Microenvironmental network of clonal CXCL13+CD4+ T cells and Tregs in pemphigus chronic blisters

Abstract

BACKGROUNDPemphigus, a rare autoimmune bullous disease mediated by antidesmoglein autoantibodies, can be controlled with systemic medication like rituximab and high-dose systemic corticosteroids combined with immunosuppressants. However, some patients continue to experience chronically recurrent blisters in a specific area and require long-term maintenance systemic therapy.METHODSSkin with chronic blisters was obtained from patients with pemphigus. Immunologic properties of the skin were analyzed by immunofluorescence staining, bulk and single-cell RNA and TCR sequencing, and a highly multiplex imaging technique known as CO-Detection by indEXing (CODEX). Functional analyses were performed by flow cytometry and bulk RNA-Seq using peripheral blood from healthy donors. Intralesional corticosteroid was injected into patient skin, and changes in chronically recurrent blisters were observed.RESULTSWe demonstrated the presence of skin tertiary lymphoid structures (TLSs) with desmoglein-specific B cells in chronic blisters from patients with pemphigus. In the skin TLSs, CD4+ T cells predominantly produced CXCL13. These clonally expanded CXCL13+CD4+ T cells exhibited features of activated Th1-like cells and downregulated genes associated with T cell receptor-mediated signaling. Tregs are in direct contact with CXCL13+CD4+ memory T cells and increased CXCL13 production of CD4+ T cells through IL-2 consumption and TGF-β stimulation. Finally, intralesional corticosteroid injection improved chronic blisters and reduced skin TLSs in patients with pemphigus.CONCLUSIONThrough this study we conclude that skin TLSs are associated with the persistence of chronically recurrent blisters in patients with pemphigus, and the microenvironmental network involving CXCL13+CD4+ T cells and Tregs within these structures plays an important role in CXCL13 production.TRIAL REGISTRATIONClinicalTrials.gov NCT04509570.FUNDINGThis work was supported by National Research Foundation of South Korea (NRF-2021R1C1C1007179) and Korea Drug Development Fund, which is funded by Ministry of Science and ICT; Ministry of Trade, Industry, and Energy; and Ministry of Health and Welfare (grant RS-2022-00165917).

Keywords: Autoimmune diseases; Autoimmunity; Dermatology; Skin; T cells.

Figure 1. Skin tertiary lymphoid structures in chronic blisters of patients with pemphigus.

(A) Skin biopsy sample showing tertiary lymphoid structures (TLSs) from a patient with PV (Supplemental Figure 1A, no. 3) stained with H&E and specific antibodies against CD20, CD4, CD138, and PNAd. Scale bar: 100 μm. (B) Schematic of the experiment. (C) Duration of skin blisters in TLS-negative (n = 8) and TLS-positive (n = 23) lesions. Student’s t tests were used to compare means for 2 groups. Data are shown as the mean ± SD. **P < 0.005. (D) Representative immunofluorescence staining for DSG-specific B cells (triangles) and plasma cells (arrows). Tissues were stained with CD138 (red), CD20 (blue), rDSG1 or rDSG3 (green), and DAPI (light gray). Dotted lines indicate the margin of TLSs. Scale bar: 50 μm. PF, pemphigus foliaceus; PV, pemphigus vulgaris; rDSG1, recombinant desmoglein 1; rDSG3, recombinant desmoglein 3.

Figure 2. CD4⁺ T cells are the major producer of CXCL13 in skin TLSs in pemphigus.

(A) Bulk RNA-Seq of tertiary lymphoid structure–positive (TLS-positive) and –negative samples (n = 5 each). (B) Gene ontology (GO) analysis and (C) volcano plot depicting upregulated (red dots) and downregulated DEGs (blue dots). (D) Representative immunofluorescence staining for CXCL13 (green) and CD20 (red) in skin TLSs from a patient with pemphigus. Nuclei were stained with DAPI (light gray). Scale bar: 100 μm. (E) Representative immunofluorescence staining for coexpression of CD4 (red) and CXCL13 (green) in skin TLSs. White arrowheads indicate CXCL13⁺CD4⁺ cells. Nuclei were stained with DAPI (light gray). Scale bar: 50 μm. (F) Percentage of CD4⁺, CD8⁺, CD20⁺, CD138⁺, FDC⁺, and HLA-DR⁺ cells in CXCL13⁺ cells from immunofluorescence images (n = 5). Data are shown as the mean ± SD. (G) Gene set enrichment analysis of Th1, Th2, Th17, and Tfh cell gene signatures using the transcriptome of TLS-positive versus TLS-negative samples. (H) Percentage of top 10 most frequently occurring clones of TCRβ chain from bulk TCR-Seq in skin TLSs from 5 patients (no. 1–5).

Figure 3. CXCL13⁺CD4⁺ T cells are clonally expanded and activated with Th1-like features.

