Phenome-wide causal proteomics enhance systemic lupus erythematosus flare prediction: A study in Asian populations

TL;DR

The integration of proteomic and clinical data significantly improves flare prediction in Asian SLE patients and identification of key proteins and their causal relationships with flare-related clinical markers provides valuable insights for proactive SLE management and personalized therapeutic approaches.

Abstract

Objective: Systemic lupus erythematosus (SLE) is a complex autoimmune disease characterized by unpredictable flares. This study aimed to develop a novel proteomics-based risk prediction model specifically for Asian SLE populations to enhance personalized disease management and early intervention. Methods: A longitudinal cohort study was conducted over 48 weeks, including 139 SLE patients monitored every 12 weeks. Patients were classified into flare (n = 53) and non-flare (n = 86) groups. Baseline plasma samples underwent data-independent acquisition (DIA) proteomics analysis, and phenome-wide Mendelian randomization (PheWAS) was performed to evaluate causal relationships between proteins and clinical predictors. Logistic regression (LR) and random forest (RF) models were used to integrate proteomic and clinical data for flare risk prediction. Results: Five proteins (SAA1, B4GALT5, GIT2, NAA15, and RPIA) were significantly associated with SLE Disease Activity Index-2K (SLEDAI-2K) scores and 1-year flare risk, implicating key pathways such as B-cell receptor signaling and platelet degranulation. SAA1 demonstrated causal effects on flare-related clinical markers, including hemoglobin and red blood cell counts. A combined model integrating clinical and proteomic data achieved the highest predictive accuracy (AUC = 0.769), surpassing individual models. SAA1 was highlighted as a priority biomarker for rapid flare discrimination. Conclusion: The integration of proteomic and clinical data significantly improves flare prediction in Asian SLE patients. The identification of key proteins and their causal relationships with flare-related clinical markers provides valuable insights for proactive SLE management and personalized therapeutic approaches.

Figures (6)

• Figure 1: Study design and patient follow-up schema. Schematic representation of the recruitment process, inclusion and exclusion criteria, and longitudinal assessment timeline for SLE patients over the 48-week study period.
• Figure 2: Comparative analysis of clinical variable importance in predicting SLE flares. (A) Ranking of clinical variables based on LR coefficients. (B) Relative importance of clinical variables determined by the RF algorithm, measured by the mean decrease in Gini impurity.
• Figure 3: Multi-dimensional proteomic profiling of SLE flares. (A) Volcano plot illustrating the distribution and statistical significance of 102 differentially expressed proteins between flare and non-flare groups ('the absolute value of the base-2 logarithm of the fold change' $\ge$ 1, $p < 0.05$). (B) Hierarchical clustering heatmap displaying normalized plasma protein expression levels, with a color gradient (blue to red) representing relative protein abundance. (C) Reactome pathway enrichment analysis of flare-associated proteins, with statistical significance indicated by -log10(p-value). (D) KEGG pathway analysis of proteins correlated with SLEDAI-2K scores. (E) STRING-based protein-protein interaction network of SLEDAI-2K-associated proteins, where node size reflects the degree of interaction. (F) Correlation matrix showing the relationships between upregulated proteins positively associated with SLEDAI-2K scores.
• Figure 4: Mendelian randomization analysis of SAA1's causal effects on SLE flare risk factors. Forest plots from random-effects inverse-variance weighted (IVW) analyses depicting the genetic causal relationships between SAA1 and (A) anti-inflammatory and antirheumatic medication use, (B) hemoglobin levels, (C) red blood cell count, and (D) hematocrit. Odds ratios and 95% confidence intervals are shown.
• Figure 5: Sensitivity analysis of Mendelian randomization results using the leave-one-out method. (A) Anti-inflammatory and antirheumatic medication use. (B) Hemoglobin levels. (C) Red blood cell count. (D) Hematocrit. This analysis evaluates the robustness of the causal relationships between SAA1 and four key outcomes: anti-inflammatory and antirheumatic medication use, hemoglobin levels, red blood cell count, and hematocrit. Each subplot (A-D) represents one of these outcomes, with individual points showing the Mendelian randomization estimate after excluding one SNP from the analysis, while the vertical line indicates the estimate using all SNPs. This approach helps identify potential outlier SNPs that may disproportionately influence the overall results.