Altered O-Glycans in stimulated whole saliva from patients with primary Sjögren’s syndrome and non-pSS sicca Authors: Sarah Kamounah, Kristina A Thomsson, Christiane Elisabeth Sørensen, Eric Paul Bennett, Niclas G Karlsson, Anne Marie Lynge Pedersen Published: 2024 ABSTRACT AbstractTo investigate if salivary O-linked glycans are altered in primary Sjögren’s syndrome (pSS), and thus contributing to explain symptoms of oral dryness, and an impaired oral mucosal barrier function leading to changes in microbial metabolism and colonization by both pathogenic and commensal microorganisms and increased prevalence of oral diseases. O-linked oligosaccharides from stimulated whole saliva (SWS) samples from 24 patients with pSS, 38 patients with non-pSS sicca, and 23 healthy controls were analyzed using liquid chromatography mass spectrometer (LC-MS). Non-fractionated reduced and alkylated saliva was dot-blotted to PVDF-membrane and O-linked oligosaccharides were released using reductive beta-elimination. The 50 most abundant glycans were identified and their intensity compared for each sample, reflecting the relative abundance of individual monosaccharide residues. Comparison of the compositions of O-glycans in SWS samples revealed higher relative levels of sialic acid (NeuAc) and lower levels of neutral amino-monosaccharides (HexNAc) in pSS and non-pSS sicca patients than in the healthy controls. MS2 fragmentation analysis of salivary O-glycans suggests that altered sulfation, fucosylation, sialylation and distribution of core types may all contribute to the observed alteration, directly or indirectly. Additionally, the short disaccharide sialyl-Tn was most abundant in the saliva samples from patients with pSS. Our findings indicate that the salivary mucin-type O-glycan profile is altered in pSS, reflecting a dysfunction of the post-translational modification of salivary mucins leading to rheological changes of saliva, oral dryness symptoms, and impaired oral mucosal barrier function. The pathophysiological significance of the aberrant O-glycosylation needs further elucidation.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-024-79473-1.Keywords: O-glycans, Primary Sjögren’s syndrome, Hyposalivation, Sialyl-TnSubject terms: Glycosylation, Autoimmune diseases, Rheumatic diseases 1. Introduction Primary Sjögren’s syndrome (pSS) is a chronic, autoimmune disease that targets exocrine glands, particularly the lacrimal and salivary glands, leading to ocular and oral dryness. The pathophysiology remains unclear but is characterized by lymphocytic infiltration of the salivary glands and an abnormal B- and T-cell response to autoantigens, SSA, causing degeneration of the glandular tissue and subsequent dysfunction1,2. PSS is the second most common autoimmune inflammatory rheumatic disease, only surpassed in prevalence by rheumatoid arthritis3. Xerostomia is the term used to describe the subjective sensation of dry mouth. The most commonly used diagnostic tool to assess salivary gland hypofunction is sialometry, which objectively measures the unstimulated whole saliva (UWS) and chewing-stimulated whole saliva (SWS) flow rates. While this method offers a non-invasive and rapid assessment, there are limitations as studies have shown that in up to 50% of individuals who experience dry mouth, the sensation is not related to hyposalivation4. Xerostomia may be caused by many different factors including intake of certain medications, radiotherapy to the head and neck region, and Sjögren’s syndrome, and it usually occurs as a result of decreased salivary secretion, and/or changes in the composition of saliva5. It is well-known that the composition of mucins in saliva is closely linked to dry mouth6. Mucin glycoproteins are crucial for the rheological characteristics of saliva, such as viscosity, elasticity, and adhesiveness, which enable surface coating and lubrication. Mucins form the lubricating thin mucosal membrane that lines the oral cavity, and acts as a physical and immunological barrier protecting the epithelial tissue against the external environment7–11. Spinnbarkeit, which measures the stringiness of saliva shows a positive correlation with the viscosity of unstimulated whole saliva and salivary mucins12,13. Human saliva primarily contains two types of mucins, MUC5B and MUC7. MUC5B is a large, multimeric glycoprotein with a high carbohydrate content. It comprises a protein backbone and numerous heterogeneous O-linked carbohydrate side chains, ranging from 2 to over 20 residues. This structure endows MUC5B with gel-forming properties that contribute to lubrication, moisture retention, and viscoelasticity14. The mucins have a negative charge due to the high number of sialic acid residues and sulfate groups, enabling them to retain water, which is essential for their hydrating and rheological properties15. A recent study has shown that removing these negatively charged groups compromised their lubricating qualities, which could contribute to the oral dryness experienced in pSS16. A