Machine learning approach to single cell transcriptomic analysis of Sjogren's disease reveals altered activation states of B and T lymphocytes.

Machine Learning Approach to Single Cell Transcriptomic Analysis of Sjogren’s Disease Reveals Altered Activation States of B and T Lymphocytes - PMC Skip to main content An official website of the United States government Here's how you know Here's how you know Search Log in Primary site navigation Search Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice J Autoimmun . Author manuscript; available in PMC: 2026 Jun 1. Published in final edited form as: J Autoimmun. 2025 May 2;154:103419. doi: 10.1016/j.jaut.2025.103419 Machine Learning Approach to Single Cell Transcriptomic Analysis of Sjogren’s Disease Reveals Altered Activation States of B and T Lymphocytes Maxwell McDermott 1 , Wenyi Li 1 , Yin-Hu Wang 1 , Allen Y Chen 2 , Rodrigo Lacruz 3 , Bettina Nadorp 1, 2 , Stefan Feske 1 Author information Article notes Copyright and License information 1 Department of Pathology, New York University Grossman School of Medicine, New York, NY 10016, USA; 2 Department of Medicine, New York University Grossman School of Medicine, New York, NY 10016, USA; 3 Department of Molecular Pathobiology, New York University College of Dentistry, New York, NY 10010, USA Author Contributions . M.M., Y.-H.W., W.L., R.S.L., B.N. and S.F. designed experiments; M.M., Y.-H.W. and W.L. conducted experiments; M.M., Y.-H.W., W.L., Y.C., B.N., and S.F. analyzed data and interpreted the results. M.M. and S.F. wrote the manuscript. All authors read and approved the final version of the manuscript. ✉ Correspondence: stefan.feske@nyulangone.org Issue date 2025 Jun. PMC Copyright notice PMCID: PMC12170166  NIHMSID: NIHMS2074107  PMID: 40318561 The publisher's version of this article is available at J Autoimmun Abstract Sjogren’s Disease (SjD) is an autoimmune disorder characterized by salivary and lacrimal gland dysfunction and immune cell infiltration leading to gland inflammation and destruction. Although SjD is a common disease, its pathogenesis is not fully understood. In this study, we conducted a single-cell transcriptome analysis of peripheral blood mononuclear cells (PBMC) from patients with SjD and symptomatic non-SjD controls to identify cell types and functional changes involved in SjD pathogenesis. All PBMCs populations showed marked differences in gene expression between SjD patients and controls, particularly an increase in interferon (IFN) signaling gene signatures. T and B cells of SjD patients displayed a depletion of ribosomal gene expression and pathways linked to protein translation. SjD patients had increased frequencies of naive B cells, which featured a unique gene expression profile (GEP) distinct from controls and had hallmarks of B cell hyperactivation. Non-negative matrix factorization (NMF) also identified several non-overlapping GEPs in CD4 + and CD8 + T cells with differential usage in SjD patients and controls. Of these, only the Th1 activation GEP was enriched in T cells of SjD patients whereas the other two GEPs were depleted in T cells, emphasizing the important role of Th1 cells in SjD. Our study provides evidence for aberrant and unique gene expression patterns in both B and T lymphocytes of SjD patients that point to their altered activation states and may provide new insights into the pathogenesis of SjD. Keywords: Autoimmunity, Sjogren’s Disease, Machine Learning, Single-cell RNA sequencing, CD4 + T cells, Th1 1. INTRODUCTION Sjogren’s Disease (SjD) is a systemic autoimmune disease primarily affecting women and the second most common after rheumatoid arthritis[ 1 ]. The hallmark characteristic of SjD is the inflammation and dysfunction of salivary glands (SG) and lacrimal gland (LG), which results in dry mouth and dry eyes, also known as Sicca syndrome[ 2 ]. The inflammation of exocrine glands is caused by lymphocytic infiltration with T and B lymphocytes and the release of proinflammatory cytokines into the tissues. SjD is characterized by circulating autoantibodies, namely anti-Ro/SS-A and anti-La/SS-B antibodies[ 3 ]. While early symptoms and current classification criteria of SjD are predominantly focused on salivary and lacrimal glands, many patients develop extra glandular symptoms due to the involvement of the lungs, kidneys, and other organs in their lifetimes[ 4 ]. Patients with SjD also have an increased risk of developing non-Hodgkin lymphoma (NHL) compared to the general population[ 5 , 6 ]. While SjD has not been found to decrease life expectancy, it results in a poor quality of life for those afflicted. Even with this clear clinical need, our understanding of SjD pathogenesis and the corresponding treatment options remain limited. Currently, a combination of environmental and genetic factors are believed to trigger the onset of SjD[ 7 ]. The exact mechanisms underlying SjD pathogenesis remain unidentified, which has also made developing treatment modalities difficult. Most clinical trials have focused on targeting B cells, as not only are they responsible for producing autoantibodies, but they also infiltrate the gland tissues during early disease development[ 8 , 9 ]. The clinical trials targeting B cells have shown promising results in alleviating symptoms of SjD. However, SjD is a complex disorder, with more than one cell type responsible for disease pathogenesis. Both CD4 + and CD8 + T cells infiltrate the salivary and lacrimal glands of patients with SjD, and in the early stages of disease, they are more abundant than their B cell counterparts[ 10 ]. An important role of CD4 + T cells in SjD is thought to be the activation and maturation of B cells, but growing evidence indicates that they also play B cell-independent roles in SjD pathogenesis, which are both pro-inflammatory and anti-inflammatory[ 11 – 15 ]. Other cell types besides T and B cells have also been implicated in SjD pathogenesis, including plasmacytoid dendritic cells (pDCs) and monocytes/macrophages[ 16 – 19 ]. pDCs are a specialized subset of DCs that produce large quantities of type I interferon (IFN), and recent studies have noted that expression of the IFN inducible genes is enriched in both SGs and peripheral blood mononuclear cells (PBMC) of patients with SjD[ 20 ]. Similarly, serum interferon levels are elevated in SjD patients, which is currently being studied as a biomarker for SjD. Moreover, the type I IFN pathway is being investigated as a target for SjD treatment[ 21 ]. In the past, SjD studies have focused on many individual cell types, focusing on their functional status individually[ 22 ]. With the advancements in computational biology over the last several years and the emergence of single cell RNA sequencing (scRNA-seq), we can now in an unbiased manner analyze gene expression in different cell types at the same time in a high throughput fashion with granular precision. In fact, scRNA-seq has been used to study many autoimmune diseases including lupus erythematosus (SLE), rheumatoid arthritis (RA), and multiple sclerosis (MS)[ 23 – 25 ]. Recent studies have utilized scRNA-seq to study SjD to find disease-specific cell types and molecular signatures[ 24 , 26 – 32 ]. As scRNA-seq becomes a common platform to investigate disease pathophysiology, the analysis of such data has become more advanced. Machine learning is now being applied to scRNA-seq to understand the underlying mechanisms of complex diseases. Deep learning methods are being used to improve several key steps in scRNA-seq analysis, including batch effect removal, cellular clustering, gene expression imputation and denoising to increase the accuracy of downstream analyses[ 33 ]. Non-negative matrix factorization (NMF) uses weighted vectors to model gene expression profiles (GEP) to discover patterns in gene expression in a manner that other methods may not detect. These GEPs can be superimposed onto other datasets, facilitating comparisons across various disease states[ 34 ]. Here, we applied NMF to identify potential mechanisms underlying the pathogenesis of SjD in an unbiased manner. We performed scRNA-seq of PBMCs from 9 patients with SjD and 8 symptomatic controls who did not meet the classification criteria for SjD. NMF was used to assess GEPs in PBMCs samples of SjD patients and controls and to detect changes in these GEPs in patients with greater sensitivity than standard pathway analyses. PBMCs of patients with SjD were characterized by an enlarged B cell compartment and reduced overall numbers of CD4 + T cells. Gene expression was strongly dysregulated in all major immune cell populations of SjD patients, in particular in CD4 + and CD8 + T cells, B cells, and monocytes. Specific GEPs in CD14 + and CD16 + monocytes and CD8 + T cells were characterized by reduced expression of genes related to the function of these cells. We identified a GEP in B cells that was characterized by increased expression of genes required for B cell function and proliferation, and that was used more heavily by patients with SjD. CD4 + T cells were characterized by two distinct functional GEPs, a Th1 activation GEP and a naive CD4 + T cell signal transduction GEP. The usage of the former was significantly enhanced in SjD patients, whereas the latter was decreased. We show that machine learning approaches are a useful tool to identify unique GEPs associated with altered functional states of immune cells in autoimmune diseases such as SjD. The unique GEPs we identified provide evidence suggesting enhanced B cell and Th1 cell function in patients with SjD. 2. MATERIALS AND METHODS 2.1. Patient samples Patient data and peripheral blood mononuclear cell (PBMC) samples of SjD patients and non-SjD controls were obtained from the Sjögren’s International Collaborative Clinical Alliance (SICCA)[ 35 , 36 ]. The age, gender, autoantibody status, focus score and the 2016 ACR-EULAR SjD classification criteria score are shown in Table S1 . The collection of patient samples by SICCA was approved by the University of California, San Francisco IRB committee (IRB 10-02551). Participants gave informed consent to participate in the study before taking part in the study. 2.2. Single-cell RNA sequencing of PBMCs PBMCs were thawed and stained with Annexin-V (Biolegend, Cat. 640912) and LIVE/DEAD (ThermoFisher, Cat. L23105 ) in Annexin-V binding buffer (Biolegend, Cat. 422201). Live PBMCs were sorted as Annexin-V – Live/Dead – on an Aria II (BD Biosciences) with a 70 μm nozzle. Live cells were collected in RPMI1640 medium (Corning, Cat. 10040CV) with 10% fetal bovine serum. Sorted cells were stained using TotalSEQ Type B hashing antibodies (BioLegend, Cat. 394633 to 394649) for patient identification post multiplexing ( Table S2 ). Collected cells were quantified using a Bio-Rad TC20 automated cell counter. Hash-Tag antibody labeled cells were encapsulated into emulsion droplets using the 10x Genomics platform for library preparation. scRNA-seq and antibody multiplexing libraries were constructed using the Chromium Single Cell 3’ v3.1 Reagent Kit and 3’ Feature Barcode Kit (PN-1000268 and PN-1000262) according to the manufacturer’s protocols. Individual libraries were diluted to 2 nM and pooled for sequencing. Pooled batches were sequenced with 100 cycle run kits (28 bp Read1, 8 bp Index1 and 91 bp Read2) on the NovaSeq6000 Sequencing System (Illumina). Amplified cDNA was evaluated with an Agilent BioAnalyzer 2100 using a High Sensitivity DNA Kit (Agilent Technologies). Final libraries were evaluated with an Agilent TapeStation 4200 using High Sensitivity D1000 ScreenTape (Agilent Technologies). 2.3. Single-cell transcriptome assembly For each batch of patients, reads from the Illumina platform were converted to fastq files using the bcl2fastq package. Fastq files were processed into transcriptome assemblies using 10x Genomics Cell Ranger 7.0.0 and assembled using the GRCh38 reference genome[ 37 ]. Hashtag values were assigned to each cell in tandem to assembly, and demultiplexing was completed for each batch prior to integration. All analyses were performed using R Statistical Software (v4.1.2; R Core Team 2021) unless otherwise noted. Downstream single cell analysis was performed using Seurat (v5.0.1)[ 38 ]. Visualization and figure generation was performed using ggplot2 (v3.4.4) unless otherwise noted[ 39 ]. 