Mitochondrial Morphology Dynamics during Apoptosis - An integrative modeling approach

Mitochondria are central to many important cellular functions. The entire mitochondrial population is in constant flux, driven by continuous fusion and division of mitochondria. Defects in mitochondrial dynamics can cause deficits in mitochondrial respiration, morphology and motility leading to apoptosis under extreme conditions. An important and still unresolved question is how the heterogeneity of mitochondrial morphology, distribution and function are mechanistically realized in the cell. Importantly, to what extent is mitochondrial morphology dependent or affects cell fate decisions. Despite the intense focus on unraveling connections between mitochondrial morphology and severe human pathologies, the analysis and systematic description of mitochondrial phenotypes remains an open challenge. Current approaches to study mitochondrial morphology are limited by low data sampling coupled to manual identification and simplistic classification of complex morphological phenotypes. The overall goal of this work was to elucidate the nature of the relationship between mitochondria morphology and apoptotic events. Diverse apoptotic triggers were systematically tested and data concerning mitochondrial phenotypes and injury was collected to infer cause and consequence relationships. Therefore, high-resolution fluorescence imaging was employed to extract high-content information essential to identify and quantify spatial and conformational events in the single cell. These included monitoring of mitochondrial membrane permeability and quantification of Bax activation under matched conditions to assess mitochondrial stress. Experimentally, mitochondrial morphological transitions were followed in human breast carcinoma MCF-7 cells by tagging a mitochondrial inner membrane protein with a fluorescent probe. We made use of apoptotic conditions that have been previously reported to cause mitochondrial fragmentation or swelling. Wide field microscopy allowed for the collection of images containing cells with mostly networked, fragmented or swollen mitochondria. Next, image analysis was performed to extract several mitochondrial features that better characterize each class. These were grouped and used to build a decision tree-based classifier that automatically classifies individual mitochondria into the correspondent phenotypic class. Our population-based classifier accounts for intracellular sub-classes, intermediate mitochondrial stages and reproduces intercellular variances with high accuracy. Our results show that distinct apoptotic stimuli lead to subtle but significant differences in mitochondrial morphology within cell population that can be specific to a particular insult. Interestingly, there was no direct relation between the induced-mitochondrial classes and the analyzed apoptotic steps. In fact, some apoptotic drugs, which are known to cause similar mitochondrial damage, showed distinct mitochondrial morphology. Therefore, the observed heterogeneous response of mitochondria to stress strongly suggests that more complex, non-linear interactions exist. Here, we propose an integrated mechanistic and data-driven modeling approach to analyze heterogeneously quantified datasets and infer hierarchical interactions between mitochondrial morphology and apoptotic events. Our modeling results suggest that Bax activation leads to mitochondrial fragmentation, which is strongly associated with mitochondrial membrane depolarization events. In turn, the loss of mitochondrial membrane potential is closely related to mitochondrial swelling. Our model predictions are in accordance with previous published results and thereby validate our modeling approach that can now be easily extended to include new datasets. Surprisingly, mitochondrial fragmentation was not the most prominent phenotype, even under conditions where Bax activation was considerably high. Instead, swollen-mitochondria seem to be closer related to mitochondrial-associated death pathways. Next steps include the extension of our pipeline in a time-resolved manner and combined datasets acquisition in order to further investigate this hypothesis. In summary, we have established and validated a platform for mitochondrial morphological and functional analysis that offers results in an unbiased, systematic and statistically relevant manner. We believe the developed platform is suitable to be extensively used in the investigation of specific molecular targets. Possible applications include RNAi screens (e.g. morphology proteins) or extended compound libraries in a high-throughput mode. Importantly, it can now be further adjusted to other studies relevant to mitochondrial programmed cell death that will hopefully lead into the better understanding of mitochondrial role in physiology and disease progression