results and these published studies.

Previous studies have established GDF-15 as a diagnostic and prognostic markers in several malignancies, including pancreatic cancer, colorectal cancer, clear cell renal cell carcinoma and liver cancer31–34. In this study, our data further demonstrated that GDF-15 expression correlated strongly with clinical parameters including T stage, N stage, M stage, overall clinical stage and gender in a subset of tumors. Additionally, our analyses revealed that across multiple tumor types, GDF-15 expression was correlated with prognostic indicators such as OS, DSS, DFI and PFI. These consistent associations strongly support the potential of GDF-15 as a clinically valuable tumor marker for diagnosis. Furthermore, our results highlighted its utility as a diagnostic marker across a broader spectrum of tumors, most notably CHOL, COAD, ESCC, GBM, LAML, READ, and THCA. Collectively, these observations suggested that monitoring GDF-15 levels could provide valuable insights into the dynamics of tumor progression in these specific cancer types.

Tumor stemness, a key characteristic of certain tumor cells, endows them with self-renewal and differentiation abilities, which paly crucial roles in tumor progression by driving initiation, growth, metastasis, and therapy resistance35. Mounting studies have increasingly focused on how abnormal gene expression affected tumor stemness features to influence tumor progression36,37. For instance, DUSP3 suppressed osteosarcoma stemness and inhibited tumor progression by regulating the EGFR/STAT3/SOX2 axis38. Notably, several published articles have specifically investigated GDF-15’s impact on tumor stemness39. In breast cancer, for instance, GDF15 facilitated radio-resistance through enhancing cancer cell stemness39. Our study further elucidated the relationships between GDF-15 expression and DNAss, EREG-METHss, DMPss, ENHss, RNAss and EREG.EXPss across various tumors, thereby providing additional evidence for GDF-15’s role in modulating tumor stemness.

Genomic instability drives tumor progression by inducing gene mutations and chromosomal aberrations, generating continuous variation that facilitates tumor cell evolution and enables the transition from early to advanced stages40. This instability further endows tumor cells with invasive, metastatic and treatment-resistant capabilities, accelerating tumor dissemination, complicating clinical management, and profoundly impacting disease outcomes41,42. A key mechanism influencing tumor behavior involves the regulation of genomic instability by aberrantly expressed molecules. For examples, DDX18 modulated chemotherapy sensitivity in colorectal cancer via modulating genomic stability43. PICH was responsible for maintaining genomic stability, thus possessing the capacity to facilitate tumor growth44. However, the role of GDF-15 in tumor genomic stability remains unclear. To address this gap, our study analyzed the relationship between GDF-15 expression and parameters of genomic stability including TMB, MATH, MSI, NEO, purity, ploidy, HRD and LOH across various cancers. Our results suggested that GDF-15 expression could influence genomic stability in tumors suggesting its potential involvement in tumor progression through this mechanism.

Epigenetics, and particularly RNA methylation, plays a critical role in tumorigenesis and development45. By regulating gene expression and disrupting the normal growth and differentiation programs of tumor cells at the molecular level, it promotes tumor formation46. In our study, we found that GDF-15 expression in multiple cancer types exhibited significant correlations with RNA methylation-related genes including m1A, m5C, m6A modifications. This association was especially prominent in cancers such as THYM, OV, HNSC, GBMLGG, TGCT, NB, THCA, KIPAN, LUAD and BRCA. These findings not only shed new light on GDF-15’s role in RNA methylation-mediated epigenetic regulation during tumor progression, but also deepen our understanding of its mechanistic involvement in tumor-related RNA methylation modifications.

Tumors are complex entities, and in most cases, key genes rarely regulate tumor progression independently. Instead, they act through molecular interactions that collectively drive oncogenic processes47. For instances, GDF-15 has been shown to bind directly to the von Willebrand factor type C domain of CCN2, inhibiting CCN2-mediated αVβ3 integrin clustering and FAK activation, thereby blocking tube formation in human umbilical vein endothelial cells, and contributing to the progression of tumors such as prostate cancer48. To further elucidate such collaborative mechanisms, our study explored molecules interacting with GDF-15 using multiple bioinformatics platforms, including EMBL-EBI, GENEMANIA and String. These analyses enhanced our understanding of GDF-15’s cooperative regulatory roles in tumorigenesis.

