Development and Validation of Predictive Quantitative Structure-Activity Relationship Models for Estrogenic...

Area of Science:

• Environmental toxicology
• Endocrinology
• Computational chemistry

Background:

• Hydroxylated polychlorinated biphenyls (OH-PCBs) are suspected endocrine disruptors.
• Nuclear receptor binding is a proposed mechanism for OH-PCB endocrine disruption.
• Experimental data on OH-PCB nuclear receptor binding are limited due to cost and time.

Purpose of the Study:

• To develop predictive models for the estrogenic activities of OH-PCBs.
• To support the hypothesis that OH-PCBs mediate endocrine disruption via nuclear receptor binding.
• To provide tools for assessing the risks of untested OH-PCB congeners.

Main Methods:

• Development of two quantitative structure-activity relationship (QSAR) models.
• Model I: Prediction of estrogenic activity for the estrogen receptor alpha (ERα) dataset using five 2D descriptors.
• Model II: Prediction of estrogenic activity for the estrogen receptor beta (ERβ) dataset using three 2D and three 3D descriptors.

Main Results:

• Model I incorporated descriptors from autocorrelation, Burden modified eigenvalues, chi path, and atom type electrotopological state classes.
• Model II utilized descriptors from autocorrelation, atom type electrotopological state, and Radial Distribution Function classes.
• Both models demonstrated robust, reliable internal and external validation metrics.

Conclusions:

• The developed QSAR models are suitable for predicting the estrogenic activities of novel OH-PCB congeners.
• These models can help prioritize OH-PCBs for further experimental investigation.
• The findings contribute to understanding the environmental toxicology and endocrine-disrupting potential of OH-PCBs.