Therapeutic Effects of Upadacitinib on Experimental Autoimmune Uveitis: Insights From Single-Cell Analysis

Abstract

Purpose: This study aimed to elucidate the impact of upadacitinib, a Janus kinase 1 (JAK1)-specific inhibitor, on experimental autoimmune uveitis (EAU) and explore its underlying mechanisms.

Methods: We utilized single-cell RNA sequencing (scRNA-seq) and single-cell assay for transposase-accessible chromatin sequencing (scATAC-seq) to investigate the JAK/signal transducer and activator of transcription (STAT) pathway in peripheral blood mononuclear cells (PBMCs) of 12 patients with Vogt-Koyanagi-Harada (VKH) disease and cervical draining lymph node (CDLN) cells of EAU. After treating EAU with upadacitinib, we analyzed immune cell gene expression and cell-cell communication by integrating scRNA data. Additionally, we applied flow cytometry and western blot to analyze the CDLN cells.

Results: The JAK/STAT pathway was found to be upregulated in patients with VKH disease and EAU. Upadacitinib effectively alleviated EAU symptoms, reduced JAK1 protein expression, and suppressed pathogenic CD4 T cell (CD4TC) proliferation and pathogenicity while promoting Treg proliferation. The inhibition of pathogenic CD4TCs by upadacitinib was observed in both flow cytometry and scRNA data. Additionally, upadacitinib was found to rescue the interferon-stimulated gene 15 (ISG15)+ CD4TCs and CD8 T and B cell ratios and reduce expression of inflammatory-related genes. Upadacitinib demonstrated the ability to inhibit abnormally activated cell-cell communication, particularly the CXCR4-mediated migration pathway, which has been implicated in EAU pathogenesis. CXCR4 inhibitors showed promising therapeutic effects in EAU.

Conclusions: Our findings indicate that the JAK1-mediated signaling pathway is significantly upregulated in uveitis, and upadacitinib exhibits therapeutic efficacy against EAU. Furthermore, targeting the CXCR4-mediated migration pathway could be a promising therapeutic strategy.

Figure 1.

Identification of cell types and JAK-mediated inflammation during EAU and VKH. (A) Representative GO terms and pathways enriched in upregulated DEGs of CDLN cells in the EAU and control comparison groups. (B) Representative GO terms and pathways enriched in upregulated DEGs of immune subsets in the EAU and control comparison groups. (C) Volcano plot illustrates the DEGs of CDLN cells in the EAU and control comparison groups; upregulated genes are shown in red and downregulated genes are shown in blue. (D) A box plot displays the JAK/STAT signaling scores in immune subsets between the EAU and control groups. (E) Representative GO terms and pathways enriched in upregulated nearest genes of differentially accessible chromatin regions (DARs) of immune subsets in the VKH and HC comparison groups. (F) Genome browser tracks showing single-cell chromatin accessibility in the JAK1, JAK3, STAT4, and STAT1 loci in CD4TCs, CD8TCs, monocytes, and BCs, respectively.

Figure 2.

Inflammation induced by EAU rescued by UPA clinically and histologically. (A, B) Day 14 fundus images following immunization in EAU mice and UPA-treated mice are presented as a representative example (A), with white arrows highlighting the presence of inflammatory exudation and linear lesions. Clinical scores of the two groups (n = 5/group) are compared using column charts (B) (C, D) Day 14 H&E-stained images following immunization in EAU mice and UPA-treated mice are presented as representative example (C), with black arrows highlighting retinal folding. Pathological scores of the two groups (n = 5/group) are compared using column charts (D). Data are shown as mean ± SD from three independent experiments. Data were analyzed using unpaired two-tailed Student’s t-tests, ****P < 0.0001. (E, F) Elevated protein levels of JAK1 were detected in CDLN cells of EAU compared with normal mice but they decreased in UPA-treated mice (n = 5/group). Data are shown as mean ± SD of three independent experiments. Data were analyzed using two-way ANOVA, ****P < 0.0001. (G–L) At day 14 following immunization, the percentages of CD4⁺ cells expressing IFN-γ (G, H), IL-17A (I, J), and Foxp3 (K, L) in the CDLN cells were evaluated in both EAU mice and those treated with UPA (n = 5/group). (M, N) Day 14 fundus images following injection of IRBP_1–20-specific CD4⁺ T cells treated with or without UPA are presented as representative example (M), with white arrows highlighting the presence of inflammatory exudation and linear lesions. Clinical scores of the two groups (n = 5/group) are compared using column charts (N). (O, P) Day 14 H&E-stained images of the two groups are presented as representative example (O), with black arrows highlighting retinal folding. Pathological score of the two groups (n = 5/group) are compared using column charts (P). Data are shown as mean ± SD from three independent experiments. Data were analyzed using unpaired two-tailed Student’s t-tests, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 3.

