DataSheet2_Dynamic Modeling of Mitochondrial Membrane Potential Upon Exposure to Mitochondrial Inhibitors.PDF

Mitochondria are the main bioenergetic organelles of cells. Exposure to chemicals targeting mitochondria therefore generally results in the development of toxicity. The cellular response to perturbations in cellular energy production is a balance between adaptation, by reorganisation and organelle biogenesis, and sacrifice, in the form of cell death. In homeostatic conditions, aerobic mitochondrial energy production requires the maintenance of a mitochondrial membrane potential (MMP). Chemicals can perturb this MMP, and the extent of this perturbation depends both on the pharmacokinetics of the chemicals and on downstream MMP dynamics. Here we obtain a quantitative understanding of mitochondrial adaptation upon exposure to various mitochondrial respiration inhibitors by applying mathematical modeling to partially published high-content imaging time-lapse confocal imaging data, focusing on MMP dynamics in HepG2 cells over a period of 24 h. The MMP was perturbed using a set of 24 compounds, either acting as uncoupler or as mitochondrial complex inhibitor targeting complex I, II, III or V. To characterize the effect of chemical exposure on MMP dynamics, we adapted an existing differential equation model and fitted this model to the observed MMP dynamics. Complex III inhibitor data were better described by the model than complex I data. Incorporation of pharmacokinetic decay into the model was required to obtain a proper fit for the uncoupler FCCP. Furthermore, oligomycin (complex V inhibitor) model fits were improved by either combining pharmacokinetic (PK) decay and ion leakage or a concentration-dependent decay. Subsequent mass spectrometry measurements showed that FCCP had a significant decay in its PK profile as predicted by the model. Moreover, the measured oligomycin PK profile exhibited only a limited decay at high concentration, whereas at low concentrations the compound remained below the detection limit within cells. This is consistent with the hypothesis that oligomycin exhibits a concentration-dependent decay, yet awaits further experimental verification with more sensitive detection methods. Overall, we show that there is a complex interplay between PK and MMP dynamics within mitochondria and that data-driven modeling is a powerful combination to unravel such complexity.

线粒体是细胞的核心生物能量细胞器。靶向线粒体的化学物质暴露通常会诱发毒性效应。细胞对能量生成扰动的应答，是通过结构重组与细胞器生物发生实现的适应性反应，与以细胞死亡为形式的牺牲性反应之间的动态平衡。在稳态条件下，有氧线粒体能量生成需要维持线粒体膜电位（mitochondrial membrane potential, MMP）。化学物质可扰动该膜电位，其扰动程度同时取决于化学物质的药代动力学（pharmacokinetics, PK）以及下游MMP动态变化过程。
本研究针对已部分公开的高内涵成像延时共聚焦成像数据集，开展数学建模分析，定量解析了HepG2细胞在24小时内暴露于多种线粒体呼吸抑制剂时的线粒体适应机制，重点关注其MMP动态变化。研究中共使用24种化合物对MMP进行扰动，这些化合物可分为两类：解偶联剂，以及分别靶向线粒体复合物I、II、III、V的复合物抑制剂。
为表征化学物质暴露对MMP动态变化的影响，我们对已有的微分方程模型进行优化，并将模型拟合至观测得到的MMP动态数据。相较于复合物I相关数据集，本模型对复合物III抑制剂的数据拟合效果更佳。为实现对解偶联剂FCCP的精准拟合，需将药代动力学衰减过程纳入模型。此外，同时纳入药代动力学衰减与离子渗漏参数，或采用浓度依赖性衰减策略，可提升对寡霉素（复合物V抑制剂）的模型拟合精度。
后续质谱检测结果显示，FCCP的药代动力学特征存在显著衰减，与模型预测结果一致。此外，检测得到的寡霉素药代动力学特征表现为：高浓度下仅存在有限衰减，而低浓度下该化合物在细胞内的浓度低于检测限，这与寡霉素存在浓度依赖性衰减的假说相符，但仍需借助更灵敏的检测方法开展进一步实验验证。
总体而言，本研究证实线粒体内部的药代动力学与MMP动态变化之间存在复杂的相互作用，而数据驱动建模与实验数据的结合是解析这类复杂生物学机制的有力手段。