Background: Primary Sjögren’s syndrome (pSS) is an autoimmune disease, known for its disabling effect and chronic course. One of the peculiar symptoms is the lachrymal glands dryness associated often but not limited to dry mouth, dental disorders, joint pain, fatigue, and, in severe cases, systemic complications. The most relevant clinical feature is the infiltration of lymphocytes within the salivary glands and the development of an autoimmune endocrinopathy that can overstimulate lymphocytes until the development of lymphoma in 5% of the patients. pSS treatments are limited, and a deeper disease understanding is mandatory. Recently, Soret et al ¹ proposed a novel classification of pSS patients, in line with the PRECISESADS project ² , aiming to reclassify the autoimmune diseases based on their biology more than the clinical features.

Objectives: The ‘interferon’ cluster 1 (C1), ‘healthy-like’ cluster 2 (C2), ‘lymphoid’ cluster 3 (C3) and ‘inflammatory’ cluster 4 (C4) are analysed with novel datasets and omics from the same patients of the Soret et al study, including RNA-seq data from B lymphocytes and metabolomics data from plasma and urine. The multi-omics data integration by the MOFA algorithm is applied to extract factors able to catch the common variance from the novel and older omics. This study aims to extend the previous work and identify metabolomics markers easily obtainable with routine analysis to classify new pSS patients and provide the best care.

Methods: Bioinformatics analyses were performed on the PRECISESADS datasets, including transcriptomics, metabolomics, methylomics and clinical data from over 300 pSS patients. The B-cell transcriptome was analysed using DESeq and GSEA. Plasma and urine metabolomics peak changes were quantified, statistically tested, and annotated using the Ceu Mass Mediator database. Common sources of variation among all the databases were identified using the MOFA integration analysis for each cluster, and the factor tested to be significantly discriminant to CTRLs. The clustering was performed in B-cell, plasma and urine data by linear discriminant analysis (LDA).

Results: The B cell transcriptome highlighted the clusters C1 and C3 as the most affected by the interferon pathway, while C2 and C4 showed few differences compared to CTRLs. The cluster C4, marked by lymphopenia, had a low contribution of B lymphocytes in driving this patient cluster. Glycosylation genes (GALNTL6, MGAT3 and ENOSF1) contributed to the C2 and C4 differences among the clusters, while C1 and C3 by interferon signalling. Metabolomics analysis shed light on differences only in the plasma C1 cluster, where Lysophosphatidylcholine (LysoPC), phosphatidylinositol (PI) and neutral sphingolipids were upregulated, together with metabolites related to protein and nucleotide degradation. All clusters had a MOFA factor linked to interferon except the C2, where a single significant factor driven by B cell genes was associated with epigenetic modifications. Cluster 4 showed a factor associated with apoptosis in line with the lymphopenia, and carnitine complex showed a protective role in C1, C3, and C4 clusters, always contributing against their phenotype. LDA unveiled the drivers of the cluster differences, including interferon for B lymphocytes and cholines-associated lipids and phosphatidylinositol for plasma.

Conclusion: This study provided novel details about the clustering of pSS patients observed in other studies ^1,2 . B lymphocytes in cluster C4 showed little difference compared to CTRLs, while glycosylation, interferon signalling and epigenetics are proposed as drivers in B cell alteration in the other Sjogren clusters. PI, choline lipids and carnitine were identified in plasma as discriminant markers in the pSS clustering prediction, making them promising for their easy clinical measurement.

REFERENCES: [1] Soret, P. A new molecular classification to drive precision treatment strategies in primary Sjögren’s syndrome. Nat Commun 12, 3523 (2021).

[2] Barturen, G. et al. Integrative Analysis Reveals a Molecular Stratification of Systemic Autoimmune Diseases. Arthritis Rheumatol. Hoboken NJ 73 , 1073–1085 (2021).

Acknowledgements: NIL.

Disclosure of Interests: CRISTIAN IPERI: None declared, Alvaro Fernández-Ochoa: None declared, Jacques-Olivier Pers: None declared, Nathan Foulquier: None declared, Guillermo Barturen: None declared, Marta Alarcon-Riquelme As part of the public European project PRECISESADS from the Innovative Medicines Initiative Joint Undertaking under Grant Agreement Number 115565. Innovative Health initiative from the European Union with in-kind contributions from the pharmaceutical industry (Sanofi, Roche, GSK, BMS, Novartis, Janssen, Tekada, Astra Zeneca and Pfizer. Payments are within the project and only BMS has made direct payments to her institution for personnel., Divi Cornec: None declared, Anne Bordron: None declared, Christophe Jamin: None declared.