The reconstitution by UPA of immunocytes configuration during EAU. (A) Experimental design for this study. (B) The t-distributed stochastic neighbor embedding (t-SNE) clustering of CDLN cells. (C) The cell ratios of different cell types were comparatively analyzed across the three groups, and the log2 fold change values of the relative changes in cell ratios are indicated on the right. (D–F) Line charts depict the proportions of ISG15⁺ CD4TCs (D), ISG15⁺ CD8TCs (E), and ISG15⁺ BCs (F) across three distinct groups derived from scRNA-seq data. (G) Representative GO terms and pathways enriched in downregulated DEGs of CDLN cells in the UPA and EAU comparison groups. P value was derived by a hypergeometric test.

Figure 4.

EAU-associated gene expression alteration in lymph nodes rescued by UPA. (A) Venn diagrams illustrate the number of DEGs for EAU, UPA, and rescue conditions. The intersecting regions indicate the number of upregulated rescue DEGs (top) and downregulated rescue DEGs (bottom). (B) Rose diagram displays the proportion of rescue DEGs compared to EAU DEGs. (C) UpSetR plots visualize the downregulated rescue DEGs across different cell types, with key genes labeled at the top. UpSet plots are an alternative to Venn diagrams used to deal with more than three sets. (D) Representative GO terms and pathways enriched in downregulated rescue DEGs in CDLN cells. P value was derived by a hypergeometric test. (E) Representative GO terms and pathways enriched in downregulated rescue DEGs in immune subsets. P value was derived by a hypergeometric test.

Figure 5.

The cell type–specific gene expression impacted by EAU and UPA. (A–C) Network plots depict the EAU (A), UPA (B), and rescue (C) DEGs in various cell types. Plots were generated using Cytoscape software. The internal node color corresponds to the cell type, and the gray edges represent the DEGs. The lines connecting the nodes indicate the attribution of DEGs to their respective cell types. (D–F) Bar plots display the count of rescue DEGs and EAU DEGs, as well as the rescue and EAU DEGs in each subtypes. The red line represents the average proportion of rescue DEGs and EAU DEGs. (G–H) Bar plots demonstrate the frequency of upregulated (G) and downregulated (H) rescue DEGs.

Figure 6.

Anomalous intercellular communication in EAU attenuated by UPA. (A) Number of inferred interactions among the three groups. (B) Interaction strength among the three groups. (C) The relative information flow of signaling pathway among the control, EAU, and UPA groups. (D) Chord diagrams display the upregulated signaling ligand–receptor pairs of CDLN cells in the EAU and control comparison groups. (E) Chord diagrams display the downregulated signaling ligand–receptor pairs of CDLN cells in the UPA and EAU comparison groups. (F) Chord diagrams display the upregulated signaling ligand–receptor pairs of immune subsets in the EAU and control comparison groups. (G) Chord diagrams display the downregulated signaling ligand–receptor pairs of immune subsets in the UPA and EAU comparison groups. (H) A violin plot is presented to display CXCR4 in immune subsets among the three groups.

Figure 7.

The key role of CXCR4 in therapy during EAU. (A, B) Day 14 fundus images following immunization in EAU mice and MSX-122–treated mice are presented as representative example (A), with white arrows highlighting the presence of inflammatory exudation and linear lesions. Clinical score of the two groups (n = 5/group) are compared using column charts (B). (C, D) Day 14 H&E-stained images following immunization in EAU mice and MSX-122–-treated mice are presented as representative example (C), with black arrows highlighting retinal folding. Pathological score of the two groups (n = 5/group) are compared using column charts (D). (E–H) Lymphocytes from CDLN cells of EAU mice stimulated by IRBP_1–20 with or without MSX-122 for 72 hours. Cells were measured by flow cytometry on the gate of CD4⁺ T cells. MSX-122 decreased the expression of IFN-γ (E, F) and IL-17A (G, H) (n = 5/group). Data are shown as mean ± SD from three independent experiments. Data were analyzed using unpaired two-tailed Student's t-tests, **P < 0.01, ****P < 0.0001.