Mathematical Modeling of the Pituitary–Thyroid Feedback Loop: Role of a TSH-T3-Shunt and Sensitivity Analysis

Julian Berberich, Johannes W. Dietrich, Rudolf Hoermann, Matthias A. MüllerFrontiers in Endocrinology 2018 · DOI

TL;DR

This paper introduces a refined mathematical model of the human pituitary-thyroid feedback loop by incorporating a novel "TSH-T3-shunt" within the thyroid gland. This extension better explains observed clinical phenomena like circadian FT3 rhythms and offers new insights into the sensitivity of thyroid hormone regulation, potentially improving diagnostic approaches for thyroid disorders.

What Problem Does This Paper Address?

Diagnosing thyroid disorders, particularly subclinical hypothyroidism, can be challenging due to the vague upper limit of the normal range for Thyroid-Stimulating Hormone (TSH). Furthermore, a significant portion of hypothyroid patients on levothyroxine (L-T4) replacement therapy still experience poor quality of life despite seemingly normal TSH levels. This suggests that current models of the hypothalamic-pituitary-thyroid (HPT) axis may be too simplistic and fail to capture crucial dynamic aspects and regulatory mechanisms. Specifically, the role of triiodothyronine (T3), a key thyroid hormone, and its production pathways, especially within the thyroid itself, have not been fully integrated into dynamic models.

Methods

The authors extend a previously established mathematical model of the HPT axis. The core innovation is the incorporation of a "TSH-T3-shunt" within the thyroid. This shunt represents a direct pathway for T3 production stimulated by TSH, in addition to the traditional conversion of thyroxine (T4) to T3 via deiodinases. This new mechanism is modeled using nonlinear Michaelis-Menten-Hill kinetics.

The model's parameters were largely adopted from previous work, with new parameters related to the TSH-T3-shunt (GT3 and k) and the activity of type 1 deiodinase (GD1) being re-estimated. This estimation involved a least-squares fitting of the model's predicted equilibrium FT3 levels to FT3 measurements from a clinical study.

A significant portion of the paper is dedicated to a sensitivity analysis. This analysis quantifies how changes in specific model parameters (e.g., thyroid secretory capacity GT) affect the model's output (e.g., TSH, T4, and FT3 concentrations) over time. This is achieved by solving a system of differential equations that describe the sensitivity functions alongside the main model equations. The authors also analyze both dynamic and stationary sensitivities.

Datasets

The primary dataset used for parameter identification was a collection of FT3 measurements from 1,121 untreated patients, as described in Ref. (11). The authors do not introduce a new benchmark dataset but rather leverage existing clinical data to validate and parameterize their model.

Key Insights and Ideas

  • The TSH-T3-Shunt is Crucial for FT3 Dynamics: The inclusion of the TSH-T3-shunt is shown to be essential for reproducing the observed circadian rhythm in free triiodothyronine (FT3) concentrations, a phenomenon not captured by models lacking this pathway.
  • Reduced FT3 Delay: The TSH-T3-shunt significantly reduces the delay between TSH and FT3 oscillations, bringing the model's predictions closer to observed clinical data. This suggests that intrathyroidal T3 production plays a more dominant role in the rapid dynamics of FT3 than previously modeled.
  • Sensitivity Reveals Regulatory Robustness: Sensitivity analysis highlights that the HPT axis exhibits different sensitivities to parameter changes depending on the physiological state. For instance, TSH is more sensitive to changes in the thyroid's secretory capacity (GT) in hypothyroid states (low GT) than in hyperthyroid states (high GT).
  • Ultra-Short Feedback Enhances Robustness: The "ultra-short feedback" loop within the pituitary gland (related to TSH feedback on TRH) enhances the robustness of TSH production against fluctuations in thyrotropin-releasing hormone (TRH) from the hypothalamus.
  • Clinical Implications for Diagnosis: The model's findings suggest that the interpretation of TSH levels, especially near the upper limit of the reference range, needs careful consideration due to these varying sensitivities. This could help explain why diagnosing subclinical hypothyroidism can be more prone to error than diagnosing subclinical hyperthyroidism.

Key Results

  • The extended model successfully reproduces a circadian rhythm in FT3, which was absent in the previous model without the TSH-T3-shunt.
  • The TSH-T3-shunt reduces the delay between TSH and FT3 from approximately 13 hours to about 6 hours, aligning better with clinical observations.
  • Sensitivity analysis shows that the FT3 response to changes in the thyroid's secretory capacity (GT) is less pronounced when the TSH-T3-shunt is included, indicating a more stable FT3 production pathway.
  • The model demonstrates that TSH is significantly more sensitive to variations in GT at lower levels of GT (hypothyroidism), providing a potential explanation for diagnostic challenges.

Limitations and Future Work

  • Model Identifiability: The authors acknowledge that the model is not fully identifiable with the available stationary FT3 measurements. This non-uniqueness in parameter estimation means that different combinations of GT3 and GD1 can yield similar FT3 outputs.
  • Parameter Simplification: The assumption of equal deiodinase activities (GD1 and GD2) for both peripheral and thyroidal conversion might not hold true in reality.
  • Lack of Membrane Transport: The model does not explicitly account for membrane transport processes between thyroidal and peripheral tissues, which could further refine its accuracy.
  • Future Work: The authors suggest that personalized model identification using dynamic hormone concentration trajectories would be highly valuable. Further research could also focus on identifying thyroidal and peripheral deiodinase activities separately and incorporating membrane transport models. Extending the model to include non-classical thyroid hormone signaling and developing compartment models for membrane transport are also proposed avenues. Ultimately, the goal is to improve diagnostic strategies and personalized medication for thyroid disorders.

Summary updated

Abstract

Despite significant progress in assay technology, diagnosis of functional thyroid disorders may still be a challenge, as illustrated by the vague upper limit of the reference range for serum thyrotropin (<i>TSH</i>). Diagnostical problems also apply to subjects affected by syndrome T, i.e., those 10% of hypothyroid patients who continue to suffer from poor quality of life despite normal <i>TSH</i> concentrations under substitution therapy with levothyroxine (<i>L</i>-<i>T</i><sub>4</sub>). In this paper, we extend a mathematical model of the pituitary-thyroid feedback loop in order to improve the understanding of thyroid hormone homeostasis. In particular, we incorporate a <i>TSH</i>-<i>T</i><sub>3</sub>-shunt inside the thyroid, whose existence has recently been demonstrated in several clinical studies. The resulting extended model shows good accordance with various clinical observations, such as a circadian rhythm in free peripheral triiodothyronine (<i>FT</i><sub>3</sub>). Furthermore, we perform a sensitivity analysis of the derived model, revealing the dependence of <i>TSH</i> and hormone concentrations on different system parameters. The results have implications for clinical interpretation of thyroid tests, e.g., in the differential diagnosis of subclinical hypothyroidism.
BibTeX
@inproceedings{berberich2018mathematical,
  author = {Julian Berberich and Johannes W. Dietrich and Rudolf Hoermann and Matthias A. Müller},
  title = {{Mathematical Modeling of the Pituitary–Thyroid Feedback Loop: Role of a TSH-T3-Shunt and Sensitivity Analysis}},
  booktitle = {Frontiers in Endocrinology},
  year = {2018},
  doi = {10.3389/fendo.2018.00091},
}
Figure