Integrated intraocular-plasma proteomics reveals conserved biomarkers for diabetic retinopathy progression: a multi-fluid biopsy study.
Summary
High-throughput proteomics of aqueous humour identified conserved protein trajectories during diabetic retinopathy progression, with neurofilament light chain (NFL) emerging as a cross-compartment biomarker. Plasma NFL in UK Biobank distinguished prevalent retinopathy and predicted incident retinopathy and vascular complications over 12 years, improving risk model performance (NRI and IDI).
Key Findings
- Temporal proteomics identified 40 candidates with monotonic changes across retinopathy progression; 25 were directionally conserved on validation (all p<0.05).
- Single-cell mapping localized 15 candidates, including NFL, to retinal neurons and glia.
- Baseline plasma NFL differentiated prevalent retinopathy (OR 1.98, 95% CI 1.61–2.42) and predicted incident retinopathy (HR 2.01, 95% CI 1.48–2.73).
- Plasma NFL predicted vascular complications (microvascular HR 2.28; macrovascular HR 1.49) over a median 12-year follow-up.
- NFL improved predictive performance of a conventional model (NRI 0.194; IDI 0.015).
Clinical Implications
Plasma NFL testing could support earlier identification and monitoring of diabetic retinopathy and stratify micro- and macrovascular complication risk, informing referral, surveillance intervals, and potential trial enrichment.
Why It Matters
Establishes NFL as a minimally invasive, pan-stage biomarker for diabetic retinopathy with external validation and prospective prognostic value, enabling improved risk stratification.
Limitations
- Discovery cohort was small (n=32), and observational analyses may be confounded by comorbid neuropathies influencing NFL.
- Generalizability across diverse populations and clinical settings, and clinical utility thresholds, require further study.
Future Directions
Prospective clinical utility studies to define actionable thresholds, evaluate integration into screening pathways, and test responsiveness to intervention in trials.
Study Information
- Study Type
- Cohort
- Research Domain
- Prognosis
- Evidence Level
- II - Prospective cohort analyses with external validation but without randomization
- Study Design
- OTHER