DMF ameliorated the symptoms and reversed transcriptional alterations of EAU

An EAU model was established using IRBP1 − 20, PTX, and CFA as previously reported [6]. After immunization, the EAU mice were daily received oral administration of DMF for 14 days. DMF administration efficiently ameliorated EAU symptoms and reduced the clinical and pathological scores (Fig. 1A and B). Additionally, to test the therapeutic effect of DMF on established EAU, we administered established EAU mice (day 10 after immunization) with DMF for 5 days. We found that the therapeutic effect still existed (Additional file 1: Fig. S1B). In the following experiments, we started DMF administration on the first day of modeling in the DMF group.

Fig. 1

scRNA-seq analysis of immune cells during EAU and post-DMF administration. A. Representative fundus images and clinical scores of normal, EAU, and DMF-treated EAU mice. Arrows indicate inflammatory exudation and vascular deformation. Each group contained six mice. The values represent the mean ± SD. Significance was determined using Kruskal Wallis Test. B. Representative fundus hematoxylin and eosin staining plots and pathological scores of normal, EAU, and DMF-treated EAU mice. Arrows indicate retinal folding and inflammatory cell infiltration. Scale bars, 20 mm. Each group contained six mice. The values represent the mean ± SD. Significance was determined using Kruskal Wallis Test. C. Scheme of the overall study design. IRBP1 − 20: interphotoreceptor retinoid-binding protein 1–20; PTX: pertussis toxin; CFA: complete Freund’s adjuvant. D. UMAP plots of immune cell clusters from the CDLNs of all the mouse groups. E. UMAP plots of canonical markers of immune cell clusters from all the mouse groups. F. Line charts showing the percentages of major immune cell types in normal, EAU, and DMF-treated EAU mice. G-H. Volcano plots showing upregulated and downregulated DEGs of all immune cell types in the DMF-treated EAU/EAU (G) and EAU/Normal (H) comparison groups. The red and blue dots indicate upregulated and downregulated DEGs, respectively. I. Venn diagrams showing the numbers of rescue DEGs of all immune cells. The overlapping parts indicate the numbers of downregulated rescue DEGs (top) and upregulated rescue DEGs (bottom). J. Bar plot showing GO terms enriched by downregulated or upregulated rescue DEGs of all immune cells

To identify the genomic variations associated with DMF administration, scRNA-seq was performed on isolated CDLN cells derived from normal mice, EAU mice, and EAU mice administered with DMF (Fig. 1C). After initial processing, including quality filtering, integration, and batch effect correction, the cells were divided into 8 distinct clusters. Uniform manifold approximation and projection (UMAP) was used for visualization (Fig. 1D). The cell clusters were annotated via their expression of marker genes (Fig. 1D and E, Additional file 1: Fig. S1C). 8 classical immune cell types, namely, T cells, B cells, natural killer cells (NKs), conventional dendritic cells (cDCs), monocytes (Monos), plasmacytoid dendritic cells (pDCs), macrophages (Macros), and neutrophils (Neus), were identified. We then explored the proportional alterations of immune cells post-DMF administration. The proportional changes during EAU were partly reversed by DMF treatment in several immune cell types, including T cells, B cells, monocytes, pDCs, and macrophages (Fig. 1F). To explore the overall effect of DMF on the transcriptome of EAU mice, we performed DEG analysis of all immune cells between the DMF and EAU groups as well as between the EAU and normal groups (Fig. 1G and H). DMF reduced the expression of genes related to AP-1 (Fos, Fosb, and Jun), antigen presentation (B2m), interferon signaling (Ifi27l2a and Isg15), lymphocyte differentiation (Pim1), and cell migration (Cxcr4) (Fig. 1G). These genes were included in the DEGs which were upregulated during EAU compared to the normal condition (Fig. 1H). Meanwhile, DMF treatment upregulated two genes involved in RNA metabolism, Rps28 and Rps29, which were downregulated during EAU (Fig. 1G and H). These results indicated that DMF partially reversed the EAU-induced transcriptional changes. To investigate the specific beneficial mechanisms of DMF on EAU, we further identified “rescue DEGs” to better elucidate the EAU-induced transcriptional alterations that were reversed following DMF treatment (Fig. 1I). The genes upregulated in EAU and downregulated in the DMF group were identified as downregulated rescue DEGs. Similarly, the genes downregulated in EAU and upregulated in the DMF group were identified as upregulated rescue DEGs (Fig. 1I). Subsequently, a GO analysis was performed to characterize the biological significance of these rescue genes. Pathways related to IL-17 signaling, antigen processing and presentation, IL-6 signaling, and Th17 cell differentiation were enriched by the downregulated rescue DEGs (Fig. 1J). Meanwhile, pathways related to RNA metabolism, G protein-coupled receptor signaling, and endocytosis were enriched by the upregulated rescue DEGs (Fig. 1J). These results indicated that DMF partly reversed EAU-induced transcriptional alterations. We also conducted DEG analysis between the DMF group and the normal group (Additional file 1: Fig. S1D). We observed that the expression of Fos, Ifi27l2a, Isg15, and Cxcr4 was lower in DMF group compared to normal group (Additional file 1: Fig. S1D). In addition, the expression of genes related to the function of macrophages and neutrophils (Ccl5, Ctsb, and Grn) was higher in the DMF group compared to the normal group. GO analysis was conducted to annotate these DEGs. The downregulated DEGs in the DMF group compared to the normal group were enriched in pathways related to electron transport chain and ATP process, while the upregulated DEGs were enriched in pathways related to neutrophil degranulation and innate immune response (Additional file 1: Fig. S1E).

