Rapamycin improves Graves' orbitopathy by suppressing CD4+ cytotoxic T lymphocytes

Abstract

CD4+ cytotoxic T lymphocytes (CTLs) were recently implicated in immune-mediated inflammation and fibrosis progression of Graves' orbitopathy (GO). However, little is known about therapeutic targeting of CD4+ CTLs. Herein, we studied the effect of rapamycin, an approved mTOR complex 1 (mTORC1) inhibitor, in a GO mouse model, in vitro, and in patients with refractory GO. In the adenovirus-induced model, rapamycin significantly decreased the incidence of GO. This was accompanied by the reduction of both CD4+ CTLs and the reduction of orbital inflammation, adipogenesis, and fibrosis. CD4+ CTLs from patients with active GO showed upregulation of the mTOR pathway, while rapamycin decreased their proportions and cytotoxic function. Low-dose rapamycin treatment substantially improved diplopia and the clinical activity score in steroid-refractory patients with GO. Single-cell RNA-Seq revealed that eye motility improvement was closely related to suppression of inflammation and chemotaxis in CD4+ CTLs. In conclusion, rapamycin is a promising treatment for CD4+ CTL-mediated inflammation and fibrosis in GO.

Keywords: Autoimmune diseases; Autoimmunity; Endocrinology; Immunotherapy; T cells.

Figure 1. Rapamycin significantly decreases CD4⁺ CTL accumulation and suppresses their function in GO mice.

(A) Representative multicolor immunofluorescence image showing CD4⁺GZMB⁺ T cells (indicated by red arrows) accumulate in the vicinity of CD90⁺ fibroblasts (indicated by green arrows) in orbital tissue from patients with GO. (B) Represe ntative multicolor immunofluorescence image showing the CD90⁺ fibroblasts with GZMB visible within the cytosol, suggesting that fibroblasts may represent targets of CD4⁺ CTL–directed cytotoxicity. (C) Multicolor immunofluorescence images of cCasp-3⁺ and HLA-DR⁺CD90⁺ fibroblasts (indicated by arrows) in orbital tissue from patients with GO. (D) Animal study design. Control mice were immunized with Ad-EGFP, GO mice were immunized with Ad-TSHRA, and rapamycin intervention mice were immunized with Ad-TSHRA while fed a diet containing rapamycin from week 11 until the end of the experiment (14 ppm, 23 weeks). (E) Representative example of thyroid and orbital tissue from the 3 groups of mice by fluorescent multiplex IHC showing coexpression of CD4 and GZMB (indicated by white arrows). (F) Bar plots exhibited counts of CD4⁺GZMB⁺ T cells in thyroid and orbital tissue from the 3 groups of mice (n = 6 in each group). Values represent the mean ± SEM. ***P < 0.001, by 2-tailed, unpaired Mann-Whitney-Wilcoxon rank test for F.

Figure 2. Rapamycin significantly ameliorates orbitopathy in GO mice.

(A) Representative microscopy images of orbital fibrosis, adipogenesis, and inflammation in mice (indicated by black arrows). ON, optic nerve; OB, orbital bones. (B and C) Scatterplots of fibrosis volume and adipogenesis area in orbits from the 3 groups of mice (n = 8 for each group). The dashed line shows the mean + 2 SDs of the value for control mice, which was considered the normal range. (D) The incidences of fibrosis, adipogenesis, and orbitopathy in GO mice and rapamycin intervention mice. The incidences of fibrosis and adipogenesis: the mean + 2 SDs of control mice was regarded as the normal range, and higher values indicated significant positivity. The incidence of orbitopathy: existing orbital pathogenesis including fibrosis and adipogenesis. (E) Bar plots exhibited counts of CD3⁺ T cells in orbital tissue from the 3 groups of mice (n = 6 for each group). Values represent the mean ± SEM. *P < 0.05, **P < 0.01, and NS P > 0.05, by 1-way ANOVA and post-ANOVA, pairwise, 2-group comparisons with Tukey’s method for B, C, and E.

Figure 3. Rapamycin significantly ameliorates hyperthyroidism in GO mice.

(A and B) Scatterplots of TRAb and TT4 levels of the 3 groups of mice (n = 8 for each group). The dashed line shows the mean + 2 SDs of the value for control mice, which was regarded as the normal range. (C) Percentage bar plots which indicated the distribution of thyroid morphologic types in the 3 groups of mice (n = 8 for each group). The thyroid morphology was divided into 3 types: normal, hyperplastic (hyper), and heterogeneous (hetero). The typical hyperplastic changes: hyperplasia of follicular cells, which were cuboidal or tall columnar cells and even led to papillary folds and protrusions in the follicular cavity. Some of the glands showed a heterogeneous morphology with a mixture of normal and hyperplastic areas. (D) Representative microscopy images for morphology of thyroid in the 3 groups of mice. The morphology of GO model mice exhibited typical hyperplastic changes. CD3⁺ cells in the thyroid are indicated by black arrows, indicating inflammation in the thyroid. (E) Bar plots exhibited infiltrated area of CD3⁺ T cells in the thyroid from the 3 groups of mice (n = 6 for each group). (F) The ratio of BW to weekly food intake of mice in the 3 groups during the experiment. The statistical differences between the GO model with control group and the rapamycin intervention group with control group were labeled at the 10th and 30th week, respectively. Values represent the mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, and NS P > 0.05, by 2-tailed, unpaired Mann-Whitney-Wilcoxon rank test for A and E and 1-way ANOVA and post-ANOVA, pairwise, 2-group comparisons with Tukey’s method for B and F.

