Inhibition of JAK-STAT pathway corrects salivary gland inflammation and interferon driven immune activation in Sjögren's Disease

Abstract

Objectives: Inflammatory cytokines that signal through the JAK- STAT pathway, especially interferons (IFNs), are implicated in Sjögren's Disease (SjD). Although inhibition of JAKs is effective in other autoimmune diseases, a systematic investigation of IFN-JAK-STAT signaling and effect of JAK inhibitor (JAKi) therapy in SjD-affected human tissues has not been reported.

Methods: Human minor salivary glands (MSGs) and peripheral blood mononuclear cells (PBMCs) were investigated using bulk or single cell (sc) RNA sequencing (RNAseq), immunofluorescence microscopy (IF), and flow cytometry. Ex vivo culture assays on PBMCs and primary salivary gland epithelial cell (pSGEC) lines were performed to model changes in target tissues before and after JAKi.

Results: RNAseq and IF showed activated JAK-STAT pathway in SjD MSGs. Elevated IFN-stimulated gene (ISGs) expression associated with clinical variables (e.g., focus scores, anti-SSA positivity). scRNAseq of MSGs exhibited cell-type specific upregulation of JAK-STAT and ISGs; PBMCs showed similar trends, including markedly upregulated ISGs in monocytes. Ex vivo studies showed elevated basal pSTAT levels in SjD MSGs and PBMCs that were corrected with JAKi. SjD-derived pSGECs exhibited higher basal ISG expressions and exaggerated responses to IFNβ, which were normalized by JAKi without cytotoxicity.

Conclusions: SjD patients' tissues exhibit increased expression of ISGs and activation of the JAK-STAT pathway in a cell type-dependent manner. JAKi normalizes this aberrant signaling at the tissue level and in PBMCs, suggesting a putative viable therapy for SjD, targeting both glandular and extraglandular symptoms. Predicated on these data, a Phase Ib/IIa randomized controlled trial to treat SjD with tofacitinib was initiated.

Keywords: Interferon; Janus kinases; STAT; Sjögren’s Disease; Tofacitinib.

Figure 1:. Bulk sequencing of minor salivary gland and IFN signature

(A) Overview of MSG biopsy and whole transcriptomic analysis using RNAseq from 22 SjD and 11 HV. (B) Heatmap illustrating the top 250 DEGs in MSG between SjD and HV highlighting multiple ISGs, cytokines, and interleukins, in the DEG. Immune pathway genes, including ISGs (e.g., IFI44L, IFI44, MX1, CXCL13), were upregulated whereas canonical salivary genes (e.g., BPIFB2, PIP, ZG16B) were down regulated in SjD. (C) Similarly, a volcano plot showing the DEGs between SjD and HV, in which some representative genes were highlighted. MSGs from SjD subjects were transcriptionally more active with 2773 upregulated versus 468 downregulated genes. (D) Pathway enrichment analysis identified JAK-STAT Pathway as one of the top three differentially utilized pathway amongst the 25 significantly enriched pathways in SjD at a p-adj<0.01 (E, G) Calculated Type-I and Type-II IFN scores revealed 2 through 2.5-fold mean increases in SjD MSGs compared to HV. Differences in mean values were compared using the Mann-Whitney U-test at a p < 0.05 deemed significant. (F, H) The activated IFN signature noted in our bulk RNAseq positively correlated with FS in the glands. Spearman correlation analysis was used to assess the significance between correlated values at a p <0.05. MSG, minor salivary gland; RNAseq, RNA sequencing; SjD, Sjogren’s Disease; HV, healthy volunteer; DEGs, differentially expressed genes; ISG, interferon stimulated genes; FS, focus score.

