Differentially Expressed Gene Pathways in the Conjunctiva of Sjögren Syndrome Keratoconjunctivitis Sicca

Sjögren syndrome (SS) is an autoimmune condition that targets the salivary and lacrimal glands, with cardinal clinical signs of dry eye (keratoconjunctivitis sicca, KCS) and dry mouth. The conjunctiva of SS patients is often infiltrated by immune cells that participate in the induction and maintenance of local inflammation. The purpose of this study was to investigate immune-related molecular pathways activated in the conjunctiva of SS patients. Female SS patients (n=7) and controls (n=19) completed a series of oral, ocular surface exams. Symptom severity scores were evaluated using validated questionnaires (OSDI and SANDE). All patients fulfilled the ACR/EULAR criteria for SS and the criteria for KCS. Fluorescein and lissamine green dye staining evaluated tear-break-up time (TBUT), corneal and conjunctival disease, respectively. Impression cytology of the temporal bulbar conjunctiva was performed to collect cells lysed and subjected to gene expression analysis using the NanoString Immunology Panel. 53/594 differentially expressed genes (DEGs) were observed between SS and healthy controls; 49 DEGs were upregulated, and 4 were downregulated (TRAF5, TGFBI, KLRAP1, and CMKLRI). The top 10 DEGs in descending order were BST2, IFITM1, LAMP3, CXCL1, IL19, CFB, LY96, MX1, IL4R, CDKN1A. Twenty pathways had a global significance score greater or equal to 2. Spearman correlations showed that 29/49 upregulated DEGs correlated with either TBUT (inverse) or OSDI or conjunctival staining score (positive correlations). Venn diagrams identified that 26/29 DEGs correlated with TBUT, 5/26 DEGs correlated with OSDI, and 16/26 correlated with conjunctival staining scores. Five upregulated DEGs (CFB, CFI, IL1R1, IL2RG, IL4R) were uniquely negatively correlated with TBUT. These data indicate that the conjunctiva of SS patients exhibits a phenotype of immune activation, although some genes could be inhibitory. Some of the DEGs and pathways overlap with previous DEGs in salivary gland biopsies, but new DEGs were identified, and some of these correlated with symptoms and signs of dry eye. Our results indicate that gene analysis of conjunctiva imprints is a powerful tool to understand the pathogenesis of SS and develop new therapeutic targets.


INTRODUCTION
Patients with Sjögren syndrome (SS) develop an ocular surface disease called keratoconjunctivitis sicca (KCS). Pathological changes in the conjunctiva in KCS include altered epithelial differentiation with increased expression of cornified envelope precursors (1) that are found in the epidermis (involucrin, SPRR-2G) (2,3) and dysfunction and loss of the goblet cells that secrete gel-forming mucin that is essential for maintaining tear film stability (4). Evidence from clinical and mouse model studies indicates innate, and adaptive immune mediators and cells contribute to these pathological changes based on clinical correlation with disease severity and improvement following topical treatment with the immunomodulatory agent cyclosporine A (5)(6)(7)(8). Current methods to diagnose SS KCS include the use of dyes that stain conjunctival epithelial cells lacking a mucin coating or impression cytology to measure conjunctival goblet cell density, but they do not measure disease-relevant biomarkers. Additionally, the mechanism(s) by which this epithelial pathology develops is not entirely understood. An increased number of activated antigenpresenting cells (APCs) and increased interferon-g expression have been found in the conjunctival epithelium in KCS (1,6,(8)(9)(10). Interferon IFN-g disrupts cholinergic mediated secretion and causes unfolded protein response (UPR) and apoptosis of conjunctival goblet cells (11,12). Additional inflammatory mediators have been detected in the conjunctiva (1,13,14).
The purpose of this study was to compare the expression of immune-related molecular factors in the conjunctiva of normal subjects and SS KCS using the NanoString ® Immunology panel.

