Background: Systemic lupus erythematosus (SLE) is a highly heterogenous multi-system autoimmune condition, which causes significant morbidity and mortality. This heterogeneity has contributed to the lack of success of clinical trials, resulting in few targeted treatments being approved. Improved stratification of SLE patients using biomarker profiles, particularly at the protein level, may enable more successful development of targeted treatments. However, to date, few studies have examined associations between individual proteins measured using untargeted proteomics, protein pathways and detailed SLE clinical phenotypes.

Objectives: We aimed to generate comprehensive serum profiles in patients with SLE using untargeted mass spectrometry (MS) based proteomics, and to explore associations between individual proteins, protein pathways and SLE clinical phenotypes and outcomes.

Methods: Unbiased quantitative proteomics was performed using liquid chromatography-tandem MS on serum samples from 60 adult SLE patients with detailed longitudinal clinical annotation. Proteins with missing values in >50% of patients were excluded, and probabilistic minimum imputation was performed in those with missing values in ≤50% of patients. Gene set variation analysis was performed using the Hallmark gene set, to quantify protein pathways present in the patient cohort. Associations between clinical features and protein pathways, as well as individual proteins, were assessed using differential abundance analysis (DA) with linear regression models. Age, sex, ethnicity and renal impairment (eGFR <60 ml/min) were included as covariates in multivariable linear regression models.

Results: 1,092 individual proteins and 23 Hallmark protein pathways were quantified in 58 SLE patients (1 end-stage renal failure patient and 1 additional outlier excluded). The median (IQR) age was 42 (34-52) years. 91% of patients were female, 48% were of Asian ethnicity, 52% were of European ethnicity and 48% had active disease at the time of blood sample collection, defined as SLEDAI-2K >4. Multiple significant associations between protein pathways and clinical phenotypes were identified using DA. Using multivariable linear regression, three protein pathways were positively associated with organ damage, including Kirsten rat sarcoma viral oncogene homolog (KRAS) signalling up (adjusted β-coef 0.16; 95% CI 0.06, 0.26; p=0.024), complement (adjusted β-coef 0.15; 95% CI 0.04, 0.25; p=0.049) and xenobiotic metabolism (adjusted β-coef 0.19; 95% CI 0.07, 0.30; p=0.024). Using multivariable DA, a single individual protein, that forms part of the complement and KRAS signalling up pathways, Cathepsin S, was positively associated with organ damage (adjusted β-coef 1.01; 95% CI 0.53, 1.48; p=0.030). The myogenesis pathway was positively associated with a history of biopsy-proven lupus nephritis (LN) (adjusted β-coef 0.023; 95% CI 0.11, 0.35; p=0.010). The apoptosis pathway was positively associated with prednisolone use (adjusted β-coef 0.20; 95% CI 0.10, 0.29; p=0.003). No individual proteins were associated with biopsy-proven LN or prednisolone use. Two protein pathways were associated with Asian ethnicity using univariable regression only: complement (β-coef -0.16; 95% CI -0.25, -0.06; p=0.002) and Il2 STAT 5 signalling (β-coef -0.16; 95% CI -0.27, -0.05; p=0.004). One individual protein that forms part of the complement pathway, fibronectin, but none from the IL-2 STAT 5 signalling pathway was associated with Asian ethnicity (adjusted β-coef 0.89; 95% CI 0.41, 1.36; p=0.029).

Conclusions: Untargeted serum proteomics identified distinct protein pathways associated with organ damage, biopsy-proven LN, prednisolone use and ethnicity in patients with SLE. Corresponding individual proteins forming part of these pathways were identified in association with organ damage and ethnicity. This highlights the potential for untargeted serum proteomics to advance our understanding of SLE pathogenesis and to elicit clinically useful biomarkers.

REFERENCES: NIL.

Acknowledgments: NIL.

Disclosure of Interests: Laura Eades Research funding from Janssen-Cilag, Hossein Valipour Kahrood: None declared, Joel Steele: None declared, Joanna Kent CSL Seqirus, Arthur R Kitching: None declared, Ralf Schittenhelm: None declared, Rangi Kandane-Rathnayake Research grants from BMS, GSK, Janssen and Novartis, Alberta Hoi AstraZeneca, Novartis, Seqirius, Janssen, Roche, GSK, Recordati, UCB, AstraZeneca, Merck Serono, GSK, BMS, Eric Morand AstraZeneca, Merck Serono, Novartis, DragonFly Tx, AstraZeneca, Biogen, Bristol Myers Squibb, Galapagos, Gilead, Genentech-Hoffman La Roche, GlaxoSmithKline, Merck Serono, Novartis, Priovant, IgM, Janssen, Quell, Remegen, Takeda,

Zenas, AbbVie, Amgen, AstraZeneca, Biogen, Bristol Myers Squibb, Eli Lilly, Merck Serono, Genentech-Hoffman La Roche, GSK, Janssen, Novartis, Takeda, Union Chimique Belge, Fabien Vincent Research funding from Janssen-Cilag, CSL, Astra-Zeneca, BMS.