Inhibition of adenosine/A2A receptor signaling suppresses dermal fibrosis by enhancing fatty acid oxidation

Abstract

Background: Skin fibrosis presents a major challenge for clinicians treating conditions like systemic sclerosis (SSc) due to its progressive course and limited treatment options. While the role of metabolism in fibrosis has gained increasing attention, the crucial alterations of metabolic pathway and the underlying signaling of metabolic interconnections in regulating SSc-related skin fibrosis remain largely elusive.

Methods: Metabolomic analysis was performed on plasma samples from 35 SSc patients to identify metabolic alterations. In bleomycin (BLM)- and hypochlorous acid (HOCL)-induced skin fibrosis mouse models, we assessed the impact of global A2a receptor knockout on skin fibrosis. Single-cell RNA sequencing of mouse skin was utilized to investigate the role of A2A in fibroblasts during fibrotic challenge. Human dermal fibroblasts were used in in vitro experiments, employing RNA sequencing and Seahorse assays, to assess the relationship between A2A signaling and fatty acid oxidation (FAO). Finally, fibroblast-specific conditional A2a knockout mice were used to test the effects of specifically targeting A2A in dermal fibroblasts.

Results: Adenosine-centered nucleotide metabolism was elevated in the plasma of SSc patients. Mechanistically, by stimulating dermal fibroblasts with key pathogenic cytokines associated with SSc, we observed significant changes in adenosine receptor A2A expression in response to IL-1β. Immunofluorescence revealed upregulation of A2A expression in dermal fibroblasts of SSc patients. Further, global A2a knockout significantly attenuated skin fibrosis in both BLM- and HOCL-induced skin fibrosis mouse models. Single-cell RNA sequencing of mouse skin revealed significant alterations in fatty acid metabolism in fibroblasts from A2a-deficient mice following fibrotic challenge. RNA sequencing, Seahorse assays and in vitro experiments showed that A2A inhibition promotes FAO by upregulating CPT1A expression via suppressing CREB phosphorylation, alleviating fibrosis in human primary dermal fibroblasts. Furthermore, targeted intervention of A2a specifically in fibroblasts improves outcomes and increases CPT1A expression in BLM-induced skin fibrosis mouse model.

Conclusion: Our study highlights the crucial interplay between adenosine metabolism-A2A receptor axis and FAO in SSc-associated skin fibrosis, suggesting that targeting the adenosine receptor A2A-FAO metabolic axis offers a promising therapeutic strategy for skin fibrosis.

Fig. 1

Nucleotide metabolism centered around adenosine is significantly elevated in SSc. (A) Clinical characteristics of SSc patients (n = 35). (B-D) Orthogonal projections to latent structures-discriminant analysis (OPLS-DA) showing group distribution (B), volcano plot representing differentially expressed metabolites (C) and Z-score plot showing significantly upregulated and downregulated metabolites (D) of plasma samples from SSc patients (n = 35) and HC (n = 14) groups. (E) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis based on significantly upregulated metabolites of SSc patients. (F) Differential metabolites in the nucleotide metabolism pathway of SSc patients compared to HC. (G) Schematic representation of ATP to adenosine hydrolysis pathway. Arrowheads indicate significantly upregulated metabolites in SSc patients. (H) Plasma adenosine concentrations in SSc patients (n = 20) and HC (n = 15). Error bar represents mean +/- SD range. *P < 0.05; HC, healthy control