METHODS
The study was conducted following the Declaration of Helsinki, and the Baylor College of Medicine Institutional Review Board approved the protocol and informed consent form before study initiation. Written informed consent was obtained from all participants after explaining the purpose and possible consequences of the study. Female control subjects and patients with SS were enrolled in the study from January 2019 to January 2021. SS patients were recruited from the multispecialty SS clinic at Baylor College of Medicine (BCM) and had a complete ocular, oral, and rheumatologic evaluation, including a panel of serum autoantibodies, and met proposed the American College of Rheumatology/European League Against Rheumatism diagnostic criteria for SS (15).
Symptom assessment in dry eye (SANDE) and Ocular Surface Disease Index (OSDI) symptom questionnaires, fluorescein tear break-up time (TBUT), Schirmer I test, cornea fluorescein and conjunctival lissamine green dye staining, and tear meniscus height (TMH) measurement using optical coherence tomography (OCT) were performed as previously described (16,17). The ocular surface clinical parameters were all measured by the same observer (S.C.P.). Dry Eye Workshop (DEWS) criteria were used to grade clinical severity (18).
Healthy control subjects had no eye irritation, a TBUT ≥ 7 s, Schirmer 1 ≥ 10 mm, corneal fluorescein score ≤2, conjunctival lissamine score ≤ 2, and no meibomian gland disease. Subjects were excluded if they had prior laser-assisted in situ keratomileusis or corneal transplantation surgery, cataract surgery in the past year, punctal occlusion with plugs or cautery, a history of contact lens wear, use of topical medications other than preservative-free artificial tears, or chronic use of systemic medications known to reduce tear production. They were instructed not to instill any eye drops on the day of the evaluation.
All SS KCS patients had a severity score ≥of 3, and healthy controls had a severity score = 0.

Conjunctival Impression Cytology and RNA Extraction
Cells were obtained by impression cytology of the temporal bulbar conjunctiva using the EyePrim ™ device (Opia Tech, Paris, France) that applies a porous membrane with uniform pressure to the conjunctival surface. After collection, the membrane was placed in 0.5 mL RNA lysis buffer (Qiagen, Valencia, CA, USA) containing 1% 2-mercaptoethanol and immediately stored at -80°C. Total RNA was isolated using an RNeasy Mini Kit (Qiagen) following the manufacturer's protocol. The RNA was eluted in 15 mL RNase-free water-The RNA concentration was measured by its absorption at 260 nm using a spectrophotometer (NanoDrop 2000; Thermo Scientific, Wilmington, DE, USA) and Agilent Bioanalyzer.

NanoString ® and Data Analysis Using ROSALIND ® NanoSstring Gene Expression Methods
Five hundred ninety-four transcripts were quantified with the NanoString ® nCounter multiplexed target platform using the Human Immunology V2 panel (www.nanostring.com). nCounts of mRNA transcripts were normalized using the geometric means of 15 housekeeping genes (ABCF1, ALAS1, EEF1G,  G6PD, GAPDH, GUSB, HPRT1, OAZ1, POLR1B, POLR2A,  PPIA, RPL19, SDHA, TBP, TUBB). Data were analyzed by ROSALIND ® (https://rosalind.onramp.bio/), with a HyperScale architecture developed by ROSALIND, Inc. (San Diego, CA). Read Distribution percentages, violin plots, identity heatmaps, and sample MDS plots were generated as part of the QC step. Normalization, fold changes, and p-values were calculated using criteria provided by NanoString ® . ROSALIND ® follows the nCounter ® Advanced Analysis protocol of dividing counts within a lane using the same lane's geometric mean of the normalizer probes. Housekeeping probes to be used for normalization are selected based on the geNorm algorithm as implemented in the NormqPCR R library (19). The abundance of various cell populations is calculated on ROSALIND using the NanoString ® Cell Type Profiling Module. ROSALIND performs filtering of Cell Type Profiling results to include scores with a pvalue lower than or equal to 0.05. Fold changes and P-values are calculated using the fast