Quantitative Systems Toxicology Modeling for Neuronal Adverse Outcome

Abstract

Human beings are getting exposed to the environmental chemicals on the daily basis affecting their nervous system directly or indirectly. Understanding the neurotoxicity caused by these chemicals is a significant challenge due to limited toxicological information. Many chemicals are often not documented as neurotoxicants due to insufficient knowledge on the human kinetics and dynamics. The kinetic data plays a pivotal role in establishing a scientifically relevant hazard evaluation and human risk assessment. The toxicokinetic factors like metabolizing enzymes with age, excretion rate, protein binding, and BBB permeation are some of the important aspects affecting the nervous systems. The human brain is a complex organ consisting of the bloodbrain barrier (BBB), which protects the entry of xenobiotics into the nervous system. At the early ages of life, BBB is still not fully developed and in older age, the integrity of BBB is not so intact being affected by many other factors. It becomes evident to understand the kinetic of environmental chemicals with age to evaluate the risk at the population level. The environmental chemicals affect the neuronal cells through various mechanisms most common being induction of oxidative stress through ROS production leading to damage to the nervous system and ultimately neurotoxicity. It is not possible to conduct experimental studies for thousands of chemicals and such studies cannot be undertaken in humans due to ethical limitations, especially in sensitive populations. To improve the neurotoxicity risk assessment, along with the kinetic, it is also important to quantify chemicals effect on the molecular dynamics at the target site and consequently response from cellular to population level by establishing dose-response relationship. Another emphasis was to reduce the use of animals for risk assessment following 3R principle suggested by the REACH guideline and integrate the in-vitro data for computational models. Many computational models are already available for several environmental chemicals, so our approach was not to reinvent the model but rather to continue with the same model and improve it further for risk assessment. However, for some chemicals, novel computational models were also developed. The objective of the research was to develop the quantitative systems toxicology (QST) model by integrating chemical exposure, physiology, chemical kinetics, molecular dynamics and cellular response (brain). For achieving this, the physiologically based pharmacokinetic models (PBPK) were used which are mathematical models based on ordinary differential equations (ODE) incorporating dynamic human physiology for defining pharmacokinetics (PK) like absorption, distribution, metabolism and excretion (ADME) inside human body. Another model is systems biology (SB) which defines the risk at molecular and cellular level by modeling the complex biological system. In the thesis, dynamic physiology based on age, sex, and subpopulation was incorporated in the conventional PBPK to improve the tissue dosimetry-based risk assessment. Multiple chemicals like Bisphenol A (BPA), organophosphate flame Retardants (OPFRs), Perfluorooctane sulfonic acid (PFOS) and perfluorooctanoic acid (PFOA) were used for developing computational models. Moreover, the IVIVE (in-vitro to in-vivo extrapolation) brain PBPK model was built by integrating BBB mechanism and dividing the brain in several compartments like hippocampus, cortex, rest brain, and CSF, an important step toward predicting neuronal adverse outcome. Since the conventional method is to use partition- coefficient for BBB permeation, we used IVIVE approach calculating permeability coefficient based on the in-vitro study conducted with BBB cell lines. Furthermore, coupling the PBPK to a systems biology reactive oxygen species model allowed us to predict the response at molecular, cellular and population level. The dynamic age-specific PBPK model was used to predict neuronal risk based on age of the individual for long-acting chemicals like PFOS and PFOA. PBPK model provided a mechanistic approach for studying the PK and helped in translating the risk to humans due to chemical exposure. PBPK for OPFRs showed prolonged disposition in brain indicating a potential neurotoxic risk. The sex-specific PBPK model helped to investigate the higher possibility of risk in girls than boys for BPA exposure due to extended half-life concluding the importance of incorporating sex in risk assessment. Reconstructed BPA exposure using the pediatric PBPK model exceeded total daily intake (TDI) set by EFSA by thousand times pointing out that children may be at higher risk. Results from this pediatric PBPK model helped us to understand the need for creating specialized PBPK models to predict the risk in sensitive populations. Dynamic age-based PBPK model implicated PFOS disposition varies in organs, specifically bone marrow and adipose tissue which was found to be a major site for PFOS accumulation. The exposure was found to be more in the geriatric population due to reduced GFR and variation in blood flow in several organs. Brain PBPK model showed that generalizing the risk and considering all sub-compartment of the brain as equal is an under predictability of risk. The integrated approach enabled us to predict the downstream targets alteration like antioxidant regulated genes, mitochondrial damage, and consequently cellular toxicity. Oxidative stress was found to be more in the geriatric population than in adults implicating neurodegenerative risk. Overall this thesis produced a successful example for improved neuronal risk assessment by integrating PBPK/PD model along with in-vitro data. It is important to consider personalized risk by regulatory bodies and under-representation of sex and age should be overcome by introducing improved computational models.