Celastrol directly targets LRP1 to inhibit fibroblast-macrophage crosstalk and ameliorates psoriasis progression

Abstract

Psoriasis is an incurable chronic inflammatory disease that requires new interventions. Here, we found that fibroblasts exacerbate psoriasis progression by promoting macrophage recruitment via CCL2 secretion by single-cell multi-omics analysis. The natural small molecule celastrol was screened to interfere with the secretion of CCL2 by fibroblasts and improve the psoriasis-like symptoms in both murine and cynomolgus monkey models. Mechanistically, celastrol directly bound to the low-density lipoprotein receptor-related protein 1 (LRP1) β-chain and abolished its binding to the transcription factor c-Jun in the nucleus, which in turn inhibited CCL2 production by skin fibroblasts, blocked fibroblast-macrophage crosstalk, and ameliorated psoriasis progression. Notably, fibroblast-specific LRP1 knockout mice exhibited a significant reduction in psoriasis like inflammation. Taken together, from clinical samples and combined with various mouse models, we revealed the pathogenesis of psoriasis from the perspective of fibroblast-macrophage crosstalk, and provided a foundation for LRP1 as a novel potential target for psoriasis treatment.

Keywords: CCL2; Celastrol; Drug target; Fibroblast; LRP1; Macrophage; Psoriasis; c-Jun.

Graphical abstract

Figure 1

Single-cell multi-omics reveals fibroblasts' importance in psoriasis. (A) Schematic diagram of single-cell transcriptomics, spatial transcriptomics, and scATAC-seq studies in normal individuals and psoriatic patients' skin. (B, C) UMAP of scRNA-seq cells recovered from 11 normal individuals and 17 psoriatic patients (GEO databases: GSE230842 and GSE173706). (D) Proportional map of each cell subpopulation in skin tissue of psoriasis patients and normal individuals. (E) Strength of interactions between skin cell types of normal individuals (left) and psoriasis patients (right). (F) GSEA enrichment analysis of psoriasis-related terms. (G) T-SNE plot showing the joint clustering of scATAC-seq data in skin tissues. Cells in the T-SNE are annotated by scRNA-seq data T-SNE plots colored by clusters, labeled by scRNA-seq data. (H) The heatmap showing the gene expression and chromatin accessibility of CCL family genes in fibroblasts of psoriasis patients and normal individuals skin tissues. (I) Expression Gene score of CCL2 in skin tissues of psoriasis patients and normal individuals. Based on scATAC-seq data. (J) Expression of CCL2 on scRNA-seq data. (K) Distribution of fibroblast subtype on the UMAP. (L) VlnPlot showing the expression of CCL2 in each fibroblast subtype. (M) Pseudotime trajectory of fibroblast subtypes visualized using Monocle2. The trajectory is color-coded from left to right by pseudotime, cell states, subtypes, and groups. (N) Heatmap displaying the functional pathways that are both highly and poorly enriched in the six cell states of fibroblast subtypes, as determined by GSEA. ∗∗P value < 0.01.

Figure 2

Ccl2 inhibition improves imiquimod-induced psoriasis-like skin inflammation in mice. (A, B) The process for screening of Ccl2 inhibitors. Primary skin fibroblasts derived from C57BL/6 mice, untreated or pre-treated with 1896 natural small molecule compounds (MedChemExpress, Library ID: HY-LD-000004273) and then stimulated by TNFα and IL17A for 6 h qPCR analysis of Ccl2 mRNA expression. The blue dotted line represents the 90% inhibition threshold. Celastrol is labelled, and inhibited Ccl2 expression by 97.0%. C57BL/6 female mice (n = 6/group) were treated intraperitoneally with the indicated doses of celastrol (CE) for four days (C–H). (C) Phenotypic representation (top) and H&E staining (bottom) of back skin sections of Sham, IMQ-induced and celastrol-treated mice. Skin epidermal thickness statistics for each group on the right. (D) Clinical scores of Sham, IMQ-induced and celastrol-treated mice. (E) Body weight during the disease process. (F–H) qPCR analysis of mRNA encoding cytokines from the IL-23/Th17 cell axis (F), pro-inflammatory cytokines (G), and keratin formation-related cytokines (H) in the skin. The Data are represented as mean ± SEM. The P-values are determined by a two-tailed unpaired Student's t-test for (C, F–H) or Tukey's multiple-comparison test for (D, E). ∗P < 0.05, ∗∗P < 0.01.

Figure 3

The scRNA-seq analysis demonstrates that celastrol blocks fibroblast-myeloid cell crosstalk in mice. (A) Workflow of dorsal skin scRNA-seq analysis for Sham, IMQ, and IMQ + CE groups. (B, C) UMAP of scRNA-seq cells recovered from Sham, IMQ and IMQ + CE groups. (D) Dot plot for expression of marker genes in skin cell types. (E) Projection of gene expression of known cell type specific markers across cell groups. (F) Proportional map illustrating the distribution of each cell subpopulation within the skin tissue of the Sham, IMQ, and IMQ + CE groups. (G) Number of interactions between IMQ induced C57BL/6 mouse skin cell types, untreated (left) or treated with CE (right). (H) Expression of Ccl2 in skin cell types. (I) Spatial distribution and proportion of CCL2 in psoriasis patients and normal individuals skin tissues. (J) Ccr2 expression in Sham, IMQ, and IMQ + CE groups. (K) Expression of Ccl2 and Ccr2 in Sham, IMQ and IMQ + CE groups. The data are represented as mean ± SEM. The P-values are determined by two-tailed unpaired Student's t-test. ∗∗P < 0.01.

