**Background:** Diabetic retinopathy (DR) is a progressive, vision-threatening complication of diabetes, with numbers projected to reach 191 million by 2030. Current diagnostic methods (fundus photography, fluorescein angiography, OCT) are semi-quantitative, subjective, and detect disease only after vascular lesions appear, missing early molecular changes. There is a critical need for early, quantifiable biomarkers. Retinol binding protein 3 (RBP3), also known as interphotoreceptor retinoid-binding protein (IRBP), is the most abundant soluble protein in the interphotoreceptor matrix (IPM), secreted exclusively by photoreceptors. It shuttles retinoids between photoreceptors and retinal pigment epithelium (RPE), transports docosahexaenoic acid (DHA), and has antioxidant properties. Multiple studies show RBP3 levels decrease in diabetic retinas and vitreous, inversely correlating with DR severity. Overexpression or intravitreal injection of recombinant RBP3 in rodents inhibits DR-related events, including VEGF-mediated retinal vascular permeability. Two-photon excitation fluorescence (TPEF) uses near-infrared light for deep, non-invasive retinal imaging with high contrast and subcellular resolution, and has been safely used in human eyes. This review argues for using TPEF to quantify RBP3 in vivo as an early DR biomarker.
**Methods:** This is a narrative review synthesizing evidence from published studies on RBP3 structure, function, and its link to DR, as well as TPEF technology and its ophthalmic applications. The authors discuss the molecular mechanisms of DR, including hyperglycemia-induced oxidative stress, inflammation, advanced glycation end products (AGEs), protein kinase C (PKC) activation, and VEGF upregulation. They highlight that oxidative stress originates in the avascular outer retina, particularly in photoreceptors and RPE cells, and that RBP3 may act as a free radical scavenger due to its multiple cysteine residues. The review cites studies showing RBP3 mRNA and protein are reduced in human retinas from diabetic donors, and that RBP3 levels in vitreous are 1000- to 5000-fold higher than in serum, making vitreous a more reliable sample. TPEF is described as capable of distinguishing retinoids and quantifying biomolecules in vivo, with fluorescence lifetime imaging (phasor analysis) enabling separation of overlapping fluorophores.
**Key Results:** The review does not present new experimental data but summarizes key findings from prior work: (1) RBP3 levels in the retina and vitreous inversely correlate with DR severity and progression. (2) Intravitreal injection of recombinant RBP3 in diabetic mice reduces VEGF-mediated retinal vascular permeability in a dose-dependent manner, decreases mRNA levels of Vegfa and Il-6, and increases ERG amplitudes (oscillatory potentials OP2, OP3, and B waves). (3) Overexpression of RBP3 via lentiviral vector or rhodopsin promoter also protects against DR. (4) RBP3 has a half-life of approximately 11 hours in the IPM and is rapidly turned over via RPE endocytosis. (5) TPEF has been used to image retinosomes, RPE storage compartments, and retinoid dynamics in vivo, with safety established for human use. (6) RBP3 binds retinoids with preference: 11-cis-retinal > all-trans-retinol > all-trans retinal > 11-cis retinol, and also binds DHA and oleic acid. (7) RBP3 is composed of four modules (~300 amino acids each), each with a hydrophobic cleft for ligand binding, allowing multiple ligands per molecule.
**Clinical Implications:** The authors propose that TPEF-based quantification of RBP3 in the IPM could serve as a non-invasive, objective, early diagnostic tool for DR, detecting molecular imbalances before clinical signs (microaneurysms, hemorrhages, neovascularization) appear. This would enable earlier lifestyle and dietary interventions, potentially preventing progression to vision-threatening DR. The approach could reduce reliance on subjective image interpretation and costly specialized personnel, as the readout would be a quantifiable number. The review acknowledges that the method currently requires injection of modified ligands or antibodies (micro-invasive), but compares this favorably to existing invasive treatments (anti-VEGF injections, laser photocoagulation). The authors emphasize that RBP3 is a 'clean' biomarker, exclusively produced by photoreceptors, and its decrease reflects oxidative stress and photoreceptor dysfunction. They call for testing the hypothesis that RBP3 measured in vivo can indicate early DR development, potentially transforming screening and management of this epidemic disease.