Multimodal analyses of early, untreated systemic sclerosis skin identify a proinflammatory vascular niche of macrophage-fibroblast signaling
Helen C. Jarnagin, Rezvan Parvizi, Zhiyun Gong, Rosemary Gedert, Xianying Xing, Lam (Alex) C. Tsoi, Rachael Bogle, Madeline J. Morrisson, Laurent Perreard, Patricia A. Pioli, Fred Kolling IV, Johann E. Gudjonsson, Dinesh Khanna, Michael L. Whitfield
Helen C. Jarnagin, Rezvan Parvizi, Zhiyun Gong, Rosemary Gedert, Xianying Xing, Lam (Alex) C. Tsoi, Rachael Bogle, Madeline J. Morrisson, Laurent Perreard, Patricia A. Pioli, Fred Kolling IV, Johann E. Gudjonsson, Dinesh Khanna, Michael L. Whitfield
View: Text | PDF
Research Article Dermatology

Multimodal analyses of early, untreated systemic sclerosis skin identify a proinflammatory vascular niche of macrophage-fibroblast signaling

Abstract

Uncovering the early interactions and spatial distribution of dermal fibroblasts and immune cells in treatment-naive patients with diffuse cutaneous systemic sclerosis (SSc) is critical to understanding the earliest events of skin fibrosis. We generated an integrated multiomic dataset of early-stage, treatment-naive diffuse cutaneous SSc skin. Skin biopsies were analyzed by single-nuclei multiome sequencing (snRNA-Seq and snATAC-Seq) and two spatial transcriptomic methods to comprehensively determine molecular changes. We identified an immunomodulatory niche within the papillary, hypodermis, and vascular regions enriched for activated myeloid cells and fibroblasts characterized by expression of genes such as CXCL12, APOE, and C7. Pathway analyses showed significant enrichment of PI3K/AKT/mTOR signaling pathway expression in these cellular niches, driven by profibrotic growth factor signaling networks. Macrophage subclustering showed SSc-specific macrophage activation of IL-6/JAK/STAT signaling and enrichment of oxidative phosphorylation pathways. Ligand-receptor analysis revealed that SSc macrophages secrete PDGF and TGF-β to activate SSc-dominant fibroblast subclusters. Spatial transcriptomic analyses showed monocyte-derived MRC1+ macrophages express PDGF near PDGFRhiTHY1hi fibroblasts. Multiomic data integration and spatial transcriptomic neighborhood analysis revealed the colocalization of fibroblasts, macrophages, and T cells around the vasculature. These data suggest that interactions between activated immune cells and immunomodulatory fibroblasts around vascular niches are an early event in scleroderma pathogenesis.

Authors

Helen C. Jarnagin, Rezvan Parvizi, Zhiyun Gong, Rosemary Gedert, Xianying Xing, Lam (Alex) C. Tsoi, Rachael Bogle, Madeline J. Morrisson, Laurent Perreard, Patricia A. Pioli, Fred Kolling IV, Johann E. Gudjonsson, Dinesh Khanna, Michael L. Whitfield

×

Figure 6

In vitro studies highlight the fibroblast PDGF response and demonstrate the efficacy of PDGFR and PI3K inhibitors.

Options: View larger image (or click on image) Download as PowerPoint
In vitro studies highlight the fibroblast PDGF response and demonstrate ...
(A) Experimental design of PDGF exposure to the SSc-isolated dermal fibroblasts. (B) Box-and-whisker plots of enrichment scores for Hallmark PI3K/AKT/mTOR pathway. Comparison to baseline shows a significant increase in GSVA score for most time points, with the largest differences after 2 and 8 hours of PDGF exposure (Wilcoxon’s rank-sum test, P value = 0.0091). The center line of the box plot indicates the median value; the box spans the first and third quartiles; whiskers show the range, excluding outliers. (C) GO term analysis of gene sets shows the molecular functions overrepresented in fibroblast lines 8 hours after PDGF treatment versus the baseline. Sorted by adjusted P value (GSEA, permutation test). (D) Experimental design of dermal fibroblasts treated with wortmannin. (E) Volcano plot with differentially expressed genes comparing 3,000 nM wortmannin treatment with DMSO-treated dermal fibroblasts. The most downregulated and upregulated genes by LogFC and adjusted P value are listed (Wald’s test). (F) Experimental design of 3D skin-like tissues treated with nintedanib from GEO GSE289407. (G) Volcano plot with differentially expressed genes comparing all nintedanib-treated tissues with control (vehicle-treated). The top most downregulated and upregulated genes by LogFC and adjusted P value are listed (Wald’s test). (H) Reactome gene sets enriched compared by sample of origin and 3D skin-like tissue treatment. Only gene sets with significant differences by ANOVA are displayed.