Dual-Faced Role of GDF6 in Cancer: Mechanistic Insights into Its Context-Dependent Regulation of Metastasis and Immune Evasion Across Human Malignancies

Abstract

Growth differentiation factor 6 (GDF6), a member of the TGF-β superfamily, plays multifaceted roles in tumorigenesis, yet its molecular mechanisms and cancer-type-specific regulatory networks remain poorly defined. This study investigates GDF6's context-dependent functions through pan-cancer multi-omics integration and functional validation. Transcriptomic data from TCGA (33 cancers, n = 10,535) and GTEx were analyzed to assess GDF6 dysregulation. Co-expression networks, pathway enrichment (KEGG/GO), and epigenetic interactions (m6A, m5C, m1A) were explored. Functional assays included siRNA knockdown, wound healing, and validation in immunotherapy cohorts. GDF6 exhibited bidirectional expression patterns, with downregulation in 23 cancers (e.g., GBM, BRCA) and upregulation in 7 malignancies (e.g., KIRC, PAAD). Mechanistically, GDF6 activated the PI3K-Akt/VEGF pathways, thereby promoting angiogenesis and metastasis. It modulated epigenetic regulation through interactions with m6A readers and erasers. Additionally, GDF6 reshaped the immune microenvironment by recruiting myeloid-derived suppressor cells (MDSCs) and cancer-associated fibroblasts. Notably, GDF6's dual role extended to immunotherapy: it suppressed anti-PD1 efficacy but enhanced anti-PD-L1 sensitivity, linked to differential MHC-II and hypoxia-response regulation. This study deciphers GDF6's context-dependent molecular networks, revealing its dual roles in metastasis and immune evasion. These findings highlight GDF6 as a central node in TGF-β-mediated oncogenic signaling and a potential therapeutic target for precision intervention.

Keywords: GDF6; PI3K-Akt pathway; TGF-β superfamily; epigenetic regulation; immune microenvironment.

Figure 1

The expression of GDF6 in various human normal tissues and tumor tissues. (A) The mRNA expression of GDF6 in normal human tissues. (B) GDF6 expression in tumors and healthy tissues (TCGA database). Tumor types with significantly downregulated GDF6 expression (C) and those with significantly upregulated expression (D) in the TCGA + GTEx databases. * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001.

Figure 2

GDF6 as a multifaceted clinical biomarker: correlation with advanced staging, survival prognosis, and immunotherapy stratification. (A) Correlation between GDF6 expression and pathological stages of all TCGA cancers. (B) Relationship between GDF6 gene expression and overall survival. (C–F) Kaplan–Meier curves showing significant survival analysis results of (B). (G) Relationship between GDF6 gene expression and disease-free survival. (H–K) Kaplan–Meier curves showing significant survival analysis results of (G). (L) PFS of patients without any immunotherapy. PFS of patients treated with all anti-PD1 (M), anti-CTLA-4 (N), and anti-PD-L1 (O). * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001.

Figure 3

Correlation analysis of GDF6 with immune regulatory genes, immune cell infiltration, and tumor stemness score. (A) Correlation between GDF6 expression and the majority of immune regulatory genes. (B) Correlation between GDF6 and the infiltration levels of six types of immune cells, including B cells, T cells (CD4+ and CD8+), neutrophils, macrophages, and dendritic cells (DCs). (C) Correlation analysis between GDF6 expression and tumor stemness score. * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001.

Figure 4

Immune infiltration analysis. (A) Analysis of GDF6 expression and immune infiltration scores in tumor type. (B) Correlation between GDF6 expression and immune infiltration levels in all TCGA cancers using the TIMER2 algorithm.

Figure 5

Analysis of the relationship between GDF6 expression and cancer genome instability. (A) Analysis of GDF6 gene alteration features (mutations, amplifications, and deep deletions) in 32 different tumors from the TCGA database. (B) Pan-cancer GDF6 SNV landscape, including missense mutations, frameshift insertions, and splice site mutations. Pan-cancer analysis of the correlation between GDF6 and TMB (tumor mutational burden) (C) and MSI (microsatellite instability) (D). (E) Analysis of the correlation between genomic HRD (homologous recombination deficiency) and GDF6 expression. (F) Correlation analysis of the expression of GDF6 and 44 class III RNA modification genes (m1A (10), m5C (13), m6A (21)) in each sample. * p < 0.05; ** p < 0.01.

Figure 6

GDF6-related gene enrichment analysis and drug sensitivity analysis. (A) Expression of the top 13 GDF6-related target genes in cancers. (B) KEGG pathway analysis based on GDF6 interactions and related genes. (C) GO analysis based on GDF6 interactions and related genes. (D) Correlation between GDF6 mRNA expression and drug sensitivity in cancer cell lines (GDSC database).

Figure 7

Functional validation and clinical biomarker potential of GDF6 (A) GDF6 knockdown in Htr8/svneo cells was confirmed by qRT-PCR. (n = 3). (B,C) Wound-healing assay demonstrated that GDF6 silencing significantly inhibited cell migration (si-NC group, n = 5; si-GDF6 group, n = 4, scale bar = 1000 μm). (D) GDF6 expression was higher in recurrent versus non-recurrent UCEC patients, with ROC analysis showing predictive accuracy (recurrent group, n = 24; non-recurrent group, n = 25, AUC = 0.66). (E) In STAD, GDF6 levels were elevated in patients with OS < 5 years (OS < 5 years group, n = 28; OS ≥ 5 years group, n = 5, AUC = 0.66). (F,G) GDF6 expression stratified immunotherapy response: high GDF6 correlated with anti-PD1 resistance (non-response group, n = 38; response group, n = 31, AUC = 0.643) but enhanced anti-PD-L1 sensitivity (non-response group, n = 20; response group, n = 8, AUC = 0.506). ROC curves and boxplots (median ± IQR) are shown for predictive performance. Mean ± SD; Unpaired t test; * p < 0.05.