method described in the nCounter ® Advanced Analysis 2.0 User Manual. The Gene Set Analysis (GSA) module from NanoString ® is incorporated into ROSALIND to summarize the global significance score and the directed global significance score. GSA summarizes the change in regulation within each defined gene set relative to the baseline, as described in the nCounter ® Advanced Analysis 2.0 User Manual. P-value adjustment is performed using the Benjamini-Hochberg method of estimating false discovery rates (FDR). Clustering genes for the final heatmap of differentially expressed genes was done using the PAM (Partitioning Around Medoids) method using the FPC R library (20) that considers the direction and type of all signals pathway the position, role, and type of every gene. Hypergeometric distribution was used to analyze the enrichment of pathways, gene ontology, domain structure, and other ontologies. The topGO R library (20) was used to determine local similarities and dependencies between GO terms to perform Elim pruning correction. Several database sources were referenced for enrichment analysis, including Interpro (21), NCBI5 (22), MSigDB (23,24), REACTOME (25), WikiPathways (26). Enrichment was calculated relative to a set of background genes relevant for the experiment. Data analyzed in ROSALIND was downloaded, and Volcano plots of differential expression data were plotted using the −log10 (P-value) and log2 fold change using GraphPad Prism (Version 9, GraphPad Software, San Diego, CA). Heatmaps were also constructed using GraphPad Prism. Venn diagrams were made using Venny 2.1 (27). Over-represented pathway investigation was performed using www.innatedb.com (28).

Statistical Analysis
Comparisons of OSDI, SANDE, TMH, TBUT, corneal and conjunctival scores between HC and SS subjects were performed using non-parametric Mann Whitney U test using GraphPad Prism version 9.0 (GraphPad Sofware). Nonparametric Spearman correlations using GraphPad Prism investigated the correlations of OSDI, TBUT, corneal and conjunctival score with up or down-regulated DEGs.

Patient Demographics
Nineteen control subjects and seven age-and-sex-matched SS KCS patients were enrolled from January 2019 to January 2021 ( Table 1). The control subjects had minimal eye irritation symptoms and no clinical signs of KCS, while SS KCS patients had significant symptoms and clinical signs. Statistical comparison of OSDI and SANDE scores, tear meniscus height (TMS), tearbreak-up time (TBUT), corneal and conjunctival staining scores between controls and SS KCS are shown in Table 2.

Gene Expression Analysis in the Conjunctiva of SS KCS Identifies Upregulated Targets
The conjunctiva of SS KCS shows epithelial metaplasia, goblet cell loss, T cell infiltration, and increased expression of HLA-DR and IFNG mRNA (6, 10, 29-32). We and others have shown that impression cytology of the conjunctiva collects a mix of epithelial and immune cells (8,14,33). To gain insights into the molecular mechanisms involved locally, impression cytology of the conjunctiva was used to collect conjunctival cells, which were then lysed and processed for RNA analysis using a NanoString ® nCounter panel enriched for immune genes (Immunology V2). This panel allows for the simultaneous evaluation of 594 genes. Gene expression analysis was quantified in the conjunctiva of healthy and SS KCS patients using nCounter and ROSALIND software, as described in methods. Out of the 594 genes, eight genes were not detected; 53 were significantly modulated in SS KCS patients: 4 genes were downregulated (TRAF5, TGFBI, KLRAP1, and CMKLRI), and 49 genes were upregulated using a stringent 1.5-fold increase or decrease threshold (volcano plot in Figure 1A). Significance was plotted against fold change (log2 values) using a -log10 of the adjusted p-value for each dot. The top 20 upregulated differentially expressed genes (DEGs) in descending order were BST2, IFITM1, LAMP3, CXCL1, IL19, CFB, LY96, MX1, IL4R, CDKN1A, SERPING1, HLA-DRB3,  S100A8, IRF7, ICAM1, C4A/B, TNFAIP3, CD74, HLA-B, and TAP1 (Table 3). A complete list of DEGs can be found in Supplementary Table 1.

