Identification and validation of RNA-binding protein SLC3A2 regulates melanocyte ferroptosis in vitiligo by integrated analysis of single-cell and bulk RNA-sequencing

Abstract

Background: The pathogenesis of vitiligo remains unclear. The genes encoding vitiligo-related RNA-binding proteins (RBPs) and their underlying pathogenic mechanism have not been determined.

Results: Single-cell transcriptome sequencing (scRNA-seq) data from the CNCB database was obtained to identify distinct cell types and subpopulations and the relative proportion changes in vitiligo and healthy samples. We identified 14 different cell types and 28 cell subpopulations. The proportion of each cell subpopulation significantly differed between the patients with vitiligo and healthy groups. Using RBP genes for unsupervised clustering, we obtained the specific RBP genes of different cell types in vitiligo and healthy groups. The RBP gene expression was highly heterogeneous; there were significant differences in some cell types, such as keratinocytes, Langerhans, and melanocytes, while there were no significant differences in other cells, such as T cells and fibroblasts, in the two groups. The melanocyte-specific RBP genes were enriched in the apoptosis and immune-related pathways in the patients with vitiligo. Combined with the bulk RNA-seq data of melanocytes, key RBP genes related to melanocytes were identified, including eight upregulated RBP genes (CDKN2A, HLA-A, RPL12, RPL29, RPL31, RPS19, RPS21, and RPS28) and one downregulated RBP gene (SLC3A2). Cell experiments were conducted to explore the role of the key RBP gene SLC3A2 in vitiligo. Cell experiments confirmed that melanocyte proliferation decreased, whereas apoptosis increased, after SLC3A2 knockdown. SLC3A2 knockdown in melanocytes also decreased the SOD activity and melanin content; increased the Fe²⁺, ROS, and MDA content; significantly increased the expression levels of TYR and COX2; and decreased the expression levels of glutathione and GPX4.

Conclusion: We identified the RBP genes of different cell subsets in patients with vitiligo and confirmed that downregulating SLC3A2 can promote ferroptosis in melanocytes. These findings provide new insights into the pathogenesis of vitiligo.

Keywords: Bulk RNA sequencing; Ferroptosis; RNA-binding protein; SLC3A2; Single-cell sequencing; Vitiligo.

Fig. 1

Schematic of sample preparation, scRNA-seq data processing, and data analysis

Fig. 2

ScRNA-seq analysis of skin from healthy donors and patients with vitiligo identified distinct cell types. (A, B) UMAP plot of composite single-cell transcriptomic profiles from all 15 skin samples from healthy donors and patients with vitiligo. Colors indicate cell clusters along with annotations. Mono. Phagocyte: mononuclear phagocyte. (C) UMAP plot split by different sample groups. (D) Bar plot comparing the proportions of cell populations of each cell type within each sample group. (E) Rank order based on decreasing values of the relative frequency ratio between two sample groups. (F) Difference in the proportion of cell clusters between two sample groups

Fig. 3

Single-cell analysis revealed heterogeneity and regulatory module of cell-specific RBPs in healthy donors and patients with vitiligo. (A) UMAP plot of scRNA-seq profile. Cells are colored according to cell clusters based on RBP-expression module. (B) Stacked bar plot showing relative proportions of cell populations based on RBP-expression module in keratinocytes, Langerhans, and melanocytes. (C) Unsupervised clustering heatmap showing relative expression (z score, row scaled) levels of top 5 representative markers in each cell population based on RBP-expression model. (D) Gene Ontology enrichment analysis of biological processes of top 50 marker genes of each cell type. Top three terms were selected for each cluster, and heatmap shows enrichment q-value of these terms (scaled by column). (E) Gene Ontology biological process of top 50 marker genes in Fig. 2D cluster8 based on RBP-expression module. (F) Dot plot showing expression of eight RBPs involved in apoptotic and immune system process terms from D in melanocytes. Dot color represents mean expression in each cluster; dot size indicates the percentage of cells expressing marker genes in each cluster

Fig. 4

Functional RBPs are differentially regulated in melanocytes of patients with vitiligo and healthy donors. (A) Dot plot showing several differentially expressed RBPs in each cell type between vitiligo and healthy sample groups. (B) Scatter plot displaying differentially expressed RBPs in C2 melanocytes. Red color indicates significantly altered RBPs. (C) Gene expression levels of S100A4 in different sample groups represented in UMAP plot. (D) Violin plot of BST2 in each cell type in different sample groups. (E) Cytoscape shows co-expression networks comprising SLC3A2 from C2 melanocyte cell clusters. Edges connect RBP-target gene pairs, while nodes represent genes. RBPs are displayed in larger font and red color. (F) Enriched biological processes of target genes for SLC3A2 from E are shown

Fig. 5

Integration of bulk RNA-seq data reveals significant differentially expressed RBPs in vitiligo melanocytes. (A) Expression heatmap of all 352 significant differentially expressed RBPs between PIG3V (vitiligo) and PIG1 (normal) melanocyte cell lines. (B) Venn diagram showing co-upregulated RBPs in bulk RNA-seq and C2 melanocyte cell clusters based on ScRNA-seq. (C) Venn diagram showing co-downregulated RBPs in bulk RNA-seq and C2 melanocyte cell clusters based on ScRNA-seq. (D) Box plot showing expression level of SLC3A2

Fig. 6

Downregulation of SLC3A2 results in obvious oxidative damage and relates to melanocyte ferrocytosis. (A) Expression level of SLC3A2 was reduced, and SLC3A2 was downregulated in SLC3A2-interference group. Gel images were cropped for display. (B) Inhibition of melanocytes in G1 phase after SLC3A2 interference. (C) SLC3A2-interference group showed significant decrease in cell proliferation rate and significant increase in apoptosis rate and ROS content. In the SLC3A2-interference group, contents of Fe²⁺, MDA, GSH, and GPX4 expression in melanocytes increased, and SOD activity and TYR and COX2 expression levels decreased. Note: SLC3A2-interference vs. control, **: P < 0.01, ***: P < 0.001; SLC3A2-interference vs. SLC3A2-interference NC, ##: P < 0.01, ###: P < 0.001