Molecular insights and ADMET profiling of iridoids as candidate leads for erectile dysfunction therapy

Erectile dysfunction (ED) significantly affects male psychosocial health and has an increasing incidence. Although phosphodiesterase 5 (PDEi-5) inhibitors are the first-line therapy, the occurrence of visual side effects resulting from cross-inhibition of the PDE-6 isoform limits their adoption. In this context, the plant Veronica anagallis-aquatica L. emerges as a promising source of bioactive therapeutic compounds. This work investigated the inhibitory potential and selectivity of eight iridoid glycosides from this species against PDE-5 and PDE-6. The results identified catalposide as the most potent ligand, exhibiting an affinity energy of -8.6 kcal/mol against PDE-5, surpassing the reference drug Vardenafil (-8.3 kcal/mol) and showing high selectivity towards PDE-6. Trajectory analyses confirmed the robustness of the catalposide-PDE-5 complex, which exhibited smaller structural fluctuations than the commercial control. Despite initial predictions indicating low oral bioavailability, structural optimization through increased sp3 fraction proved to be a viable strategy for enhancing membrane permeability. In conclusion, the findings position catalposide as a promising prototype for the development of more selective and safe drugs for the treatment of ED.

The two-dimensional structures of the ligands were constructed in Marvin JS and subjected to energy minimization using the semi-empirical PM7 method in MOPAC software. The preparation of macromolecules (addition of polar hydrogens and Gasteiger and Kollman charges) was performed in AutoDockTools. Virtual screening by molecular docking was performed in AutoDock Vina, with exhaustiveness set to 64. The stability of the protein-ligand complexes was evaluated through 500 ns molecular dynamics (MD) simulations using GROMACS 2020.4 software. The systems were solvated with the three-point water model (TIP3P) in a cubic box and described by the CHARMM36 force field. The ligand parameterization was generated via the SwissParam server. The simulations employed the LeapFrog integrator, temperature maintained constant at 310 K via V-rescale thermostat, and pressure coupling maintained at 1 bar using the Parrinello-Rahman method. For predictions of physicochemical properties, pharmacokinetic optimization, and ADMET profile analysis (including hERG toxicity), the predictive servers ADMETlab 3.0, optADMET, ADMET-AI, XenoSite, StopTox, and Pred-hERG were used. Three-dimensional structural complexity was evaluated from the calculation of the MCE-18 function.