Overall, DMF treatment efficiently ameliorated EAU symptoms and partly reversed EAU-induced transcriptional alterations.

DMF partly reversed the transcriptional alterations among the major immune cells

Next, we examined the influence of DMF on transcriptomic profiles of the major immune cell types. DEG analysis was conducted, and rescue DEGs were identified for each major immune cell type (Fig. 2A-C). Most rescue DEGs were found in T cells, B cells, and monocytes (Fig. 2C). Subsequently, GO analysis was performed. In T cells, their downregulated rescue DEGs were involved in pathways associated with IL-17 signaling, Th17 cell differentiation, TNF signaling, and interferon production, whereas the upregulated ones were involved in pathways associated with G protein-coupled receptor signaling and peptidyl-amino acid modification (Fig. 2D). For B cells, pathways associated with cellular responses to stress and stimuli were down-rescued, whereas pathways associated with the negative regulation of immune system processes, regulation of cytoskeleton organization, and actin filament organization were up-rescued (Fig. 2E). For NKs, cDCs, and monocytes, pathways associated with the cellular response to interferon and the innate immune response were down-rescued, whereas pathways associated with RNA processing and metabolism were up-rescued (Fig. 2F and G). Additionally, the pathway related to cytotoxicity and the pathway associated with inflammatory response was down-rescued in NKs and monocytes, respectively (Fig. 2G). The number of rescue DEGs of pDCs, macrophages, and neutrophils was limited; therefore, we performed GO analysis on the downregulated or upregulated DEGs in the DMF group compared to the EAU group. In pDCs and macrophages, pathways associated with cellular responses to stress and stimuli, and antigen processing and presentation were downregulated, whereas pathways associated with RNA processing were upregulated after DMF treatment (Additional file 1: Fig. S1F and G). In neutrophils, the downregulated DEGs post-DMF administration were associated with the cytokine signaling-related pathway. The upregulated DEGs of neutrophils were involved in pathways associated with ribonucleoside metabolic and biosynthetic processes (Additional file 1: Fig. S1F and G). These results indicated that DMF might exert its effects through downregulation of several autoimmune- and inflammation-related genes that were upregulated in EAU.

Fig. 2

scRNA-seq analysis of the rescue DEGs in the major immune cell types. A-B. Rose diagrams showing the numbers of upregulated and downregulated DEGs of the major immune cell types in the DMF-treated EAU/EAU (A) and EAU/Normal comparison groups (B). C. Bar plot showing the numbers of upregulated and downregulated rescue DEGs of the major immune cell types. D-E. Bar plot showing GO terms enriched by downregulated or upregulated rescue DEGs in T cells (D) and B cells (E). F-G. Dot plot showing GO terms enriched in downregulated (F) and upregulated (G) rescue DEGs in NKs, cDCs, and monocytes. H. UpSet plot showing the downregulated rescue DEGs in eight types of cells. Genes in both T and B cells are labeled at the top. I-J. Violin plots showing the expression of Pim1 (I) and Cxcr4 (J) in T and B cells from normal, EAU, and DMF-treated EAU mice