reduction in residual mucosal saliva (RMS), forming the thin film of residual saliva on mucosal surfaces, is associated with hyposalivation, and a deficiency in RMS has been found related to xerostomia17. Reduced RMS is also observed in patients with normal whole saliva flow rates, suggesting that the mucosal surfaces are not adequately moisturized, despite normosalivation18. These findings highlight the necessity for qualitative analysis of saliva, in addition to measuring glandular output, to understand the changes linked to xerostomia. In this study, we hypothesized that potential disturbances in the synthesis of glycoproteins impair their rheological properties, which can contribute to explain the extensive perception of oral mucosal dryness in patients with pSS. We therefore investigated whether the composition of O-linked glycans in saliva differs between patients with pSS and non-pSS sicca and healthy controls. A further understanding of an abnormal glycosylation pattern in patients with pSS may improve diagnostics and aid in potential glycan-mediated therapeutic interventions. Methods Patient characteristics The study cohort has previously been described in detail19. In brief, 85 patients were included in this study; 24 patients fulfilling the American College of Rheumatology/European League against Rheumatism (ACR-EULAR) classification criteria for pSS20, 38 patients with symptoms of oral and ocular dryness but who did not fulfill the classification criteria for pSS (non-pSS sicca), and 23 age- and gender-matched healthy control subjects with no history of autoimmune disease. The patients with pSS and non-pSS sicca were all referred with a suspicion of pSS by rheumatology and ophthalmology outpatient clinics or by dentists in private practice. The study was performed following the Declaration of Helsinki and approved by the Ethical Committees for the Region of Copenhagen, Denmark (H-18042385), the Danish Data Protection Agency, and the General Data Protection Regulation (GDPR). Written informed consent was obtained from all participants. Interview and clinical examination All participants were assessed by a single investigator (SK) at the Clinic for Oral Medicine, at the Department of Odontology, Faculty of Health and Medical Sciences, University of Copenhagen. The investigator was assisted by a senior investigator (AMLP). A comprehensive, standardized interview was conducted with each participant to gather information about their general health, medication use, and consumption of tobacco and alcohol. To evaluate the severity of xerostomia, participants completed the Bother Index 1 (BI1), Bother Index 5 (BI5), and Xerostomia Inventory (XI) questionnaires21. The Clinical Oral Dryness Score (CODS) was used to assess objective signs of oral dryness. This evaluation includes 10 clinical features that are indicative of hyposalivation22. The clinical examination included measurements of periodontal pocket depth, gingival index, and plaque index on six index teeth, as well as assessments of decayed, missing, and filled teeth (DMF-T) and surfaces (DMF-S) (for details23). Each participant also underwent a minor salivary gland biopsy of the lower lip for histopathological evaluation and focus scoring, as previously described19,23,24. Measurements of spinnbarkeit Prior to the whole saliva sialometry, UWS was collected for 2 min, and 100 µl of the UWS sample was loaded onto the Neva Meter (IMI-0501, Ishikawa Ironworks Co., Kobe, Japan). Three measurements of spinnbarkeit were obtained. The average of the three measurements was then calculated for each patient as previously described4,12. Measurements of residual mucosal saliva The residual mucosal saliva (RMS) was collected from the anterior hard palate, tongue, and buccal mucosa using SialoPaper strips (Oraflow Inc., USA) applied on each surface for 10 s. The volumes of fluid absorbed by the sialostrips were measured using Periotron 8000®(Oraflow Inc., USA)4,25. Measurements of whole saliva flow rates Unstimulated whole saliva (UWS) was collected over 15 min using the draining method, while chewing-stimulated whole saliva (SWS) was collected over 5 min as previously described19,23. The saliva samples were immediately placed on ice, then aliquoted and stored at -80 °C for later analysis. The UWS samples were centrifuged at 10.000 x g, at 4oC, for 10 min and separated into supernatant and pellet before aliquoting and freezing. Measurement of total protein concentration Measurement of the total protein concentration (mg/ml) in SWS samples was performed by the method of Lowry et al.26. Each measurement of the saliva samples was measured in duplicate. The total protein output (mg/min) was then calculated by multiplying each concentration with the individual stimulated flow rate. Pilot study We initially performed a pilot study to analyze O-glycans in UWS and SWS samples from four patients with pSS and four healthy controls from the study cohort. For O-glycan analyses, 50 µl of non-fractionated reduced and alkylated