Using Cells and Mathematical Models to Interpret Thyroid Hormone Disruption

May 25, 2021

by Jeff Fisher

This story begins in 1998 when I reviewed a technical document for the chemical perchlorate, now known to be widely distributed in the environment on earth and has been found in high concentrations on Mars.  In 1998, the subject of this document was a proposed safe drinking water level for perchlorate based on historic medical use of perchlorate as a probe of thyroid function,  in the diagnostic called the thyroid discharge test.  This document changed my career to this day. Before this day in 1998 I was happily working with chlorinated solvents, halon replacements, an occasional explosive, and jet fuels.  I was part of a team involved in the early development and application of physiologically based pharmacokinetic (PBPK) models in toxicology.  It was fun.    

After reading more about perchlorate, I decided that it would be a great chemical for PBPK modeling. We could easily provide better health-based risk estimates for safe levels of perchlorate than the 1998 document.  Perchlorate readily enters the body, is not metabolized, interferes with thyroidal uptake of iodine, and is cleared from the body in urine.  It was so simple, the absence of metabolism made this chemical so appealing!  And so started my quest to understand better how the hypothalamic-pituitary-thyroid (HPT) axis functions on a quantitative basis.  I was moving into pharmacodynamics and  not looking back.  Many people were involved in the perchlorate research over the years and their stories may be different than mine. This is my story, and I am sticking to it.  

Years of research on perchlorate ensued along with struggles to stay current with scientific advances in understanding the complexity of the HPT axis in the thyroid gland and brain.  At some point, probably after all the work was finished to describe perchlorate’s inhibition of thyroidal uptake of iodide by the sodium iodide symporter protein, the challenges were not about perchlorate, per se, but developing pharmacodynamic models that would describe the perchlorate dose related HPT axis perturbations.   These models were called biologically based dose (BBDR) HPT axis models.  BBDR models are a combination of PBPK models coupled with simple or complex mathematical representations of biology and  pharamcodynamic responses. Google BBDR and perchlorate.    

I was intrigued as I talked with thyroid endocrinologists and asked about ‘how things work’.  I heard many different answers.  We (myself and students) decided to use the models to help understand how the HPT may work by setting boundary conditions for many moving parts (model parameters) of the HPT model.  In one case, it became clear (at least to me) we could not explain why rats dosed with perchlorate became “hypothyroid” (increased serum thyroid stimulating hormone (TSH) shortly after dosing.  The predicted thyroidal iodide stores were still substantial immediately after dosing with perchlorate.  Perchlorate’s primary mode of action is blocking thyroid uptake of iodide.  We still don’t have an answer for why the rat’s HPT axis responds so quickly after administration of perchlorate.   BBDR HPT axis modeling work for perchlorate continued into 2015 with a focus on humans and pregnancy and the interactions of dietary iodine status and exposure to perchlorate.  I worked with many wonderful people bridging the disciplines of developmental neurotoxicology, endocrinology, and biological based modeling.  It was fun.  

Fast forward to 2021, 23 years after I read the 1998 document. I am thrilled to still be using pharmacodynamic models as a tool to understand dose-response characteristics of endocrine disruption and its implications on human health. The questions never stop.  Why do serum T3 levels go up? Or down?  Why does TSH remain unchanged but serum thyroid hormones are lowered?  The list is long. Does chemical induced hepatic UGT induction in laboratory animals translate to human HPT axis pertubations?    

More  in vitro tools are now available, including molecular biology.  We can perform in vitro studies to examine specific modes of action related to HPT disruption, such as hepatic enzyme induction.  We can interpret the in vitro results by conducting whole body thyroid hormone simulations to determine the potential for endocrine disruption.  Many questions remain, however, on how to link HPT axis disturbances with apical effects or biomarkers of effect.  I believe that these BBDR HPT models are key to connecting the dots for empirical data and will be valuable for advancing safety science.   

Are you pondering how to address toxicology questions related to disturbances in thyroid hormones? Let’s chat. Email me at:  Be sure to check out our recent white paper on our approach to modeling thyroid hormone disruption.