To further identify the critical mediators of effects of DMF, we generated an UpSet plot to visualize the common and specific downregulated rescue DEGs among the major immune cell types (Fig. 2H). We observed that 10 genes (Jun, Hsp90aa1, Bst2, Cxcr4, Pim1, Hsph1, Hspe1, Rpl23a, Klf6, and Dnaja1) were uniquely down-rescued in T and B cells (Fig. 2H). Among these genes, Hsp90aa1, Hsph1, Hspe1, and Dnaja1 encode molecular chaperones and cochaperones for protein folding [46, 47]. Rpl23a encodes a ribosomal protein. Studies on Jun, Bst2, and Klf6 have indicated the involvement of these molecules in a variety of tumors [48,49,50,51]. Pim1 has been reported to regulate the differentiation of T and B cells [52, 53]. Pim1 also promotes Cxcr4 cell surface expression, thus regulating cell migration [54]. Intriguingly, Pim1 and Cxcr4, two genes actively involved in lymphocyte differentiation and migration [52,53,54] were down-rescued in T and B cells by DMF (Fig. 2I and J). These results indicated that DMF might exert its effect on uveitis by regulating Pim1 and Cxcr4 expression.

DMF reversed the proportional and transcriptional alterations of T-cell subsets during EAU

Considering the critical role of T cells in EAU pathogenesis, we reclustered T cells and identified 8 cell clusters, including naïve CD4 + T (NCD4) cells, naïve CD8 + T (NCD8) cells, T follicular helper (Tfh) cells, Th1 cells, Th17 cells, Treg cells, cytotoxic T lymphocytes (Ctl), and proliferative T (Prot) cells (Fig. 3A and B, Additional file 1: Fig. S2A). CD4 + T cells are actively involved in EAU development. In our data, DMF counteracted the elevated frequencies of Th1 and Th17 cells during EAU, while concurrently promoting an expansion in the Treg cell population. In addition, the proportion of effector CD8 + T cells, namely Ctl cells, was not rescued by DMF treatment (Fig. 3C). Considering the altered proportions in CD4 + T cell subsets, we conducted trajectory analysis to explore the effect of DMF on CD4 + T cell differentiation (Fig. 3D). We observed an increased trend from NCD4 cells to Th1 and Th17 cells during EAU, which was reversed post-DMF administration (Fig. 3D). Meanwhile, the trend toward Treg cells also enhanced by DMF treatment (Fig. 3D). Subsequently, we assessed the expression levels of Pim1 and Cxcr4 within individual CD4 + T cell subpopulations. The expression of Pim1 was significantly upregulated during EAU in Th1 cells, Th17 cells, and Treg cells, and down-rescued by DMF (Fig. 3E). Considering the role of Pim1 in regulating T cell differentiation, DMF may reverse the Teff/Treg cell imbalance by regulating Pim1 during uveitis. Moreover, the expression of Cxcr4 was upregulated in EAU and repressed by DMF treatment in all CD4 + T-cell subpopulations (Fig. 3F).

Fig. 3

DMF partly reversed the proportional and transcriptional alterations in T-cell subsets during EAU. A. UMAP plots of T-cell subsets from the CDLNs of all the mouse groups. B. UMAP plots of canonical markers of T-cell subsets from all the mouse groups. C. Bar plots showing the percentages of T-cell subsets among CDLNs from normal, EAU, and DMF-treated EAU mice. D. Pseudotime trajectory analysis of CD4 + T cells. Cells are colored according to pseudotime (left) or cell type (right). E-F. Line plots showing the expression of Pim1 (E) and Cxcr4 (F) in each CD4 + T-cell subset. G. Rose diagram showing the numbers of upregulated and downregulated rescue DEGs in each T-cell subset. H. Dot plot showing GO terms enriched by downregulated rescue DEGs in NCD4, Tfh, Th1, Th17, Treg, Prot, Ctl, and NCD8 cells. I. Dot plot showing GO terms enriched by upregulated rescue DEGs in NCD4, Th1, Th17, Treg, and NCD8 cells