2.4. Demultiplexing and quality control of cells Prior to combining the three batches of scRNA-seq data into one Seurat object, each batch was subjected to quality control and demultiplexing. Hashtag counts were normalized using a centered log-ratio method, and positive cells were defined via Seurat’s HTOdemux function, using the 0.99 quantile as the positive cut-off. Cells identified as doublets were filtered out from further downstream analysis. After doublet removal, cells were further filtered based on mitochondrial content and feature content. Cells with greater than 15% of features being mitochondrial genes were removed; likewise, cells with less than 1,000 unique molecular identifiers (UMIs) were removed. Cells falling into the 98 th and 2 nd percentiles for the number of genes covered per cell were also filtered as doublets and negatives, respectively. After quality control for doublet, singlet, and mitochondrial percentage, the remaining cells were integrated using Seurat with good mixing of data, and a robust number of variable genes. A total of 105,152 cells were successfully demultiplexed, and subsequently passed quality control filters, with a mean of 6,185 cells per sample with a similar distribution across all batches. 2.5. Normalization and integration of samples Gene expression levels were normalized per cell by log normalization and then scaled by a factor of 10,000. To correct for potential batch effects and sampling effects, each patient was individually integrated using variance stabilizing transformation (SCTv2), setting both batch and mitochondrial percentages as variables to regress out. After transformation, samples were integrated by reciprocal principal component analysis (RPCA) using 4,000 highly variable features[ 40 ]. To address potential disease state bias, one non-SjD control sample and one SjD patient sample were selected as integration anchor references. Both samples were from the same batch, and they were selected for having the highest average UMI per cell. 2.6. Cell clustering and cell type annotation To identify different immune cell types, we used the integrated assay of all combined samples and performed a PCA calculating the first 75 components. We then generated a uniform manifold approximation projection (UMAP) and a t-distributed stochastic neighbor embedding (t-SNE) plot using the first 60 dimensions and 20 nearest neighbors. Cell clusters were calculated with the same settings as both dimensional reduction plots, with a resolution of 1.4. Calculated clusters were annotated using SingleR, using the Database Immune Cell Expression Data as a reference[ 41 , 42 ]. The Azimuth reference was used as a second database with similar cell type identification[ 38 ]. Separate clusters that matched to the same cell type were combined into a single cluster. These combined clusters were then confirmed using known marker genes for each cell. 2.7. Analysis of cell type annotation using cell type markers To confirm that we had properly identified each cluster and their corresponding cell types, we grouped the cells on two different levels of granularity. The first, broader-level annotation grouped cells into seven broad cell types, defined as CD4 + and CD8 + T cells, other T cells that were neither CD4 + nor CD8 + , natural killer (NK) cells, B cells, monocytes, and dendritic cells (DCs). We used canonical cell type markers listed in the Azimuth database for cell type clustering as follows: Monocytes ( CTSS, FCN1, NEAT1, LYZ, PSAP, S100A9) , DCs (CD74, HLA-DPA1, HLA-DPB1, HLA-DQA1, CCDC88A, HLA-DRA, HLA-DMA ), B cells ( CD79A, RALGPS2, CD79B, MS4A1, BANK1, CD74 ), NK cells ( GNLY, GZMB, NKG7, KLRF1, PRF1 ), CD4 + T cells ( IL7R, MAL, LTB, CD4, LDHB ), CD8 + T cells ( CD8A, CD3D, TMSB10, HCST, CD3G, LINC02446 ), and other T cells ( TRDC, GZMK, IL7R, ANK3 ). The cellular composition of each sample was calculated using the total number of filtered cells per sample, and the mean cellular frequency was calculated for each disease state. Within each of these seven broad annotations, more granularly resolved cell types were confirmed using the appropriate marker genes, as illustrated in the corresponding supplemental figures ( Figure S2A , B and Figure S3A , B ). Marker genes for both broad annotations and granular subtypes are examples and are non-exhaustive; additional canonical markers were used to confirm proper cluster identification. Results are included in Supplemental Data File 1A . 2.8. Cellular communication analysis To determine if cell-to-cell communication between the seven major immune cell subsets within PBMCs was perturbed in SjD patients, we used the Cellchat package(v1.6.1)[ 43 ]. We accessed the CellchatDB for the collection of ligand-receptor pairs using the secreted signaling and ECM Receptor sections. Cell-to-cell communication was calculated independently for SjD and non-SjD control cohorts using the same pipeline, database, and default thresholds. The false discovery threshold for cellular interactions for each disease state was set at a Bonferroni-adjusted value (P adj ) ≤ 0.01. The bottom 50 th percentile of weighted discovered ligand-receptor pairs was removed due to low gene expression levels. After calculating cell-to-cell communication for each condition, we subtracted the normalized values for each ligand-receptor pair to determine differential communication. 2.9. Non-negative matrix (NMF) factorization To identify unique cell-type functions and their changes between SjD patients and non-SjD controls, we performed consensus non-negative matrix factorization (cNMF) as described in[ 44 ]. cNMF is based on the NMF implementation in scikit-learn v0.20.0. We filtered out genes that were not expressed in at least 20 cells across all samples and removed any cells that had less than 500 genes. No cells were lost via this filtering. We then calculated the top 3,000 overdispersed genes and used only those to perform NMF analysis. To determine the number of components (k) that would be most accurate, we calculated k over a range of 10 to 50, with each value having 200 iterations. We calculated the silhouette score and reconstruction error as implemented in cNMF. We found that k = 34 is the smallest, most stable solution. The final consensus solution was determined using a density threshold of 0.1 to exclude outlier solutions. We identified a total of 20 gene expression profile (GEP) usages with a mean score of greater than 0.1 for any cell type. The remaining 14 usages (of 34 total) were not readily used by any cell type or groups of cell types, with average usage scores of less than 0.02, and did not correlate with either batch nor individual samples. Marker genes for each GEP were calculated using multiple least squares regression of normalized z-scored gene expression against the consensus GEP usage matrix as implemented by cNMF. A functional enrichment test for differentially utilized GEP was subsequently performed using the top 100 marker genes per GEP using the EnrichR platform[ 45 ]. cNMF analysis was performed in Python (v.3.7.0) using scanpy (v.1.6.0), pandas (v.1.1.3), numpy (v.1.19.2), matplotlib (v.3.3.2)[ 46 , 47 ]. After cNMF analysis, the cell × usage (K × G) matrix was superimposed as metadata back onto the scanpy object with corresponding per cell values for each GEP. This object was converted back into Seurat for subsequent analysis and quantification. 2.10. Differential gene expression analysis To identify differentially expressed genes (DEGs) between SjD patients and non-SjD donors, we divided each identified cell type cluster by disease state. All cell types were confirmed to have equal representation in batches. We then calculated differential gene expression between the disease states using the FindMarkers function in Seurat. The RNA assay was utilized for expression analysis as per Seurat publisher’s recommendation[ 48 ]. When calculating DEGs we used the MAST algorithm implemented by Seurat[ 49 ]. We considered all genes found in at least 10% of cells with an absolute log 2 fold-change (FC) ≥ 0.1 and P adj ≤ 0.01 to be significant. Non-significant expression changes were calculated as well to act as background genes for downstream gene set enrichment analysis (GSEA). 2.11. Functional enrichment analysis To determine the functional enrichment or depletion of pathways associated with DEGs in each cell type, we performed pathway analyses. Using DEGs between SjD patients and controls, we performed a functional enrichment test (FET) using the GSEA function of ClusterprofileR 4.0[ 50 ]. Databases queried for our GSEA analysis included a combination of the canonical databases of molecular signature database, along with all 3 of the gene ontologies, and the Hallmark GSEA pathways datasets[ 51 , 52 ]. Pathways were filtered out if they contained ≥ 400 or ≤ 10 genes. Pathways were considered to be significant if they had a P adj ≤ 0.01. GSEA plots were made using the Gseaplot2 function of ClusterprofileR. Differentially activated upstream regulators were assessed using Qiagen’s ingenuity pathway analysis (IPA) tool (QIAGEN, https://digitalinsights.qiagen.com/IPA ). Upstream regulators that had an absolute z-score of < 0.2 or < 20% gene coverage were filtered out. Outputs of the Qiagen IPA platform were converted into a tab-delimited format, which was used to create bar plots using the R ggplot2 package. 2.12. Datamining of single cell RNA-seq of SjD labial gland biopsies. A publicly available dataset of labial gland biopsies of SjD patients and non-SjD controls was obtained from the CellxGene platform as an R data object, including all cell annotations provided by the original authors [ 53 , 54 ]. As the data was stored using Ensembl gene names, we converted them to gene symbols and summed duplicate gene symbols to create a gene expression matrix using the biomaRt package v3.2 [ 55 ]. Next, we applied gene expression programs (GEPs) determined by NMF analysis of PBMCs samples to this dataset using a modular-based scoring system, implemented via the AddModuleScore function in Seurat with 100 control genes. GEP modular scores were quantified using the same methodology as in the PBMCs dataset, and differential expression was calculated following the same approach. 2.13. Statistical analyses All analyses were performed using R Statistical Software (v4.1.1; R Core Team 2021). All bar plots represent mean values unless otherwise noted. Error bars represent 1.5 times the interquartile range between the 25th and 75th percentiles unless otherwise noted. Statistical significance between experimental groups was determined using a Wilcoxon rank-sum test and corrected with the Benjamini-Hochberg method unless otherwise noted in the figure legends. Differences were considered significant with P values < 0.05 (denoted by *), P < 0.01 (**), and P < 0.001 (***). The number of patients per experimental group is indicated in figure legends. 3. RESULTS 3.1. Enrichment of B cells and depletion of CD4 + T cells within PBMC of SjD patients. Changes in cellular composition of immune cells within circulating PBMCs and in glandular tissues have been linked to SjD pathogenesis. Whereas most studies have focused on the numbers of immune cell populations, less information is available about their functional states. Here we used single cell transcriptomics data from PBMCs of patients with SjD to obtain insights into gene expression and thus, indirectly the function, of different immune cell populations ( Figure 1A ). With one exception, all nine SjD patients analyzed were positive for anti-SSA/Ro antibodies and had ACR/EULAR scores ranging from 5–9 ( Table S1 ). The control cohort contained eight patients with sialadenitis and sicca symptoms but without SjD (anti-SSA/Ro negative; ACR/EULAR scores of 0 or 1). A total of 196k cells from 17 samples total and three experiments were assembled. Following dimensional reduction and clustering, seven cell types were annotated manually, using conserved marker genes for each major subpopulation ( Figure 1B ), which were validated by reference-based[ 42 , 56 ] annotation in more detail( Figure S1A , S1B ). The seven identified main immune cell populations included CD4 + and CD8 + T cells, other T cell subsets, B cells, NK cells, monocytes and dendritic cells (DCs). Of these, only the frequencies of CD4 + T cell and B cell populations showed statistically significant quantitative differences between SjD patients and the control group ( Figure 1C ). Whereas the frequency of B cells was increased in the SjD patient group, that of CD4 + T cells was reduced. An increase in B cell numbers and plasmablasts has been reported in the blood of SjD patients [ 57 , 58 ], as have reduced numbers of CD4 + T cells [ 59 , 60 ]. Likewise, alterations in the ratio of B and T cells has been demonstrated in patients with SjD although not using scRNA-seq for analysis[ 61 ]. Fig. 1. Dysregulated gene expression in immune cell populations within PBMC of SjD patients. Open in a new tab A) Aggregate UMAP plot of immune cell subsets identified in 105k peripheral blood mononuclear cells (PBMC) from 9 SjD patients and 8 non-SjD controls analyzed across 3 batches. B) Heatmap of conserved canonical markers genes of 7 distinct major immune cell subsets. C) Median frequency of each major immune cell subset in SjD patients and non-SjD controls. Each dot represents the cell type frequency of one patient. D) Volcano plots of differentially expressed genes (DEGs) in SjD patients and non-SjD controls for each major immune cell subset. The top 5 dysregulated genes are annotated. E) Dot plot of major dysregulated pathways in each immune cell subset analyzed by gene set enrichment using gene sets from the Reactome database. Shown are overlapping pathways that are dysregulated in at least 3 of 7 cell types. Z scores in (B) were calculated using the normalized average expression per cell in each cell type across all samples. Statistical analyses in (C), (D), and (E) were performed using Wilcoxon rank sum test. Genes in (D) were considered significant if the adjusted P value (P adj ) was <0.05. Pathways in (E) were considered overlapping if they were found to have a P adj < 0.01 in 3 or more cell types. Significance was adjusted using the Benjamini-Hochberg method. *P < 0.05; **P < 0.01; ***P < 0.001. 