Search for a command to run...
Abstract Importance Cardiovascular disease (CVD) is a major cause of morbidity/mortality in juvenile-onset systemic lupus erythematosus (JSLE), yet no reliable tools exist to stratify CVD-risk. Objective To identify serum biomarkers associated with atherosclerosis progression and response to atorvastatin. Design/Setting We used data/samples from a sub-cohort of the APPLE trial (2009) which investigated atorvastatin vs. placebo to reduce atherosclerosis progression in JSLE, measured by change in carotid intima-media thickness (CIMT), and conducted a baseline autoantibody diagnostic-accuracy biomarker study. Participants/Exposure APPLE trial participants (randomized 1:1 to atorvastatin vs. placebo) with matched baseline serum samples and stratified based on 36-month CIMT progression were included in the analysis. Main Outcomes and Measures Baseline serum autoantibodies were profiled using a functional proteomic platform (Sengenics, N=94). Empirical Bayes moderated t-test and Receiver Operating Characteristic (ROC) based logistic regression were applied to identify autoantibody signatures predictive of high vs. low atherosclerosis progression. Results Ninety-four children and young people with JSLE (age mean [SD] =15.3 [2.4] years; 73 [78%] female, 8 [8.5%] Asian, 23 [24.5%] Black, 43 [45.7%] White, and 20 [21.3%] Other) were evaluated. Autoantibody levels against six novel autoantigens (STK24, RAD23B, HDAC4, STAT4, SEPTIN9, NFIA) classified high vs. low CIMT progression in the placebo arm (combined AUC 0.87, 95% CI −0.75 to 0.96). In the atorvastatin arm, autoantibodies to eight autoantigens (ABI1, ATP5B, CSNK2A2, NRIP3, PRKAR1A, PDK4, BATF, NUDT2), distinguished the statin responders vs. non-responders (combined AUC 0.96, 95% CI −0.88 to 1). An additional 27-autoantibody signature predicted response/partial response to atorvastatin (AUC 0.88, 95% CI – 0.76 to 0.97). Protein–protein interaction analysis identified endothelial disruption and lipid infiltration as key atherosclerosis mechanisms in atorvastatin non-responders. Combining the autoantibody prediction models with disease parameters and a metabolic signature did not increase model performance in either placebo (AUC 0.81, 95% CI – 0.68 to 0.94 vs. 0.87, 95% CI −0.75 to 0.96) or sttin arms (AUC 0.84, 95% CI −0.73 to 0.95 vs. 0.88, 95% CI −0.76 to 0.97). Conclusions and Relevance This study identified novel autoantibody signatures for atherosclerosis progression and statin response in JSLE, with potential utility for precision medicine approaches for CVD-risk management. Key Points Question Can functional proteomic analyses identify autoantibody signatures predictive of atherosclerosis progression and response to statin treatment in children and young people with juvenile-onset systemic lupus erythematosus? Findings Using baseline samples from the APPLE trial (1:1 RCT of atorvastatin vs placebo), we identified novel autoantibody profiles that accurately distinguished individuals with high versus low carotid intima-media thickness progression over three years in both placebo (AUC 0.87, 95% CI-0.75 to 0.96) and atorvastatin groups (AUC 0.96, 95% CI-0.88 to 1). Meaning Autoantibody signatures show strong potential for early risk stratification and for identifying those most likely to benefit from statin therapy.