Single-cell profiling reveals pathogenic role and differentiation trajectory of granzyme K+CD8+ T cells in primary Sjögren's syndrome

Abstract

Primary Sjögren's syndrome (pSS) is a systemic autoimmune inflammatory disease mainly defined by T cell-dominated destruction of exocrine glands. Currently, CD8+ T cells are thought to be involved in the pathogenesis of pSS. However, the single-cell immune profiling of pSS and molecular signatures of pathogenic CD8+ T cells have not been well elucidated. Our multiomics investigation showed that both T cells and B cells, especially CD8+ T cells, were undergoing significant clonal expansion in pSS patients. TCR clonality analysis revealed that peripheral blood granzyme K+ (GZMK+) CXCR6+CD8+ T cells had higher a proportion of clones shared with CD69+CD103-CD8+ tissue-resident memory T (Trm) cells in labial glands in pSS. CD69+CD103-CD8+ Trm cells featured by high expression of GZMK were more active and cytotoxic in pSS compared with their CD103+ counterparts. Peripheral blood GZMK+CXCR6+CD8+ T cells with higher CD122 expression were increased and harbored a gene signature similar to Trm cells in pSS. Consistently, IL-15 was significantly elevated in pSS plasma and showed the capacity to promote differentiation of CD8+ T cells into GZMK+CXCR6+CD8+ T cells in a STAT5-dependent manner. In summary, we depicted the immune profile of pSS and further conducted comprehensive bioinformatics analysis and in vitro experimental investigations to characterize the pathogenic role and differentiation trajectory of CD8+ Trm cells in pSS.

Keywords: Autoimmune diseases; Autoimmunity; Immunology; Rheumatology; T cells.

Figure 1. Single-cell transcriptional signatures of LGs and PBMCs in pSS patients.

(A) Overview of the experimental approach. (B and C) The UMAP of total cells from LGs and PBMCs, colored by cluster. (D) Expression heatmap shows differentially expressed genes of each cluster. Selected genes are indicated to the left, and complete lists of top genes are available in Supplemental Table 5. (E) Visualization of LG flow cytometry data by t-SNE. The experiment was performed 3 times.

Figure 2. Clonal expansion of CD8⁺ T cells in pSS patients.

(A) Clone size of PBMCs and LG T cells from HCs and pSS patients. (B) Clonal expansion percentage (at least 2 or more cells were included in each clonotype) of PBMCs and LG memory T cells in HCs and pSS patients. (C) Bar graph shows clonal size of CD4⁺ or CD8⁺ T cells in pSS patients. (D) Clonal expansion percentages of all CD4⁺ or CD8⁺ memory T cells from HCs and pSS patients. (E) UMAP plot shows the clonal distribution of CD4⁺ T cells and CD8⁺ T cells of pSS patients. (F) The Chao and ACE index scores for CD8⁺ T cells in HCs and pSS patients. (G) Morisita index of shared clones in CD8⁺ T cells of PBMCs and LGs from the same patient. (H) Alluvial plot shows the shared clones in CD8⁺ T cells of LGs and PBMCs from pSS patients. In the box-and-whisker plots in B, D, and F, each data point represents 1 individual, horizontal lines indicate medians, bounds of the boxes represent IQRs, and whiskers extend to the farthest data points. **P < 0.01; ***P < 0.001 by 1-way ANOVA with Dunnett’s multiple-comparison test (B and D).

Figure 3. CD103^–CD8⁺ Trm cells are more cytotoxic than CD103⁺CD8⁺ Trm cells.

(A) Violin plots displaying Trm_Up or Trm_Down scores of each T cell cluster. PB_CD4T_M includes CD4⁺ Tcm, GZMB⁺CD4⁺ Tm, and GZMK⁺CD4⁺ Tm. PB_CD8T_M includes GZMB⁺CD8⁺ Tm and GZMK⁺CD8⁺ Tm. CD103n_CD8T refers to CD103^–CD8⁺ T cells. CD103p_CD8T refers to CD103⁺CD8⁺ T cells. (B) CellChat-derived the type 2 IFN signaling network among major cell types in LGs from pSS patients. Top: The shade level of the bar represents the scaled expression of genes, and the circle size represents the percentage of cells expressing the genes. Bottom: CellChat scoring according to IFNG and IFN-γ receptor expression level. (C) The GSEA plot shows the enrichment score of CD8⁺ cytotoxic and IFN-γ signaling pathways between CD103^–CD8⁺ Trm and CD103⁺CD8⁺ Trm cells. P values are shown on the right of the plot. (D) Flow cytometry results (left) and statistical analysis (right) of HLA-DR expression level of CD8⁺ Trm subpopulations in LGs of pSS patients (n = 5). The blue dots represent CD103⁺CD8⁺ Trm and red CD103^–CD8⁺ Trm cells. (E) Statistical analysis compares the difference in expression level of HLA-DR in CD103^–CD8⁺ Trm cells between pSS patients (n = 5, red bar) and HCs (n = 5, blue bar). (F) Volcano plot shows the differentially expressed genes between CD103^–CD8⁺ Trm and CD103⁺CD8⁺ Trm cells. (G) Flow cytometry result (left) and statistical analysis (right) of the expression level of GZMK in different LG Trm cells of pSS patients. (H) Statistical analysis comparing the difference in expression level of GZMK in different LG Trm cells of pSS patients. Data are presented as mean ± SD. *P < 0.05; **P < 0.01; ***P < 0.001 by 2-tailed, paired Student’s t test (D), 2-tailed, unpaired Student’s t test (E), or 1-way ANOVA with Dunnett’s multiple-comparison test (G and H).

