MCMC output files for: Quantitative characterization of population-wide tissue- and metabolite-specific variability in perchloroethylene toxicokinetics in male mice

Quantification of inter-individual variability is a continuing challenge in risk assessment, particularly for compounds with complex metabolism and multi-organ toxicity. Toxicokinetic variability for perchloroethylene (perc) was previously characterized across three mouse strains and in one mouse strain with various degrees of liver steatosis. To further characterize the role of genetic variability in toxicokinetics of perc, we applied Bayesian population physiologically-based pharmacokinetic (PBPK) modeling to the data on perc and metabolites in blood/plasma and tissues of male mice from 45 inbred strains from the Collaborative Cross (CC) mouse population. After identifying the most influential PBPK parameters based on global sensitivity analysis, we fit the model with a hierarchical Bayesian population analysis using Markov chain Monte Carlo simulation. We found that the data from three commonly used strains were not representative of the full range of variability in perc and metabolite blood/plasma and tissue concentrations across the CC population. Using inter-strain variability as a surrogate for human inter-individual variability, we calculated dose-dependent, chemical-, and tissue-specific toxicokinetic variability factors (TKVFs) as candidate science-based replacements for the default uncertainty factor for human toxicokinetic variability of 100.5. We found that TKVFs for glutathione conjugation metabolites of perc showed the greatest variability, often exceeding the default, whereas those for oxidative metabolites and perc itself were generally less than the default. Overall, we demonstrate how a combination of a population-based mouse model such as the CC with Bayesian population PBPK modeling can reduce uncertainty associated with toxicokinetic human variability by deriving the chemical-specific adjustment factors needed to increase accuracy and precision in quantitative risk assessment.

Methods
This database includes large output files of MCMC simulations in PBPK modeling in a manuscript titled 'Quantitative Characterization of Population-wide Tissue- and Metabolite-specific Variability in Perchloroethylene Toxicokinetics in Male Mice' submitted to Tox.Sci.

The hierarchical Bayesian population statistical model was applied for PBPK model calibration and estimation of model parameters and their uncertainty and variability as previously described (Bois 2000a, 2000b; Weihsueh A. Chiu et al. 2009; Hack 2006) and the updated conceptual representation described in  Dalaijamts et al. (2018, 2020) was used.

PBPK modeling in conjunction with statistical modeling, including the MC/MCMC simulations, was performed using GNU MCSim v.5.6.5 software (Bois 2009). The modeling processes are complex comprised with different steps and components.

Hierarchical Metropolis-Hastings algorithms within the Gibbs sampler (Gelfand and Adrian 1990; Geman and Geman 1984) was used in MCMC simulation. The MCMC simulation generates posterior parameter values at the population level, and parameter values for the strains used in the experiments.

PBPK model was applied to predict of population variability and uncertainty of internal dose metrics, including AUCs of perc, TCA, and GSH-conjugation metabolites and perc disposition dose metrics, at 36 h post oral exposures to single doses of 10, 100, and 1000 mg perc per kg b.w. using parameter posteriors of population-generated random strains

Stages of the PBPK modeling run on softwares as follow:

MCMC simulations of preliminary and final models were run on MCSim v.5.6.5 software on Linux/Unix environment in Terra cluster in the High Performance Research Center, Texas A&M University.
Global Sensitivity analysis was run on 'pksensi' version 1.2.0 R package developed by our lab.
Evaluation of model fits were run on MCSim v.5.6.5 software on Windows.
Dose metric predictions were run on MCSim under R on Windows.

个体间变异的量化在风险评估领域始终是一项核心挑战，对于代谢复杂且具有多器官毒性的化合物而言尤为突出。此前已有研究针对3种小鼠品系以及1种存在不同程度肝脂肪变性的小鼠品系，表征了全氯乙烯（perchloroethylene，简称perc）的毒物动力学变异特征。为进一步阐明遗传变异在全氯乙烯毒物动力学中的作用，本研究将贝叶斯群体生理学药代动力学（PBPK）模型应用于来自协作杂交（CC）小鼠种群的45个近交品系雄性小鼠的血液/血浆及组织中全氯乙烯及其代谢物的相关数据。在基于全局敏感性分析识别出对PBPK模型影响最大的参数后，我们采用马尔可夫链蒙特卡洛（Markov chain Monte Carlo，MCMC）模拟进行分层贝叶斯群体分析以拟合该模型。研究发现，3种常用品系的相关数据无法代表CC种群中全氯乙烯及其代谢物在血液/血浆及组织中的浓度变异全貌。以品系间变异作为人类个体间变异的替代指标，我们计算了剂量依赖性、化学物特异性及组织特异性的毒物动力学变异因子（TKVFs），以此作为基于科学依据的候选替代物，替代人类毒物动力学变异默认不确定因子100.5。研究发现，全氯乙烯谷胱甘肽结合代谢物的TKVFs变异程度最大，通常超出默认不确定因子；而氧化代谢物及全氯乙烯本身的TKVFs则普遍低于该默认值。总体而言，本研究展示了如何将协作杂交这类基于群体的小鼠模型与贝叶斯群体PBPK模型相结合，通过推导化学物特异性调整因子来提升定量风险评估的准确性与精密度，从而降低与人类毒物动力学变异相关的不确定性。

### 方法
本数据集包含一篇投稿至《Tox.Sci.》、题为《雄性小鼠全氯乙烯毒物动力学中群体水平组织及代谢物特异性变异的定量表征》的手稿中，PBPK建模相关的MCMC模拟大型输出文件。

本研究采用分层贝叶斯群体统计模型进行PBPK模型校准，以及模型参数及其不确定性与变异的估计，相关方法参照此前发表的研究（Bois 2000a、2000b；Weihsueh A. Chiu等2009；Hack 2006），并采用了Dalaijamts等（2018、2020）更新的概念性表征框架。

本研究结合PBPK建模与统计建模（包括MC/MCMC模拟），采用GNU MCSim v5.6.5软件完成（Bois 2009）。建模流程较为复杂，涵盖多个步骤与组成部分。

MCMC模拟采用了吉布斯采样器（Gibbs sampler）中的分层Metropolis-Hastings算法（Gelfand与Adrian 1990；Geman与Geman 1984）。MCMC模拟可生成群体水平的后验参数值，以及实验中所用品系的参数值。

本研究采用群体生成的随机品系参数后验值，针对单次经口暴露剂量分别为10、100、1000 mg全氯乙烯每千克体重的实验，在暴露后36小时预测了群体变异与内暴露剂量指标的不确定性，其中内暴露剂量指标包括全氯乙烯、三氯乙酸（TCA）及谷胱甘肽（GSH）结合代谢物的曲线下面积（AUC），以及全氯乙烯处置相关剂量指标。

PBPK建模在软件上的执行流程如下：
1. 初步模型与最终模型的MCMC模拟均在德克萨斯农工大学高性能研究中心Terra集群的Linux/Unix环境下，通过MCSim v5.6.5软件完成。
2. 全局敏感性分析采用本实验室开发的pksensi v1.2.0 R包完成。
3. 模型拟合优度评估在Windows系统下通过MCSim v5.6.5软件完成。
4. 剂量指标预测在Windows系统下通过R环境中的MCSim完成。