3.2. Dysregulated gene expression in all immune cell populations of SjD patients. We next analyzed if changes in B cell and CD4 + T cell frequencies are related to gene expression that might provide clues about their activation state. We analyzed differentially expressed genes (DEGs) between SjD and control samples for each of the main immune cell populations individually using the MAST function of Seurat ( Figure 1D )[ 49 ]. We found strongly dysregulated gene expression in each of the seven immune cell types between SjD and control samples with approximately similar numbers of up- or downregulated genes. The largest numbers of DEGs were found in CD8 + T cells, followed by CD4 + T cells, B cells, and monocytes. Next, we performed a functional enrichment test using the gene set enrichment analysis (GSEA) function of ClusterprofileR[ 50 ]. We found a significant increase in interferon signaling in all seven immune cell populations of SjD patients, with both IFN-γ and IFN-α/β signaling pathways enriched ( Figure 1E ). Increased type I and II IFN signatures have been reported and are a common molecular feature of SjD[ 62 ]. We also observed a strong depletion of pathways related to translational initiation and elongation, rRNA processing, and nonsense-mediated mRNA decay. Depletion of these pathways was restricted to B cells and the three T cell populations and not present in other PBMC subsets. Of note, this analysis focused on shared pathways that are dysregulated in at least 3 of the seven main immune cell populations, whereas differentially regulated pathways in individual immune cell populations of SjD patients will be reported further below. Our data indicate that gene expression is markedly altered in all immune cells within PBMCs of patients with SjD and that shared dysregulated pathways mainly affect IFN signaling and ribosome-associated pathways. 3.3. Enhanced intercellular communication of DCs with other immune cells in patients with SjD Having found that gene expression is markedly altered in all seven immune cell populations of SjD patient PBMCs, we hypothesized that these changes may affect intercellular communication. We therefore, calculated inferred cellular communication between immune cell populations of SjD patients and controls using the Cellchat program[ 43 ]. Cell-to-cell communication was analyzed in a weighted manner to accommodate different cell frequencies. Comparing intercellular communication in samples of SjD patients and controls, we found that communication between all cell populations was reduced in samples of SjD patients ( Figure S1C ). B cells, in particular, showed decreased cellular communication with all other immune cell populations in SjD patients despite their increased frequencies based on scRNA-seq data ( Figure 1C ). The only cell type that showed increased communication with other PBMC subsets, including all three T cell populations and NK cells, were DCs ( Figure S1C , D ). We hypothesized that the increased communication of DCs in SjD patients might correlate with a robust increase in gene expression. A differential gene expression analysis between DCs from SjD patients and non-SjD controls using the Seurat FindMarkers function, however, showed only a small number of genes with increased expression, especially compared to other immune cell types, with less than 50 enriched genes ( Figure S1E ). We noted a strong upregulation of genes related to IFN signaling in SjD patients. This included MX1 and MX2 and many IFN-stimulated and IFN-inducible genes. While we were able to identify DEGs in DCs of SjD patients and controls, it is noteworthy that the frequency of DCs in PBMC samples is very small (< 1%) and the number of DEGs much lower than in other immune cell populations (< 50). To avoid missing functional changes based on differential gene expression alone and to better understand the functional status of DCs in patients with SjD, we used non-negative matrix factorization (NMF), a machine learning approach suitable to identify differences between datasets of limited size in an unbiased manner [ 44 ]. NMF allowed us to identify and quantify the functional state of each cell individually while still sequestering the robust IFN signature identified by GSEA. Using NMF, we identified two distinct gene expression profiles (GEP) utilized by the DC population. The first GEP was predominantly used by cells in the pDC compartment, and the second GEP was utilized by all myeloid cells ( Figure 2A , 2B ). When we calculated the usage scores for the first GEP, we found that it was not significantly enriched in pDCs (or conventional DCs) of patients compared to controls, although there was a trend towards preferential usage in the SjD patients ( Figure 2A ). This conclusion, however, is limited by the very small frequency of circulating pDCs. The usage of the second, All myeloid cell GEP by DCs was not different between patients and controls either. We did however find a significant, yet small, increase in the usage of this GEP by monocytes of SjD patients ( Figure 2B ). We conclude that although DCs of SjD patients show a marked increase in expression of IFN-stimulated genes, their usage of an All myeloid cell GEP is not different from that of non-SjD controls. Fig. 2. Identification of functional gene expression profiles in dendritic cells (DC) and monocytes of SjD patients. Open in a new tab A) UMAP plot of PBMCs expressing a gene expression profile (GEP) associated with plasmacytoid dendritic cells (pDC) that was identified by NMF. The red box on the right is a zoomed in version of the pDC population shown in the UMAP plot on the left. The boxplot shows a quantification of the usage scores for this GEP in pDCs and conventional DCs (cDCs) in SjD patients and non-SjD controls. B) UMAP plot of PBMCs colored by expression of a GEP associated with myeloid lineage cells identified by NMF. The boxplot shows a quantification of the usage scores for this GEP in all DCs and all monocytes in SjD patients and non-SjD controls. C) Top 15 dysregulated pathways based on DEGs in all monocytes of SjD patients and non-SjD controls analyzing the canonical pathways of the IPA platform. D) Top 20 upstream regulators of DEGs in all monocytes of SjD patients and non-SjD controls using the IPA regulator database. E) UMAP plot of PBMCs colored by expression of a GEP associated with CD16 + monocytes identified by NMF. The boxplot shows usage scores for this GEP in CD14 + and CD16 + monocytes in SjD patients and non-SjD controls. F) UMAP plot of PBMCs colored by expression of a GEP associated with CD14 + monocytes identified by NMF. The boxplot shows usage scores for this GEP in CD14 + and CD16 + monocytes in SjD patients and non-SjD controls. G) Top 10 Wikipathways pathways associated with the CD14 + monocyte GEP identified by EnrichR analysis. H) Heatmap of DEGs contained in the CD14 + monocyte GEP in CD14 + monocytes of SjD patients and non-SjD controls. Colors depict log2 fold changes (FC) in expression. Statistical analyses in (A), (B), (E), and (F) were completed using a Wilcoxon rank sum test. The significance of pathways and regulators shown in (C),(D) and (G) was calculated using a right-tailed Fisher’s exact test. The significance in (A-G) was adjusted using the Benjamini-Hochberg method. Scales in (A), (B), (E), and (F) represent the usage scores per cell for each GEP with a range between 0.2 (low) and 0.7 (high). *P < 0.05; **P < 0.01; ***P < 0.001. 3.4. Depletion of distinct functional gene expression profiles in CD14 + and CD16 + monocytes of SjD patients. Monocytes and macrophages have not been the focus of much research in SjD, but they may play a significant role in causing salivary and lacrimal gland inflammation through cytokine production, antigen presentation, or phagocytosis[ 63 ]. The analysis of transcriptomes of monocytes from SjD patients compared to controls showed the dysregulated expression of more than 2,000 genes ( Figure 1D ). Ingenuity pathway analysis (IPA) of DEGs in monocytes identified several pathways that are enriched in monocytes of SjD patients. These include ATP synthesis and oxidative phosphorylation (with a concomitant depletion of mitochondrial dysfunction-associated genes), MHC class I processing and presentation, IL-1 family signaling, and noncanonical NFκB signaling ( Figure 2C ). Enrichment of the latter two pathways was driven, however, by the presence of many PSM genes encoding proteasome subunits in both pathways and not bona fide IL-1 or NFκB signaling genes (not shown). An IPA upstream regulator analysis revealed IFN-related signaling as the main cause of differential gene expression in monocytes of SjD patients ( Figure 2D ). Monocytes could be further segregated into two populations that are distinguished by the expression of CD14 ( Figure S2A ) and CD16 ( Figure S2B ) marker genes. These populations largely overlapped with two functional GEPs that we identified by NMF analysis ( Figure 2E , F ). The GEP associated with CD16 + cells was only used by this monocyte subset but not by CD14 + monocytes. The usage of this GEP was reduced in CD16 + monocytes from patients with SjD ( Figure 2E ). Conversely, the GEP associated with CD14 + monocytes was only used by this monocyte subset. Its usage was also reduced in SjD patients compared to controls ( Figure 2F ). When we performed a functional enrichment test of genes in the CD16 + monocyte associated GEP, we were unable to identify specific pathways that were dysregulated in SjD patient samples compared to controls, nor were we able to identify distinct patterns of altered genes upon manual inspection ( Figure S2C ). Although the usage of the CD16 + monocyte GEP is depleted in SjD patients, the biological relevance of this depletion is unclear. By contrast, a functional enrichment test of the CD14 + monocyte GEP, which is also depleted in SjD patients, revealed that this GEP is linked to IL-1 and NFκB signaling ( Figure 2G ). Genes belonging to this GEP and depleted in SjD patients include the IL-1 signaling-associated genes IL1B (encoding IL-1β), IRAK2 , and IRAK3 (encoding interleukin-1 receptor-associated kinases 2 and 3), TNFAIP3 (encoding the IL-1 signaling inhibitor A20), IL1RN (encoding the interleukin-1 receptor antagonist, IRAP), and the NFκB signaling associated genes NFKB1 (encoding the NFκB p50 subunit), NFKBIA (encoding IκBα), and NFKBIZ (encoding IκBζ, which is induced by IL-1 signaling). Although NFκB signaling-related genes in the CD14 + monocyte GEP were depleted, these genes encompassed both positive and negative regulators of NFκB signaling, and the functional consequences of their altered expression remain to be elucidated ( Figure 2H , Figure S2D ). It is noteworthy that IL-1 signaling has previously been linked to SjD pathogenesis by controlling systemic or local immune responses and regulating the function of gland epithelial cells[ 64 ]. 