Differential Pathways Are Activated in the Conjunctiva of SS KCS Patients
Previous studies have identified a significant role for inflammation in the conjunctiva of SS KCS patients (6,13,14,34). To gain insight into the pathways involved, we analyzed which DEGs have been annotated or identified in immune pathways using the ROSALIND software and nCounter software, using the gene set analysis. This algorithm calculates a T-statistic for each gene against each covariate in the model. A gene set's global significance score for a covariate measures the cumulative evidence for the differential expression of specific genes in a pathway. Out of the 32 immune pathways included in the panel, the top 20 pathways are shown in Figure 1B (2). As expected, not all pathways in the panel were identified in our results. Some DEGs overlapped in more than one pathway. For example, the proteasome subunits beta type-8 (PSMB8) and type-9 (PSMB9) were annotated in multiple pathways (type I interferon, MHC Class I antigen presentation, B cell receptor signaling, cytokine signaling, innate immune system, adaptive immune system, NFkB, immunometabolism, and apoptosis). Both PSMB8 and PSMB9 are integral parts of the immunoproteasome and, together with TAP1 and TAP2, are involved in the MHC class I presentation. TAP1 and TAP2 are transporters (-1,-2, respectively) and ATP binding cassette subfamily B members. Similar to PSMB8 and PSMB9, TAP1 and TAP2 were also annotated in multiple pathways (MHC class I presentation, phagocytosis, host-pathogen interaction, adaptive immune response, Figures 2-4). However, other DEGs were uniquely annotated. An example is LAMP3 (lysosome-associated membrane glycoprotein 3 or DC-LAMP), which was uniquely annotated in the phagocytosis pathway.
It has been demonstrated that SS patients have a Type I and II Interferon signature in their serum (35)(36)(37)(38). Accordingly, we observed many DEGs involved in Type I and II interferon responses in the conjunctiva, such as BST2, IFITM1, IRF7, and MX1 ( Figure 2). BST2 (bone marrow stromal antigen 2) was the most upregulated DEG in our results. Elevated BST2 levels are found in labial gland SS biopsies compared to controls and these levels correlated with rheumatoid factor and B2M serum levels (39). IFITM1 (interferon-induced transmembrane protein 1), IRF7 (interferon regulatory factor 7), and MX1 (interferoninduced GTP-binding protein Mx1) are also interferonresponsive genes (40) that were upregulated with greater significance in our results. IFITM1 was the second most upregulated DEGs in our results, and it is involved in Type I interferon and B cell receptor signaling, which also showed upregulation of PSMB8 and PSMB9.