saliva was dot-blotted to a PVDF membrane and visualized with Alcian Blue staining. O-glycans were released from the proteins on the PVDF membrane using reductive beta elimination and analyzed on a graphitized carbon column connected to an ion trap mass spectrometer (LC-MS) (further detailed below). Glycan sequencing was performed by manual examination of the MS/MS spectra and when possible, compared to published spectra and spectra available online (unicarb-db.expasy.org) (Supplementary data 1). Sample preparation Sample preparation was performed using a modified protocol published elsewhere27 Saliva (50 µl) was reduced in 10 mM dithiothreitol in 50 µl urea (8 M urea with 0.1% SDS) for 2 h at 95 °C, followed by alkylation in 25 mM iodoacetamide in dark, at room temperature and overnight. Samples were divided in two aliquots, and 50 µl of non-fractionated reduced and alkylated saliva was dot-blotted to a PVDF membrane (Immobilon P, Millipore), followed by visualization using Alcian Blue stain. PVDF membrane spots were excised and placed in test tubes (two spots/sample), followed by 5 × 15 min destain/washes in MeOH. The O-glycans were released from the protein with 40 µl beta elimination solution (1 M NaBH4 in 0.1 M NaOH) at 50 °C in a water bath. Samples were neutralized with 1–2 µl glacial acetic acid, followed by desalting using cation exchange media (AG50WX8 (Biorad)) in C18 zip tips (Millipore), two zip tips/sample, and dried with vacuum centrifugation. Borate residuals were eliminated by repeated additions of MeOH (5 times 50 µl) and dried in between. LC-MS analyses Reduced O-glycans were resuspended in 15 µl of water and injected (2 µl) onto a liquid chromatography-electrospray ionization tandem mass spectrometry (LC-ESI/MS). The oligosaccharides were separated on a column (10 cm × 250 μm) packed in-house with 5 μm porous graphite particles (PGC, Hypercarb, Thermo-Hypersil, Runcorn, UK) and a flow rate of 5 µl/min. The oligosaccharides were eluted with the following gradient: 0–46 min 0–45% B, wash 46–54 min 100% B, then equilibration between 54 and 78 min with 0% B. Buffer A was 10 mM ammonium bicarbonate (ABC) and buffer B was 10 mM ABC in 80% acetonitrile. A 40 cm × 50 μm inner diameter fused silica capillary was used as transfer line to the ion source. The samples were analyzed in negative ion mode on a LTQ linear ion trap mass spectrometer (Thermo Electron, San José, CA), with an IonMax standard ESI source equipped with a stainless-steel needle kept at − 3.5 kV. Compressed air was used as nebulizer gas. The heated capillary was kept at 270 °C, and the capillary voltage was -50 kV. Full scan (m/z 380–1800, two microscan, maximum 100 ms, target value of 30,000) was performed, followed by data-dependent MS2 scans (two microscans, maximum 100 ms, target value of 10,000) with normalized collision energy of 30%, isolation window of 2.5 units, activation q=0.25 and activation time 30 ms). The threshold for MS2 was set to 300 counts. Data acquisition was conducted with Xcalibur® software (Version 2.0.7). Data evaluation The raw files were submitted to Progenesis software for peak quantitation, generating an output list of m/z values in Excel format. Further filtering of the data was performed by excluding ions which were not associated with MS2 spectra. The 50 major glycans in each sample were selected for further evaluation, which was estimated to include between 95 and 100% of the intensity of the total glycan pool (based on a calculation from the six samples in the pilot study). These 50 peaks were identified manually by comparing the mono- and double-charged ions to a list of glycans in a ‘master spreadsheet’, compiled from analyses of saliva samples in the pilot study. The ‘master file’ was continuously extended when going through the dataset, and when new compositions were identified. MS2 spectra were double-checked manually when required, for example when retention times did not seem reasonable, or when more than one monosaccharide composition was possible for the same m/z. Ions with the same mass as N-glycans found in the pilot study were excluded after manual inspection of the MS2 spectra. In three of the 85 samples analyzed (1 pSS and 2 non-pSS sicca), only 38, 44 and 34 glycans could be detected. The relative abundance of the 50 glycans was summed up to 100%. The individual glycans were then reduced to their compositions of HexNAc, Hex, Fuc, NeuAc and SO3, and then multiplied by percentage relative abundance. Each parameter was then summed up over the whole sample to give the ‘MSAC metric’ (mass spectrometric average composition) as previously described28 (Supplementary data 2). Secretor status was determined by presence of blood-group H glycan at m/z 530 (equivalent to Fucα1-2Galβ1-3GalNAcol) in the MS data. It should be noted that the saliva samples were not fractionated with SDS-PAGE or other filters and may thus be a mixture of all salivary proteins, but previous studies support that the most abundant glycoproteins are the mucins MUC7 and