To investigate the transcriptional alterations within individual T-cell subpopulations, GO analysis was performed on the rescue DEGs within each T-cell subpopulation. Most rescue DEGs were detected in the NCD4 and NCD8 cells (Fig. 3G). Pathways associated with protein folding, and cellular responses to stress and stimuli were down-rescued in most T cell subpopulations (Fig. 3H). The pathway related to Th17 cell differentiation was down-rescued in NCD4 and Th17 cells (Fig. 3H). In addition, the pathways associated with TNF signaling and T-cell activation were down-rescued in NCD4 and NCD8 cells (Fig. 3H). For GO analysis of upregulated rescue genes, we observed that pathways related to RNA processing and metabolism were up-rescued in most T cell subsets and the pathway annotated as “peptidyl-amino acid modification” was up-rescued in NCD4 cells (Fig. 3I). The amount of upregulated rescue DEGs of Tfh and Prot cells was limited. Thus, we conducted GO analysis on their upregulated DEGs in the DMF group compared with the EAU group and identified enriched pathways related to peptidyl-amino acid modification and RNA processing (Additional file 1: Fig. S2B and C). Collectively, the altered expression of Pim1 and Cxcr4 in CD4 + T cell subpopulations during EAU and post-DMF administration indicated the potential involvement of these two molecules in the beneficial effects of DMF administration. In addition, the results of GO analysis demonstrated that DMF reversed the transcriptional alterations of T-cell subsets during EAU.

DMF reversed the Teff/Treg imbalance by inhibiting the PIM1-AKT-FOXO1 pathway

Augment Teff (Th1 and Th17) cells and insufficient Treg cells, namely, the Teff/Treg imbalance, are critical pathogenic factors in AU [55, 56]. Our above analysis indicated that DMF might regulate CD4 + T-cell differentiation by regulating PIM1. To verify our hypothesis, flow cytometry was performed on CDLN cells isolated from EAU group or DMF group. DMF treatment reduced PIM1 expression in CD4 + T cells during EAU (Fig. 4A). Additionally, the proportions of Th17 cells (CD4 + IL17A + T cells) and Th1 cells (CD4 + IFN-γ + T cells) were significantly decreased, whereas the proportion of Treg cells (CD4 + Foxp3 + T cells) was significantly enhanced in the DMF group compared to the EAU group (Fig. 4B-D). Therefore, DMF might reverse the imbalance of Teff/Treg by inhibition of PIM1 expression.

Fig. 4

DMF reduced PIM1 expression in CD4 + T cells and reversed the Teff/Treg imbalance. A-D. The proportions of PIM1 + cells (A), Th17 cells (B), Th1 cells (C), and Treg cells (D) among CD4 + T cells from CDLNs of EAU mice and DMF-treated EAU mice were measured via flow cytometry after immunization on day 14. Each group contained six mice. The data are expressed as the mean ± SD. Significance was determined using unpaired two-tailed Student’s t-test

To further identify the role of DMF in CD4 + T cell subsets, in vitro experiments were also performed. We isolated CDLN cells of EAU mice and cocultured these cells with IRBP1 − 20 or IRBP1 − 20 plus DMF. The treatment with DMF resulted in a reduction in the proportions of Th17 and Th1 cells, along with an augmentation in Treg cell proportion (Fig. 5A-C). Meanwhile, upon transferring these CD4 + T cells into naive mice, we noted that DMF treatment diminished the capacity of autoreactive CD4 + T cells to instigate EAU (Fig. 5D).

Fig. 5

DMF reversed the Teff/Treg imbalance by regulating PIM1-AKT-FOXO1. A-C. CDLN cells from EAU mice were cultured with IRBP1-20 or IRBP1-20 plus DMF. Flow cytometry was performed to determine the proportions of Th17 cells (A), Th1 cells (B), and Treg cells (C). Each group contained six mice. The data are expressed as the mean ± SD. Significance was determined using an unpaired two-tailed Student’s t-test. D. Representative fundus images and clinical scores of mice transferred with CD4 + T cells cultured with IRBP1-20 or IRBP1-20 plus DMF after 14 days of modeling. White arrowheads indicate inflammatory exudation and vascular deformation. Each group contained six mice. The data are expressed as the mean ± SD. Significance was determined using Mann-Whitney U test. AT: Adoptive transfer. E-G. CDLN cells from the EAU group were cultured with IRBP1-20 alone or with IRBP1-20 plus DMF for 72 h. Flow cytometry showing the proportions of PIM1 + cells (E), pAKT + cells (F), and pFOXO1 + cells (G) among the CD4 + T cells. The data are presented as the means ± SD from six independent experiments. Significance was determined using one-way ANOVA

Next, we delved deeper into elucidating the regulatory mechanisms by which DMF impacts the Teff/Treg cell balance. PIM1 is able to modulate CD4 + T-cell differentiation through phosphorylation of the AKT-FOXO1 pathway [53]. The transcriptional factor, FOXO1, promotes Treg differentiation via induction of FOXP3 expression but represses Th1 and Th17 differentiation via inhibition of T-bet and RORγ [57,58,59,60]. PIM1 enhances AKT phosphorylation and activates its kinase activity [53]. FOXO1 phosphorylated by AKT translocates from the nucleus to the cytosol, resulting in loss of its regulatory activity [57, 61]. Thus, we performed flow cytometry to determine the influence of DMF on the PIM1-AKT-FOXO1 pathway. Upon stimulation with IRBP1 − 20, the expression of PIM1 and the phosphorylation levels of AKT and FOXO1 were enhanced in CD4 + T cells. These alterations induced by IRBP1 − 20 were reversed by DMF treatment (Fig. 5E-G). Thus, DMF might reverse the Teff/Treg imbalance by regulating the PIM1-AKT-FOXO1 pathway.