3.5. Increased abundance and functional signature of naive B cells in patients with SjD. B cells play an important role in SjD by producing autoantibodies, particularly anti-SSA and anti-SSB antibodies. Having determined that the B cell frequencies are increased in SjD patients compared to non-SjD controls, we hypothesized that this increase could be restricted to a specific B cell subset. Among the naive, intermediate, and memory B cell subsets identified by scRNA-Seq ( Figure S1A ), only the naïve subset was significantly expanded in SjD patients, although all subsets showed a trend towards enrichment ( Figure 3A ). Plasmablasts, although detectable, were too infrequent to allow for meaningful biological conclusions; therefore, they were not considered for further downstream analysis. Differential gene expression analysis had shown profound up- and downregulation of a total of 2,774 genes in all B cell subsets combined. A pathway analysis of these DEGs showed enrichment of pathways such as pre-mRNA processing, protein ubiquitination, ATP synthesis, oxidative phosphorylation, non-canonical NF-kB signaling, and BCR signaling, and depletion of pathways including eIF2 signaling, which is crucial for the regulation of protein synthesis ( Figure 3B ). To better understand the causes of dysregulation of such a large number of genes in B cells of SjD patients, we conducted an upstream regulator analysis of DEGs in all B cells of SjD patients and controls using the IPA platform ( Figure 3C ). We found enrichment of several interferons, including IFN-α and IFN-γ as upstream regulators in B cells of patients with SjD. Also enriched was La Ribonucleoprotein 1 (LARP1), which regulates the translation of 5’TOP mRNAs, including those encoding ribosome proteins. Non-phosphorylated LARP1 interacts with 5’ and 3’ UTRs of ribosome-encoding mRNAs and inhibits their translation[ 65 , 66 ], which is consistent with the decreased activity of pathways associated with translational initiation and elongation observed in B cells ( Figure 1E ). Depleted upstream regulators in B cells of SjD patients were the proto-oncogenes MYC and MYCN, which is also consistent with the dysregulation of ribosome-associated pathways as MYC regulates ribosome biogenesis and protein synthesis. Fig. 3. Enrichment of a disease specific GEP in B cells of patients with SjD. Open in a new tab A) Median frequencies of B cell subsets (of all PBMCs) in SjD patients and non-SjD controls. Each dot represents the cell type frequency of one patient. B) Top 15 dysregulated pathways based on DEGs in all B cells of SjD patients and non-SjD controls using the canonical pathways of the IPA platform. C) Top 20 upstream regulators of DEGs in all B cells of SjD patients and non-SjD controls using the IPA regulator database. D-F) UMAP plots colored by expression of the GEPs associated with B cell identify (D) , naive B cells (E) , and SjD specific genes in B cells (F) identified by NMF. The boxplots show usage shown for these GEP in naive, intermediate and memory B cells of SjD patients and non-SjD controls. G) Top 10 MGI (mouse genome informatics) phenotypes associated with the SjD B cell GEP identified by EnrichR analysis. H) Heatmap of DEGs contained in the SjD B cell GEP in all B cells of SjD patients and non-SjD controls. Colors depict log 2 fold changes (FC) in expression. Statistical analysis in (A), (D), (E), and (F) was performed using a Wilcoxon rank sum test. The significance of pathways and regulators in (B), (C) and (G) was calculated using a right-tailed Fisher’s exact test. The significance in (A-F) was adjusted using the Benjamini-Hochberg method. Pathways shown in (G) are the top 10 ranked by P value identified using a right-tailed Fisher’s exact test. Scales shown in (D-F) represent the usage scores per cell for each GEP with a range between 0.2 (low) and 0.7 (high). *P < 0.05; **P < 0.01; ***P < 0.001. To identify functional gene expression profiles that correlate with specific B cell subsets and are used specifically by patients with SjD, we again conducted an NMF analysis of all B cell subsets of patients and controls. We identified three GEPs, of which one was ubiquitously used by all three B cell subsets, and we therefore called B cell identity GEP ( Figure 3D ). It contains the canonical B cell markers we used to identify B cells by scRNA-seq ( Figure 1A ). Naive B cells of SjD patients showed a significant usage of this GEP compared to non-SjD controls, although the differences were modest. A second GEP we identified in B cells was predominantly used by naïve B cells of both patients and controls, whereas intermediate and memory B cells do not utilize this GEP ( Figure 3E ). Usage of this Naive B cell GEP was significantly depleted in naïve B cells from SjD patients compared to non-SjD controls. A third GEP was found in all three B cell subsets and different from the other two B cell GEPs because its usage was significantly increased in naive, intermediate, and memory B cells of SjD patients compared to controls ( Figure 3F ). The overall usage of this SjD B cell GEP was highest in naive B cells. Because naive B cells were also the only subsets whose numbers were significantly increased in SjD patients ( Figure 3A ), we hypothesized that the genes contained in this GEP might be responsible for the increased B cell frequencies in patients. A functional enrichment test of the genes contained in the SjD B cell GEP using EnrichR identified genes that are associated with phenotypes related to B cell physiology, proliferation, and morphology, impaired humoral immune responses, and decreased transitional B cell numbers ( Figure 3G ). We analyzed the differential expression of genes belonging to the SjD B cell GEP in naive B cells from SjD patients and controls ( Figure 3H ). The vast majority of DEGs are enriched in B cells of SjD patients with the most highly enriched genes including ROR1 (encoding Receptor Tyrosine Kinase-Like Orphan Receptor 1), which plays a role in the development of B cells, and GAB1 (GRB2-Associated Binder 1), a protein that regulates proliferation via PI3K/AKT and MAPK/ERK pathways, although its role in B cells is not known. Enriched genes also include known regulators of B cell fate and function, including MZB1 (encoding marginal zone B and B1 cell-specific protein), SYK and LYN (encoding kinases required for B cell receptor signaling), PAX5 and BACH2 (transcription factors regulating different stages of B cell development). Taken together, we find an expansion of naive B cells in patients with SjD and identify increased usage of a GEP in naive B cells of patients that contains many genes involved in B cell development and function. These data provide indirect evidence that naive B cells from SjD patients are activated in a disease-specific manner. 3.6. Depletion of a signal transduction GEP in CD8 + T cells of SjD patients. CD8 + T cells contribute to injury of acinar or ductal epithelial cells in the exocrine glands of patients SjD by a variety of mechanisms, including the production of IFN-γ and induction of apoptosis [ 67 ]. A recent study showed that granzyme K-producing CD8 + T cells target seromucous acinar cells in patients with SjD [ 53 ]. CD8 + T cells showed the strongest differential gene expression with more than 4,500 genes up- or downregulated in SjD patients compared to controls ( Figure 1D ), suggesting that this T cell subset is significantly functionally dysregulated in patients’ PBMC. Moreover, CD8 + T cells from patients showed a profound depletion of translation and ribosome-associated pathways without associated changes in the frequencies in CD8 + T cells within PBMCs of SjD patients ( Figure 1C , E ). To understand the functional state of CD8 + T cells, we first conducted a pathway analysis of DEGs in SjD patients compared to controls using the IPA platform. We observed a strong enrichment of type I and type II IFN signaling pathways, as well as TCR signaling, Th1 pathway, and death receptor signaling in the patients’ CD8 + T cells ( Figure 4A ). Of note, while depletion of ribosome-associated pathways was again observed, it did not appear as one of the top dysregulated pathways. Consistent with the pathway analysis, we identified many interferons and IFN signaling-related factors as drivers of gene expression in CD8 + T cells of SjD patients using an upstream regulator analysis ( Figure 4B ). These included IFN-α, IFN-γ, IFNL1, IRF7 and STAT1. These findings related to enhanced IFN signaling are similar to those in other immune cell subsets ( Figure 1E ) and likely are secondary to the overall IFN-driven inflammatory milieu in SjD patients. To identify additional changes in gene expression that are more specific to CD8 + T cells and that may provide insight into their functional state, we again used NMF. We identified a GEP that is specific to a subpopulation that corresponds to naive CD8 + T cells and is not used by CD4 + T cells or other T cells either ( Figure 4C ). We therefore named it Naive CD8 + T cell GEP. The usage of this GEP was significantly depleted in CD8 + T cells of patients with SjD. A functional enrichment analysis of the genes in this GEP revealed that many of them are associated with TCR signaling ( Figure 4D ). Specific pathways these genes were linked to included beta-catenin, canonical NF-kB, calcineurin-NFAT, and Ras signaling. As expected from the reduced usage of this GEP, the expression of almost all genes was reduced in CD8 + T cells of SjD patients ( Figure 4E , Figure S4A ). Some of the genes in this GEP whose expression was most strongly depleted and that have been associated with CD8 + T cell function before were S100B, CD37, CCR7 , and RCAN3 . Many of the depleted genes have either positive or negative regulatory roles in TCR signaling and CD8 + T cell function, and it is, therefore, difficult to determine the functional consequences resulting from the reduced usage of this GEP by CD8 + T cells of SjD patients. Fig. 4. Depletion of genes in a GEP associated with naive CD8 + T cells in patients with SjD. Open in a new tab A) Top 15 dysregulated pathways based on DEGs in all CD8 + T cells of SjD patients and non-SjD controls using the canonical pathways of the IPA platform. B) Top 20 upstream regulators of DEGs in all CD8 + T cells of SjD patients and non-SjD controls using the IPA regulator database. C) UMAP plot of PBMCs colored by expression of a GEP associated with naive CD8 + T cells identified by NMF. The boxplot shows usage scores for this GEP in all CD4 + T cells, CD8 + T cells and other T cells of SjD patients and non-SjD controls. D) Top 10 NCI Nature pathways associated with the naive CD8 + T cell GEP identified by EnrichR analysis. E) Heatmap of DEGs contained in the naive CD8 + T cell GEP in all CD8 + T cells of SjD patients and non-SjD controls. Colors depict log2 foldchanges (FC) in expression. The significance of pathways and regulators shown in (A), (B) and (D)was calculated using a right-tailed Fisher’s exact test. The statistical analysis in (C) was done using a Wilcoxon rank sum test. The significance in (A-D) was adjusted using the Benjamini-Hochberg method. The scale in (C) represents the usage scores per cell for the GEP with a range between 0.2 (low) and 0.7 (high). *P < 0.05; **P < 0.01; ***P < 0.001. 