Cinnamaldehyde protects against early diabetic retinopathy: an integrative In vivo and computational study

Diabetic retinopathy (DR) is a leading cause of vision impairment, and current treatments are limited to advanced stages. Cinnamaldehyde (CA), the primary bioactive constituent of cinnamon, has demonstrated anti-apoptotic and anti-diabetic properties, but its role and mechanism in early DR remain unexplored. This study investigated whether CA protects against early DR and whether the effect is associated with phosphoinositide 3-kinase (PI3K)/protein kinase B (AKT) pathway activation.

This study investigated whether CA exerts anti‑apoptotic effects in early DR and whether these effects are associated with PI3K/AKT pathway phosphorylation. Network pharmacology, molecular docking, molecular dynamics simulations, and ADMET prediction were employed as hypothesis‑generating and supportive computational tools to predict potential key targets, validate binding stability, and assess drug‑likeness.

Diabetes was induced in rats by streptozotocin. After 8 weeks of hyperglycemia, diabetic rats received daily oral CA (100 mg/kg) or vehicle until week 20. Retinal morphology and apoptosis were assessed by H&E and TUNEL staining. Expression of PI3K/AKT pathway components and apoptosis‑related markers was analyzed by immunohistochemistry, RT‑qPCR and Western blot. Network pharmacology was used to predict CA targets and DR‑related genes, followed by GO and KEGG enrichment. Molecular docking was performed on 10 targets (Bcl‑2, Bax, XIAP, CASP3, IL6, ERBB2, TNF‑α, EGFR, BDNF, AKT2), with subsequent 100‑ns molecular dynamics simulations and MM‑GBSA binding free energy calculations for four key targets (ERBB2, AKT2, Bcl‑2, TNF‑α). ADMET properties were predicted using ADMETlab 3.0.

CA did not lower blood glucose or prevent weight loss in diabetic rats (p > 0.05 vs. diabetic controls). However, CA was associated with ameliorated retinal histological changes and a reduction in TUNEL-positive cells in the ganglion cell layer (from 37.33 ± 7.09% to 18.67 ± 2.25%, p < 0.05). CA treatment was associated with upregulation of Bcl-2 and XIAP and downregulation of Bax and Caspase-3 at both protein and mRNA levels (all p < 0.05), and with increased PI3K/AKT phosphorylation (p < 0.05) without altering total PI3K or AKT. Re-analysis of animal-level densitometry showed that p-PI3K/PI3K and p-AKT/AKT in the D group were 27.1% and 32.5% of N levels, respectively, and that CA treatment increased these ratios to 213.5% and 227.2% of D levels. Network pharmacology predicted 77 common targets between CA and DR, with PI3K‑AKT among the top enriched pathways. Molecular docking revealed favorable binding of CA to ERBB2 (-6.275 kcal/mol), AKT2 (-5.678 kcal/mol), EGFR (-5.712 kcal/mol), TNF‑α (-5.560 kcal/mol), and Bcl‑2 (-5.503 kcal/mol). Single-trajectory 100-ns MD simulations were consistent with retention of CA in the AKT2 and ERBB2 binding pockets, with protein-aligned ligand displacements of ∼4-5 Å, whereas the ligand left its initial binding site in the Bcl-2 and TNF-α trajectories. ADMET prediction was consistent with favorable oral bioavailability, Caco-2 permeability, and predicted BBB permeability, but does not establish blood-retinal barrier penetration.

CA treatment was associated with amelioration of retinal histological changes and a reduction in TUNEL-positive cells in the GCL under persistent hyperglycemia, without a detected improvement in fasting glucose. The observed changes were associated with PI3K/AKT phosphorylation and modulation of apoptotic regulators. Computational analyses provide exploratory, hypothesis-generating evidence. The correlational nature of the PI3K/AKT findings, the absence of causal pathway intervention, and the exploratory nature of the computational predictions should be considered when interpreting these results.