Figure 2:. Immunofluorescence microscopy

(A) Overview of MSG biopsy, image acquisition and flow cytometry. (B) MSG IF image showing upregulated expression of JAK1 and JAK3 in SjD epithelial and infiltrating immune cells. JAK3 was especially enrich in ductal cells. Cellular proportion changes in the SjD MSGs showed greater numbers of immune cells, less numbers of epithelial cells. (C) Mean fluorescence of KRT18, JAK1 and JAK3 in SjD and non-SjD showing cellular proportion in the MSGs. Expression of JAK1 localized to immune cells (+6%; p<0.0001, Mann-Whitney Test) but JAK3 expression was seen in both epithelial and infiltrating immune cells (35% and 15%, respectively; p<0.0001, Mann-Whitney Test). (D) Cellular population change in 7 SjD and 6 non-SjD was characterized by flow cytometry showed reduced numbers of epithelial cells and increased numbers of infiltrating immune cells in SjD MSGs compared to controls. (E) Flow cytometry of MSG represented the frequency of pSTAT proteins: pSTAT1 (2.1-fold, p=0.017), pSTAT3(Ser727) (2.2-fold, p=0.031), and pSTAT6 (3.3-fold, p=0.112) were higher at baseline in SjD epithelial cells compared to non-SjD. Although not reaching the threshold of statistical significance, the frequency of pSTAT proteins (i.e., pSTAT3(Ser727) (p=0.056)) on CD45+ cells showed a similar frequency of pSTAT proteins in the SjD MSG. P value was calculated using Welch’s test. MSG, minor salivary gland; IF, immune fluorescent; SjD, Sjogren’s Disease; KRT18, Keratin-18.

Figure 2:. Immunofluorescence microscopy

(A) Overview of MSG biopsy, image acquisition and flow cytometry. (B) MSG IF image showing upregulated expression of JAK1 and JAK3 in SjD epithelial and infiltrating immune cells. JAK3 was especially enrich in ductal cells. Cellular proportion changes in the SjD MSGs showed greater numbers of immune cells, less numbers of epithelial cells. (C) Mean fluorescence of KRT18, JAK1 and JAK3 in SjD and non-SjD showing cellular proportion in the MSGs. Expression of JAK1 localized to immune cells (+6%; p<0.0001, Mann-Whitney Test) but JAK3 expression was seen in both epithelial and infiltrating immune cells (35% and 15%, respectively; p<0.0001, Mann-Whitney Test). (D) Cellular population change in 7 SjD and 6 non-SjD was characterized by flow cytometry showed reduced numbers of epithelial cells and increased numbers of infiltrating immune cells in SjD MSGs compared to controls. (E) Flow cytometry of MSG represented the frequency of pSTAT proteins: pSTAT1 (2.1-fold, p=0.017), pSTAT3(Ser727) (2.2-fold, p=0.031), and pSTAT6 (3.3-fold, p=0.112) were higher at baseline in SjD epithelial cells compared to non-SjD. Although not reaching the threshold of statistical significance, the frequency of pSTAT proteins (i.e., pSTAT3(Ser727) (p=0.056)) on CD45+ cells showed a similar frequency of pSTAT proteins in the SjD MSG. P value was calculated using Welch’s test. MSG, minor salivary gland; IF, immune fluorescent; SjD, Sjogren’s Disease; KRT18, Keratin-18.

Figure 3:. Single cell RNAseq and pSTATs frequencies of MSG

(A) UMAP embedding of the entire dataset colored by generated clusters labelled by cell type annotation. From all profiled MSG samples from 7 SjD and 5 non-SjD, Leiden clustering identified 11 different cell clusters corresponding to mucous (MUC5B) and seromucous acinar cells (MUC7), ductal cells (S100A2), plasma cells (IGHA, IGHG1), fibroblasts (COL1A2), myoepithelial (KRT14), pericytes (ACTA2), B-cells (CD79A), antigen presenting cells (HLA-DRA, CD68), T-lymphocytes (CD3D), and erythrocytes (HBB) (n=51736 cells). (B) Differential of cell density showing increased immune cell infiltration in SjD. (C) The top ten DEG in each of the cell types were dominated by ISGs including increased expression B2M, HLA-B, SAA1, IL32, and MGP. (D) Differential expression of IFN score in SjD and non-SjD, immune cells showed the biggest fold-changes in IFN score were in infiltrating immune cells (e.g., plasma cells: 700-fold, APCs: 400-fold; T cells: 300-fold; seromucous cells: 250-fold; ductal epithelial cells: 250-fold) exhibited higher IFN scores. (E) Fold change expression of JAK-STAT genes on all cell types. JAK1 was the most ubiquitously expressed in MSGs and Seromucous acinar cells showed increased expression of all JAKs. UMAP, Uniform manifold approximation and projection; MSG, minor salivary gland; SjD, Sjogren’s Disease; DEG, differentially expressed genes.