MUC5B29. Statistics The data was analyzed using GraphPad Prism v10 and R v4.1.2. The data was tested for normal distribution using the Shapiro-Wilk normality test. The clinical data from the three groups (pSS, non-pSS sicca, and healthy controls) were compared using two-sample t-tests for normally distributed data, and non-parametric Mann-Whitney U-tests for data that were not normally distributed. Two-way ANOVA test followed by Tukey’s multiple comparisons correction were used for comparing the MSAC metrics of O-glycans between the three groups. Non-parametric Kruskal Wallis test and Dunn’s multiple comparisons test were used to compare the MSAC metric ratios among the three groups. Correlation analyses were performed using Pearson correlation and Spearman correlation for parametric and non-parametric data, respectively. A p-value of ≤ 0.05 was considered statistically significant. Results Table 1provides a summary of the demographic, clinical, and pSS-related characteristics of the study population. At the time of diagnosis, the median age of the patients in the pSS group was 58 years. There were no significant differences between the pSS and non-pSS groups in terms of age, gender, tobacco use, or severity of xerostomia. Additionally, the UWS and SWS flow rates, as well as the number of patients with hyposalivation, were comparable between the two groups as previously reported19. However, patients with pSS had a significantly higher prevalence of keratoconjunctivitis sicca, circulating serum anti-SSA antibodies, and focus score ≥ 1 in the labial salivary gland biopsy than those with non-pSS sicca. Both patients with pSS and non-pSS sicca patients had significantly lower UWS and SWS flow rates than the healthy controls. Furthermore, DMF-T (p = 0.04) and DMF-S (p = 0.05) scores were significantly higher in the patients with pSS than in the healthy controls, while no significant differences were observed between non-pSS sicca patients and healthy controls. The output of total protein (mg/min) in SWS was significantly lower in patients with pSS (p = 0.007) and non-pSS sicca patients (p = 0.04) than in the healthy controls, while no significant difference was found between the two patient groups. The total protein output was inversely correlated with BI-5 (r = -0.53, p = 0.01), and positively correlated with UWS (r = 0.66, p = 0.001) and SWS (r = 0.84, p < 0.001) flow rates in the pSS group. The total protein concentration (and output) was also positively correlated with UWS (r = 0.40, p = 0.01) and SWS (r = 0.80, p < 0.001) flow rates in the non-pSS sicca group. Table 1Demographic and clinical characteristics of the study population.pSS (n = 24)non-pSS (n = 38)Healthy controls (n = 23)P-valueAge (years)58 ± 1456 ± 1359 ± 14N.S.Female23 (96%)36 (95%)21 (91%)N.S.Secretors(presence of blood-group H glycan at m/z 530)21 (88%)32 (84%)19 (83%)N.S.Current smoker6 (25%)12 (32%)1 (4%)N.S.*, **,0.01***Hyposalivation (UWS ≤ 0.10 ml/min)9 (38%)14 (37%)0 (0%)N.S.* < 0.002** 0.002*** Keratoconjunctivitis sicca22 (92%)11 (29%)N.A. 0.003* N.A**,***Positive serum anti-SSA/Ro antibody20 (83%)10 (26%)N.A.< 0.001*, **N.A**,***Focus score ≥ 1 in LSG biopsy17 (71%)0 (0%)0 (0%)< 0.001*, ** N.S.*** UWS (ml/min)Median (ml/min)0.18 ± 0.160.180.20 ± 0.180.170.51 ± 0.340.41N.S.*< 0.001**, ***SWS (ml/min)Median (ml/min)1.12 ± 0.810.841.46 ± 0.981.002.33 ± 0.862.29N.S.*< 0.001**, ***Bother Index 1Median6.13 ± 2.336.55.47 ± 2.9760.04 ± 0.210N.S.*< 0.001**, ***Bother Index 5Median5.17 ± 3.1745.42 ± 3.7350.09 ± 0.290N.S.*< 0.001**, ***Xerostomia Inventory score39 (16–53)38 (11–55)13 (11–18)N.S.*< 0.001**, ***CODSMedian2.3 ± 2.221.53 ± 1.621.50.30 ± 0.560N.S.*< 0.001**, ***RMS (µL)Hard palateTongueBuccal mucosa0.60 ± 0.722.69 ± 1.401.99 ± 0.970.69 ± 0.902.96 ± 1.182.82 ± 0.871.58 ± 1.313.71 ± 03.59 ± 0.30N.S*< 0.001**, ***N.S*< 0.001**, *** 0.002* < 0.001**, ***Spinnbarkeit (mm)Median (mm)12.09 ± 11.966.3417.69 ± 16.3210.0716.89 ± 14.309.19N.S*, **, ***Total protein output SWS (mg/min)Median (mg/min)6.71 ± 4.985.488.22 ± 5.697.139.47 ± 3.258.88N.S*0.007**0.04***DMF-TMedian score16.75 ± 7.351714.76 ± 7.2613.512.09 ± 6.0012N.S*, ***0.04**DMF-SMedian score54.63 ± 38.6248.5044.13 ± 33.1731.5033.91 ± 23.2529.00N.S*, ***0.05**Comparison analyses between pSS and non-pSS*, pSS and HC**, non-pSS and HC***.CODS, clinical oral dryness score; DMF-T, Decayed, Missing, Filled Teeth; DMF-S, Decayed, Missing, Filled Surfaces; LSG, labial salivary gland; RMS, residual mucosal saliva; UWS, unstimulated whole saliva; SWS, stimulated whole saliva. Demographic and clinical characteristics of the study population. Secretors (presence of blood-group H glycan at m/z 530) N.S.*, **, 0.01*** N.S.* < 0.002** 0.002*** 0.003* N.A**,*** < 0.001*, ** N.A**,*** < 0.001*, ** N.S.*** UWS (ml/min) Median (ml/min) 0.18 ± 0.16 0.18 0.20 ± 0.18 0.17 0.51 ± 0.34 0.41 N.S.* < 0.001**, *** SWS (ml/min) Median (ml/min) 1.12 ± 0.81 0.84 1.46 ± 0.98 1.00 2.33 ± 0.86 2.29 N.S.* < 0.001**, *** Bother Index 1 Median 6.13 ± 2.33 6.5 5.47 ± 2.97 6 0.04 ± 0.21 0 N.S.