To further elucidate the potential biological functions of GDF-15 in tumors, we performed GO and KEGG analyses on interacting molecules retrieved from the STRING database. The results revealed that GDF-15 might be associated with iron ion homeostasis, cellular iron ion homeostasis and acute-phase response in BP, signaling receptor activator activity and receptor ligand activity in MF, and cellular senescence in KEGG. Published studies have already documented the effects of GDF-15 on tumor cell characteristics. For instance, exosomes enriched with GDF-15 from 5-fluorouracil-resistant colon cancer cells have been shown to promote angiogenesis49. Besides, in prostate cancer, GDF-15 secreted by osteocytes stimulated the growth and invasion of cancer cells50. In this study, GDF-15 predominantly exhibited positive correlations with angiogenesis, apoptosis, differentiation, hypoxia, inflammation, metastasis and stemness. Conversely, it showed negative associations with cell cycle, DNA repair, EMT, invasion and proliferation. Notably, the relationship between GDF-15 and DNA damage was tumor-type dependent, with both positive and negative correlations observed across different cancers. Collectively, these findings underscored the multifaceted role of GDF-15 in tumor progression.

The tumor immune microenvironment plays a pivotal yet complex role in tumor progression, functioning as both a natural antagonist and a potential accomplice. Under normal conditions, immune cells and cytokines collaborate to recognize and eliminate tumor cells, thereby suppressing tumor growth. However, when this microenvironment becomes imbalanced with immunosuppressive factors dominating, it could instead facilitate tumor progression51. Immune-related genes, immune checkpoint molecules and immune cell infiltration were deeply intertwined with this process and intimately linked to tumor development52,53. Notably, GDF-15 has been well-documented to play a role in regulating tumor immunity12,54. For instance, tumor-derived GDF-15 hindered T-cell recruitment and impaired responses to anti-PD-1 therapy54. During early tumor development, NF-κB-modulated GDF-15 expression restrained macrophage surveillance55 and in malignant gliomas, GDF-15 promoted the proliferation and facilitated immune escape56. Besides, Green Hong et al. found that in patients with advanced NSCLC receiving PD-1/PD-L1 inhibitor therapy, high plasma GDF15 levels correlated with fewer PD-1+/CD8+ T cells, suggesting GDF15 may inhibit CD8+ T-cell-mediated antitumor effects48. In colorectal adenocarcinoma, GDF15 also promoted Treg differentiation, which suppressed CD8+ T-cell activation and impaired anti-tumor immunity57. Our findings further reinforced these observations, demonstrating strong associations between GDF-15 and immune-related genes, immune checkpoint molecules and immune cell infiltration in certain tumors. This underscored GDF-15’s critical involvement in tumor immune regulation.

Current research GDF-15’s role in LUAD remains complex, with existing findings revealing notable contradictions. For instance, some studies indicated that GDF-15 inhibited LUAD cell growth and bone metastasis58, while others suggested that downregulating GDF-15 enhanced chemosensitivity and suppressed metastasis by blocking the TGF-β/Smad pathway59. In this study, bioinformatics analysis of databases showed that high GDF-15 mRNA expression was associated with better OS. However, IHC validation using a tissue microarray of 228 LUAD cases revealed that only 36.41% of patients exhibited high GDF-15 expression (SI > 4). Further analysis of 160 samples with survival data demonstrated that high GDF-15 expression correlated with poor OS when stratified by gender or age (using 60 years as the cutoff), a result was diametrically opposed to the bioinformatics findings. This contradiction may stem from several factors: (1) Spatiotemporal heterogeneity in expression regulation between mRNA and protein levels, where post-translational modifications or microenvironmental differences could lead to distinct functional outcomes; (2) Biases in clinical characteristics (e.g., treatment history, tumor stage) between database samples and our IHC cohort; (3) The potential dual role of GDF-15 itself in LUAD. To resolve these issues, future research will focus on the following aspects: First,, designing in vitro cell experiments and in vivo animal models to investigate how exogenous modulation of GDF-15 expression affects LUAD progression. Second, expanding the clinical sample size and integrating single-cell sequencing with spatial transcriptomics to deeply analyze the association between GDF-15 expression patterns and tumor heterogeneity.

