**Background:** Diabetic retinopathy (DR) is a leading cause of blindness among working-age adults, with projections estimating that by 2045, approximately 161 million adults will be affected by DR, 45 million by vision-threatening DR, and 29 million by diabetic macular edema. The disease is a complex microvascular complication of diabetes mellitus, but growing evidence indicates that neurodegeneration is an early event, leading the American Diabetes Association to define DR as a highly specific neurovascular complication. The exact mechanisms driving DR progression are not fully understood, but multiple biochemical pathways are implicated, including the polyol pathway, protein kinase C (PKC) activation, accumulation of advanced glycation end products (AGEs), the hexosamine pathway, and oxidative stress. Recently, non-coding RNAs (ncRNAs)—including microRNAs (miRNAs), long non-coding RNAs (lncRNAs), and circular RNAs (circRNAs)—have emerged as key regulators of these pathways and as potential biomarkers for early diagnosis and monitoring of DR.
**Methods:** This is a narrative review that synthesizes findings from a wide range of published studies on the biochemical pathways and ncRNAs involved in DR. The authors discuss epidemiological data, risk factors (e.g., disease duration, HbA1c, hypertension, obesity, sleep apnea), and the classification of DR into non-proliferative (NPDR) and proliferative (PDR) stages. They then detail the molecular mechanisms of hyperglycemia-induced damage, including the polyol pathway (increased conversion of glucose to sorbitol and fructose via aldose reductase and sorbitol dehydrogenase), the diacylglycerol (DAG)/PKC pathway (activation of PKC-β and -δ isozymes leading to pericyte apoptosis and basement membrane thickening), growth factor dysregulation (e.g., VEGF, IGF-1, HGF, PDGF, and pro-inflammatory cytokines such as IL-1β, IL-6, IL-8, TNF-α), and oxidative stress (overproduction of ROS via NADPH oxidase, mitochondrial dysfunction, and mtDNA damage). The review also covers the role of inflammation and leukostasis, noting that diabetic non-transgenic mice exhibit a three-fold increase in adherent leukocytes compared to controls. Finally, the authors summarize evidence on dysregulated ncRNAs: miRNAs (e.g., miR-21, miR-126, miR-409-5p, miR-27b-3p, miR-320a-3p), lncRNAs (e.g., XIST, MIAT, RNCR3, MALAT1, ANRIL, SOX2OT, ENST00000505731, NR-126161), and circRNAs (e.g., circ_0005015, circHIPK3, circZNF609, circZNF532, circEhmt1, circPWWP2A). Data are drawn from in vitro studies, animal models (e.g., STZ-induced diabetic mice, db/db mice), and human samples (plasma, serum, vitreous humor, aqueous humor, retinal tissues, and fibrovascular membranes).
**Key Results:** The review reports that hyperglycemia activates the polyol pathway, leading to sorbitol accumulation and increased AGE formation, which in turn activates PKC and oxidative stress. PKC-δ signaling is specifically implicated in pericyte degeneration. Oxidative stress, driven by mitochondrial ROS and NADPH oxidase (especially Nox2), causes mtDNA damage, particularly in the D-loop region, and impairs mitochondrial replication/repair enzymes (POLG1, POLG2, helicase). This results in a positive feedback loop of superoxide overproduction. Inflammation is characterized by leukostasis, with a three-fold increase in adherent leukocytes in diabetic mice and a two-fold increase in RAGE transgenic mice. Regarding ncRNAs, the review highlights that miR-409-5p is up-regulated in vitreous fluid of PDR patients and in retinal tissues of diabetic mice, and its knock-down suppresses VEGF-induced neovascularization in vitro and reduces acellular capillaries in vivo. miR-27b-3p and miR-320a-3p are significantly associated with high risk of DR, targeting thrombospondin-1. lncRNA MIAT is overexpressed in plasma of DR patients and in high-glucose-treated endothelial cells, regulating angiogenesis. lncRNA RNCR3 is up-regulated in diabetic retinas and fibrovascular membranes of PDR patients; its knock-down reduces retinal microvascular leakage, inhibits migration and tube formation, and decreases cytokines (IL-2, IL-3, IL-4, IL-5, IL-9, IL-13, IL-17, MCP-1, VEGF, TNF-α). lncRNA MALAT1 is up-regulated in diabetic retinas and vitreous humor, and its knock-down downregulates inflammatory mediators (IL-6, TNF-α) and modulates antioxidant defense via NRF2. circRNA circ_0005015 is up-regulated in diabetic retinas, vitreous, plasma, and fibrovascular membranes, promoting endothelial cell proliferation, migration, and tube formation. circZNF532 is up-regulated in pericytes and acts as a miR-29a-3p sponge; its overexpression protects against pericyte degeneration and vascular dysfunction.
**Clinical Implications:** The review underscores that ncRNAs hold promise as non-invasive biomarkers for early diagnosis, staging, and monitoring of DR. For example, five miRNAs (hsa-miR-195-5p, hsa-miR-20a-5p, hsa-miR-20b-5p, hsa-miR-27b-3p, hsa-miR-451a) have been validated for stratification of DR stages, and lncRNAs ENST00000505731 and NR-126161 show high sensitivity and specificity in distinguishing PDR from NPDR. Targeting specific ncRNAs (e.g., anti-miR-409-5p therapy, RNCR3 shRNA, MALAT1 siRNA) may offer novel therapeutic strategies to reduce neovascularization, inflammation, and neurodegeneration. However, the authors caution that further research is needed to validate these findings in large, diverse populations, standardize isolation and quantification methods, and fully elucidate the mechanisms of ncRNA secretion and uptake. Ultimately, integrating ncRNA profiling into clinical practice could enable earlier intervention and personalized treatment for DR, reducing the substantial socio-economic burden of this sight-threatening disease.