The complement system participates in the lysis of infectious organisms by facilitating the binding of antibodies and phagocytic cells to microorganism membranes. It also promotes inflammation and immune clearance. There were five upregulated DEGs in our results, as shown in Figure 3 (CFB, CFI, C2, C4A/B, and SERPING1). CFB, a component of the alternative pathway of complement, is also increased in cerebrospinal fluid of patients with SS and fatigue (41). Both C2, C4A, and C4B proteins are known to participate in the generation of the classical complement pathway. The complement pathway is also part of the innate immune system, which showed many DEGs, such as CEACAM-1, CXCL1, S100A8, S100A9, PIGR, and MUC1. CEACAM-1 (carcinoembryonic antigen-related cell adhesion molecule 1) is a co-inhibitory molecule in mucosal immunology (42). CXCL1 is an antimicrobial protein that is a chemoattractant for neutrophils, which have been implicated in dry eye (43,44). PIGR is a polymeric immunoglobulin receptor (also known as secretory component) that mediates polymeric IgA and IgM transport across mucosal epithelial cells. It is highly expressed in the lacrimal gland and tears (45) and, once secreted, can act as an antimicrobial agent on its own (46). The involvement of the major inflammatory pathway NFkB in SS is well established in the literature (47,48). Figure 3 shows the DEGs involved in this pathway: ICAM-1, IL1R1, Ly96, TNFAIP3 (encoding tumor necrosis factor, alpha-induced protein 3 or A20 protein), TNFSF13B, and downregulation of TRAF5. TNFSF13B encodes BAFF, B cell activation factor, a protein involved in B cell survival. BAFF SNPs and mutations have been shown in SS patients (49,50). TRAF5, TNF receptor-associated factor 5, is a scavenger protein involved in TNF and NFkB activation and mediates IL-17-ACT1 interactions. TRAF5 SNPs showed an association with rheumatoid arthritis compared to control subjects (51). Genome-wide association studies have shown that TNFA1P3 is upregulated in Chinese SS patients (52). TNFAIP3, HLA-DMA, HLA-DR genes (HLA-DRB5 and HLA-DRB1), CDKN1B (cyclin-dependent kinase inhibitor 1B), IFITM1, NKFB1 (nuclear factor of kappa light polypeptide gene enhancer in B-cells 1), and IRF7 also elevated in labial salivary biopsies (53). Other small molecules and drugs that target the NFkB pathway in SS are undergoing clinical trials [reviewed in (54)].
The immunometabolism pathway had two DEGS; CMKLR1 and RARRES3 (-7.1-and 1.6-fold change, respectively, Figure 4). CMKLR1 encodes chemerin chemokine-like receptor 1, also known as resolvin E1 receptor, a critical molecule in inflammation resolution (63). A CMKLR1 agonist antibody has shown promise in pre-clinical models of intestinal mucosal inflammation (64). RARRES3, retinoic acid receptor responder protein 3, is an intracellular retinoid receptor responder that belongs to RNA pattern recognition receptors. RARRES3 has been shown to interact with the immunoproteasome and upregulate IRF1 in breast cancer cells, and to stimulate the proliferation of keratinocytes (65).
TGBI and ICAM-1 were differentially modulated in our dataset and are annotated in the cell adhesion pathway (down and upregulated DEGs, respectively, Figure 4). Mutations of TGFBI (transforming growth factor beta-induced) have been implicated in cornea dystrophies (66). ICAM-1, intercellular adhesion molecule-1, is a critical molecule in the adaptive immune system that facilitates cell adhesion and promotes cell adhesion. ICAM-1 is elevated in SS serum and mouse models of dry eye disease (67)(68)(69)(70). Lifitegrast, which blocks the immunological synapse of lymphocyte function-associated antigen/ICAM-1, is an FDA-approved eye drop for dry eye disease (71)(72)(73). Pathway overrepresentation analysis using the InnateDB tool (28) showed that the "Antigen processing and presentation" pathway was over-represented in our results (10% of genes, P<0.00001, p-adjusted value), using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. Genes involved in this pathway were B2M, CD74, HLA-A, HLA-B, HLA-DMA, TAP1, and TAP2. These same genes were annotated in Type I and II interferon responses, MHC class I and II antigen presentation, phagocytosis, and degradation in the NanoString ® pathways ( Figure 2). B2M (beta-2 microglobulin), involved in MHC Class I antigen presentation to CD8 T cells, has shown to be a hub for protein interaction in SS (74). HLA-A and HLA-B are major histocompatibility complexes that form a heterodimer with B2M.
These results demonstrate the pleiotropic abilities of specific genes, which participate in multiple pathways.

Correlation of Upregulated DEGs With Symptoms Severity Score and Signs
We performed Spearman correlation analysis between the normalized counts vs. OSDI, or TBUT, or corneal staining score, or conjunctival score to investigate which DEGs  (Table 4), and some DEGs correlated with more than one clinical sign. Only upregulated DEGs showed any significant correlation. Two representatives Spearman correlation graphs are shown in Figure 5A: 1) CD74, a critical step in MHC II pathway (75), inversely correlated with TBUT and positively with conjunctival staining score, 2) LAMP3, a lysosomal protein involved in autophagy (76), which correlated with OSDI, TBUT, and conjunctival staining score. Increased levels of CD74 have been previously reported in minor salivary gland SS biopsies (62,77).