Figure 4. Pathogenic role of GZMK⁺CD8⁺ T cells in pSS.

(A) The GZMK gene expression level in LGs of pSS patients and non-SS patients with sicca. (B and C) Correlation between the expression level of GZMK with focus score and the percentage area fraction of CD45⁺ cells. For A–C, the TPM value of GZMK was transformed from GEO GSE173808 microarray count data. (D) Representative multiplex staining of CD8 (green), CD103 (yellow), and GZMK (red) in LGs from pSS patient. Scale bar: 30 μm. The experiment was performed 2 times. (E) Correlation between GZMK⁺CD103^–CD8⁺ T cells and erythrocyte sedimentation rate (ESR) in D (n = 20). (F) Correlation between GZMK⁺CD103^–CD8⁺ T cells and the level of serum IgA in D (n = 20). (G and H) Correlation between the number or frequency of GZMK⁺CD103^–CD8⁺ T cells and focus score of LGs from pSS patients in D (n = 20). Data are presented as mean ± SD. ***P < 0.001 by 2-tailed, unpaired Student’s t test (A). P values in B, C, and E–H were derived using Pearson’s correlation.

Figure 5. GZMK⁺CXCR6⁺CD8⁺ T cells are Trm precursors.

(A) UMAP plot shows cluster distribution of CD8⁺ T cells in LGs and PBMCs from pSS patients. (B) UMAP plot shows the tissue distribution of CD8⁺ T cells. (C) Trm score of the PBMC CD8⁺ T cell cluster and LG Trm cells. LG_CD8T_RM includes CD103⁺CD8⁺ Trm and CD103^–CD8⁺ Trm cells. (D) Proportion of clone types shared with CD103⁺CD8⁺ Trm or CD103^–CD8⁺ Trm cells. (E) The distribution of clones shared with CD103^–CD8⁺ Trm cells in GZMK_CD8T_M. (F) Monocle3-derived pseudotime and trajectory line of CD8⁺ T cells on UMAP plot. (G) Volcano plot shows the differentially expressed genes in cells with shared clones of GZMK_CD8T_M. (H) Frequency of CXCR6⁺GZMK⁺CD8⁺ T cells in PBMC CD8⁺ T cells from pSS patients (n = 22) and HCs (n = 9). Blue bar represents cells from HCs and red pSS patients. (I) The geometric mean fluorescence intensity of HLA-DR in CXCR6⁺GZMK⁺CD8⁺ T cells, CXCR6^–GZMK⁺CD8⁺ T cells, GZMB⁺CD8⁺ T cells, and GZMB^–GZMK^–CD8⁺ T cells in PBMCs from pSS patients (n = 15). Red dots represent GZMK⁺CXCR6⁺CD8⁺ T cells and blue GZMB⁺CD8⁺ T cells. (J) Representative flow cytometry plots show the EOMES expression of CXCR6⁺GZMK⁺CD8⁺ T cells and GZMK⁺CD103^–CD8⁺ Trm cells. The experiment was performed 3 times. Data are presented as mean ± SD. **P < 0.01; ***P < 0.001 by 2-tailed, unpaired Student’s t test (H) or 1-way ANOVA with Dunnett’s multiple-comparison test (I).

Figure 6. IL-15 primes CD8⁺ T cells to differentiate into Trm precursors in pSS.

(A) The geometric mean fluorescence intensity of CD122 in CXCR6⁺GZMK⁺CD8⁺ T cells, GZMB⁺CD8⁺ T cells, and GZMB^–GZMK^–CD8⁺ T cells in PBMCs from pSS patients (n = 10). Red dots represent GZMK⁺CXCR6⁺CD8⁺ T cells, blue GZMB⁺CD8⁺ T cells, and green GZMB^–GZMK^–CD8⁺ T cells. (B) IL-15 concentration in blood plasma from pSS patients (n = 19) and HCs (n = 15). Blue bar represents cells from HCs and red pSS patients. (C) Correlation between frequency of CXCR6⁺GZMK⁺CD8⁺ T cells and blood plasma IL-15 concentration in pSS patients with paired PBMCs and blood plasma data. (D) Representative flow cytometry plots show the CD45RA and CCR7 expression of CXCR6⁺GZMK⁺CD8⁺ T cells after cytokine treatment. (E) Frequency of CXCR6⁺GZMK⁺ cells among CD8⁺ T cells after treating with IL-2, IL-15, and STAT5 inhibitor (STAT5-IN-1) for 48 hours. (F) Flow cytometry result (left) and statistical analysis (right) shows the frequency of CXCR6⁺GZMK⁺ cells among CD8⁺ Tcm cells after treating with IL-2, IL-15, and STAT5 inhibitor (STAT5-IN-1) for 48 hours. (G) Representative multiplex immunohistochemical staining of CD8 (green) and Ki67 (red) in LGs of pSS patients. Scale bar: 30 μm. The experiment was performed 3 times. (H) Violin plot showing the gene CXCL16 expression level in the main clusters of LGs from pSS patients. (I) Representative images of CXCL16 RNAscope in situ hybridization (red) and CD8 staining (green). The experiment was performed 2 times. Scale bar: 50 μm. Data are presented as mean ± SD. *P < 0.05; ***P < 0.001 by 1-way ANOVA with Dunnett’s multiple-comparison test (A, E, and F), 2-tailed, unpaired Student’s t test (B), or Pearson’s correlation (C).