Characterizing the role of mitochondrial dynamics during Drosophila convergent extension using NADH fluorescence lifetime imaging

Mitochondria are dynamic organelles that can fragment or fuse to support different bioenergetic demands (e.g., glycolysis vs. oxidative phosphorylation) and distinct cell behaviors (e.g., mitosis or migration). While the role of mitochondrial dynamics in wound healing and metabolic disorders has received significant attention, the role of mitochondrial fission and fusion during normal embryonic development is less well understood--in part due to the difficulty of studying such processes in vivo. Combined with the depth-resolved imaging capabilities of multiphoton microscopy, fluorescence lifetime imaging (FLIM) of the mitochondrial cofactor NADH can be used to simultaneously visualize mitochondrial network morphology and infer certain aspects of cellular bioenergetics (e.g., glycolysis vs. oxidative phosphorylation) in a label-free, non-invasive manner. Here we demonstrate that NADH FLIM can be used to accurately track the subcellular localization and topology of mitochondrial networks in live Drosophila embryos. We used this technique to assess whether cells show changes in NADH lifetime during convergent extension (CE)--a conserved process of tissue remodeling in which thousands of germband cells undergo coordinated intercalation to drive elongation of the head-to-tail axis. Contrary to our expectations, we did not observe significant changes in NADH lifetime or network appearance during CE in wild-type embryos, suggesting that germband cells do not need to alter their baseline metabolism to fuel cell intercalation during normal development. To directly assess the role of mitochondrial fission and fusion during CE, we used RNA interference to disrupt the fission mediator Drp1 and the fusion mediator Opa1. Consistent with expectations, inhibiting mitochondrial fission in Drp1-knockdown embryos led to hyper-fused networks and significantly longer NADH lifetimes, indicating a shift towards oxidative phosphorylation. Conversely, inhibiting mitochondrial fusion in Opa1-knockdown embryos led to more hyper-fragmented networks and significantly shorter NADH lifetimes, indicating a shift towards glycolysis. Interestingly, inhibiting either fission or fusion altered tissue elongation and greatly increased the rate of cell intercalation errors, suggesting that a precise network topology is required for proper CE. We hypothesize that the CE defects in Drp1-knockdown embryos are primarily due to incorrect basal subcellular localization of mitochondria, whereas the CE defects in Opa1-knockdown embryos are due to deficient ATP and/or ROS production. These experiments demonstrate the utility of FLIM-based applications for characterizing the role of mitochondria during normal embryonic development, which could yield a better understanding of the metabolic underpinnings of various pathologies that involve epithelial remodeling, including spina bifida, defective wound healing, and cancer metastasis.