In conclusion, our study demonstrated that GDF-15 exhibits differential expression patterns across tumor types and influenced multiple oncogenic processes, including clinical parameters, prognostic significance, diagnostic potential, tumor stemness, genomic instability, RNA methylation, molecular interactions, immune-related genes, immune checkpoints and immune cell infiltration. These comprehensive findings provided valuable insights into the multifaceted roles of GDF-15 in cancer biology and establish a theoretical framework for future development of GDF-15-targeted therapies. However, this study has certain limitations that warrant consideration. First, the exclusive reliance on open-access databases introduced inherent heterogeneity among study samples, potentially contributing to systematic bias. Second, while some bioinformatics-derived results aligned with published studies, many observations require experimental verification. Third, although database predictive analysis indicated that GDF-15 was closely associated with tumor biological activities such as proliferation and cell cycle regulation and despite existing experimental evidence linking GDF-15 to proliferation and cell cycle dynamics in various tumor cells, including clear cell renal cell carcinoma, colorectal cancer, gastric cancer, melanoma, and cervical cancer33,60–63, this study did not perform experimental verification of these associations. Future studies will focus on generating robust experimental evidence to substantiate these findings and address current limitations through rigorous functional investigations.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1 (20.6KB, docx)
Supplementary Material 2 (5.2MB, docx)

Acknowledgements

The main data statistical analysis of this study was based on online statistical analysis tools. We appreciate the hard work of all tool developers.

