Inflammation, a defense mechanism, happens during infections and injuries, involving complex immune responses. Medication using non-steroidal anti-inflammatory (NSAID) as Cyclooxygenase-2 (COX-2) inhibitors is clinically effective. It, however, has cardiovascular side effects. Thus, screening of COX-2 inhibitor from the nature, applying in silico study that is an effective and economical technique to reduce time-consuming in drug discovery process, is needed to be carried out to find drug candidates for anti-inflammatory with lower toxicity to cardiovascular. The virtual screening of ninety-two (92) compounds from Ardisia plants, which are Ardisia humilis Vahl and Ardisia elliptica, and current drugs for inhibiting COX-2 enzyme had been carried out using Molegro Virtual Docker v 7.0.0 (MVD), applying human COX-2 enzyme as target (PDB ID: 5IKQ). It, next, was visualized by performing PyMol software version 2.5.1. Lastly, a web-based prediction tools towards human ether-a-go-go related gene (hERG1) was used to study the cardiovascular toxicity, evaluating four best scores from Ardisia compounds and existing drugs. The results showed that the top four score compounds had been obtained and passed the Lipinski rule, including Isorhamnetin 3-sulfate, Isorhamnetin, Cornudentanone, and Rapanone, scoring of re-rank 115.27 kcal/mol, -105.87 kcal/mol, -105.10 kcal/mol and -104.38 kcal/mol, respectively. In addition, these compounds had shown comparative result with existing drugs such as Celecoxib, Rofecoxib, Meloxicam, Etodolac, and Meclofenamic acid. In addition, referring to the toxicity prediction, Isoharmnetin 3-sulfate and Rapanone showed a low-risk cardiovascular toxicity compared to existing drugs (Celecoxib and Etodolac). In silico study of compounds in Ardisia Plant, therefore, might assist further research to obtain potential candidates as anti-inflammatory drug with inferior side effect on cardiovascular.