Figure 4:. Single cell RNAseq and pSTATs frequencies of PBMC

(A) Overview of using patient’s serum and PBMC for several assays. (B) Somalogic aptamer-based 1.3K target proteomics analysis revealed most of proteins were significantly upregulated in serum from SjD in volcano plot showing protein expression, in which some representative genes were highlighted. (C) IFN regulated proteins were found to be upregulated in SjD serum (p<0.0001). P value was calculated using Kruskal-Wallis test. (D) UMAP embedding of the entire dataset colored by generated clusters labelled by cell type annotation. Leiden clustering identified 10 different cell clusters from all profiled PBMC samples from 8 SjD and 6 non-SjD (n=206687 cells). (E) The upregulated genes were associated with immune and inflammatory responses revealed by functional annotation analysis from PBMC scRNAseq. (F) Differential expression of IFN score, monocytes had the highest expression of ISGs, followed by dendritic cells and then NK cells. SjD, Sjogren’s Disease; UMAP, Uniform manifold approximation and projection; RNAseq, RNA sequencing; ISG, interferon stimulated genes.

Figure 5:. Basal pSTATs frequencies in PBMCs

(A-D) Flow cytometry analysis revealed basal pSTATs levels in PBMCs were upregulated in 21 SjD compared to 10 HV. P value was calculated using Welch’s test. SjD, Sjogren’s Disease; HV, healthy volunteer.

Figure 6:. Treatment effects of tofacitinib in PBMCs

(A) Overview of using PBMC treated by tofacitinib for scRNAseq and FACS. PBMCs were treated with or without 5mM tofacitinib for 1 hour prior to IFNβ treatment for 30 minutes or 6 hours, respectively. (B) Treatment with tofacitinib blocked STATs phosphorylation status induced by IFNβ stimulation in PBMCs from SjD. P value was calculated using Mann-Whitney test. (C) UMAP embedding of the entire dataset colored by generated clusters labelled by 4 general cell type annotations. (D) Differential utilization of IFN signature of each condition showed tofacitinib abolished the IFNβ-induced IFN score to baseline level. (E) Volcano plot showing DEGs by pseudobulk analysis across SjD and HV, in which some representative genes were highlighted. T cells (129 genes) and monocytes (61 genes) exhibited greater numbers of downregulated ISGs in the context of tofacitinib. ISGs (e.g., IFIT1, IFIT3) were among the top DEG in all subsets after IFNβ and tofacitinib normalizing this response. RNAseq, RNA sequencing; SjD, Sjogren’s Disease; DEG, differentially expressed genes; HV, healthy volunteer; ISG, interferon stimulated genes.

Figure 7:. Treatment effects of tofacitinib in pSGECs

(A) Overview of using pSGEC treated by tofacitinib for IF and RT-qPCR. pSGEC were derived from fresh MSG biopsies. (B, C) Differential expression of pSTAT1 in pSGECs. Tofacitinib treatment mitigated IFNβ-induced STAT1 phosphorylation in both the nucleus and cytosol. P value was calculated using Kruskal-Wallis test. (D) Expression change of ISGs (i.e., CXCL10, ISG15, MX1) on pSGECs showing dose-dependently decreased IFNβ-induced ISGs mRNA expression. 5 SjD and 5 HV (n=5 individuals’ lines, respectively). 1 SjD and 1 HV samples were eliminated from MX1 result. P value was calculated using Mann-Whitney test and Kruskal-Wallis test. pSGEC, primary salivary gland epithelial cells; SjD, Sjogren’s Disease; HV, healthy volunteer; ISG, interferon stimulated genes.

Figure 7:. Treatment effects of tofacitinib in pSGECs

(A) Overview of using pSGEC treated by tofacitinib for IF and RT-qPCR. pSGEC were derived from fresh MSG biopsies. (B, C) Differential expression of pSTAT1 in pSGECs. Tofacitinib treatment mitigated IFNβ-induced STAT1 phosphorylation in both the nucleus and cytosol. P value was calculated using Kruskal-Wallis test. (D) Expression change of ISGs (i.e., CXCL10, ISG15, MX1) on pSGECs showing dose-dependently decreased IFNβ-induced ISGs mRNA expression. 5 SjD and 5 HV (n=5 individuals’ lines, respectively). 1 SjD and 1 HV samples were eliminated from MX1 result. P value was calculated using Mann-Whitney test and Kruskal-Wallis test. pSGEC, primary salivary gland epithelial cells; SjD, Sjogren’s Disease; HV, healthy volunteer; ISG, interferon stimulated genes.