Fig. 2

Adenosine receptor A2a deficiency reduces fibrosis in both BLM- and HOCL-induced skin fibrosis mouse models. (A) RT-qPCR analysis of adenosine receptors expression in human dermal fibroblasts treated with TGF-β, IL-1β, IL-4, IL-6 or IL-12 for 48 h. Statistical significance was determined by comparison with the NC group. (B) Representative immunofluorescent staining of A2A (green) and VIMENTIN (red), and quantification of VIMENTIN⁺ A2A⁺ cells in the skin sections from HC and SSc patients (n = 5, Scale bar: 100 μm). (C-H) Quantification of plasma adenosine concentrations (C); representative H&E staining (D) and Masson’s trichrome staining (E), quantification of dermal thickness (μm) (F) and hydroxyproline content (G), and representative image of immunohistochemistry staining of α-SMA, along with quantification of α-SMA-positive myofibroblasts in the skin tissues from WT and A2a^-/- mice treated with PBS (n = 4) or BLM (n = 8). (I) RT-qPCR analysis of Acta2, Col1a1, Col1a2 and Il1b in the skin from WT and A2a^-/- mice treated with PBS or BLM (n = 6). (J-M) Representative H&E staining (J), Masson’s trichrome staining (K), quantification of dermal thickness (μm) (L), and representative image of immunohistochemistry staining of α-SMA, along with quantification of α-SMA-positive myofibroblasts (M) in the skin tissues from WT and A2a^-/- mice treated with PBS (n = 4) or HOCL (n = 8) (Scale bar: 200 μm). RT-qPCR of human dermal fibroblasts was normalized against GAPDH, and RT-qPCR results of murine skin were normalized against βactin. n = 4–8/group. Error bars represent mean +/- SD range. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001; ns, not significant, HC, healthy control

Fig. 3

Single-cell RNA-seq revealed altered fatty acid metabolism in fibroblasts after BLM and HOCL challenge in A2a^−/− mice. (A) UMAP plot of main cell types in single-cell RNA-seq data from skin tissue of WT and A2a^−/− mice treated with PBS or BLM or HOCL. (B) Dot plot illustrating marker gene expression levels for the different cell types defined in (A). (C) The proportion of different cell types in skin. (D) Number of differentially expressed genes (DEGs) in different cell types of WT and A2a^−/− mice treated with BLM (upper) or HOCL (lower). Red bars indicate upregulated genes, and blue bars indicate downregulated genes. (E-F) Gene Ontology (GO) enrichment analysis of DEGs in skin fibroblasts between WT mice treated with PBS or BLM (E), and between WT and A2a^−/− mice treated with BLM (F). (G-H) GO enrichment analysis of DEGs in skin fibroblasts between WT mice treated with PBS or HOCL (G), and between WT and A2a^−/− mice treated with HOCL (H)

Fig. 4

A2A mediates fibrotic response by regulating FAO in human dermal fibroblasts. (A-B) Immunoblot analysis of fibronectin and α-SMA (A) and RT-qPCR analysis of ACTA2, COL1A1 and FN1 mRNA (B) in dermal fibroblasts treated with TGF-β, with or without 1–20 μM SCH442416, for 48 h. Statistical significance was determined by comparison with the TGF-β group. (C-D) Volcano plot of the significant DEGs (C) and GO enrichment analysis of DEGs (D) in dermal fibroblasts treated with or without TGF-β for 48 h (left panel), and between the TGF-β-treated dermal fibroblasts treated with or without SCH442416 (10μM) for 48 h (right panel). (E) Heatmap showing the expression of FAO-related genes in the dermal fibroblasts treated with TGF-β, with or without SCH442416, for 48 h. (F) Gene set enrichment analysis (GSEA) of fatty acid oxidation pathway in dermal fibroblasts treated with or without TGF-β (left panel) for 48 h, and in TGF-β-treated dermal fibroblasts treated with or without SCH442416 (right panel) for 48 h. (G) Oxygen-consumption rate (OCR) of dermal fibroblasts pretreated with TGF-β, with or without SCH442416 (10μM), for 48 h. (H-J) Quantification of basal OCR, maximal OCR and ATP production from (G). (K-L) Quantification of acetylcarnitine and palmitoylcarnitine in cell lysate (K) and cell culture supernatant (L) from dermal fibroblasts treated with TGF-β, with or without SCH442416, for 48 h. (M) Quantification of plasma acetylcarnitine and palmitoylcarnitine concentrations from SSc patients (n = 20) and healthy controls (n = 15). (N-O) Immunoblot analysis of fibronectin and α-SMA (N), and RT-qPCR analysis of ACTA2, COL1A1 and FN1 mRNA (O), in dermal fibroblasts treated with TGF-β in combination with different concentrations of IVA337 for 48 h. (P-Q) Immunoblot analysis of fibronectin and α-SMA (P), and RT-qPCR analysis of COL1A1, COL1A2 and FN1 mRNA (Q) in dermal fibroblasts treated with TGF-β, CGS21680 (10μM), and IVA337 (20μM) for 48 h. RT-qPCR results were normalized against GAPDH. Error bars represent mean +/- SD range. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001; Oligo, oligomycin, FCCP, Carbonyl cyanide-4 (trifluoromethoxy) phenylhydrazone, Rot/AA, Rotenone/antimycin, Eto, etomoxir, ns, not significant; HC, healthy control