* < 0.001**, *** Bother Index 5 Median 5.17 ± 3.17 4 5.42 ± 3.73 5 0.09 ± 0.29 0 N.S.* < 0.001**, *** N.S.* < 0.001**, *** CODS Median 2.3 ± 2.2 2 1.53 ± 1.62 1.5 0.30 ± 0.56 0 N.S.* < 0.001**, *** Hard palate Tongue Buccal mucosa 0.60 ± 0.72 2.69 ± 1.40 1.99 ± 0.97 0.69 ± 0.90 2.96 ± 1.18 2.82 ± 0.87 1.58 ± 1.31 3.71 ± 0 3.59 ± 0.30 N.S* < 0.001**, *** N.S* < 0.001**, *** 0.002* < 0.001**, *** Spinnbarkeit (mm) Median (mm) 12.09 ± 11.96 6.34 17.69 ± 16.32 10.07 16.89 ± 14.30 9.19 Total protein output SWS (mg/min) Median (mg/min) 6.71 ± 4.98 5.48 8.22 ± 5.69 7.13 9.47 ± 3.25 8.88 N.S* 0.007** 0.04*** DMF-T Median score 16.75 ± 7.35 17 14.76 ± 7.26 13.5 12.09 ± 6.00 12 N.S*, *** 0.04** DMF-S Median score 54.63 ± 38.62 48.50 44.13 ± 33.17 31.50 33.91 ± 23.25 29.00 N.S*, *** 0.05** Comparison analyses between pSS and non-pSS*, pSS and HC**, non-pSS and HC***. CODS, clinical oral dryness score; DMF-T, Decayed, Missing, Filled Teeth; DMF-S, Decayed, Missing, Filled Surfaces; LSG, labial salivary gland; RMS, residual mucosal saliva; UWS, unstimulated whole saliva; SWS, stimulated whole saliva. Structural mapping study - oligosaccharide features in saliva Overall, the UWS samples stained much less than the SWS samples when evaluating the Alcian Blue staining, regardless of whether they originated from patients with pSS or healthy controls. All 16 samples contained O-glycans, although, during inspection of the LC-MS profiles, it was clear that the four UWS samples from patients with pSS contained the lowest levels of O-glycans. Due to the extensive time required for sequence assignment, we proceeded with six SWS samples, and sequenced and semi-quantified all components using the MS signal intensity (Supplementary data 1). The O-glycan features are summarized in Table 2. Table 2Summary of O-glycan features in the pilot study.HC01HC02HC03SS01SS02SS03PROPOSED SECRETOR STATUSNONSECSECSECSECSECSEC% Acidic glycans (NeuAc + SO3)69.139.661.654.954.748.4% Glycans with SO3 epitopes41.48.023.015.120.12.9% Glycans with blood group H epitopes0.012.622.138.238.126.1% Glycans with blood group A epitopes0.016.215.00.00.24.3% Glycans with Lewis epitopes34.519.420.528.631.623.6O-glycan (N-glycans and peeling products have been excluded in these calculations).No obvious differences related to healthy controls (HC samples) or patients with primary Sjögren’s syndrome (SS samples) were observed. Summary of O-glycan features in the pilot study. O-glycan (N-glycans and peeling products have been excluded in these calculations). No obvious differences related to healthy controls (HC samples) or patients with primary Sjögren’s syndrome (SS samples) were observed. Including glycoforms, low amounts of peeling products, and nine N-glycans which may also be released during this procedure, we identified 232 distinct glycan structures, each comprising up to 14 monosaccharide residues. Not all glycans were found in all samples, partly due to their blood group and secretor status. Approximately 80% of the Caucasian population express fucosyltransferase 2 (FUT2) and can make blood group H/A/B epitopes on the glycans in mucosal secretions30. Quantitative comparison study in stimulated whole saliva Decreased relative amounts of core 3/4 and increased NeuAc in patient samples In the larger dataset (Supplementary data 2) including all patients, we observed a significant decrease in amino monosaccharides (HexNAc), in pSS and non-pSS sicca patients, which may reflect lower expression of core 3 and core 4 glycans. MSAC metrics summed up over all samples is shown in Fig. 1A; Table 3, demonstrating a statistically significant decrease in HexNAc (amino monosaccharides) in both patients with pSS and non-pSS sicca patients compared to healthy controls. MSAC metrics of O-glycans in stimulated whole saliva samples by LC-MS. A, B: Summary of all analyzed samples (A), and with non-secretors (determined by MS) excluded (B). Two-way ANOVA with Tukey’s multiple comparisons correction. Mean with standard deviation is shown. **p < 0.01, *p < 0.05.C, D: Average glycan size (sum of MSAC metrics for the individual sample (C) and acidity across the three groups (D). Non-parametric Kruskal Wallis test with Dunn’s multiple comparisons test. Mean with standard deviation is shown. Table 3Mean ± SD values of the MSAC metrics of O-glycans in stimulated whole saliva samples.pSS(n = 24)Non-pSS, sicca(n = 38)Healthy controls (n = 23)P-values(*pSS vs. HC)(**non-pSS vs. HC) MSAC (all) HexNAc2.27 ± 0.342.30 ± 0.242.48 ± 0.340.014*, 0.025**Hex1.94 ± 0.231.97 ± 0.222.07 ± 0.21N.S.Fuc0.85 ± 0.290.91 ± 0.280.83 ± 0.33N.S.NeuAc0.41 ± 0.230.45 ± 0.270.23 ± 0.150.049*, 0.0035**Sulf0.24 ± 0.160.27 ± 0.250.37 ± 0.24N.S. pSS (n = 21) Non-pSS, sicca (n= 32) Healthy controls (n = 19) MSAC (secretors excluded) HexNAc2.23 ± 0.352.26 ± 0.232.43 ± 0.350.017*, 0.027**Hex1.90 ± 0.201.91 ± 0.172.02 ± 0.18N.S.Fuc0.87 ± 0.290.96 ± 0.260.91 ± 0.30N.S.NeuAc0.37 ± 0.200.42 ± 0.250.21 ± 0.13N.S.