Background: Histone deacetylases (HDACs) are attractive therapeutic targets for the treatment
of cancer and other diseases. There are numerous published patent applications till 2017. It was
claimed that novel HDACIs were optimized as potential drug candidates, designed for regional or systemic
release, and created as significant inhibitors.
Objective: In the present study, 3D-QSAR and molecular docking were used to provide a theoretical
basis for finding highly potent anti-tumor drugs.
Methods: QSAR was used to generate models and predict the HDAC1 inhibitory activity using the
Sybyl program (x1.2 version). Biaryl benzamides (n=73) as selective HDAC1 inhibitors were selected
as our data set, which was split randomly into training (n=63) and test sets (n=10). Docking was carried
out using the MOE software. Partial least square was used as QSAR model-generation method. External
validation and cross-validation (leave-one-out and leave-10-out) were used as validation methods.
Results: Both CoMFA (q2, 0.663; rncv
2 , 0.909) and CoMSIA models (q2, 0.628; rncv
2 , 0.877) for training
set yielded significant statistical results. The predictive ability of the derived models was examined
by a test set of 10 compounds and external validation results displayed rpred
2 and rm
2 values of 0.767 and
0.664 for CoMFA and 0.722 and 0.750 for CoMSIA, respectively.
Conclusion: The obtained models showed a good predictive ability in both internal and external validation
and could be used for designing new biaryl benzamides as potent HDAC1 inhibitors in cancer
treatment. The amido and amine groups of benzamide part as scaffold and the bulk groups as a hydrophobic
part were key factors to improve inhibitory activity of HDACIs.