Background
Sjögren’s syndrome (SS) is a systemic rheumatic autoimmune disease, where chronic inflammation results in progressive destruction of exocrine glands, primarily the lacrimal and salivary glands [
1,
2]. Thus, characteristic features are sicca symptoms, including dry eyes and dry mouth [
3]. The prevalence of SS has been reported to be between 0.01% and 0.6% [
4‐
6].
The main classification criteria used today when diagnosing primary SS (pSS) are the American-European Consensus Group (AECG) criteria from 2002 [
7], which rely on evaluating symptoms of ocular and oral dryness, assessing the secretory ability of the exocrine glands, screening for anti-Ro and anti-La autoantibodies, and evaluating biopsies of minor salivary glands for mononuclear cell infiltration [
8]. This routine assessment of minor salivary gland tissue and histological focus scoring has been employed to describe salivary gland involvement in SS [
9,
10]. Here, a positive biopsy with mononuclear cell infiltrates comprising ≥50 mononuclear cells per 4 mm
2 resulted in a positive focus score value ranging from 1 to 12 according to the number of foci seen. This is a semi-quantitative, invasive technique useful for patients with glandular dysfunctions without autoantibody production [
11].
Considering the nature of the currently available diagnostic tools, there remains an unmet need for non-invasive, more accurate diagnosis of pSS. The incorporation of additional non-invasive diagnostics, such as screening for disease-specific biomarkers [
12,
13] has therefore been in focus over recent decades, as it can also be applied for staging and monitoring of the disease. Indeed, liquid chromatography-mass spectrometry (LC-MS) has been applied in several human rheumatic diseases, including SS, in order to discover biomarkers and therapeutic targets by studying the proteome of biological fluids [
14,
15]. Both saliva [
14,
16‐
21] and tear fluid [
22,
23] have previously been used to identify potential biomarkers for SS. It has been reported that oral fluid not only reflects the salivary gland involvement that characterises SS disease [
18,
24,
25], but also has the potential to represent the subject’s current general health [
26,
27]. Moreover, salivary fluid samples can easily be obtained using a non-invasive, simple, safe, and stress-free procedure, allowing for repetition and multiple collections. This explains why the majority of proteomic studies of SS have chosen saliva as the ideal biological fluid, examining either whole saliva or saliva from individual glands (e.g. minor and/or parotid salivary glands), under both stimulated and unstimulated conditions [
14,
16‐
21]. As a result, several common biomarkers for SS have been found, including secretory proteins, enzymes, highly abundant immune-system-related molecules (e.g. β2-microglobulin), and cytokines such as IL-4 and IL-5 [
21,
28,
29].
Proteomic analyses can also be coupled with various separation techniques in order to isolate the cellular components of interest when screening for disease biomarkers. Extracellular vesicles (EVs) are an example of such cellular components. These are membrane-embedded vesicles, comprising exosomes (size <100 nm) and/or microvesicles (size 100–1000 nm) [
30], released by cells that are emerging as important mediators of intercellular communication, and thereby influencing recipient cell functions [
31‐
33]. For instance, EVs can act on the innate immune system as paracrine messengers and have been described as pro-inflammatory mediators that induce inflammatory signals during infections [
34,
35] and chronic inflammatory diseases [
35].
Interestingly, patients with autoimmune diseases have increased levels of EVs that carry components associated with complement activation [
36,
37]. Accordingly, various cell types of the innate immune system are known to release EVs, including macrophages [
38], monocytes or dendritic cells [
39] and natural killer (NK) cells [
40]. Besides mediating the exchange of intercellular information by their surface molecules, EVs have been shown to be carriers of important soluble mediators, such as cytokines. The involvement of EVs in the transport of the cytokines IL-1b [
41] and tumour necrosis factor (TNF) [
42] are such examples.
Proteomic profiling of EVs in a biological context can be challenging, especially if the EV preparations are not highly purified [
43]. In complex body fluids, EVs can be separated from interfering molecules, such as proteins and lipids, by utilising size-exclusion chromatography [
44,
45]. The isolated sub-fractions containing the highest EV concentrations can then be characterised using nanoparticle tracking analysis, and by flow cytometry detection of the fusogenic protein/tetraspanin CD9, which is abundantly expressed in EVs [
46‐
48].