Next, we used the data in Table 4 to make Venn diagrams to identify the unique and shared correlations ( Figure 5B). We identified that 26/29 DEGs showed an inverse correlation with TBUT; 5 of these 26 DEGs showed a positive correlation with OSDI, and 16/26 showed a positive correlation with conjunctival staining score. Five upregulated DEGs (CFB, CFI, IL1R1, IL2RG, IL4R) were uniquely inversely correlated with TBUT, while PML (promyelocytic leukemia protein) was uniquely correlated with conjunctival staining score ( Table 4). PML is upregulated by STAT1 and participates in TNF-a and IFN-a induction of angiogenesis (78). Detection of PML in liver biopsies has been used as a diagnostic tool for autoimmune liver diseases, including primary biliary cirrhosis (79,80). In addition to LAMP3, only an additional 6 DEGs correlated with OSDI (CDKN1A, PRDM1, HLA-B, MX1, PSMB8, and TMEM173). The disconnection between symptoms and signs in SS KCS is well described in the literature.

DISCUSSION
The pathogenesis of SS has been a subject of intense investigation. The cornea and the conjunctiva develop pathological changes in SS KCS (81). Because it is not possible to obtain cells from the cornea noninvasively, we performed multiplex gene analysis of cells obtained from the conjunctiva using impression cytology using an Immunology panel. Impression cytology of conjunctiva has been used to investigate conjunctival goblet cell density, epithelial metaplasia, and obtain cells for flow cytometry (10,(82)(83)(84)(85)(86)(87). It has also been used to investigate gene expression between different forms of dry eye disease and comparisons between healthy controls and SS KCS, as in our studies (6,13,14).
A careful investigation of pathways in our immunology panel confirmed some findings in literature but also unraveled new pathways and DEGs that had not been previously associated with SS. Pathways that have been previously associated with SS include Type I and Type II Interferon, MHC Class I and II presentation, Innate immune system, Phagocytosis, NFkB, B cell receptor to name a few. Increased expression of the interferoninducible genes is consistent with our findings of increased IFN-g expression in the conjunctiva of patients with SS KCS (1,6) and that interferon-g produced by NK and T cells increases chemokine expression and promotes goblet cell loss in conjunctival of mouse SS models (11,(88)(89)(90)(91)(92). Interestingly, the B cell receptor signaling pathway was upregulated in our results involving the conjunctiva. The involvement of B cells in SS is well established, as SS patients have B cell infiltration in salivary and lacrimal glands in humans and rodents. SS patients have an increased risk for non-Hodgkin lymphomas (93)(94)(95)(96). The role of B cell receptor signaling in the conjunctiva of SS KCS patients has not been thoroughly investigated.
There were several upregulated genes in the host-pathogen, NLR signaling, and TLR signaling pathways. While no ocular surface pathogen has been identified in the pathogenesis of SS, there is a link between intestinal dysbiosis (with increased proteobacteria genera) and the severity of SS KCS (97,98). Additionally, antibiotic-induced dysbiosis in mice increased the ocular surface inflammatory response to topically applied lipopolysaccharide (99) and led to a worse dry eye phenotype after desiccating stress (97). Upregulated genes in these pathways can heighten the TLR signaling (e.g., LY96) and innate and adaptive immune responses to microbial products that can be released into the circulation with intestinal dysbiosis. Several of the upregulated DEGs (IL-4R, IL-19, and DUSP) may have compensatory immunosuppressive functions. Th2 cytokines, such as IL-4 and IL-13, are usually implicated in allergic conjunctival inflammation (100), and IL-13 signaling through its heterodimeric receptor, which includes the IL-4R, stimulates goblet cell proliferation and mucin production in the conjunctiva (101,102). Similarly, IL-19 is an anti-inflammatory cytokine that belongs to the IL-10 family and participates in the modulation of macrophage from M1 to M2, and DUSP4 negatively regulates phosphorylation of several MAPKs (103)(104)(105).
