# In-silico assessment of phytochemical derivatives generated using CHEESE webserver for advancement of druggable candidate in pancreatic cancer therapy

## TL;DR

This study uses the CHEESE webserver to generate and evaluate phytochemical derivatives for their potential to inhibit EZH2, a target in pancreatic cancer therapy.

## Contribution

The study demonstrates the effectiveness of the CHEESE webserver in generating phytochemical derivatives with improved binding affinity to EZH2.

## Key findings

- Several phytochemical derivatives showed higher binding affinities to EZH2 than the original compounds.
- Molecular dynamics simulations confirmed the stability of key protein-ligand complexes.
- The derivatives met pharmacokinetic and drug-likeness criteria, suggesting potential for further experimental validation.

## Abstract

Research efforts for pancreatic cancer (PC) therapy has led to investigations of numerous therapeutic targets, yet there are still limited efficacy outcomes. In 2021, the annual PC cases of 508,533 resulted in mortality of 505,752 for both sexes according to GLOBOCAN. In order to address this burden, enhancer of zeste homolog 2 (EZH2); an epigenetic regulator implicated in various cancers, has been an attractive target, due to promising tumor-suppressive effects in both preclinical and clinical studies. This effect was observed from drugs that have strong affinity towards EZH2, but there is need to improve the structural moieties for better interaction towards this protein. In this study, we employed a structure-based drug discovery approach using CHEESE webserver for rapid ligand-based screening to identify and evaluate phytochemical derivatives for their potential to bind EZH2. Five phytochemicals, namely Moracin P, Naringenin 5-rhamnoside, Pinostrobin 5- glucoside, Phytocassane A, and Sakuranin with best performance against EZH2-PPARs from our previous study was used to generate top ten new derivatives each. The derivatives were subjected to molecular docking, pharmacokinetic, and toxicity predictions. The top-performing derivative interacting with EZH2 were further subjected to ADMET profiling with favorable pharmacokinetic, toxicity properties, and meeting key drug-likeness criteria. Molecular docking results revealed that several derivatives of Moracin P, Naringenin 5-rhamnoside, and Phytocassane A displayed higher predicted binding affinities (− 6.4 to − 8.2 Kcal/mol) compared to the parent template previously assessed for EZH2, engaging critical residues through hydrogen bonds and hydrophobic interactions. Molecular dynamic simulation (MDS) over 200 ns further confirmed the stability of selected protein-ligand complexes, with Moracin_P7 and Pinostrobin 5-glucoside_5 exhibiting high degrees flexibility within the first 100 ns, but remained stable for the last 100 ns, while contact analysis highlighted consistent interactions with the active site residues Gln653, Asp657, Asp664, Ser669, Asn673, Phe678, His711 and Tyr731. Conclusively, these findings provide evidence that CHEESE webserver is suitable to generate phytochemical derivatives that can be explored as feasible candidate in EZH2 inhibitory study and laying the groundwork for further in-vitro and in-vivo validation.

The online version contains supplementary material available at 10.1007/s40203-025-00536-w.