Abbreviations

GDF-15

Growth differentiation factor 15

HPA

Human protein atlas

GETx

Genotype-tissue expression

TCGA

The Cancer Genome Atlas Program

UCSC

University of California Santa Cruz

IHC

Immunohistochemistry

ROC

The receiver operating characteristic

OS

Overall survival

DSS

Disease-specific survival

DFI

Disease-free interval

PFI

Progression-free interval

T

Tumor

N

Node

M

Metastasis

G

Grade

DNAss

DNA stemness score

EREG-METHss

Epigenetic regulation of DNA methylation

DMPss

Differential methylation probe

ENHss

Enhancer element/DNA methylation

RNAss

RNA stemness score

EREG.EXPss

Epigenetic regulation of RNA expression

TMB

Tumor mutation burden

MATH

Mutant-allele tumor heterogeneity

MSI

Microsatellite instability

NEO

Neoantigen

HRD

Homologous recombination deficiency

LOH

Heterozygosity

EMBL-EBI

European Molecular Biology Laboratory’s European Bioinformatics Institute

STRING

The search tool for the retrieval of interacting genes/proteins

GENEMANIA

Gene function using a multiple association network integration algorithm

GO

Gene ontology

KEGG

Kyoto encyclopedia of genes and genomes

m1A

N1-methyladenosine

m5C

5 methylcytidine

m6A

N6-methyladenosine

BP

Biological process

MF

Molecular function

ACC

Adrenocortical carcinoma

BLCA

Bladder urothelial carcinoma

BRCA

Breast invasive carcinoma

CESC

Cervical squamous cell carcinoma and endocervical adenocarcinoma

CHOL

Cholangiocarcinoma

COAD

Colon adenocarcinoma

COAD

Colon adenocarcinoma

READ

Rectum adenocarcinoma esophageal carcinoma

DLBC

Lymphoid neoplasm diffuse large B-cell lymphoma

ESCA

Esophageal carcinoma

GBM

Glioblastoma multiforme

HNSC

Head and neck squamous cell carcinoma

KICH

Kidney chromophobe

KIRC

Kidney renal clear cell carcinoma

KIRP

Kidney renal papillary cell carcinoma

LAML

Acute myeloid leukemia

LGG

Brain lower grade glioma

LIHC

Liver hepatocellular carcinoma

LUAD

Lung adenocarcinoma

LUSC

Lung squamous cell carcinoma

MESO

Mesothelioma

OV

Ovarian serous cystadenocarcinoma

PAAD

Pancreatic adenocarcinoma

PCPG

Pheochromocytoma and Paraganglioma

PRAD

Prostate adenocarcinoma

READ

Rectum adenocarcinoma

SARC

Sarcoma

STAD

Stomach adenocarcinoma

SKCM

Skin cutaneous melanoma

STES

Stomach and esophageal carcinoma

TGCT

Testicular germ cell tumors

THCA

Thyroid carcinoma

THYM

Thymoma

UCEC

Uterine corpus endometrial carcinoma

UCS

Uterine carcinosarcoma

UVM

Uveal melanoma

Author contributions

Conceptualization: Min Fang, Jian Xiao Data curation: Ziyu Wang, Zhu Wu, Yuxin Zhou, Fengrui YangFormal analysis: Min Fang, Zhu Wu, Jian XiaoFunding acquisition: Jian XiaoInvestigation: Yuxin Zhou, Zhu Wu, Ziyu WangMethodology: Min Fang, Ziyu Wang, Fengrui YangProject administration: Jian XiaoResources: Jian Xiao Software: Min Fang, Ziyu Wang, Jian XiaoSupervision: Jian XiaoValidation: Min FangVisualization: Ziyu Wang, Fengrui Yang, Yuxin ZhouWriting – original draft: Min FangWriting – review & editing: Ziyu Wang, Zhu Wu, Yuxin Zhou, Fengrui Yang, Jian Xiao.

Funding

This study is partly supported by the National Natural Science Foundation of China-Youth Science Fund (No. 82204453).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval and consent to participate

Shanghai Outdo Biochip Company had previously completed the ethical approval process with the approval numbers YB M-05-01 and YB M-05-02 for tissue microarrays, thus ensuring compliance with ethical norms and regulations. The other data are publicly available and considered as exempt under the ethical board review of the corresponding author’s institution.