One pathway that was discovered in our panel was immunometabolism, with two unique DEGs (CMKLR1 and RARRES3). CMLKR1 also participates in the resolution of inflammation. Topically applied resolvin E1 analogs improved corneal staining, goblet cell density, and improved tear secretion while decreasing CD11b + and CD4 + infiltration to the cornea in a  murine model of dry eye disease (106,107). A protective role for resolvin D has been shown in the salivary gland of SS animal models (108). Interestingly, RARRES1, another member of the same family, is elevated in SS labial salivary gland biopsy microarray studies (53). This is the first time that RARRES3 is associated with the conjunctiva of SS KCS patients.
Our results showing increased XBP-1 and STAT1 levels agree with the literature and with our previous publication that showed increased XBP-1 and GRP78 in the SS KCS conjunctiva compared to healthy controls (11). Increased levels of STAT1 have been previously reported in humans and animal models, also in the conjunctiva (13,14,62,109). XBP-1 (X-box binding protein-1) participates in ER stress and UPR, decreasing salivary gland protein secretion (110). In the conjunctiva, XBP-1 participates in the IFN-g-induced UPR by decreasing Muc5ac protein secretion from cultured conjunctival goblet cells (11).
Correlation analysis of the DEGs with clinical signs identified that 53% of DEGs correlated with at least one clinical sign. Immune Pathways in SS KCS IFITM1, the second most upregulated DEG in our results, correlated with TBUT and conjunctival staining score. Elevated IFITM1 levels have been reported in the conjunctiva of dry eye patients (111). PSMB8 correlated with OSDI, TBUT, and conjunctival staining, while TAP1 and TAP2 correlated with TBUT and conjunctival staining. Epigenetic changes in PSMB8 and TAP1 have been reported in human labial gland biopsies in SS patients from the SICCA repository (112). Increased mRNA and protein expression in PSMB8 and PSMB9 in epithelial and immune cells of salivary glands have been reported (77,113). LAMP3 was one of the only five DEGs correlated with three ocular parameters (OSDI, TBUT, and conjunctival staining score). LAMP3 is a marker of mature DCs in humans (114), and we and others have shown that the conjunctiva of SS KCS patients is rich in antigen-presenting cells (8,33,115,116). LAMP3+ patients have increased levels of autoantibodies in their serum, and in-vitro, overexpression of LAMP3 induces apoptosis of minor salivary gland epithelial cells (117). Adenovirus transfection with LAMP3 in mice induces an SS-like phenotype accompanied by decreased salivary function (118). CDKN1A, cyclin-dependent kinase inhibitor 1, p21, correlated with OSDI and conjunctival staining score. Elevated senescence markers CDKN1A (119) and CDKN1B have been found in minor salivary gland SS biopsies (62). Genetic deletion of p21 in lupus-prone mice promoted apoptosis of long-lived T and B cells, decreasing autoimmunity (120). Our results are the first to show elevated CDKN1A, PSMB8, and PSMB9 in the conjunctiva of SS patients. Our studies agree with Kessal and colleagues, who reported increased C2 and CFB (complement factor B) in the conjunctiva of SS KCS patients (13,14). The complement system is activated at the ocular surface at low levels, where it is thought to participate in immunosurveillance (121). However, functional studies of the complement system in SS are primarily concentrated in the serum, where decreased complement levels have been found (122). Variants in C4 genes (which segregate with MHC II genes) increase the odds ratio of SS (123). A systematic review in the literature showed that a deficiency of C1-inhibitor (encoded by SERPING1) was present in autoimmune diseases (124). Our study adds CFI (complement factor I), C4/B, and SERPING1 to the list of upregulated genes in the complement cascade in the conjunctiva of SS KCS patients. The role of the complement cascade