In addition, we orally administered DMF to normal mice for 14 days to investigate the potential effects of DMF on normal mice. We observed that normal mice administered with DMF didn’t show any retina lesions in the funduscopic examinations and the eyeball pathological sections on day 14 after treatment initiation (Additional file 1: Fig. S3A and B). Furthermore, flow cytometry analysis showed no significant differences in the proportions of Th1, Th17, and Treg cells among CDLN cells between normal mice and those treated with DMF (Additional file 1: Fig. S3C-E). These results indicated that DMF administration had no significant effect on the proportion of Th1, Th17, and Treg cells in normal mice.

DMF reversed the proportional and transcriptional alterations in B-cell subsets during EAU

Currently, the role of B cells has received increasing attention in immune-related disorders. Next, we studied the effect of DMF on B cells. We reclustered B cells and identified 3 cell clusters, namely, naïve B cells (NBCs), germinal center B cells (GCs), and plasma cells (PCs), according to their expression of classical markers (Additional file 1: Fig. S4A- C). During EAU, there was an increase in the proportions of PCs and GCs, which were reversed by DMF treatment (Additional file 1: Fig. S4D). To identify the impact of DMF on the differentiation of B cells, we conducted trajectory analysis on B-cell subsets. The trend toward GCs and PCs was increased during EAU and reduced post-DMF administration (Additional file 1: Fig. S4E). Pim1 has been reported to support B-cell differentiation [53]. Next, we evaluated Pim1 expression in B-cell subsets. Among all B-cell subsets, the expression of Pim1 increased in EAU and was repressed by DMF administration (Additional file 1: Fig. S4F). Further, flow cytometry was conducted to validate the effect of DMF on B cells. DMF treatment reduced PIM1 expression and PC proportion among B cells (Additional file 1: Fig. S4G and H). Thus, DMF might reduce the proportion of PC by repressing PIM1 expression. We also evaluated Cxcr4 expression. DMF reversed the enhanced Cxcr4 expression in NBCs during EAU but did not reverse the altered Cxcr4 expression in GCs and PCs (Additional file 1: Fig. S4I).

Subsequently, the influence of DMF on the transcriptomes of B-cell subpopulations was explored. DEG analysis was conducted. The greatest number of rescued DEGs were detected in NBCs (Additional file 1: Fig. S4J). GO analysis of the downregulated rescue DEGs revealed that in NBCs and PCs, pathways associated with cellular responses to stress and stimuli and regulation of lymphocyte activation were down-rescued (Additional file 1: Fig. S4K). The amount of downregulated rescue DEGs of GCs was limited. Thus, we performed GO analysis on the downregulated DEGs of GCs in DMF group compared to EAU group and identified enriched pathways related to translation and peptide biosynthetic process (Additional file 1: Fig. S4L). Meanwhile, GO analysis of the upregulated rescue DEGs showed that the pathway related to RNA metabolism was up-rescued in both NBCs and GCs (Additional file 1: Fig. S4M). There were no upregulated rescue DEGs in PCs. GO analysis of the upregulated DEGs of PCs in DMF group compared to EAU group revealed enrichment in the pathway related to intracellular signaling by second messengers (Additional file 1: Fig. S4N). Collectively, DMF reversed the proportional and transcriptional alterations in B-cell subsets during EAU.