Footnotes

Publisher’s note

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

References

  • 1.Li, X. P. et al. The emerging role of Super-enhancers as therapeutic targets in the digestive system tumors. Int. J. Biol. Sci.19 (4), 1036–1048 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Bray, F. et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J. Clin.74 (3), 229–263 (2024). [DOI] [PubMed] [Google Scholar]
  • 3.Pan-cancer analysis of whole genomes. Nature578 (7793), 82–93. (2020). [DOI] [PMC free article] [PubMed]
  • 4.Liu, Y. et al. mirna-383-5p functions as an Anti-oncogene in glioma through the akt/mtor signaling pathway by targeting VEGFA. Curr. Cancer Drug Targets. 24 (4), 463–475 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Ock, C. Y. et al. Pan-Cancer Immunogenomic perspective on the tumor microenvironment based on PD-L1 and CD8 T-Cell infiltration. Clin. Cancer Res.22 (9), 2261–2270 (2016). [DOI] [PubMed] [Google Scholar]
  • 6.Bootcov, M. R. et al. MIC-1, a novel macrophage inhibitory cytokine, is a divergent member of the TGF-beta superfamily. Proc. Natl. Acad. Sci. U S A. 94 (21), 11514–11519 (1997). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Uhlén, M. et al. Proteomics. Tissue-based map of the human proteome. Science347 (6220), 1260419 (2015). [DOI] [PubMed] [Google Scholar]
  • 8.Tsai, V. W. W. et al. The MIC-1/GDF15-GFRAL pathway in energy homeostasis: implications for obesity, cachexia, and other associated diseases. Cell. Metab.28 (3), 353–368 (2018). [DOI] [PubMed] [Google Scholar]
  • 9.Stacchiotti, S. et al. GDF-15 predicts epithelioid hemangioendothelioma aggressiveness and is downregulated by sirolimus through ATF4/ATF5 suppression. Clin. Cancer Res.30 (22), 5122–5137 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Hasanpour Segherlou, Z. et al. GDF-15: diagnostic, prognostic, and therapeutic significance in glioblastoma multiforme. J. Cell. Physiol.236 (8), 5564–5581 (2021). [DOI] [PubMed] [Google Scholar]
  • 11.Vocka, M. et al. Growth/differentiation factor 15 (GDF-15) as new potential serum marker in patients with metastatic colorectal cancer. Cancer Biomark.21 (4), 869–874 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Melero, I. et al. Neutralizing GDF-15 can overcome anti-PD-1 and anti-PD-L1 resistance in solid tumours. Nature (2024). [DOI] [PMC free article] [PubMed]
  • 13.Hao, Z. et al. Upregulated BAP31 links to poor prognosis and tumor immune microenvironment in breast cancer. Int. J. Mol. Sci.26 (13), (2025). [DOI] [PMC free article] [PubMed]
  • 14.Uhlen, M. et al. A pathology atlas of the human cancer transcriptome. Science357 (6352), (2017). [DOI] [PubMed]
  • 15.Chen, D. et al. Sangerbox 2: enhanced functionalities and update for a comprehensive clinical bioinformatics data analysis platform. Imeta3 (5), e238 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Kanehisa, M. et al. KEGG: biological systems database as a model of the real world. Nucleic Acids Res.53 (D1), D672–d677 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Kanehisa, M. & Goto, S. KEGG: Kyoto encyclopedia of genes and genomes. Nucleic Acids Res.28 (1), 27–30 (2000). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Chen, J. et al. Downregulation of the circadian rhythm regulator HLF promotes multiple-organ distant metastases in non-small cell lung cancer through PPAR/NF-κb signaling. Cancer Lett.482, 56–71 (2020). [DOI] [PubMed] [Google Scholar]
  • 19.Cui, Y. et al. Reversal of epithelial-mesenchymal transition and Inhibition of tumor stemness of breast cancer cells through advanced combined chemotherapy. Acta Biomater.152, 380–392 (2022). [DOI] [PubMed] [Google Scholar]
  • 20.Chen, F. et al. Moving pan-cancer studies from basic research toward the clinic. Nat. Cancer. 2 (9), 879–890 (2021). [DOI] [PubMed] [Google Scholar]
  • 21.Li, Y. et al. Proteogenomic data and resources for pan-cancer analysis. Cancer Cell.41 (8), 1397–1406 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Chen, G. et al. GPC2 is a potential diagnostic, immunological, and prognostic biomarker in Pan-Cancer. Front. Immunol.13, 857308 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Kannan, S., O’Connor, G. M., Bakker, E. Y. Molecular mechanisms of PD-1 and PD-L1 activity on a pan-cancer basis: A bioinformatic exploratory study. Int. J. Mol. Sci.22 (11). (2021).   [DOI] [PMC free article] [PubMed]