Fig. 5

A2A mediates FAO through regulating CPT1A expression. (A-B) Immunoblot analysis of CPT1A (A) and RT-qPCR analysis of CPT1A (B) in dermal fibroblasts treated with TGF-β, with or without 1–20 μM SCH442416, for 48 h. (C-D) Representative immunofluorescent staining of CPT1A (green) and VIMENTIN (red) (C), and quantification of VIMENTIN⁺ CPT1A⁺ cells in the skin tissues from HC and SSc patients (D) (n = 5, scale bar: 20 μm). (E-F) Immunoblot analysis of fibronectin and α-SMA (E), and RT-qPCR analysis of COL1A1, COL1A2 and FN1 mRNA (F), in dermal fibroblasts treated with TGF-β, SCH442416 (10μM), and etomoxir (5μM) for 48 h. (G) RT-qPCR analysis of CPT1A mRNA in dermal fibroblasts after silencing of CPT1A. (H-I) Immunoblot analysis of fibronectin, α-SMA and CPT1A (H), and RT-qPCR analysis of ACTA2 and FN1 mRNA (I), in dermal fibroblasts after silencing of CPT1A and incubation with TGF-β for 48 h. (J-K) Immunoblot analysis of fibronectin and α-SMA (J), and RT-qPCR analysis of ACTA2 and FN1 mRNA (K), in dermal fibroblasts after silencing of CPT1A and incubation with TGF-β and SCH442416 (10μM) for 48 h. RT-qPCR results were normalized against GAPDH. Error bars represent mean +/- SD range. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001; HC, healthy control

Fig. 6

A2A regulates CPT1A expression via p-CREB/CREB pathway. (A) Immunoblot analysis of p-CREB and CREB in dermal fibroblasts treated with TGF-β, with or without SCH442416 (10μM), for 48 h. (B-C) Immunoblot analysis of p-CREB, CREB, fibronectin and CPT1A (B); and RT-qPCR analysis of COL1A1, FN1 and CPT1A (C) in dermal fibroblasts pretreated with or without forskolin (30μM, 30 min), followed by stimulation with or without TGF-β for 48 h. (D-E) Immunoblot analysis of p-CREB, CREB, fibronectin and CPT1A (D); and RT-qPCR analysis of COL1A1, FN1 and CPT1A (E) in dermal fibroblasts pretreated with or without forskolin, followed by stimulation with or without TGF-β and SCH442416 (10μM) for 48 h. (F-G) Immunoblot analysis of p-CREB, CREB, fibronectin and CPT1A (F), and RT-qPCR analysis of COL1A1, FN1 and CPT1A (G) in dermal fibroblasts treated with TGF-β, CGS21680 (10μM), and 666 − 15 (0.1μM) for 48 h. RT-qPCR results were normalized against GAPDH. Error bars represent mean +/- SD range. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001; HC, healthy control

Fig. 7

Fibroblast-specifically knockout of A2a ameliorates BLM-induced skin fibrosis. (A-C) Representative H&E staining (A), Masson’s trichrome staining (B) and quantification of dermal thickness (μm) of skin sections (C) from A2a^f/fCol1a2^cre− and A2a^f/fCol1a2^cre+ mice treated with PBS or BLM (Scale bar: 200 μm). (D-E) Representative image of immunohistochemistry staining of α-SMA (D), and quantification of α-SMA-positive myofibroblasts (E) in the skin (Scale bar: 100 μm). (F-G) Quantification of hydroxyproline content and RT-qPCR analysis of fibrotic genes Acta2, Col1a1 and Col1a2, and inflammation-related genes Ifng, Il1b, Il6 and Il11 in the skin. (H) Representative immunofluorescent staining of CPT1A (green) and VIMENTIN (red), and quantification of VIMENTIN⁺ CPT1A⁺ cells in the skin (Scale bar: 20 μm). RT-qPCR results were normalized against βactin. n = 5/group; error bars represent mean +/- SD range. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001; ns, not significant