Figure 4

Celastrol targets LRP1 to alleviate psoriasis in mice. (A, B) Mouse primary skin fibroblasts were analyzed using MTT assay. (C) Celastrol directly binding proteins using TRAP assay. (D) LRP1 expression in various cell populations of the scRNA-seq of psoriasis patients and normal individuals skin tissues. (E) Spatial feature plot of LRP1's expression. (F) LRP1 staining in skin sections of psoriatic patients (n = 3) and normal individuals (n = 3). (G) Expression of Lrp1 in skin tissue from Sham, IMQ and IMQ + CE groups. (H–L) Binding of celastrol to LRP1 was detected using CETSA (H), pull down (I), immunofluorescence (J), docking (K), and MST (L). (M, N) LRP1 expression in mouse primary skin fibroblasts infected with si-Ctrl, si-Lrp1, vector, or LRP1 plasmids. (O, P) qPCR detection of mRNA expression of Ccl2 and Lrp1 in knockdown (O) and overexpression (P) LRP1 mouse primary skin fibroblasts. The data are represented as mean ± SEM. The P-values are determined by Tukey's multiple comparison test for (O, P). ∗P < 0.05, ∗∗P < 0.01.

Figure 5

Celastrol prevents LRP1 from binding with c-Jun. (A) Plot of rank-sorted TF motifs and colored by the significance of their enrichment in fibroblasts, which compared the psoriasis group with the normal group and based on the scATAC-seq profiles. (B) Heatmap of CXCL family -related genes in each group. (C) Genome tracks of JUN in psoriasis group. (D, E) Phosphorylation levels of transcription factors c-Jun and c-Fos in knockdown (D) and overexpression (E) of LRP1 mouse primary skin fibroblasts. (F) Mouse primary skin fibroblasts derived from C57BL/6 mice were stimulated by TNFα and IL17A and treated with or without celastrol. Conducted subcellular separation and detect cytoplasm and nuclear proteins with respective antibodies. (G) Fibroblasts were stimulated by TNFα and IL17A and treated with or without celastrol, followed by isolation of the cytoplasm and nucleus and immunoprecipitation using anti-LRP1 beads. (H) Immunofluorescence staining of primary skin fibroblasts generated from C57BL/6 mice was stimulated by TNFα and IL17A and treated with or without celastrol. Scale bar: 20 μm.

Figure 6

LRP1 deficiency relieves the IMQ-induced psoriasiform inflammation. (A–E) Phenotypic representation (A), representative H&E staining (B), epidermal thickness (C), clinical scores (D), and body weight (E) of back skin sections of wild-type (WT) (n = 6) and Lrp1^−/− mice (n = 6) treated with IMQ for 4 days. (F, G) qPCR analysis of mRNA encoding cytokines from the IL-23/Th17 cell axis (F) and pro-inflammatory cytokines (G) in skin. (H) Immunofluorescence staining of primary skin fibroblast generated from Lrp1^−/− mice and wild-type mice. Scale bar: 20 μm. (I) Diagram of cell co-culture model. Primary skin fibroblasts derived from Lrp1^−/− and wild-type mice were pre-treated with celastrol and then stimulated by TNFα and IL17A, and the cell culture supernatants were co-cultured with mouse peritoneal macrophages (sh-Nc or sh-Ccr2 or sh-Ccr4) for 6 h. (J) Primary skin fibroblasts derived from Lrp1^−/− and wild-type mice were pre-treated with celastrol, anti-CCL2, or both celastrol and anti-CCL2, and then stimulated by TNFα and IL17A for 1 h, and removed them. After another 24 h of incubation, the cell culture supernatants were co-cultured with mouse peritoneal macrophages for 6 h to detect the mRNA expression of psoriasis-related factors in peritoneal macrophages. Results were normalized to Gapdh expression. The data are represented as mean ± SEM. The P-values are determined by Tukey's multiple comparison test for (C–G, J). ∗P < 0.05, ∗∗P < 0.01, ns, not significant.

Figure 7

Celastrol ameliorates psoriasis-like inflammation in cynomolgus monkeys. Cynomolgus monkeys (n = 6/group) were applied with celastrol on the back for 12 days. (A, B) Phenotypic representation (A) and representative H&E staining (B) of back skin sections of Sham, IMQ-induced, and celastrol-treated Cynomolgus monkeys. Scale bar: 50 μm. (C) Ultrasound imaging of Cynomolgus monkey skin. (D) Epidermal thickening ratio of Cynomolgus monkeys during the disease process. (E) Cynomolgus monkey skin vascularisation grading. (F) Skin sections from Sham, IMQ-induced, and celastrol-treated Cynomolgus monkeys were immune-stained for LRP1 together with p-c-Jun. Scale bar: 50 μm. The data are represented as mean ± SEM. The P-values are determined by Tukey's multiple comparison test for (B, D, E). ∗P < 0.05, ∗∗P < 0.01.