(A) Schematic of scRNA-Seq and scTCR-Seq for skin lesions with tertiary lymphoid structures (TLSs) in patients with pemphigus (n = 4). (B) UMAP visualization of 2,141 TCRβ⁺ T cells. (C) Violin plot showing the expression of the indicated marker genes in T cell subsets by scRNA-Seq. CD4, CD4⁺ T cells; CD8, CD8⁺ T cells. (D) Heatmap of the Morisita-Horn index quantifying overlapping TCRs among clusters. (E) The Shannon entropy calculation of the diversity of the TCR repertoire in each cluster. One-way ANOVA and Student’s t tests were used to compare means for 2 groups. **P < 0.005; ***P < 0.0001. Data are shown as the mean ± SD. (F) UMAP visualization of shared TCRs between T cells in skin TLSs and DSG3-specific, activation-induced marker (AIM)⁺ Tfh and non-Tfh memory CD4⁺ T cells in PBMCs from a patient with pemphigus vulgaris. (G) Volcano plot showing upregulated (orange dots) and downregulated (purple dots) in cluster 5 DEGs compared with cluster 1. (H) UMAP visualization showing gene signature of tissue-resident memory T (Trm) cells. (I) Dot plot showing expression of genes in the categories of costimulatory or coinhibitory receptors and glycolysis in each cluster. (J) Linear regression analyses of the expression of CXCL13 and the correlation with expression of TNFRSF18, TPI1, PGAM1, LAG3, and TIGIT in clusters 1 and 5. Pearson’s correlation analysis was used to measure the strength of relationships between variables. (K) Gene ontology analysis in cluster 5 compared with cluster 1. (L) Dot plot depicting the TCR-mediated gene set in each cluster. (M) Linear regression analysis of the correlation between CXCL13 and LCK in clusters 1 and 5. Pearson’s correlation analysis was used to measure the strength of relationships between variables.

Figure 4. Tregs are adjacent to CXCL13⁺CD4⁺ T cells.

(A) Workflow for CODEX imaging and analysis of skin tertiary lymphoid structures (TLSs) in patients with pemphigus (n = 9). (B) Six representative markers (left) for FoxP3 (red), CD4 (magenta), CD20 (cyan), HLA-DR (green), CXCL13 (white), and CD8 (blue) and a representative Voronoi diagram (right) of the TLSs after cell mapping. (C) Frequency of PD-1⁺ and ICOS⁺ cells in CXCL13⁺ versus CXCL13^– CD4⁺ Tm cells. Paired t tests were used to compare values for 2-variable plots. *P < 0.05. (D) Densities of Tregs, HLA-DR⁺ cells, CD8⁺ T cells, and B cells based on their distance from the center of CXCL13⁺ versus CXCL13^– CD4⁺ Tm cells. Wilcoxon matched-pairs signed-rank test. *P < 0.05. (E) Frequencies of Tregs, HLA-DR⁺ cells, CD8⁺ T cells, and B cells adjacent to CXCL13⁺ versus CXCL13^– CD4⁺ Tm cells in TLSs. Paired t tests were used to compare values for 2-variable plots. *P < 0.05. (F) Representative figures highlighting CXCL13⁺CD4⁺ Tm cells (red) and Tregs (yellow) in the Voronoi diagram. (G) Frequencies of marker-positive cells in Tregs adjacent to CXCL13⁺ versus CXCL13^– CD4⁺ Tm cells. Data are shown as the mean ± SD.

Figure 5. The production of CXCL13 in CD4⁺ T cells is increased by Tregs.

(A–F) CXCL13⁺CD4⁺ T cells were differentiated in with and without Treg conditions in vitro. Conventional T cells were stained with CellTrace Far Red (CTFR), and sorted CD25⁺CD127^loCD4⁺ Tregs were stained with Cell Trace Violet (CTV). (A) Representative plots and (B) graph of the relative frequencies of CXCL13⁺ cells in CTFR⁺CD4⁺ and CTFR⁺GITR⁺CD4⁺ T cells (n = 8). Paired t tests were used to compare values for 2-variable plots. ***P < 0.0001. (C) Frequencies of CTFR⁺ conventional T cells and CTV⁺ Tregs in CXCL13⁺GITR⁺CD4⁺ T cells. (D) Gene ontology analysis using downregulated DEGs and (E) gene set enrichment analysis of IL-2 pathway gene signatures from the bulk RNA-Seq of CD4⁺ T cells in the Treg-undepleted condition compared with the Treg-depleted condition (n = 5). (F) Dot plot showing expression of genes involved in the IL-2 pathway, TGF-β pathway, and IL-10 pathway in each cluster, as assessed by scRNA-Seq. (G) CXCL13⁺CD4⁺ T cells were differentiated in the presence or absence of neutralizing anti–IL-2 antibody and/or TGF-β. Relative frequencies of CXCL13⁺ cells in GITR⁺CD4⁺ T cells (n = 5). Paired t tests were used to compare values for 2-variable plots. *P < 0.05. (H) Differentiated CXCL13⁺CD4⁺ T cells were cocultured with or without induced Tregs the in presence or absence recombinant IL-2 protein and TGF-β–blocking antibody. Relative frequencies of CXCL13⁺ cells in GITR⁺CD25^–/loCD4⁺ T cells (n = 8). Paired t tests were used to compare values for 2-variable plots. **P < 0.005.

Figure 6. Intralesional corticosteroid injection ameliorates chronic blisters with skin TLSs in patients with pemphigus.

(A) Change in skin lesion areas during Intralesional corticosteroid injection (ILI) in 18 patients. (B) Representative clinical images and immunofluorescence staining of a tertiary lymphoid structure–positive (TLS-positive) chronic lesion in patient 1 before and after ILI. Tissues were costained with CD20 (red), CD138 (green), and DAPI (light gray). Scale bar: 100 μm. (C and D) Bulk RNA-Seq of paired skin lesions with TLSs for comparison of before and after ILI (n = 3). (C) Volcano plot shows upregulated (red dots) and downregulated DEGs (blue dots). (D) Bulk TCR-Seq of paired skin lesions with TLSs for the comparison of before and after ILI (patient #1 to #3). The change in the proportion of top 10 clones of the TCRβ chains is shown for each patient.