*, 0.007**Sulf0.22 ± 0.160.22 ± 0.160.31 ± 0.18N.S. pSS (n = 24) Non-pSS, sicca (n = 38) Healthy controls (n = 23) Average glycan size5.47 ± 0.595.66 ± 0.435.64 ± 0.41N.S.Acidity0.65 ± 0.310.72 ± 0.290.60 ± 0.24N.S.Ratio HexNAc/Hex1.16 ± 0.111.17 ± 0.111.20 ± 0.12N.S.Ratio Fuc/HexNAc0.39 ± 0.160.41 ± 0.140.36 ± 0.16N.S.Ratio NeuAc/HexNAc0.18 ± 0.100.20 ± 0.120.10 ± 0.070.01*, 0.001**Ratio (Fuc + Sulf)/NeuAc3.47 ± 2.234.88 ± 5.739.72 ± 13.790.002*, 0.001**Ratio Fuc/NeuAc0.79 ± 0.851.49 ± 2.646.14 ± 7.69< 0.0001*, < 0.0001**Ratio NeuAc/Hex0.21 ± 0.100.23 ± 0.130.11 ± 0.080.009*, 0.0008**Ratio Sulf/Fuc0.33 ± 0.320.54 ± 0.970.60 ± 0.61N.S.Ratio Sulf/NeuAc0.79 ± 0.851.49 ± 2.643.58 ± 6.290.007*, 0.004**Ratio Fuc/Hex0.45 ± 0.180.47 ± 0.160.42 ± 0.18N.S.Sialyl-Tn0.59 ± 0.670.32 ± 0.490.09 ± 0.260.005*, N.S.**MSAC, Mass spectrometric average composition; N.S, not significant. Mean ± SD values of the MSAC metrics of O-glycans in stimulated whole saliva samples. pSS (n = 21) pSS (n = 24) MSAC, Mass spectrometric average composition; N.S, not significant. A significant increase in NeuAc glycans in both patients with pSS and non-pSS sicca was also observed (Fig. 1A). When non-secretors were excluded, the increase in NeuAc remained significant in non-pSS sicca patients, and the decrease in HexNAc remained significant in pSS and non-pSS sicca patients. (Fig. 1B). The change in HexNAc and NeuAc in pSS and non-pSS sicca did not appear influenced by the size or the acidity of the detected glycans since the average glycan size (equal to the sum of MSAC metrics) was similar across all three groups (Fig. 1C) and the overall acidity (sialic acid and sulfate) appeared to be the same across the three groups (Fig. 1D; Table 3). Correlating the MSAC values with other measures from our patients and controls combined (Table 4), we could not find that NeuAc correlated with any of the other MSAC values. In this comparison, the level of sulfation appeared to be the factor that influenced the glycosylation the most (negatively correlated with Hex and Fuc (DeHex) and positively correlated with average size). Table 4Spearman correlation analysis.SpinnbarkitSialyl-TnHexNAcHexdeHexNeuAcSulfAverage glycan sizeAnti-SSA positivity0.08-0.040.24 0.48**** 0.210.31 -0.48**** -0.09KCS0.040.220.130.260.390.36-0.250.11Focus score ≥ 1-0.160.200.060.250.260.30-0.30-0.04CODS-0.090.11-0.18-0.20-0.080.090.08-0.05Bother Index 5-0.150.22-0.37 -0.40**** -0.260.100.310.01Bother Index 1-0.100.24-0.37-0.36-0.170.060.25-0.06Xerostomia Inventory-0.090.28 -0.43**** -0.42**** -0.200.050.25-0.09RMS palate0.01-0.170.230.32-0.02-0.04-0.210.06RMS tongue0.11-0.170.070.230.00-0.10-0.26-0.21RMS buccal0.09-0.240.170.150.00-0.17-0.010.09Spinnbarkeit1.00-0.060.140.00-0.15-0.080.100.00UWS0.22-0.110.210.270.02-0.08-0.12-0.05SWS0.13-0.140.160.180.02-0.10-0.020.01Hyposalivation UWS-0.200.00-0.08-0.18-0.010.060.220.16Hyposalivations SWS-0.160.17-0.12-0.160.02-0.060.03-0.07Sialyl-Tn-0.061.00-0.33-0.020.07-0.04-0.14-0.21DMF-T-0.120.13-0.37 -0.55**** -0.40 -0.04 0.43**** -0.07DMF-S-0.130.03 0.41**** 0.60**** 0.280.17 -0.62**** -0.08Total protein output (mg/ml)0.01-0.14 0.57**** 0.63**** 0.390.10 -0.59**** -0.04HexNAc0.14-0.331.00 0.66**** 0.150.14-0.28 0.44**** Hex0.00-0.02 0.66**** 1.000.390.10 -0.62**** 0.18deHex-0.150.070.150.391.000.13 -0.47**** 0.15NeuAc-0.08-0.040.140.100.131.00-0.130.27Sulf0.10-0.14-0.28 -0.62**** -0.47**** -0.131.00 0.42**** Average glycan size0.00-0.21 0.44**** 0.180.150.27 0.42**** 1.00The correlation analyses included the entire study population (n = 85).CODS, clinical oral dryness score; DMF-S, Decayed Missing Filled Surfaces; DMF-T, Decayed Missing Filled Teeth; KCS, keratoconjunctiva sicca; RMS, residual mucosal saliva, SWS, chewing-stimulated whole saliva flow rate; UWS, unstimulated whole saliva flow rate.Moderate correlation: 0.4–0.59, Strong correlation: 0.60–0.79.*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0005. Spearman correlation analysis. The correlation analyses included the entire study population (n = 85). CODS, clinical oral dryness score; DMF-S, Decayed Missing Filled Surfaces; DMF-T, Decayed Missing Filled Teeth; KCS, keratoconjunctiva sicca; RMS, residual mucosal saliva, SWS, chewing-stimulated whole saliva flow rate; UWS, unstimulated whole saliva flow rate. Moderate correlation: 0.4–0.59, Strong correlation: 0.60–0.79. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0005. Ratios of MSAC values for some of the monosaccharides for individual samples The MSAC metrics ratios between monosaccharides within the individual samples is shown in Fig. 2. Overall, statistically significant differences are observed related to NeuAc ratios including NeuAc/HexNAc, (Fuc + Sulf)/NeuAc, Fuc/NeuAc, NeuAc/Hex, Sulf/NeuAc (Fig. 2). These findings suggest that increased sialylation in pSS and non-pSS sicca samples compared to controls is accompanied by a decrease in sulfation and fucosylation. MSAC metric ratios between patients with pSS and non-pSS sicca and healthy controls. Non-parametric Kruskal Wallis test with Dunn’s multiple comparisons test. **** p < 0.0005; ***, p < 0.001; **p < 0.01; *p < 0.05. Mean with standard deviation is shown. Higher abundance of Sialyl Tn in patient samples Evaluation of LC-MS raw data revealed that the short glycan Sialyl-Tn (NeuAcα2-6GalNAc, detected in its reduced form as m/z 513) was present in samples across all three groups. However, it appeared significantly more abundant in the saliva samples from patients with pSS. In the majority of healthy controls (87%), S-Tn fell below the ‘Top 50’ glycans, whereas it was detected at significantly higher levels in SWS samples from patients with pSS (Fig. 3). Relative abundance of Sialyl Tn (NeuAcα2-6GalNAc) in saliva, identified among the 50 most abundant glycans, measured with LC-MS. Numbers in parenthesis indicate number of positive identifications in each group (i.e., 3 individuals of total 23 analyzed). Non-parametric Kruskal Wallis test with Dunn’s multiple comparisons test. **p < 0.01. Mean with standard deviation is shown. Spearman correlation analysis revealed no significant correlation between S-Tn and any clinical or paraclinical features (Table 4). Discussion Mucin-type O-glycans serve various functions in mucosal secretions, including hydrating and protecting underlying epithelial cells, providing resistance to protease degradation, and trapping and removing bacteria via specific receptor sites within the O-glycan chains of mucin31. In this study, we identified the compositional changes in mucin-type O-glycans in stimulated whole saliva from patients with pSS and non-pSS sicca compared to healthy controls. Our data demonstrated that both patients with pSS and non-pSS sicca patients had mean UWS flow rates exceeding the hyposalivation threshold (> 0.1 ml/min), yet significantly lower than those of healthy controls. The prevalence of hyposalivation did not differ significantly between the pSS and non-pSS sicca groups, but both groups exhibited a significantly higher prevalence of hyposalivation compared to the healthy control group. Although many pSS and non-pSS sicca patients did not present with hyposalivation, the mean residual mucosal saliva (RMS) volume was markedly lower across all sites in both groups compared to healthy control group. This finding is consistent with previous research4,15. RMS levels were lower in patients with pSS than in non-pSS sicca patients at all sites, with a statistically significant reduction observed only in the buccal mucosa. Previous studies have similarly reported site-specific reductions in RMS17,32,33. The mean spinnbarkeit was lowest in patients with pSS. However, there were no statistically significant differences in spinnbarkeit among the groups, and no correlation was found between spinnbarkeit and xerostomia, UWS, SWS, RMS, total protein, or MSAC metrics of the monosaccharides (Table 4). We also found higher levels of NeuAc in patients with pSS and non-pSS sicca. This was further reflected in the ratios between the other monosaccharides and NeuAc which distinguished patients from healthy controls indicating structural changes in O-glycans that may be more prevalent in saliva of patients with xerostomia. Our data also demonstrated significantly lower levels of HexNAc in patients with pSS and non-pSS sicca patients than in healthy controls. The accompanying increase in sialylation in pSS and non-pSS sicca patients suggests a complex relationship, where both terminal, core types, elongation, and branching can be involved. With the thorough investigation of the type of glycans present in saliva (Table 2), an increase in sialylation would decrease the amount of blood group ABH containing antigens. This transition would result in a decrease in fucosylation but should also decrease the HexNAc level in blood group A secretor individuals. Furthermore, since all cores from 1 to 4 was detected in saliva, a potential decrease in core 3 and 4 and an associated increase in core 1 and 2 would also lower the proportion of HexNAc. Mucin glycosylation in lung inflammation has previously been shown to progress via a transition from core 4 into core 2 type mucin glycans with associated increase of sialylation28. Unlike core 1 and 2, core 3 O-glycans are mainly synthesized in mucous epithelia of the gastrointestinal and respiratory tracts, and the salivary glands. Core 4 glycoprotein is produced from core 3 by the addition of GlcNAc to the O-GalNAc. Both core 3 and 4 O-glycans have been reported to repress tumor progression, although the underlying mechanisms remain unclear34. Interestingly, these glycans are synthesized in normal cells but have been found downregulated in gastric and colorectal carcinomas35–37. Further investigations of the type of mucins and oligosaccharide structural elements in pSS would reveal how mucin expression and its glycosylation is involved in pSS and pSS sicca. Unexpectedly, we found significantly higher levels of Sialyl Tn antigen (S-Tn) in the SWS samples from patients with pSS compared to healthy controls. Aberrant glycosylation in cancer often results in the accumulation of immaturely