DMF decreased T cell ocular infiltration

Uveitis is characterized by the infiltration of T cells and other leukocytes into the eyes [12]. To determine the impact of DMF on ocular infiltrating immune cells and the underlying mechanism involved, we integrated scRNA-seq data from retina samples derived from EAU mice into our present study. Retinal cells were clustered and visualized via UAMP. Fifteen cell clusters were identified, namely, rod cells (RODs), cone cells (CONEs), macroglia (MAGs), cone bipolar cells (CBCs), T cells, monocytes and macrophages (Mono&Macro), rod bipolar cells (RBCs), cDCs, neutrophils, amacrine cells (ACs), microglia, pDCs, NK cells, retinal pigmented epithelial cells (RPEs), and vascular endothelial cells (VECs). Retinal monocytes and macrophages were mixed into a cluster, namely, Mono&Macro (Fig. 6A, Additional file 1: Fig. S5A and B). Besides retinal intrinsic cells, retinal cells in our data contained quite a few infiltrated immune cells (Fig. 6B). Among the retinal immune cells, T cells, Mono&Macro accounted for the greatest percentage of cells (Fig. 6C). We then conducted CellChat to infer the intercellular communication network among retinal cells during EAU (Additional file 1: Fig. S5C). Our above results indicated decreased expression of Cxcr4 in T cells post-DMF administration. Thus, we identified the CXCL signaling pathway network among retina cells to identify the role of Cxcr4 in ocular immune cell infiltration (Fig. 6D). VECs were the major senders of CXCL signaling, whereas neutrophils and T cells were the major receivers (Fig. 6E). Contribution prediction indicated that Cxcl12-Cxcr4 was the most significant ligand-receptor pair among CXCL signaling (Fig. 6F). VECs were the major source of Cxcl12 and acted on Cxcr4 expressed on microglia and infiltrated immune cells (Fig. 6G). Thus, these results indicated that the ligand-receptor pair Cxcl12-Cxcr4 actively engaged in immune cell infiltration during EAU, which might be reversed by DMF.

Fig. 6

scRNA-seq analysis of the retina from EAU mice. A. UMAP plot of retina cells from EAU mice. B-C. Pie charts showing the cell composition (B) and immune cell composition (C) of the retina. D. The inferred CXCL signaling network between different cell clusters. The edge width represents the communication probability. E. Heatmap of the CXCL signaling network displaying the relative importance of each cell group. F. Relative contribution of each ligand-receptor pair in the communication network of the CXCL signaling pathway. G. The inferred Cxcl12-Cxcr4 intercellular network between different cell clusters. The edge width represents the communication probability. H. UMAP plots of integrated immune cells from the retina of EAU mice and CDLNs of all the mouse groups. I. Violin plots showing the expression of Pim1 and Cxcr4 in T cells from the CDLNs of normal, EAU, and DMF-treated EAU mice as well as from the retina of EAU mice. J. UMAP plots of integrated T cells from the retina of EAU mice and CDLNs of all the mouse groups. K. Bar plots showing the percentages of T-cell subsets from CDLNs of normal, EAU, and DMF-treated EAU mice as well as from the retina of EAU mice. L. The proportions of Cxcr4 + cells among CD4 + T cells from CDLNs of EAU mice and DMF-treated EAU mice were measured via flow cytometry after immunization on day 14. Each group contained six mice. The data are expressed as the mean ± SD. Significance was determined using unpaired two-tailed Student’s t-test. M-N. The proportions of Th17 cells (M) and Th1 cells (N) among ocularly infiltrated CD4 + T cells were measured by flow cytometry after immunization on day 14. Each group contained six mice. The data are expressed as the mean ± SD. Significance was determined using unpaired two-tailed Student’s t test

Peripheral T cells infiltrating into the eyes is an important pathogenic factor for uveitis [12]. Our above data showed that DMF reduced the expression of Cxcr4 and Pim1. Cxcr4 was actively involved in cell migration [62]. Pim1 promotes the cell surface expression of Cxcr4 thus regulating CXCR4-dependent cell migration [54]. To explore the impact of DMF on ocular T cell infiltration during uveitis, we integrated retinal immune cells of EAU mice with CDLN cells from each group and observed that T cells in the retina of EAU mice expressed the highest level of Cxcr4 and Pim1 (Fig. 6H and I). In addition, we reclustered T cells from CDLNs and the retina, and found that the proportions of Th1 and Th17 cells in the retina during EAU were much higher than those in CDLNs of each group (Fig. 6J and K). Subsequently, flow cytometry was utilized to exam the function of DMF on Cxcr4 expression and ocular T-cell infiltration. DMF reduced CXCR4 expression in CD4 + T cells in CDLNs (Fig. 6L). The proportions of ocular infiltrating Th17 and Th1 cells were significantly lower in the DMF group than in the EAU group (Fig. 6M and N). Thus, DMF treatment could inhibit ocular infiltration of Teff cells in uveitis and this effect might depend on its inhibition of PIM1 and CXCR4 expression.