in the conjunctiva warrants further investigation. Antigen presentation and processing, including MHC class I and class II pathways, were overrepresented in our results. Elevated HLA-A, HLA-B, HLA-DMA levels, together with other HLA-DR genes (HLA-DRB5 and HLA-DRB1), have been found in labial salivary biopsies in SS patients (53,77). Genomewide studies have shown that DEGs TAP1, TAP2, PSMB9, HLA-DR, HLA-DMA have been associated with SS in more than one study (52,125,126). LAMP3 was the third most upregulated DEG in our results, in agreement with previous studies that also reported elevated LAMP3 levels in SS (62,117). MX1 has also been found elevated in other cohorts of SS patients, including conjunctiva (13,14), labial salivary biopsies (53,62,127,128), and serum (129). When comparing our DEGs with previous studies that performed comprehensive gene analysis of the conjunctiva, our studies identified four overlapping DEGs (MX1, C2, CFB, STAT1, STAT2) (13,14). Compared to studies that performed gene analysis in labial gland biopsies from SS patients, our results identified some overlapping DEGs (62,74,77,113,127,128). Unique DEGs not previously associated with SS were HLA-DRB3, CFI, CEACAM1, CXCL1, DUSP4, IL19, IL1R1, MUC1, SERPING1, S100A8, S100A9, PIGR, XBP1, although some proteins encoded by these DEGs have been associated with dry eye [XBP1, HL-A-DR (11,33,115)]. The majority of these DEGs participate in Innate immune response. Some of these proteins are secreted by epithelial cells and demonstrate the importance of the epithelial barrier in autoimmunity (130). Elevated levels of PIGR have been found in the SS saliva (131). MUC1, mucin 1, has been shown to regulate inflammation and TLR signaling after infections in mucosal sites (132). SNPs in the MUC1 gene have been identified in SS and dry eye subjects (133)(134)(135). Interestingly, 5/50 upregulated genes in the conjunctiva of a mouse model of SS were also included in our upregulated DEGs [MX1, ICAM1, S100A9, IRF7, BATF, (136)].
The strengths of this study include robust gene expression data via NanoString and comprehensive measures of signs and symptoms of dry eye disease to allow correlation of gene expression levels with different objective measures of the disease. The limitations of our study include the relatively limited set of genes evaluated by this NanoString ® platform and the restriction to immune-related genes as compared to the more open approach of RNA sequencing, though the use of the NanoString platform allowed for a direct focus on immunerelated genes, which was the purpose of the study.
Further studies are necessary to validate these DEGs identified in our study, which identified shared and unique DEGs and pathways in the conjunctiva of SS KCS patients. These findings highlight that some therapies targeting immune mediators might be efficacious for SS KCS. The ocular surface also has unique activated pathways that warrant further investigation if these DEGs can be used as valid biomarkers modulated by therapies such as cyclosporine A and short-term steroids. Investigation of immune pathways in the conjunctiva might yield novel therapeutic targets for SS KCS.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by Baylor College of Medicine Institutional Review Board. The patients/participants provided their written informed consent to participate in this study.

AUTHOR CONTRIBUTIONS
CP, CT-V, LS, SP, ZY, and RB were involved in the conception and design of the study. CP, CT-V, LS, SP, and ZY were involved in data acquisition. CP, CT-V, LS, SCP, ZY, and RB were involved in data analysis and interpretation. CP drafted the manuscript. All authors contributed to the article and approved the submitted version.

ACKNOWLEDGMENTS
This work could not have been done without the voluntary participation of healthy subjects and SS patients. We acknowledge support from the Genomic and RNA Profiling Core at Baylor College of Medicine.