Figure 4. The mTORC1 signaling pathway is upregulated in the orbital tissue, thyrocytes, and PBMCs of patients with GO.

(A) Orbits, thyroids, and PBMCs from patients with GO were all investigated to determine the key pathway involved in GO pathogenesis. (B) A Venn diagram of differentially expressed genes (DEGs) in the GO gene set identified in the GSE58331 data set (orbital tissues; GO vs. HC) and in the GSE9340 data set (thyrocytes; GO vs. GD). A total of 245 GO-specific genes were identified. (C) Top 10 enrichment signatures revealed by KEGG analysis of the 245 GO-specific genes. The mTOR signaling pathway was the top enriched pathway in GO. P values are indicated by color. The number of DEGs in enrichment signatures were indicated by circle size. (D) Representative images of Western blot, which analyzed the key signaling molecules involved in mTORC1 signaling pathways from PBMCs of HCs and patients with GO. (E) Bar plots for the relative intensities in the Western blot of mTOR, p-mTOR, S6K, and p-S6K in the PBMCs of healthy controls and patients with GO (n = 6 for each group). Values represent the mean ± SEM. *P < 0.05 and **P < 0.01, by 2-tailed, unpaired Mann-Whitney-Wilcoxon rank test and independent-sample 2-tailed t tests for E.

Figure 5. The inhibition of mTORC1 by rapamycin decreases the frequency and functions of CD4⁺ CTLs in PBMCs of patients with GO.

(A and B) Flow cytometric analysis of key signaling molecules in mTORC1 signaling pathways in CD4⁺GZMB⁺ cells and CD4⁺GZMB^– cells from patients with GO (n = 4). FACS plots and bar plots were presented for intracellular p-mTOR (A) and its downstream molecule p-S6K (B). CD4⁺GZMB⁺ cells expressed higher levels of p-mTOR and p-S6K than CD4⁺GZMB^– cells. (C) Representative images of Western blot, which analyzed the key signaling molecules involved in mTORC1 signaling pathways from PBMCs treated with or without rapamycin. The PBMCs were obtained from patients with GO and incubated with or without rapamycin (100 nM, 48 hours). (D) Bar plots for relative intensities in the Western blot of p-mTOR and p-S6K in PBMCs treated with or without rapamycin (n = 3). (E) Flow cytometric analysis of the proportions of CD4⁺ CTLs expressing GZMB in PBMCs treated with or without rapamycin (n = 3). FACS plots and bar plots are presented. (F) Flow cytometric analysis of CCL5 in CD4⁺GZMB⁺ cells treated with or without rapamycin (n = 3). FACS plots and bar plots are presented. (G and H) Flow cytometric analysis of the proportions of CD4⁺PRF1⁺ and CD4⁺IFNG⁺ cells in PBMCs treated with or without rapamycin (n = 3). FACS plots and bar plots are presented. Values represent the mean ± SEM. *P < 0.05 and **P < 0.01, by 2-tailed, paired samples t tests for A and B and 2-tailed, independent samples t tests for D–H;

Figure 6. Rapamycin improves orbitopathy and diplopia in patients with refractory GO.

(A) CAS before and after rapamycin treatment (n = 5). (B) EOMy restriction gradings and Gorman diplopia scale of individual patients with GO. (C) Representative MRI images on T1 weight and short tau inversion recovery (STIR) sequence of good responders before and after rapamycin treatment.

Figure 7. Single-cell sequencing of CD4⁺ T cells demonstrates rapamycin suppresses mTORC1 pathway and cytotoxic and inflammatory functions of CD4⁺ CTLs in patients with GO.

(A) The graph-based clustering and uniform manifold approximation and projection (UMAP) algorithm were applied in 23,317 CD4⁺ T cells from 2 patients before and after rapamycin treatment. Clusters denoted by the same color scheme were labeled with inferred cell types: CD4⁺ T naive-like cells (CCR7⁺SELL⁺), CD4⁺ T_CM cells (SELL⁺CD27⁺), CD4⁺ T_EM cells (LIMS1⁺GPR183⁺), CD4⁺ T_regs (FOXP3⁺IKZF2⁺), CD4⁺ Th1/Th17-polarized cells (CCR6⁺CXCR3⁺), and CD4⁺ CTLs (GZMA⁺CCL5⁺). The clusters (upper right) denoted by blue and red were labeled with treatment status (before or after rapamycin treatment). (B) GSEA enrichment plots for mTORC1 signaling gene set in the transcriptome of CD4⁺GZMB^hi versus CD4⁺GZMB^lo cells (upper), and in the transcriptome of clusters 2 and 12 versus cluster 0 (bottom), respectively. (C) Heatmap of gene expression related to cell chemotaxis, the inflammatory responses, and the mTORC1 signaling pathway in CD4⁺ CTLs before and after rapamycin treatment. The expression was indicated by color. (D) Trajectory analysis of CD4⁺ CTLs. Cells on the trajectories are aligned in the order of differentiation (the arrow shape) representing the gradual transition from initial state to cell fate state. The upper trajectory shows the cells colored by cluster. The trajectory below shows the cells colored by treatment status (before or after rapamycin treatment).