Hypothesis

Innate immune memory (trained immunity), mediated by persistent epigenomic reprogramming of circulating monocytes via histone 3 lysine 4 trimethylation (H3K4me3) and histone 3 lysine 27 acetylation (H3K27ac) at pro-inflammatory gene promoters (TNF-α, IL-6, IL-1β), accumulates during RA flares and fails to fully reset during remission. We hypothesize that serial chromatin immunoprecipitation sequencing (ChIP-seq) of circulating CD14+ monocytes, quantified as a Trained Immunity Burden Score (TIBS) — a weighted composite of H3K4me3/H3K27ac peak intensities at 12 canonical trained immunity loci — predicts flare recurrence 6–14 weeks before DAS28 elevation, independently of adaptive immune markers (RF, ACPA, lymphocyte subsets).

Rationale

Current RA flare prediction relies on adaptive immune biomarkers and composite disease activity indices that detect flares only at or near clinical onset. However, the innate immune system retains "memory" through metabolic and epigenomic reprogramming of myeloid progenitors and circulating monocytes — a phenomenon termed trained immunity. In RA, synovial macrophages exhibit trained immunity signatures, but whether circulating monocyte epigenomic states serve as predictive biomarkers for flare timing remains unexplored.

Key biological premises:

1. β-glucan and oxidized LDL induce trained immunity via mTOR/HIF-1α metabolic rewiring and SET7-mediated H3K4me3 deposition
2. RA monocytes show persistent H3K27ac enrichment at inflammatory enhancers even during clinical remission (de novo trained immunity from endogenous DAMPs)
3. TIBS should capture the cumulative epigenomic "debt" that predisposes to flare re-initiation when stochastic perturbations exceed the trained threshold

Testable Predictions

1. Primary: TIBS measured from peripheral CD14+ monocytes at baseline predicts time-to-next-flare with C-statistic >0.78, adjusting for DAS28, CRP, ESR, RF, and ACPA (Cox proportional hazards)
2. Secondary: TIBS trajectory slope (ΔTIBS/month) >1.5 SD above cohort mean precedes DAS28 >3.2 by 6–14 weeks with sensitivity >75% and specificity >70%
3. Mechanistic: High-TIBS monocytes stimulated ex vivo with TLR2/4 ligands produce >2-fold TNF-α/IL-6 versus low-TIBS monocytes (functional validation)
4. Negative control: TIBS does not correlate with peripheral T/B cell activation markers (CD69, CD25, plasmablast frequency), confirming innate-specific signal

Study Design

Prospective observational cohort, n=150 RA patients in clinical remission (DAS28 <2.6), monthly blood draws for 12 months. ChIP-seq on sorted CD14+ monocytes at 12 trained immunity loci. Primary endpoint: time-to-first-flare (DAS28 >3.2 on consecutive visits). Adjusted for age, sex, disease duration, DMARD regimen, and glucocorticoid use.

Limitations

• ChIP-seq cost and technical complexity limit clinical scalability (ATAC-seq or targeted methylation panels as pragmatic alternatives)
• Trained immunity loci selection based on β-glucan/BCG models may not fully capture RA-specific epigenomic reprogramming
• Confounding by glucocorticoid-induced epigenomic changes requires careful adjustment
• Single-center design limits generalizability; multi-center validation needed
• Monocyte heterogeneity (classical vs. intermediate vs. non-classical) may dilute signal if not sorted

Clinical Significance

If validated, TIBS would represent the first innate immune-specific predictive biomarker for RA flare, enabling pre-emptive therapeutic escalation weeks before clinical relapse. This paradigm shifts flare prediction from adaptive immunity and inflammatory markers toward epigenomic innate immune surveillance, potentially identifying a therapeutic window for trained immunity modulators (e.g., mTOR inhibitors, HDAC inhibitors) as flare-prevention strategies.

LES AI • DeSci Rheumatology