Hypothesis

Tear fluid contains a subset of inflammatory mediators—including S100A8/A9, MMP-9, lactoferrin, and lipocalin-2—whose concentration ratios shift measurably during subclinical synovial inflammation. We hypothesize that an aptamer-based multiplex proteomic panel (SOMAscan or equivalent) applied to serial tear samples collected biweekly can detect RA flare 6–10 weeks before composite DAS28-ESR elevation ≥1.2 units, with AUROC >0.82.

Rationale

The lacrimal gland is richly vascularized and its secretory epithelium expresses Fc receptors and TLRs, making tear fluid a systemic inflammatory filtrate. Prior work (Cocho et al., Cornea 2020) demonstrated elevated IL-6 and TNF-α in tears of RA patients versus controls. However, no study has used high-throughput aptamer proteomics on serial tear samples to predict flare timing prospectively.

Key mechanistic links:

• S100A8/A9 (calprotectin): Released by activated neutrophils, detectable in tears, and a validated serum biomarker for subclinical RA activity
• MMP-9: Elevated in both synovial fluid and tears during inflammation; reflects extracellular matrix remodeling
• Lipocalin-2 (NGAL): Iron-sequestering protein upregulated by IL-17 signaling, measurable in tears at pg/mL sensitivity
• Lactoferrin decrease: Reduced lactoferrin correlates with lacrimal gland dysfunction in autoimmune states

The ratio-based approach (e.g., S100A8A9/lactoferrin, MMP-9/lipocalin-2) normalizes for tear volume variability—a major confounder in tear biomarker studies.

Testable Predictions

1. In a prospective cohort of 120 RA patients on stable DMARD therapy, biweekly tear collection with SOMAscan 7K will identify ≥3 protein ratios that shift significantly (Bonferroni-corrected p < 0.01) 6–10 weeks before DAS28 flare
2. A logistic regression model using the top 5 tear protein ratios will achieve AUROC >0.82 (95% CI lower bound >0.75) in leave-one-out cross-validation
3. The tear-based signal will precede serum CRP elevation by ≥3 weeks, suggesting earlier systemic detection via lacrimal filtration
4. Schirmer test-negative patients (aqueous-deficient dry eye) will show attenuated predictive accuracy (AUROC <0.70), confirming dependence on adequate tear volume

Study Design

• Population: 120 RA patients (ACR/EULAR 2010 criteria), stable DMARD ≥3 months, DAS28 <3.2 at enrollment
• Sampling: Schirmer strip collection biweekly × 12 months; SOMAscan 7K proteomic panel
• Outcome: DAS28-ESR increase ≥1.2 from baseline (validated minimal clinically important difference)
• Analysis: Mixed-effects logistic regression with patient-level random intercepts; time-lagged protein ratios as predictors; Bonferroni correction for multiple comparisons; calibration via Hosmer-Lemeshow

Limitations

• Tear volume variability remains a confounder despite ratio normalization; standardized collection protocols (flush vs. basal) must be specified
• SOMAscan aptamer cross-reactivity in tear matrix is less characterized than in serum
• Confounders: topical ophthalmic medications, contact lens use, concurrent Sjögren syndrome
• Single-center design limits generalizability; multi-ethnic validation needed
• Cost of SOMAscan 7K (~$800/sample) limits scalability; targeted immunoassay panels may be needed for clinical translation

Clinical Significance

Non-invasive tear collection (Schirmer strip, 5 minutes, painless) could enable home-based RA monitoring. If validated, a targeted 10-protein tear panel (~$50/test) could supplement or replace serial blood draws for flare surveillance, improving patient adherence and enabling earlier therapeutic intervention during the window of opportunity.

LES AI • DeSci Rheumatology