Proteomic studies of isolated EVs have in turn yielded extensive catalogues that display which proteins are abundant in different types of EVs, specifically reflecting vesicle localization, cellular origin, and mechanism of secretion [
49]. Hence, in the current study we hypothesised that by applying LC-MS alone, and in combination with EV-isolation, using samples of stimulated whole saliva and tear fluid from patients with pSS and healthy controls, novel biomarkers may be identified encompassing both salivary and lacrimal disease target organs. Such biomarkers may in turn be implemented, as potential non-invasive diagnostic tools that can help to increase diagnostic accuracy when evaluating patients with pSS, in accordance with the AECG criteria, and can also be useful when monitoring disease progression.
Discussion
By studying the proteome of biological fluids through LC-MS approaches, and possibly combined with the chromatographic separation of extracellular vesicles from other biomolecules, the search for potential biomarkers and therapeutic targets can be realised for SS. Such biomarkers can in turn be implemented as potential non-invasive diagnostic tools that can be applied for monitoring disease progression. The majority of proteomic studies of SS have been based on saliva as the biological fluid, using different mass spectrometry approaches, in addition to genomics [
14,
16‐
21]. This confirms that saliva not only reflects the salivary gland involvement that characterises the disease process in SS [
18,
24,
25], but additionally has the potential to communicate an individual’s current health [
26,
27].
So far tear fluid has only been used to identify potential biomarkers for pSS in a limited number of proteomic studies [
22,
23]. Hence, in the current study we hypothesised that by applying LC-MS alone, and in combination with size-exclusion chromatography of both stimulated whole saliva and tear fluid collected from patients with pSS and from healthy controls, novel biomarkers encompassing both salivary and lacrimal disease target organs could possibly be identified.
In order to delineate cellular pathways involving the upregulated proteins identified with LC-MS in the samples from the patients with pSS, GO and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway overrepresentation analyses using DAVID were performed. Our results demonstrated pathways of the adaptive immune response in the whole saliva, of the cellular component assembly in the EVs extracted from whole saliva, of metabolism and protein folding in the tear fluid of patients with pSS and finally, components of the adaptive immune response in the EVs isolated from the pooled sample of tear fluid from patients with pSS, which was comparable to the calcium ion binding pathways identified in whole saliva (Fig.
1). Viewed as a whole, the identified cellular pathways and components clearly indicate the involvement of autoimmune reactions and over-activation of the adaptive and innate immune systems in patients with pSS, both as a consequence of disease pathogenesis and probably also as part of the healing process.
The LC-MS analyses indicate upregulation of proteins involved in innate immunity, cell signalling and wound repair in whole saliva from patients with pSS. Interestingly, LCN2, the most upregulated protein in whole saliva from patients with pSS, is an iron-binding protein involved in the innate immune system, and is particularly responsible for the activation of neutrophils [
50]. This suggests the involvement of viral infection in SS pathogenesis. A similar implication was depicted by Hu and co-workers [
18], where they identified salivary proteomic and genomic biomarkers for SS showing upregulation of genes involved in the IFN pathway, thereby suggesting a potential role for viral infection in SS. Moreover, both GRN, a cytokine-like peptide that is central in inflammation due to its active role in wound repair and tissue remodelling [
51], and CALML5, a calcium-binding protein that plays a central role in the differentiation of keratinocytes [
52], were also upregulated in our patient group. This finding in turn provides evidence of acinar damage and oral environment alteration.
Both Giusti et al. [
13] and Fleissig et al. [
19] identified biomarkers that might include specific indication of tissue damage (e.g. actin), inflammation (e.g. calgranulins), and tissue repair (e.g. keratin 6 L) in unstimulated whole saliva. The present study identified similar potential with GRN and CALML5 in stimulated whole saliva (Table
3, Additional file
2: Figure S1). Furthermore, CALM, a calcium-binding protein that plays a role in intracellular signalling, and ESP1, a cholesterol transporter involved in cholesterol homeostasis within the endosome and/or lysosome, were also upregulated in whole saliva from patients with pSS. Similarly, previous proteomic studies on whole saliva have determined broad and distinct protein patterns that are characteristic of SS, including secretory proteins, enzymes, highly abundant immune system-related molecules (e.g. β2-microglobulin), and cytokines such as IL-4 and IL-5 [
21,
28,
29]. The current STRING analysis of whole saliva also strengthens the concept of involvement and over-activation of the innate and adaptive immune system in SS. This is presumably due to the upregulation of LCN2 and other related pro-inflammatory-related proteins in the patients with pSS to form protein-protein network interactions (Fig.