3.7. Two distinct functional GEPs in CD4 + T cells of SjD patients and signs of enhanced Th1 function. CD4 + T cells have been extensively studied in SjD, and they make up a large portion of the lymphocytic infiltrates found in SGs of patients [ 68 ]. Several different CD4 + T cell subsets, including T helper (Th) 1, Th17, and T follicular helper (Tfh) cells, have been implicated in SjD pathogenesis [ 68 ]. Another CD4 + T cell subset relevant to SjD is regulatory T cells (Treg), whose frequencies in the blood and exocrine glands of SjD patients have alternatively been reported to be increased, decreased, or unchanged [ 69 – 78 ]. We recently reported enhanced functional gene expression signatures in Th1 cells and decreased signatures in Treg cells in the blood of patients with SjD[ 79 ]. Here, we assessed differential gene expression in all CD4 + T cells of patients with SjD and non-SjD controls to identify potential imbalances in the different subsets of CD4 + T cells and gain insights into their functional state. A pathway analysis of DEGs in all CD4 + T cells revealed the enrichment of pathways related to TCR signaling, oxidative phosphorylation (OXPHOS), and the regulation of apoptosis ( Figure 5A ). These findings were confirmed by GSEA ( Figure 5B ). Enriched genes in CD4 + T cells of SjD patients included well-known molecules involved in TCR signaling, including CD3E, CD3G, LCK, ZAP70 , and LAT ( Figure S4B ). Although not detected by IPA pathway analysis, our manual analysis and GSEA of enriched genes also identified many PSM genes that encode proteasome subunits, which may indicate increased protein degradation in CD4 + T cells of SjD patients ( Figure S4B , C ). By contrast, pathways related to protein translation were strongly downregulated in patients’ CD4 + T cells ( Figure 5A ). An upstream regulator analysis of DEGs using the IPA platform identified many T cell-associated surface receptors and cytokines to be enriched in SjD patients ( Figure 5C ). These included the TCR, CD3, CD28, and CD40L, as well as IL-2, IFN-γ and IL-15. IFN-γ is the signature cytokine of Th1 cells, and IL-15 promotes the proliferation of developing Th1 cells (especially in combination with IL-12), potentially indicating enhanced polarization of CD4 + T cells into Th1 cells in SjD patients. The upstream regulator analysis also found a strong enrichment of LARP1 and depletion of MYC in CD4 + T cells of SjD patients, which is similar to findings in their B cells and correlates with the depletion of genes and pathways associated with protein translation ( Figure S5A – C ). Fig. 5. Enhanced usage of a functional GEP in CD4 + Thl cells of SjD patients. Open in a new tab A) Top 15 dysregulated pathways based on DEGs in all CD4 + T cells of SjD patients and non-SjD controls using the canonical pathways of the IPA platform. B) GSEA of DEGs in all CD4 + T cells of SjD patients and controls. Shown are enriched gene sets related to TCR signaling and OXPHOS. C) Top 20 upstream regulators of DEGs in all CD4 + T cells of SjD patients and non-SjD controls using the IPA regulator database. D) UMAP plot of PBMCs colored by expression of a GEP associated with Thl activation identified by NMF. The boxplot shows usage scores for this GEP in all CD4 + T cells, CD8 + T cells and other T cells of SjD patients and non-SjD controls. E) Top 10 Wikipathways associated with the Th1 activation GEP identified by EnrichR analysis. F) Heatmap of DEGs contained in the Thl activation GEP of all CD4 + T cells of SjD patients and non-SjD controls. Colors depict log 2 fold changes (FC) in expression. G) UMAP plot of PBMCs expressing a GEP associated with Naïve CD4 + T cell signal transduction identified by NMF. The boxplot shows usage scores for this GEP in all CD4 + T cells, CD8 + T cells and other T cells of SjD patients and non-SjD controls. H) Top 10 Wikipathways associated with the Naïve CD4 + T cell signal transduction GEP identified by EnrichR analysis. I) Tile plot of DEGs contained in the CD4 + signal transduction GEP of all CD4 + T cells of SjD patients and non- SjD controls. Colors depict log 2 fold changes (FC) in expression. Statistical analysis in (D) and (G) was performed using a Wilcoxon rank sum test. The significance of pathways and regulators shown in (A), (C), (E), and (H) was calculated using a right-tailed Fisher’s exact test. The significance in (A), (B), (C), (E), (G), and (H) was adjusted using the Benjamini-Hochberg method. Scales shown in (D) and (G) represent the usage scores per cell for each GEP with a range between 0.2 (low) and 0.7 (high). *P < 0.05; **P < 0.01; ***P < 0.001. We next conducted an NMF analysis of all CD4 + T cells to identify functional gene expression profiles that correlate with specific CD4 + T cell subsets and may be used specifically by SJD patients. We identified two major GEPs in CD4 + T cell subsets ( Figure 5D , G ) as identified by the expression of marker genes ( Figure S1A ). One was predominantly utilized by CD4 + T cells that were identified as Th1 cells by marker gene expression ( Figure S1A ). Its usage was significantly increased in CD4 + T cells (but not other T cell types) of SjD patients compared to controls ( Figure 5D ). To identify the biological function of this GEP, we performed a functional enrichment test on the genes contained within it ( Figure 5E ). Most of the significantly enriched pathways associated with genes in this GEP were linked to TCR signaling, suggesting that this GEP is related to T cell activation. Individual inspection of DEGs contained in this GEP revealed that most were enriched in SjD patients. These included IL-12RB2 (a subunit of the IL-12 receptor that is required for Th1 differentiation) and IFNG-AS1 (a long noncoding RNA, lncRNA, also called TMEVPG1, that is associated with increased Th1 responses in SjD [ 80 ]) ( Figure 5F , Figure S4D ). We thus named this GEP Th1 activation . Our findings suggest that Th1 cells are hyperactivated in the blood of patients with SjD. The other GEP we identified was expressed in different CD4 + T cell subsets ( Figure 5G , Figure S1A ). Although broadly expressed in CD4 + T cells, this GEP does not overlap with the Th1 activation GEP, indicating that they are functionally distinct from one another. The usage of the second GEP, which we named naïve CD4 + T cell signal transduction GEP, was significantly depleted in CD4 + T cells of SjD patients compared to non-SjD controls ( Figure 5G ). Upon performing a FET to determine the biological implications of these genes being depleted, we identified pathways related to signal transduction, generation of second messengers, WNT signaling, and TCF/LET target gene regulation to be linked with this GEP ( Figure 5H ). Individual inspection of genes in this GEP revealed that most of them were depleted in CD4 + T cells of SjD patients ( Figure 5I , Figure S4E ). These genes include regulators of signal transduction in T cells such as CD247 (encoding the TCRζ chain), ITK, CAMK4 , and CD28 , regulators of T cell development including THEMIS , TCF7 , and LEF1 and genes promoting T cell survival such as BCL2 and IL7R . The functional significance of gene depletion in this naive CD4 + T cell signal transduction GEP is unknown. 3.8. Enriched SjD B cell signature, as well as signal transduction and Th1 activation found in salivary glands of SjD patients. To investigate how the changes in gene expression in PBMCs of our cohort of SjD patients correlate with those in SjD patients in other studies, we examined the utilization of the identified GEPs using scRNA-seq datasets of PBMCs [ 24 ] and labial SG biopsies from SjD patients and non-SjD controls [ 53 ]. The reanalysis of gene expression data from Hong et al. [ 24 ] revealed no statistically significant enrichment or depletion of immune cell types within PBMCs of SjD patients, consistent with the original publication ( Figure S5A – C ). The utilization of GEPs we had identified in our dataset was also altered in PBMCs of the SjD patients studied by Hong et al. We confirmed reduced usage of the CD14 + monocyte GEP ( Figure S5D , E ), naïve CD8 + T cell GEP ( Figure S5H , I ), and naïve CD4 + T cell signal transduction GEP ( Figure S5J , K ) in SjD patients compared to healthy controls. Notable differences included normal usage of the SjD B cell GEP ( Figure S5F , G ) and reduced usage of the Th1 activation GEP ( Figure S5L , M ) in SjD patients reported by Hong et al. The causes of the latter discrepancies are not clear but may be related to differences in disease activity and treatment history, or SSA/Ro antibody positivity of patients in both studies. We next aimed to compare our findings in PBMCs of SjD patients with those in immune cells in the SGs of SjD patients. We examined the utilization of GEPs in scRNA-seq data from labial SG biopsies of SjD patients and non-SjD controls maintaining the immune cell cluster annotations within SGs as originally published ( Figure S6A ) [ 53 ]. Monocytes were not identified in the SG dataset, and we were thus unable to assess the utilization of the monocyte specific GEPs we had identified in PBMC. The SjD B cell GEP which is preferentially utilized in PBMC of SjD patients was not enriched in B cells infiltrating SGs of SjD patients, but we detected a significant enrichment in IgG + and IgA + plasma cells (PC) of patients compared to controls ( Figure S6B – D ). The GEP associated with naive CD4 + T cell signal transduction was also depleted in SGs of SjD patients, although the depletion was less pronounced compared to PBMC ( Figure S6E , F ). Moreover, we observed an increased utilization of the Th1 activation GEP in SGs of SjD patients compared to controls ( Figure S6G , H ), which is similar to PBMC. The fact that the utilization of these GEPs in B cells and CD4 + T cells of SjD patients is similar in PBMC and SG infiltrating immune cells suggests that gene expression changes in PBMC provide valuable insights into the activation states of cells in the glands of patients. 4. DISCUSSION Compared to its prevalence, the pathophysiology of SjD is inadequately understood. High throughput techniques such as scRNA-seq allow for unbiased analysis of the prevalence of different immune cell types in SjD patients and their functional status. These studies have revealed several potential mechanisms underlying disease pathogenesis, but also showed some limitations of traditional gene expression analysis methods to determine altered functional states of immune cells in SjD patients. Here we use NMF as a machine learning approach to identify GEPs in scRNA-Seq data of PBMCs from SjD patients and non-SjD controls that were not detected by standard pathway analyses and that may more closely correlate with the functional state of immune cells during disease. Our analyses revealed an increase in the frequency of B cells in SjD patients that was driven mostly by naive B cells, and not memory B cells whose frequency was unchanged. B cells have long been implicated in the pathophysiology of SjD as emphasized by the presence of autoantibodies and the fact that B cell infiltrates are found in SGs of patients with SjD. B cells were shown to be hyperactive in SjD, resulting in increased proliferation and antibody production[ 81 ]. It is therefore noteworthy that altered B cell numbers or phenotypes are not a prominent finding in most of the recent scRNA-seq studies of SjD patients. One study, however, reports increased numbers of effector memory and naive B cells in SGs of SjD patients and expanded IGHD + naive B cells in their blood [ 31 ], which is consistent with the expansion of naive IGHM + IGHD + B cells in our study. Using NMF, here we identify several functional GEPs in B cells whose usage is altered in SjD patients. A GEP corresponding to naive B cells is depleted in SjD patients, whereas a GEP that is specifically associated with B cells in SjD patients is increased compared to non-SjD controls. The latter SjD B cell GEP contained genes involved in both B cell activation and proliferation. Most of the genes in this B cell signature were upregulated, including the tyrosine kinases LYN and SYK which regulate the proliferation and activation of B cells[ 82 , 83 ]. In SjD, treatment with remibrutinib, a Bruton’s tyrosine kinase (BTK) inhibitor, has been shown to alleviate symptoms, and reduce serum levels of SSA/B in SjD patients[ 84 ]. These findings underscore the important role of B cells in SjD and the effectiveness of monoclonal antibodies that deplete B cells or inhibit B-cell activating factor (BAFF) signaling (belimumab, ianalumab) in SjD treatment [ 85 , 86 ]. Our findings are consistent with previous reports of B cell hyperactivation in SjD and show that they utilize a unique GEP whose constituent genes may be new therapeutic targets in SjD. CD4 + T cells play an important role in the pathophysiology of SjD. While their numbers are found to be increased in the salivary and lacrimal glands of SjD patients [ 68 ], CD4 + T lymphopenia in the blood of patients is a common laboratory finding in SjD. Reduced CD4 + T cell counts are correlated with higher systemic disease activity scores and are considered a predictive factor for lymphoma development [ 58 – 60 ]. Here we report that the frequency of CD4 + T cells among PBMC of SjD patients is decreased, which was not driven by the depletion of a single subtype of CD4 + T cells. CD4 + T cells can be divided into distinct T helper (Th) cell subsets, including Th1, Th17 and T follicular helper (Tfh) cells, all of which have been implicated in SjD [ 87 ]. As we reported recently, the frequencies of eleven CD4 + T cell subsets were comparable between SjD patients and non-SjD controls including Th1, Th17 and Tfh cells[ 79 ]. In the current study, we did not observe GEPs that were uniquely associated with Th17 or Tfh cells in SjD patients. Using NMF, we did however identify a unique GEP associated with Th1 activation that whose utilization was increased in CD4 + T cells of SjD patients. Most genes in this GEP showed increased expression in SjD patients including several that had previously been linked to Th1 function, including IL-12RB2 (a subunit of the IL-12 receptor required for Th1 differentiation), IFNG-AS1 (a