  • 24.Wei, C. et al. Pan-Cancer analysis shows that ALKBH5 is a potential prognostic and immunotherapeutic biomarker for multiple cancer types including gliomas. Front. Immunol.13, 849592 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Fang, M. et al. Diagnostic, prognostic, and immunological roles of NCAPG in pan-cancer: A bioinformatics analysis. Med. (Baltim).104 (10), e41761 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Wischhusen, J., Melero, I. & Fridman, W. H. Growth/Differentiation Factor-15 (GDF-15): from biomarker to novel targetable immune checkpoint. Front. Immunol.11, 951 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.He, Y. et al. Growth differentiation factor 15 is required for triple-negative breast cancer cell growth and chemoresistance. Anticancer Drugs. 34 (3), 351–360 (2023). [DOI] [PubMed] [Google Scholar]
  • 28.Liu, Y. et al. Knockdown of growth differentiation factor-15 inhibited nonsmall cell lung cancer through inactivating PTEN/PI3K/AKT signaling pathway. Genes Genomics. 45 (4), 507–517 (2023). [DOI] [PubMed] [Google Scholar]
  • 29.Myojin, Y. et al. Hepatic stellate cells in hepatocellular carcinoma promote tumor growth via growth differentiation factor 15 production. Gastroenterology160 (5), 1741–1754e16 (2021). [DOI] [PubMed] [Google Scholar]
  • 30.Zhang, X. et al.  KLF4 suppresses the proliferation of Perihilar cholangiocarcinoma by negatively regulating GDF15 and phosphorylating AKT. Oncol. Rep.50 (6). (2023). [DOI] [PMC free article] [PubMed]
  • 31.Shen, Y. et al. Research on values of GDF-15 level in the diagnosis of primary liver cancer and evaluation of chemotherapeutic effect. Eur. Rev. Med. Pharmacol. Sci.22 (12), 3749–3754 (2018). [DOI] [PubMed] [Google Scholar]
  • 32.Wang, J. et al. The H6D genetic variation of GDF15 is associated with genesis, progress and prognosis in colorectal cancer. Pathol. Res. Pract.211 (11), 845–850 (2015). [DOI] [PubMed] [Google Scholar]
  • 33.Yang, D. et al. Downregulation of GDF15 suppresses ferroptosis and predicts unfavorable prognosis in clear cell renal cell carcinoma. Cell. Div.18 (1), 21 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Zhu, X. et al. GDF15 and LCN2 for early detection and prognosis of pancreatic cancer. Transl Oncol.50, 102129 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Loh, J. J. & Ma, S. Hallmarks of cancer stemness. Cell. Stem Cell.31 (5), 617–639 (2024). [DOI] [PubMed] [Google Scholar]
  • 36.Staicu, I. KEAP1 maintains stemness. Nat. Immunol.26 (1), 3 (2025). [DOI] [PubMed] [Google Scholar]
  • 37.Zhou, M. et al. Wu Mei Wan suppresses colorectal cancer stemness by regulating Sox9 expression via JAK2/STAT3 pathway. J. Ethnopharmacol.338 (Pt 1), 118998 (2025). [DOI] [PubMed] [Google Scholar]
  • 38.Wei, Z. et al. DUSP3 restrains the progression and stemness property of osteosarcoma through regulating EGFR/STAT3/SOX2 axis. Int. J. Biol. Sci.21 (1), 160–174 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Zhao, X. et al. GDF15 contributes to radioresistance by mediating the EMT and stemness of breast cancer cells. Int. J. Mol. Sci.23 (18). (2022). [DOI] [PMC free article] [PubMed]
  • 40.Chen, X. et al. Chromosomal instability as a driver of cancer progression. Nat. Rev. Genet.26 (1), 31–46 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Federica, G., Michela, C. & Giovanna, D. Targeting the DNA damage response in cancer. MedComm5(11), e788 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Galick, H. A. et al. Germ-line variant of human NTH1 DNA glycosylase induces genomic instability and cellular transformation. Proc. Natl. Acad. Sci. U S A. 110 (35), 14314–14319 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Zhao, W. et al. DDX18 influences chemotherapy sensitivity in colorectal cancer by regulating genomic stability. Exp. Cell. Res.444 (1), 114344 (2025). [DOI] [PubMed] [Google Scholar]
  • 44.Li, Z. et al. PICH, A protein that maintains genomic stability, can promote tumor growth. Gene935, 149074 (2025). [DOI] [PubMed] [Google Scholar]