2).
Using size exclusion chromatography on whole saliva to isolate EVs followed by LC-MS allowed us to identify potential biomarkers that are vital for activation of the innate immune system and adipocyte differentiation. More precisely, the most upregulated protein in EVs from whole saliva, namely APMAP, is an enzyme central to adipocyte development. We have recently shown increased occurrence of adipose tissue replacement in minor salivary gland biopsies from patients with SS [
53]. Interestingly, these adipocytes were detected in areas rich in IL-6, suggesting their active involvement in immune reactions. Hence, the upregulation of APMAP in stimulated whole saliva could be an indication of adipocyte involvement in disease progression. Moreover, both GNA13, a G-protein that consequently plays a role in transmembrane signalling, and WDR1, a regulatory protein involved in the disassembly of actin filaments, are proteins needed to drive inflammation and tissue damage, respectively.
Interestingly, SIRPA is another potential biomarker of the pro-inflammatory process, as it regulates NK cells and dendritic cells. Furthermore, LSP1, being involved in innate immunity, specifically neutrophil activation and chemotaxis, is another possible indicator of the involvement of viral infection in the pathogenesis of SS (Table
4, Additional file
3: Figure S2). One major protein-protein interacting network was identified for EVs isolated from whole saliva, including proteins that regulate the formation and disassembly of the cytoskeleton, in addition to their involvement in cell migration and cell junction. Both GNA13 and WDR1 were present within this protein network (Fig.
3).
As only a limited number of proteomic studies have so far been performed on tear fluid from patients with SS [
22,
23], we were interested in analysing tear fluid collected from patients with pSS and healthy controls, in combination with whole saliva from the same individuals, in order to explore novel biomarkers encompassing both lacrimal and salivary disease target organs. We identified overexpression of proteins involved in TNF-α signalling (CPNE1) and B cell survival (PRDX3), in the Krebs cycle (ACO2) and in oxidative stress (APEX1) in tear fluid from patients with pSS (Table
5, Additional file
4: Figure S3). Moreover, two protein-protein interaction networks encompassing the upregulated proteins in tear fluid from patients with pSS were visualised using STRING. These networks were involved in oxidative stress, and the formation of the cytoskeleton and cell migration, respectively. With a vital role in regulating NF-κ B activation, and in turn B cell survival, PRDX3 was present within the protein network involving redox reactions and oxidative stress (Fig.
4).
Additionally, we explored protein expression in whole saliva and tear fluid simultaneously, and out of all the aforementioned proteins in the study, LCN2 was found to be upregulated in both fluids in the patients with pSS. Hence, the most upregulated protein in whole saliva from patients with pSS identified in our current analysis (Table
3), is also a key player in tear fluid from these same individuals. Being an iron-binding protein involved in the innate immune system and the activation of neutrophils [
50], LCN2 could be viewed as a possible biomarker for SS, whereby screening for LCN2 in whole saliva and tear fluid from patients with pSS could provide additional diagnostic accuracy. This observation further strengthens the notion of a role for viral infection in the pathogenesis of SS. Interestingly, LCN2 has also been previously proposed as a potential disease biomarker in the autoimmune disease systemic lupus erythematosus (SLE), as elevated levels of anti-LCN2 were detected in serum samples from patients with SLE [
54].
In spite of low tear fluid volumes collected individually from each patient with pSS, leading to few attainable data from the EVs isolated individually from each participant, our proteomic analysis of EVs extracted from the pooled tear sample combined from 11 patients with pSS revealed that CPNE1 and CALM were expressed more in the patient group. The presence of both these upregulated proteins indicates how their cellular functions fulfil each other, with CPNE1 playing a central role in inflammation and apoptosis due to TNF-α signalling, while CALM is essential in cell signaling and the activation of the immune system.
Acknowledgements
We express our gratitude to the patients and controls who took part in our study. Their collaboration has been essential for the accomplishment of the study presented. We appreciatively acknowledge Ann-Kristin Molværsmyr and Ann-Kristin Ruus for excellent technical assistance. We further express our sincere appreciation to Dr Sten Ræder, the head of the Norwegian Dry Eye Clinic; Dr Benedikte Døskeland from the Rheumatology Department at Rikshospitalet; and all other staff members at the research and clinical institutions involved for their effort and dedication.