lncRNA associated with increased Th1 responses in SjD [ 80 ]) and the proinflammatory cytokine IL26 . Th1 cells play important pathogenic roles in several autoimmune diseases[ 88 , 89 ] including SjD[ 90 – 93 ]. Our finding of hyperactivated Th1 cells in PBMC is of SjD patients is consistent with the increased utilization of the Th1 activation GEP in SGs of SjD patients we analyzed [ 53 ] and with a recent study showing that ~ 30% of CD4 + T cells in SGs of SjD patients express the transcription factor T-bet, identifying them as Th1 cells[ 94 ]. Moreover, the expression of IFN-γ is increased in SGs and saliva of patients with SjD[ 90 – 96 ], and IFN-γ expression correlated with the degree of lymphocytic infiltration[ 92 ]. The increased Th1 activation GEP we identify in this study makes sense in light of the increased IFN response in SjD patients and the ability of type I and II IFN to induce Th1 differentiation [ 98 , 99 ]. It is noteworthy that none of the other recent scRNA-seq studies identified a Th1 signature in CD4 + T cells of patients with SjD, either in the blood or in SGs. This is consistent with the fact that we did not detect Th1 signature in our analysis of DEG from PBMCs of SjD patients using conventional pathway analysis tools. Instead, we identified the Th1 activation GEP by using NMF emphasizing the sensitivity of this method to identify gene expression patterns. Of note, the Th1 activation GEP contained GZMB and GZMK whose expression was increased. It is possible that this GEP has some functional overlap with CD4 + CTL. This is intriguing because a recent study identified an increased proportion of CD4 + T cells in the blood of SjD patients that express GZMB marking them as potentially cytotoxic [ 24 ]. Although we also detected GZMB-expressing CD4 + CTL by scRNA-seq, their numbers were comparable in SjD patients and controls[ 79 ]. The increased utilization of the Th1 activation GEP does not represent a generally enhanced activation of all CD4 + T cells in SjD patients. We identified a second GEP in CD4 + T cells, named the naive CD4 + T cell signal transduction GEP, which was significantly depleted in SjD patients. Genes in this GEP are associated with pathways linked to signal transduction and most of these genes were depleted including CD247 (encoding the TCRζ chain), ITK, CAMK4 and CD28 . Moreover, we found genes related to T cell development (such as THEMIS , TCF7 and LEF1 ) and T cell survival (such as BCL2 and IL7R ) to be depleted. Of note, the naive CD4 + T cell signal transduction GEP was also depleted in SGs of patients with SjD [ 53 ]. Moreover, we identified one unique GEP in CD8 + T cells, which was significantly depleted in SjD patients, and which we named naive CD8 + T cell GEP. It contains many genes that are linked to TCR signaling including NF-kB, calcineurin-NFAT and Ras signaling. The expression of almost all genes within this GEP was reduced in CD8 + T cells of SjD patients including S100B, CD37, CCR7 and RCAN3 . Some of the depleted genes overlap with those found in the naive CD4 + T cell signal transduction GEP including TCF7, LEF1 and RCAN3 . Because many of the depleted genes in the naive CD4 + T cell signal transduction GEP and the naive CD8 + T cell GEP regulate TCR signaling and function, our findings suggest that circulating naive T cells in the blood of SjD patients may be dysfunctional. Of note, several recent scRNA-seq based studies of SjD patients have identified CD8 + cytotoxic T cells (CD8 + CTLs) infiltrating the SGs of patients with SjD [ 26 , 30 , 32 , 53 , 94 ]. GZMK + CXCR6 + CD8 + T cells were also increased in the blood of SjD patients and had a gene signature similar to tissue-resident memory T cells in SGs [ 32 ]. By contrast, we did not observe increased cytotoxic CD8 + T cells numbers in the blood of patients with SjD. The causes underlying the differences between our findings and those in other reports may be related to the analysis of SjD patients belonging to different disease subgroups[ 97 ], patients with different disease activity or treatment histories. Like several of the other scRNA-seq studies, we observed a strong increase in the expression of IFN induced genes and IFN signaling pathways that was apparent in all immune cell populations we analyzed[ 24 , 27 , 28 , 30 ]. These genes included IFI44L, ISG15, STAT1 and others. While this increase is likely secondary to elevated levels of type I and type II interferons in SjD patients, we were not able to identify a single cell type as the source of IFN production in the blood of patients. A unique finding of our study, which may be linked to increased IFN responses, are strong changes in gene expression that point to perturbed protein homeostasis in immune cells of SjD patients. First, we observed an increased expression of PSM genes encoding proteasome subunits in CD4 + T cells and monocytes, potentially resulting in enhanced protein degradation in SjD patients. Second, we found a strong depletion of genes associated with pathways related to translational initiation and elongation. Broadly decreased expression of ribosomal proteins has been described in pDCs of SjD patients [ 17 ], but as the numbers of pDC within PBMC of our samples was very low, we were unable to replicate those findings. The reduced expression of translation associated genes in SjD patients in our study was specific to CD4 + T cells, CD8 + T cells and B cells and not observed in monocytes, DC or NK cells. The cause of this reduction is unclear but may be related to chronic activation of autoreactive T and B cells in SjD patients. Chronic T cell activation eventually leads to exhaustion, and exhausted CD8 + T cells were shown to downregulate a large number of translation associated genes[ 98 ]. We must note, however, that expression of canonical markers of T cell exhaustion including LAG3, PD1 or TIM3 was not altered in T cells of SjD patients. In addition, type I IFN, which are the first line of the cellular defense against viruses but are also increased in SjD, result in the upregulation of IFN-stimulated genes (ISG) that control the translation of viral and host mRNAs[ 99 ]. An alternative explanation for the depletion of translation associated genes is related our identification of three depleted GEPs in T and B cells of SjD patients, namely the naive CD4 + T cell signal transduction GEP, naive CD8 + T cell GEP and naive B cell GEP. This finding suggests a reduced function of these cells, which is intriguing because defects in signal transduction in T cells were shown to lead to reduced ribosomal transcription[ 100 ]. Thus, the signal transduction defects we observed in naive T and B cells may contribute to their ribosomal depletion. It is noteworthy that ribosomal deficiencies can have downstream effects beyond global reductions in translation because the targeted depletion of individual ribosomal proteins was shown to selectively modulate the translation of mRNAs related to specific T cell functions[ 101 ]. This could explain why we detected two GEPs associated with increased CD4 + T cell and B cell function, namely the Th1 activation and SjD B cell GEPs, despite the strong depletion of ribosome-associated pathways. An important question is how the GEPs we identified in PBMC of patients with SjD can be translated into clinical practice. For example, the SjD B cell GEP and the Th1 activation GEP (or genes within these GEPs) could be useful as biomarkers for clinicians to diagnose subgroups of patients with important implications for prognosis and treatment strategies. A recent study identified distinct subgroups of SjD patients characterized by B-cell active disease with low symptom burden (BALS), high systemic disease activity (HSA), and low systemic disease activity with high symptom burden (LSAHS) [ 97 ]. We speculate that SjD subgroups such as BALS and HSA could be associated with specific enrichment of the SjD B cell GEP and the Th1 activation GEP, respectively. The assignment of patients to distinct SjD subgroups using the GEPs described above may have important implications regarding the choice of therapeutic approaches. Patients with BALS may be more responsive to B cell focused therapies [ 102 – 104 ] whereas patients with HSA may respond better to treatment with JAK inhibitors that target IFN signaling, but which are not effective for all SjD patients[ 28 , 105 , 106 ]. The assignment of patients to SjD subgroups using GEPs may also facilitate prognosis because BALS subgroup patients may develop systemic SjD manifestations over time[ 97 ]. Our study has several limitations. 88% (8/9) of the SjD patients we analyzed are positive for anti-SSA/Ro antibodies, whereas 50–70% of patients with primary SjD are SSA + [ 107 ]. Our findings may therefore best reflect the immune dysregulation in SSA + patients rather than SjD as a whole. Moreover, our study focuses on gene expression in PBMCs, and not immune cells in the SGs of patients, which might provide a more direct picture of their functional state and thus, the immune dysregulation leading to gland inflammation and dysfunction. We did not have available paired PBMC and SG samples of the same SjD patients. Using scRNA-seq data from labial SGs of SjD patients from a recently published study[ 53 ], however, we were able to confirm that several of the GEPs we found to be dysregulated in PBMC of SjD patients were similarly affected in gland infiltrating CD4 + T cells and plasma cells. Although the patient cohorts in that and our study were not perfectly matched, for instance regarding their SSA + status (70% vs 88%, respectively), the similarities in GEP usage provide assurance that the changes we observed in PBMC of SjD patients reflect those in gland infiltrating immune cells. A more general limitation of all scRNA-seq based studies is that changes in gene expression do not fully predict the functional state of cells, and that key findings require further experimental validation. Supplementary Material 1 NIHMS2074107-supplement-1.docx (16.2MB, docx) HIGHLIGHTS. Machine learning identifies unique gene expression profiles (GEP) in PBMC of patients with SjD Increased frequencies of circulating naive B cells with hallmarks of B cell hyperactivation in SjD Enrichment of a Th1 cell activation GEP in SjD Acknowledgements. Patient data and PBMC used in this study are from the Sjögren’s International Collaborative Clinical Alliance (SICCA) Next Generation Studies, funded under grant #1U01DE028891 (and previously under contracts N01DE-32636 and #HHSN26S201300057C) by the National Institute of Dental and Craniofacial Research (NIDCR). This manuscript was prepared using a publicly available SICCA data set and does not necessarily reflect the opinions or views of the SICCA investigators, the NIH or NIDCR. We acknowledge technical support from the Genome Technology Center (RRID: SCR_017929), and the Cytometry and Cell Sorting Laboratory (RRID: SCR_019179) at NYU. Funding Sources. This study was funded by National Institutes of Health (NIH) grant R01DE027981 (to R.S.L. and S.F.), an award from the Colton Center for Autoimmunity at NYU (to S.F.), and Hunan Province Graduate Student Research and Innovation Project CX20190160 from Central South University, Changsha, Hunan, China (to W.L.). Footnotes Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Conflicts of Interest . S.F. is a scientific cofounder of CalciMedica; the other coauthors declare no conflict of interest. Data Availability. All raw sequencing data has been deposited in the Gene Expression Omnibus (GEO), accession number GSE253568 . REFERENCES [1]. Patel R, Shahane A. The epidemiology of Sjogren’s syndrome. Clin Epidemiol, 2014;6:247–55. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [2]. Vivino FB. Sjogren’s syndrome: Clinical aspects. Clinical Immunology, 2017;182:48–54. [ DOI ] [ PubMed ] [ Google Scholar ] [3]. Kyriakidis NC, Kapsogeorgou EK, Tzioufas AG. A comprehensive review of autoantibodies in primary Sjögren’s syndrome: Clinical phenotypes and regulatory mechanisms. Journal of Autoimmunity, 2014;51:67–74. [ DOI ] [ PubMed ] [ Google Scholar ] [4]. Tzioufas AG, Kapsogeorgou EK, Moutsopoulos HM. Pathogenesis of Sjögren’s syndrome: What we know and what we should learn. Journal of Autoimmunity, 2012;39:4–8. [ DOI ] [ PubMed ] [ Google Scholar ] [5]. Chatzis LG, Stergiou IE, Goules AV, Pezoulas V, Tsourouflis G, Fotiadis D et al. Clinical picture, outcome and predictive factors of lymphoma in primary Sjögren’s syndrome: results from a harmonized dataset (1981–2021). Rheumatology, 2021;61:3576–85. [ DOI ] [ PubMed ] [ Google Scholar ] [6]. Flament T, Bigot A, Chaigne B, Henique H, Diot E, Marchand-Adam S. Pulmonary manifestations of Sjögren’s syndrome. European Respiratory Review, 2016;25:110–23. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [7]. Björk A, Mofors J, Wahren-Herlenius M. Environmental factors in the pathogenesis of primary Sjögren’s syndrome. Journal of Internal Medicine, 2020;287:475–92. [ DOI ] [ PubMed ] [ Google Scholar ] [8]. Du W, Han M, Zhu X, Xiao F, Huang E, Che N et al. The Multiple Roles of B Cells in the Pathogenesis of Sjogren’s Syndrome. Front Immunol, 2021;12:684999. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [9]. Mielle J, Tison A, Cornec D, Le Pottier L, Daien C, Pers J-O. B cells in Sjögren’s syndrome: from pathophysiology to therapeutic target. Rheumatology, 2019;60:2545–60. [ DOI ] [ PubMed ] [ Google Scholar ] [10]. Verstappen GM, Kroese FGM, Bootsma H. T cells in primary Sjogren’s syndrome: targets for early intervention. Rheumatology (Oxford), 2021;60:3088–98. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [11]. Li B, Xing Y, Gan Y, He J, Hua H. Labial gland-derived mesenchymal stem cells and their exosomes ameliorate murine Sjögren’s syndrome by modulating the balance of Treg and Th17 cells. Stem Cell Research & Therapy, 2021;12. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [12]. Li J, Wu L, Chen Y, Yan Z, Fu J, Luo Z et al. Anticeramide Improves Sjögren’s Syndrome by Blocking BMP6-Induced Th1. Journal of Dental Research, 2023;102:93–102. [ DOI ] [ PubMed ] [ Google Scholar ] [13]. Shi L, Wang J, Guo H-X, Han X-L, Tang Y-P, Liu G-Y. Circulating Th2 cell reduction and Th1/Th2 imbalance are correlated with primary Sjogren’s syndrome-associated interstitial lung disease. Arthritis Research & Therapy, 2022;24. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [14]. Voigt A, Bohn K, Sukumaran S, Stewart CM, Bhattacharya I, Nguyen CQ. Unique glandular ex-vivo Th1 and Th17 receptor motifs in Sjögren’s syndrome patients using single-cell analysis. Clinical Immunology, 2018;192:58–67. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [15]. Verstappen GM, Kroese FGM, Bootsma H. T cells in primary Sjögren’s syndrome: targets for early intervention. Rheumatology, 2019;60:3088–98. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [16]. Ainola M, Porola P, Takakubo Y, Przybyla B, Kouri VP, Tolvanen TA et al. Activation of plasmacytoid dendritic cells by apoptotic particles – mechanism for the loss of immunological tolerance in Sjögren’s syndrome. Clinical and Experimental Immunology, 2017;191:301–10. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [17]. Hillen MR, Pandit A, Blokland SLM, Hartgring SAY, Bekker CPJ, van der Heijden EHM et al. Plasmacytoid DCs From Patients With Sjogren’s Syndrome Are Transcriptionally Primed for Enhanced Pro-inflammatory Cytokine Production. Front Immunol, 2019;10:2096. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [18]. Wildenberg ME, Welzen-Coppens JM, van Helden-Meeuwsen CG, Bootsma H, Vissink A, van Rooijen N et al. Increased frequency of CD16+ monocytes and the presence of activated dendritic cells in salivary glands in primary Sjogren syndrome. Ann Rheum Dis, 2009;68:420–6. [ DOI ] [ PubMed ] [ Google Scholar ] [19]. Zhou J, Zhang X, Yu Q. Plasmacytoid dendritic cells promote the pathogenesis of Sjogren’s syndrome. Biochim Biophys Acta Mol Basis Dis, 2022;1868:166302. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [20]. Brkic Z, Versnel MA. Type I IFN signature in primary Sjögren’s syndrome patients. Expert Review of Clinical Immunology, 2014;10:457–67. [ DOI ] [ PubMed ] [ Google Scholar ] [21]. Yao Y, Liu Z, Jallal B, Shen N, Rönnblom L. Type I interferons in Sjögren’s syndrome. Autoimmunity Reviews, 2013;12:558–66. [ DOI ] [ PubMed ] [ Google Scholar ] [22]. Witas R, Gupta S, Nguyen CQ. Contributions of Major Cell Populations to Sjögren’s Syndrome. Journal of Clinical Medicine, 2020;9:3057. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [23]. Cheng L, Wang Y, Wu R, Ding T, Xue H, Gao C et al. New Insights From Single-Cell Sequencing Data: Synovial Fibroblasts and Synovial Macrophages in Rheumatoid Arthritis. Frontiers in Immunology, 2021;12. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [24]. Hong X, Meng S, Tang D, Wang T, Ding L, Yu H et al. Single-Cell RNA Sequencing Reveals the Expansion of Cytotoxic CD4+ T Lymphocytes and a Landscape of Immune Cells in Primary Sjögren’s Syndrome. Frontiers in Immunology, 2021;11. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [25]. Zhang F, Wei K, Slowikowski K, Fonseka CY, Rao DA, Kelly S et al. Defining inflammatory cell states in rheumatoid arthritis joint synovial tissues by integrating single-cell transcriptomics and mass cytometry. Nature Immunology, 2019;20:928–42. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [26]. Chang L, Zheng Z, Xiao F, Zhou Y, Zhong B, Ni Q et al. Single-cell clonal tracing of glandular and circulating T cells identifies a population of CD9+ CD8+ T cells in primary Sjogren’s syndrome. J Leukoc Biol, 2024;115:804–18. [ DOI ] [ PubMed ] [ Google Scholar ] [27]. Cui Y, Zhang H, Wang Z, Gong B, Al-Ward H, Deng Y et al. Exploring the shared molecular mechanisms between systemic lupus erythematosus and primary Sjogren’s syndrome based on integrated bioinformatics and single-cell RNA-seq analysis. Front Immunol, 2023;14:1212330. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [28]. Gupta S, Yamada E, Nakamura H, Perez P, Pranzatelli TJ, Dominick K et al. Inhibition of JAK-STAT pathway corrects salivary gland inflammation and interferon driven immune activation in Sjogren’s disease. Ann Rheum Dis, 2024;83:1034–47. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [29]. Hou X, Hong X, Ou M, Meng S, Wang T, Liao S et al. Analysis of Gene Expression and TCR/B Cell Receptor Profiling of Immune Cells in Primary Sjogren’s Syndrome by Single-Cell Sequencing. J Immunol, 2022;209:238–49. [ DOI ] [ PubMed ] [ Google Scholar ] [30]. Huang J, Tang J, Zhang C, Liu T, Deng Z, Liu L. Single-cell transcriptomic analysis uncovers heterogeneity in the labial gland microenvironment of primary Sjogren’s syndrome. J Transl Autoimmun, 2024;9:100248. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [31]. Xiang N, Xu H, Zhou Z, Wang J, Cai P, Wang L et al. Single-cell transcriptome profiling reveals immune and stromal cell heterogeneity in primary Sjogren’s syndrome. iScience, 2023;26:107943. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [32]. Xu T, Zhu HX, You X, Ma JF, Li X, Luo PY et al. Single-cell profiling reveals pathogenic role and differentiation trajectory of granzyme K+CD8+ T cells in primary Sjogren’s syndrome. JCI Insight, 2023;8. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [33]. Bao S, Li K, Yan C, Zhang Z, Qu J, Zhou M. Deep learning-based advances and applications for single-cell RNA-sequencing data analysis. Briefings in Bioinformatics, 2022;23:bbab473. [ DOI ] [ PubMed ] [ Google Scholar ] [34]. Kotliar D, Curtis M, Agnew R, Weinand K, Nathan A, Baglaenko Y et al. Reproducible single cell annotation of programs underlying T-cell subsets, activation states, and functions. bioRxiv, 2024. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [35]. Daniels TE, Criswell LA, Shiboski C, Shiboski S, Lanfranchi H, Dong Y et al. An early view of the international Sjogren’s syndrome registry. Arthritis Rheum, 2009;61:711–4. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [36]. Shiboski CH, Baer AN, Shiboski SC, Lam M, Challacombe S, Lanfranchi HE et al. Natural History and Predictors of Progression to Sjogren’s Syndrome Among Participants of the Sjogren’s International Collaborative Clinical Alliance Registry. Arthritis Care Res (Hoboken), 2018;70:284–94. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [37]. Zheng GXY, Terry JM, Belgrader P, Ryvkin P, Bent ZW, Wilson R et al. Massively parallel digital transcriptional profiling of single cells. Nature Communications, 2017;8:14049. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [38]. Hao Y, Hao S, Andersen-Nissen E, Mauck WM, Zheng S, Butler A et al. Integrated analysis of multimodal single-cell data. Cell, 2021;184:3573–87.e29. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [39]. Wickham H. ggplot2. Use R!, 2016. [ Google Scholar ] [40]. Hafemeister C, Satija R. Normalization and variance stabilization of single-cell RNA-seq data using regularized negative binomial regression. Genome Biology: Cold Spring Harbor Laboratory; 2019. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [41]. Aran D, Looney AP, Liu L, Wu E, Fong V, Hsu A et al. Reference-based analysis of lung single-cell sequencing reveals a transitional profibrotic macrophage. Nature Immunology, 2019;20:163–72. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [42]. Schmiedel BJ, Singh D, Madrigal A, Valdovino-Gonzalez AG, White BM, Zapardiel-Gonzalo J et al. Impact of Genetic Polymorphisms on Human Immune Cell Gene Expression. Cell, 2018;175:1701–15 e16. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [43]. Jin S, Guerrero-Juarez CF, Zhang L, Chang I, Ramos R, Kuan C-H et al. Inference and analysis of cell-cell communication using CellChat. Nature Communications, 2021;12. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [44]. Kotliar D, Veres A, Nagy MA, Tabrizi S, Hodis E, Melton DA et al. Identifying gene expression programs of cell-type identity and cellular activity with single-cell RNA-Seq. eLife, 2019;8:e43803. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [45]. Evangelista JE, Xie Z, Marino GB, Nguyen N, Daniel JB Clarke, A. Ma’ayan. Enrichr-KG: bridging enrichment analysis across multiple libraries. Nucleic Acids Research, 2023;51:W168–W79. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [46]. Harris CR, Millman KJ, van der Walt SJ, Gommers R, Virtanen P, Cournapeau D et al. Array programming with NumPy. Nature, 2020;585:357–62. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [47]. Wolf FA, Angerer P, Theis FJ. SCANPY: large-scale single-cell gene expression data analysis. Genome Biol, 2018;19:15. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [48]. Stuart T, Butler A, Hoffman P, Hafemeister C, Papalexi E, Mauck WM et al. Comprehensive Integration of Single-Cell Data. Cell, 2019;177:1888–902.e21. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [49]. Finak G, Mcdavid A, Yajima M, Deng J, Gersuk V, Shalek AK et al. MAST: a flexible statistical framework for assessing transcriptional changes and characterizing heterogeneity in single-cell RNA sequencing data. Genome Biology, 2015;16. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [50]. Wu T, Hu E, Xu S, Chen M, Guo P, Dai Z et al. clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. The Innovation, 2021;2:100141. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [51]. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA et al. Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles. Proceedings of the National Academy of Sciences, 2005;102:15545–50. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [52]. Liberzon A, Subramanian A, Pinchback R, Thorvaldsdóttir H, Tamayo P, Mesirov JP. Molecular signatures database (MSigDB) 3.0. Bioinformatics, 2011;27:1739–40. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [53]. Pranzatelli TJF, Perez P, Ku A, Matuck B, Huynh K, Sakai S et al. GZMK+CD8+ T cells Target A Specific Acinar Cell Type in Sjogren’s Disease. Res Sq, 2024. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [54]. Megill C, Martin B, Weaver C, Bell S, Prins L, Badajoz S et al. cellxgene: a performant, scalable exploration platform for high dimensional sparse matrices. bioRxiv, 2021:2021.04.05.438318. [ Google Scholar ] [55]. Durinck S, Spellman PT, Birney E, Huber W. Mapping identifiers for the integration of genomic datasets with the R/Bioconductor package biomaRt. Nature Protocols, 2009;4:1184–91. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [56]. Monaco G, Lee B, Xu W, Mustafah S, Hwang YY, Carre C et al. RNA-Seq Signatures Normalized by mRNA Abundance Allow Absolute Deconvolution of Human Immune Cell Types. Cell Rep, 2019;26:1627–40 e7. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [57]. Hamza N, Bos NA, Kallenberg CGM. B-cell populations and sub-populations in Sjögren’s syndrome. La Presse Médicale, 2012;41:e475–e83. [ DOI ] [ PubMed ] [ Google Scholar ] [58]. Mingueneau M, Boudaoud S, Haskett S, Reynolds TL, Nocturne G, Norton E et al. Cytometry by time-of-flight immunophenotyping identifies a blood Sjogren’s signature correlating with disease activity and glandular inflammation. J Allergy Clin Immunol, 2016;137:1809–21 e12. [ DOI ] [ PubMed ] [ Google Scholar ] [59]. Fessler J, Fasching P, Raicht A, Hammerl S, Weber J, Lackner A et al. Lymphopenia in primary Sjogren’s syndrome is associated with premature aging of naive CD4+ T cells. Rheumatology (Oxford), 2021;60:588–97. [ DOI ] [ PubMed ] [ Google Scholar ] [60]. Mandl T, Bredberg A, Jacobsson LT, Manthorpe R, Henriksson G. CD4+ T-lymphocytopenia--a frequent finding in anti-SSA antibody seropositive patients with primary Sjogren’s syndrome. J Rheumatol, 2004;31:726–8. [ PubMed ] [ Google Scholar ] [61]. Szabó K, Jámbor I, Szántó A, Horváth IF, Tarr T, Nakken B et al. The Imbalance of Circulating Follicular T Helper Cell Subsets in Primary Sjögren’s Syndrome Associates With Serological Alterations and Abnormal B-Cell Distribution. Frontiers in Immunology, 2021;12. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [62]. Bodewes ILA, Björk A, Versnel MA, Wahren-Herlenius M. Innate immunity and interferons in the pathogenesis of Sjögren’s syndrome. Rheumatology, 2021;60:2561–73. [ DOI ] [ PubMed ] [ Google Scholar ] [63]. Sequi-Sabater JM, Beretta L. Defining the Role of Monocytes in Sjogren’s Syndrome. Int J Mol Sci, 2022;23. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [64]. Yamada A, Arakaki R, Kudo Y, Ishimaru N. Targeting IL-1 in Sjögren’s syndrome. Expert Opinion on Therapeutic Targets, 2013;17:393–401. [ DOI ] [ PubMed ] [ Google Scholar ] [65]. Fonseca BD, Lahr RM, Damgaard CK, Alain T, Berman AJ. LARP1 on TOP of ribosome production. WIREs RNA, 2018;9:e1480. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [66]. Hong S, Freeberg MA, Han T, Kamath A, Yao Y, Fukuda T et al. LARP1 functions as a molecular switch for mTORC1-mediated translation of an essential class of mRNAs. eLife, 2017;6. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [67]. Zhou H, Yang J, Tian J, Wang S. CD8(+) T Lymphocytes: Crucial Players in Sjogren’s Syndrome. Front Immunol, 2020;11:602823. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [68]. Verstappen GM, Kroese FGM, Bootsma H. T cells in primary Sjögren’s syndrome: targets for early intervention. Rheumatology, 2021;60:3088–98. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [69]. Li X, Li X, Qian L, Wang G, Zhang H, Wang X et al. T regulatory cells are markedly diminished in diseased salivary glands of patients with primary Sjogren’s syndrome. J Rheumatol, 2007;34:2438–45. [ PubMed ] [ Google Scholar ] [70]. Liu MF, Lin LH, Weng CT, Weng MY. Decreased CD4+CD25+bright T cells in peripheral blood of patients with primary Sjogren’s syndrome. Lupus, 2008;17:34–9. [ DOI ] [ PubMed ] [ Google Scholar ] [71]. Gottenberg JE, Lavie F, Abbed K, Gasnault J, Le Nevot E, Delfraissy JF et al. CD4 CD25high regulatory T cells are not impaired in patients with primary Sjogren’s syndrome. J Autoimmun, 2005;24:235–42. [ DOI ] [ PubMed ] [ Google Scholar ] [72]. Szodoray P, Papp G, Horvath IF, Barath S, Sipka S, Nakken B et al. Cells with regulatory function of the innate and adaptive immune system in primary Sjogren’s syndrome. Clin Exp Immunol, 2009;157:343–9. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [73]. Banica L, Besliu A, Pistol G, Stavaru C, Ionescu R, Forsea AM et al. Quantification and molecular characterization of regulatory T cells in connective tissue diseases. Autoimmunity, 2009;42:41–9. [ DOI ] [ PubMed ] [ Google Scholar ] [74]. Sarigul M, Yazisiz V, Bassorgun CI, Ulker M, Avci AB, Erbasan F et al. The numbers of Foxp3 + Treg cells are positively correlated with higher grade of infiltration at the salivary glands in primary Sjogren’s syndrome. Lupus, 2010;19:138–45. [ DOI ] [ PubMed ] [ Google Scholar ] [75]. Alunno A, Petrillo MG, Nocentini G, Bistoni O, Bartoloni E, Caterbi S et al. Characterization of a new regulatory CD4+ T cell subset in primary Sjogren’s syndrome. Rheumatology (Oxford), 2013;52:1387–96. [ DOI ] [ PubMed ] [ Google Scholar ] [76]. Christodoulou MI, Kapsogeorgou EK, Moutsopoulos NM, Moutsopoulos HM. Foxp3+ T-regulatory cells in Sjogren’s syndrome: correlation with the grade of the autoimmune lesion and certain adverse prognostic factors. Am J Pathol, 2008;173:1389–96. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [77]. Li XM, Li XP, Qian L, Wang GS, Zhang H, Zhei ZM et al. [Expression of CD4+ CD25+ regulatory T cells in peripheral blood and salivary gland of patients with primary Sjogren’s syndrome]. Zhonghua Yi Xue Za Zhi, 2007;87:1034–6. [ PubMed ] [ Google Scholar ] [78]. Furuzawa-Carballeda J, Hernandez-Molina G, Lima G, Rivera-Vicencio Y, Ferez-Blando K, Llorente L. Peripheral regulatory cells immunophenotyping in primary Sjogren’s syndrome: a cross-sectional study. Arthritis Res Ther, 2013;15:R68. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [79]. Wang YH, Li W, McDermott M, Son GY, Maiti G, Zhou F et al. IFN-gamma-producing T(H)1 cells and dysfunctional regulatory T cells contribute to the pathogenesis of Sjogren’s disease. Sci Transl Med, 2024;16:eado4856. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [80]. Wang J, Peng H, Tian J, Ma J, Tang X, Rui K et al. Upregulation of long noncoding RNA TMEVPG1 enhances T helper type 1 cell response in patients with Sjogren syndrome. Immunol Res, 2016;64:489–96. [ DOI ] [ PubMed ] [ Google Scholar ] [81]. Kroese FG, Abdulahad WH, Haacke E, Bos NA, Vissink A, Bootsma H. B-cell hyperactivity in primary Sjogren’s syndrome. Expert Rev Clin Immunol, 2014;10:483–99. [ DOI ] [ PubMed ] [ Google Scholar ] [82]. Brian B. F. t., Freedman TS. The Src-family Kinase Lyn in Immunoreceptor Signaling. Endocrinology, 2021;162:bqab152. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [83]. Leveille E, Chan LN, Mirza A-S, Kume K, Müschen M. SYK and ZAP70 kinases in autoimmunity and lymphoid malignancies. Cellular Signalling, 2022;94:110331. [ DOI ] [ PubMed ] [ Google Scholar ] [84]. Thomas D, Martin K, Antónia S, Jui-Cheng T, Athena SP, Ilona P et al. Efficacy and safety of remibrutinib, a selective potent oral BTK inhibitor, in Sjögren’s syndrome: results from a randomised, double-blind, placebo-controlled phase 2 trial. Annals of the Rheumatic Diseases, 2023:ard-2023–224691. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [85]. Chen YH, Wang XY, Jin X, Yang Z, Xu J. Rituximab Therapy for Primary Sjogren’s Syndrome. Front Pharmacol, 2021;12:731122. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [86]. Mariette X, Barone F, Baldini C, Bootsma H, Clark KL, De Vita S et al. A randomized, phase II study of sequential belimumab and rituximab in primary Sjogren’s syndrome. JCI Insight, 2022;7. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [87]. Brito-Zeron P, Baldini C, Bootsma H, Bowman SJ, Jonsson R, Mariette X et al. Sjogren syndrome. Nat Rev Dis Primers, 2016;2:16047. [ DOI ] [ PubMed ] [ Google Scholar ] [88]. Imam T, Park S, Kaplan MH, Olson MR. Effector T Helper Cell Subsets in Inflammatory Bowel Diseases. Front Immunol, 2018;9:1212. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [89]. Wagner CA, Roque PJ, Goverman JM. Pathogenic T cell cytokines in multiple sclerosis. J Exp Med, 2020;217. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [90]. Kang EH, Lee YJ, Hyon JY, Yun PY, Song YW. Salivary cytokine profiles in primary Sjogren’s syndrome differ from those in non-Sjogren sicca in terms of TNF-alpha levels and Th-1/Th-2 ratios. Clin Exp Rheumatol, 2011;29:970–6. [ PubMed ] [ Google Scholar ] [91]. Kolkowski EC, Reth P, Pelusa F, Bosch J, Pujol-Borrell R, Coll J et al. Th1 predominance and perforin expression in minor salivary glands from patients with primary Sjogren’s syndrome. J Autoimmun, 1999;13:155–62. [ DOI ] [ PubMed ] [ Google Scholar ] [92]. Mitsias DI, Tzioufas AG, Veiopoulou C, Zintzaras E, Tassios IK, Kogopoulou O et al. The Th1/Th2 cytokine balance changes with the progress of the immunopathological lesion of Sjogren’s syndrome. Clin Exp Immunol, 2002;128:562–8. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [93]. van Woerkom JM, Kruize AA, Wenting-van Wijk MJ, Knol E, Bihari IC, Jacobs JW et al. Salivary gland and peripheral blood T helper 1 and 2 cell activity in Sjogren’s syndrome compared with non-Sjogren’s sicca syndrome. Ann Rheum Dis, 2005;64:1474–9. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [94]. Kaneko N, Chen H, Perugino CA, Maehara T, Munemura R, Yokomizo S et al. Cytotoxic CD8(+) T cells may be drivers of tissue destruction in Sjogren’s syndrome. Sci Rep, 2022;12:15427. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [95]. Bertorello R, Cordone MP, Contini P, Rossi P, Indiveri F, Puppo F et al. Increased levels of interleukin-10 in saliva of Sjogren’s syndrome patients. Correlation with disease activity. Clin Exp Med, 2004;4:148–51. [ DOI ] [ PubMed ] [ Google Scholar ] [96]. Maehara T, Moriyama M, Hayashida JN, Tanaka A, Shinozaki S, Kubo Y et al. Selective localization of T helper subsets in labial salivary glands from primary Sjogren’s syndrome patients. Clin Exp Immunol, 2012;169:89–99. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [97]. Nguyen Y, Nocturne G, Henry J, Ng WF, Belkhir R, Desmoulins F et al. Identification of distinct subgroups of Sjogren’s disease by cluster analysis based on clinical and biological manifestations: data from the cross-sectional Paris-Saclay and the prospective ASSESS cohorts. Lancet Rheumatol, 2024;6:e216–e25. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [98]. Wherry EJ, Ha SJ, Kaech SM, Haining WN, Sarkar S, Kalia V et al. Molecular signature of CD8+ T cell exhaustion during chronic viral infection. Immunity, 2007;27:670–84. [ DOI ] [ PubMed ] [ Google Scholar ] [99]. Li MM, MacDonald MR, Rice CM. To translate, or not to translate: viral and host mRNA regulation by interferon-stimulated genes. Trends Cell Biol, 2015;25:320–9. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [100]. Tan TCJ, Knight J, Sbarrato T, Dudek K, Willis AE, Zamoyska R. Suboptimal T-cell receptor signaling compromises protein translation, ribosome biogenesis, and proliferation of mouse CD8 T cells. Proceedings of the National Academy of Sciences, 2017;114:E6117–E26. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [101]. Zhu X, Zhang W, Guo J, Zhang X, Li L, Wang T et al. Noc4L-Mediated Ribosome Biogenesis Controls Activation of Regulatory and Conventional T Cells. Cell Reports, 2019;27:1205–20.e4. [ DOI ] [ PubMed ] [ Google Scholar ] [102]. Thalayasingam N, Baldwin K, Judd C, Ng W-F. New developments in Sjogren’s syndrome. Rheumatology, 2021;60:vi53–vi61. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [103]. Verstappen GM, Kroese FG, Meiners PM, Corneth OB, Huitema MG, Haacke EA et al. B Cell Depletion Therapy Normalizes Circulating Follicular Th Cells in Primary Sjogren Syndrome. J Rheumatol, 2017;44:49–58. [ DOI ] [ PubMed ] [ Google Scholar ] [104]. Grigoriadou S, Chowdhury F, Pontarini E, Tappuni A, Bowman SJ, Bombardieri M. B cell depletion with rituximab in the treatment of primary Sjögren’s syndrome: what have we learnt? Clin Exp Rheumatol, 2019;37 Suppl 118:217–24. [ PubMed ] [ Google Scholar ] [105]. Charras A, Arvaniti P, Le Dantec C, Arleevskaya MI, Zachou K, Dalekos GN et al. JAK Inhibitors Suppress Innate Epigenetic Reprogramming: a Promise for Patients with Sjögren’s Syndrome. Clinical Reviews in Allergy & Immunology, 2020;58:182–93. [ DOI ] [ PubMed ] [ Google Scholar ] [106]. Gandolfo S, Ciccia F. JAK/STAT Pathway Targeting in Primary Sjögren Syndrome. Rheumatol Immunol Res, 2022;3:95–102. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] [107]. Fayyaz A, Kurien BT, Scofield RH. Autoantibodies in Sjögren’s Syndrome. Rheum Dis Clin North Am, 2016;42:419–34. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials 1 NIHMS2074107-supplement-1.docx (16.2MB, docx) Data Availability Statement All raw sequencing data has been deposited in the Gene Expression Omnibus (GEO), accession number GSE253568 . ACTIONS View on publisher site PDF (2.3 MB) Cite Collections Permalink RESOURCES Similar articles Cited by other articles Links to NCBI Databases Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top

Related documents

• Single-Cell RNA Sequencing Revealed the Role of Interferon-Gamma Related Genes in Primary Sjögren's Syndrome.#440760
• Computational systems biology approaches to unveil molecular interactions in Sjögren's disease pathogenesis#468350
• Utilising Machine Learning and Single-Cell Analysis to Uncover SKCM Metastasis-Related Genes.#3732
• Application of a machine learning model integrating T cell subsets and clinical markers in 28-Day mortality prediction of sepsis patients.#483390
• Cytokines and immunologic checkpoint molecules in predicting success of allergen immunotherapy.#209408