  • 45.Chen, S. et al. RNA adenosine modifications related to prognosis and immune infiltration in osteosarcoma. J. Transl Med.20 (1), 228 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Han, M. et al. Effects of RNA methylation on tumor angiogenesis and cancer progression. Mol. Cancer. 22 (1), 198 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Dyson, H. J. & Wright, P. E. How does p53 work? Regulation by the intrinsically disordered domains. Trends Biochem. Sci.50 (1), 9–17 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Whitson, R. J., Lucia, M. S. & Lambert, J. R. Growth differentiation factor-15 (GDF-15) suppresses in vitro angiogenesis through a novel interaction with connective tissue growth factor (CCN2). J. Cell. Biochem.114 (6), 1424–1433 (2013). [DOI] [PubMed] [Google Scholar]
  • 49.Zheng, X. et al. Exosomes derived from 5-fluorouracil-resistant colon cancer cells are enriched in GDF15 and can promote angiogenesis. J. Cancer. 11 (24), 7116–7126 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Wang, W. et al. Prostate cancer promotes a vicious cycle of bone metastasis progression through inducing osteocytes to secrete GDF15 that stimulates prostate cancer growth and invasion. Oncogene38 (23), 4540–4559 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Gajewski, T. F., Schreiber, H. & Fu, Y. X. Innate and adaptive immune cells in the tumor microenvironment. Nat. Immunol.14 (10), 1014–1022 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Tay, C., Tanaka, A. & Sakaguchi, S. Tumor-infiltrating regulatory T cells as targets of cancer immunotherapy. Cancer Cell.41 (3), 450–465 (2023). [DOI] [PubMed] [Google Scholar]
  • 53.Sultan, H. et al. Neoantigen-specific cytotoxic Tr1 CD4 T cells suppress cancer immunotherapy. Nature632 (8023), 182–191 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Haake, M. et al. Tumor-derived GDF-15 blocks LFA-1 dependent T cell recruitment and suppresses responses to anti-PD-1 treatment. Nat. Commun.14 (1), 4253 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Ratnam, N. M. et al. NF-κB regulates GDF-15 to suppress macrophage surveillance during early tumor development. J. Clin. Invest.127 (10), 3796–3809 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Roth, P. et al. GDF-15 contributes to proliferation and immune escape of malignant gliomas. Clin. Cancer Res.16 (15), 3851–3859 (2010). [DOI] [PubMed] [Google Scholar]
  • 57.Husaini, Y. et al. Growth differentiation factor-15 slows the growth of murine prostate cancer by stimulating tumor immunity. PLoS One. 15 (6), e0233846 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 58.Duan, L. et al. The role of GDF15 in bone metastasis of lung adenocarcinoma cells. Oncol. Rep.41 (4), 2379–2388 (2019). [DOI] [PubMed] [Google Scholar]
  • 59.Lu, Y. et al. CDP138 Silencing inhibits TGF-β/Smad signaling to impair radioresistance and metastasis via GDF15 in lung cancer. Cell. Death Dis.8 (9), e3036 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60.Huntington, K. E. et al. GSK-3 inhibitor Elraglusib enhances tumor-Infiltrating immune cell activation in tumor biopsies and synergizes with Anti-PD-L1 in a murine model of colorectal cancer. Int. J. Mol. Sci.24 (13).  (2023). [DOI] [PMC free article] [PubMed]
  • 61.Joo, M. et al. GDF15 promotes cell growth, migration, and invasion in gastric cancer by inducing STAT3 activation. Int. J. Mol. Sci.24 (3). (2023) [DOI] [PMC free article] [PubMed]
  • 62.Li, S. et al. GDF15 promotes the proliferation of cervical cancer cells by phosphorylating AKT1 and Erk1/2 through the receptor ErbB2. J. Exp. Clin. Cancer Res.37 (1), 80 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 63.Zeng, F. et al. BET inhibitors synergize with Sunitinib in melanoma through GDF15 suppression. Exp. Mol. Med.55 (2), 364–376 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

Supplementary Material 1 (20.6KB, docx)
Supplementary Material 2 (5.2MB, docx)

Data Availability Statement

No datasets were generated or analysed during the current study.

Articles from Scientific Reports are provided here courtesy of Nature Publishing Group

RESOURCES