truncated O-glycans, such as Tn antigen. S-Tn is a glycan structure that halts further O-glycan chain elongation and is associated with tumor development and malignancy38. However, S-Tn can also be found in non-cancerous healthy and inflamed tissues. Rømer et al. detected S-Tn expression in the healthy gastrointestinal tract, primarily intracellularly or in suprabasal cells39. Therefore, it would be valuable to perform immunohistochemical staining of salivary gland tissue to determine whether the higher S-Tn expression in patients with pSS is associated with a specific glandular cell type and whether it is present at the cell surface or intracellularly. Mucins are essential for the viscoelasticity of saliva13. Early studies indicated that removing mucins weakened the rheological properties of saliva, which were restored upon re-adding the mucin fraction40. The glycan component of mucins is vital for this property, and the degradation of mucin glycans reduces the viscoelasticity of saliva41. Understanding mucins and their glycosylation is critical, particularly in pSS, where changes in mucin composition could negatively impact the mechano-physical properties of saliva. The GALNTs, a large family of glycosyltransferases, play a significant role in human health and disease. GALNT1 initiates the biosynthesis of mucin-type O-glycosylation42,43, and it has been shown that let-7i-5p expression is negatively correlated with GALNT1, and overexpression of GALNT1 can inhibit inflammation induced by let-7i-5p44. Along this line, it is interesting that Sembler-Møller et al. found a significant upregulation of let-7i-5p in patients with pSS compared to non-pSS sicca patients45. This study did not allow us to investigate the level of glycosylation, but it would be of interest to investigate whether GALNT1 is downregulated in patients with pSS in future studies. Conclusion The present study shows that alterations in O-linked glycan composition in saliva can distinguish a group of healthy controls from patients with pSS and non-pSS sicca, who present with significantly more xerostomia, lower saliva production, and higher caries experience. The monosaccharide N-acetylneuraminic acid (NeuAc) was more abundant in O-linked salivary glycans of both pSS and non-pSS sicca patients, which calls for further research into the functional significance in both conditions. Higher S-Tn abundance in saliva from pSS patients and a reduction in core 3 and 4 structures in saliva from both patient groups were further characteristics revealed by the present analyses. More studies are needed to clarify how these compositional changes affect the hydrating and protective functions, particularly of the highly abundant and heavily O-glycosylated mucins in whole saliva and residual mucosal saliva. Electronic supplementary material Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 2 Acknowledgements We would like to extend our sincere gratitude to laboratory technician Louise Rosgaard Duus for her assistance with the laboratory analyses of the samples. Chunsheng Jin is acknowledged for his help with MS glycan structure interpretation as part of the pilot study. We also deeply appreciate all the patients who volunteered to participate in this study. List of symbol Bother Index Clinical oral dryness score Decayed Missing Filled Surfaces Decayed Missing Filled Teeth European League Against Rheumatism Fucose Focus score N-acetylgalactosamine N-acetylglucosamine Hexose Liquid chromatography mass spectrometer Mass spectrometric average composition N-acetylneuraminic acid (sialic acid) Non-Sjögren’s syndrome sicca N-acetyl hexosamine Residual mucosal saliva Primary Sjögren’s syndrome Sialyl-Tn Stimulated whole saliva Unstimulated whole saliva Xerostomia inventory Author contributions Collection of samples and clinical data: S.K. and A.P. Analysis of data: K.T. and S.K. Data interpretation S.K., K.T. C.S., N.K. and A.P. Drafted the manuscript: S.K. All authors critically revised the manuscript and approved the final version. Funding This work was supported by SciLifeLab and BioMS (Swedish Research Council), the Danish Dental Association Research Foundation, and the Danish Rheumatism Association. Data availability The data supporting the findings of this study is provided in the manuscript and supplementary data. Additional experimental rawdata is available upon request from the corresponding author. Declarations Competing interests The authors declare no competing interests. Ethics approval and patient consent The study was performed following the Declaration of Helsinki and approved by the Ethical Committees for the Region of Copenhagen, Denmark (H-18042385). Written informed consent was obtained from all participants. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Supplementary Material 2 Data Availability Statement The data supporting the findings of this study is provided in the manuscript and supplementary data. Additional experimental rawdata is available upon request from the corresponding author.