Single-cell transcriptome and multi-omics integration reveal ferroptosis-driven immune microenvironment remodeling in knee osteoarthritis

Abstract

Background: Knee osteoarthritis (KOA) is a chronic inflammatory joint disorder marked by cartilage degradation and immune microenvironment dysregulation. While transcriptomic studies have identified key pathways in KOA, the interplay between ferroptosis (an iron-dependent cell death mechanism) and immune dysfunction at single-cell resolution remains unexplored. This study integrates single-cell and bulk transcriptomics to dissect ferroptosis-driven immune remodeling and identify diagnostic biomarkers in KOA.

Methods: We analyzed scRNA-seq data (GSE255460, n = 11) and bulk RNA-seq cohorts (GSE114007: 20 KOA/18 controls; GSE246425: 8 KOA/4 controls). Single-cell data were processed via Seurat (QC: mitochondrial genes >3 MAD; normalization: LogNormalize; batch correction: Harmony) and annotated using CellMarker/PanglaoDB. CellChat decoded intercellular communication, SCENIC reconstructed transcriptional networks, and Monocle2 for pseudotime trajectory mapping. Immune infiltration (CIBERSORT) and a LASSO-SVM diagnostic model were validated by ROC curves. Functional enrichment (GSEA/GSVA) and immunometabolic profiling were performed.

Results: Twelve chondrocyte clusters were identified, including ferroptosis-active homeostasis chondrocytes (HomC) (p < 0.01), which exhibited 491 DEGs linked to lipid peroxidation. HomC orchestrated synovitis via FGF signaling (ligand-receptor pairs: FGF1-FGFR1), amplifying ECM degradation and inflammatory cascades (CellChat). SCENIC revealed 10 HomC-specific regulons (e.g., SREBF1, YY1) driving matrix metalloproteinase activation. A 7-gene diagnostic panel (IFT88, MIEF2, ABCC10, etc.) achieved AUC = 1.0 (training) and 0.78 (validation). Immune profiling showed reduced resting mast cells (p = 0.003) and monocytes (p = 0.02), with ABCC10 correlating negatively with CD8+ T cells (r = -0.65) and M1 macrophages. GSEA/GSVA implicated HIF-1, NF-κB, and oxidative phosphorylation pathways in KOA progression. Pseudotime analysis revealed fibrotic transitions (COL1A1↑, TNC↑) in late-stage KOA cells.

Conclusion: This study establishes ferroptosis as one of the key drivers immune-metabolic dysfunction in KOA, with HomC acting as a hub for FGF-mediated synovitis and ECM remodeling. The diagnostic model and regulon network (SREBF1/YY1) offer translational tools for early detection, while impaired mast cell homeostasis highlights novel immunotherapeutic targets. Our findings bridge ferroptosis, immune dysregulation, and metabolic stress, advancing precision strategies for KOA management.

Keywords: diagnostic biomarkers; ferroptosis; immune microenvironment; knee osteoarthritis; single-cell transcriptomics.

Figure 1

Cellular annotation and differences in ferroptosis scores. (A) Cells were grouped into 18 clusters using UMAP algorithm based on important components available from PCA. (B) Annotation of the 18 clusters, categorized into 12 cell types: preHTC (prehypertrophic chondrocytes), FC (fibrocartilage chondrocytes), EC (effector chondrocytes), HomC (homeostasis chondrocytes), proC (proliferation chondrocytes), InfC (inflammatory chondrocytes), HTC (hypertrophic chondrocytes), preInfC (pre-inflammatory chondrocytes), RepC (reparative chondrocytes), RegC (regulator chondrocytes), preFC (prefibrocartilage chondrocytes), and cycle cells. (C) Bubble plot of 12 cell types and their corresponding cell markers in the Doplot. (D) Pie chart illustrating the proportion of 12 cell types.(E) Differential expression of ferroptosis scores across the 12 cell types.

Figure 2

Cell-cell interactions. (A) The number and strength of interactions between cell subpopulations. (B, C) Cell communication through signaling pathways, with FGF signaling as the major ligand-receptor docking pathway. (D) Signal intensity from THomC is higher than other cells. (E) Interaction network between cells and the FGF signaling pathway in chord diagram format. (F) Violin plot of the expression level of FGF signaling pathway in cells. (G) Bubble plot showing receptor-ligand interactions between cells.

Figure 3

SCENIC analysis. (A) Heatmap displaying the regulon activity scores of each cell in the HomC subpopulation. SCENIC analysis identified all regulons in this subpopulation, and their activity distribution is visualized. (B, C) Scatter plots showing the specificity ranking of transcription factors in high-score and low-score groups.

Figure 4

Construction of predictive models. (A) LASSO coefficient distribution of prognostic genes and gene combinations at the minimal lambda value. (B) Ten-fold cross-validation of the LASSO model to select the optimal lambda value. (C) Lasso coefficients for selected genes. (D, E) ROC curves for training and validation sets.

Figure 5

Immune infiltration analysis. (A) Relative percentages of immune cell subpopulations. (B) Correlation of immune cells, with blue representing negative correlation and red representing positive correlation. (C) Differences in immune cell content between control and disease samples. (D) Correlation between key genes and immune cells. ns, not significant (p > 0.05); *p < 0.05; **p < 0.01; ***p < 0.001.

Figure 6

Relationship between key genes and immune factors. (A–E) Correlation of key genes with chemokines, immunoinhibitors, immunostimulators, MHC, and receptors.

Figure 7

GSEA analysis of key genes. (A–G) KEGG signaling pathways involving key genes, including pathway regulation and associated genes.

Figure 8

GSVA analysis of key genes. (A–G) GSVA analysis for key genes, with blue representing high-expression genes involved in signaling pathways and green representing low-expression genes. The hallmark gene set is used as the background.

Figure 9

Single-cell expression. (A) Scatter plot showing the expression profile of key genes in single cells. (B) Violin plot of key gene expression in single cells.

Figure 10

Cellular developmental trajectories. (A–C) Pseudotime analysis and developmental trajectories of cells. (D) Gene expression dynamics in each pigment cell branch. (E) Relationship between key gene expression and cell developmental trajectories.