Frequency- and spike-timing-dependent mitochondrial Ca2+ signaling regulates the metabolic rate and synaptic efficacy in cortical neurons

  1. Ohad Stoler
  2. Alexandra Stavsky
  3. Yana Khrapunsky
  4. Israel Melamed
  5. Grace Stutzmann
  6. Daniel Gitler
  7. Israel Sekler
  8. Ilya Fleidervish

  • Curated by eLife

    eLife logo

    Evaluation Summary:

    This work describes how mitochondrial calcium in different regions of pyramidal neurons is controlled by action potentials and synaptic input. The authors show that calcium is controlled in a highly non-linear manner by calcium entry into cells (through voltage-dependent calcium channels) during sequences of action potentials. A particularly interesting finding is the high degree of localization of calcium rises in individual mitochondria in dendrites, and the requirement for both synaptic input and back-propagating action potentials to produce prominent rises of calcium in dendritic mitochondria. The work provides fundamental new information about how calcium entry during action potentials and synaptic input controls mitochondrial function.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 and Reviewer #2 agreed to share their name with the authors.)

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Mitochondrial activity is crucial for the plasticity of central synapses, but how the firing pattern of pre- and postsynaptic neurons affects the mitochondria remains elusive. We recorded changes in the fluorescence of cytosolic and mitochondrial Ca 2+ indicators in cell bodies, axons, and dendrites of cortical pyramidal neurons in mouse brain slices while evoking pre- and postsynaptic spikes. Postsynaptic spike firing elicited fast mitochondrial Ca 2+ responses that were about threefold larger in the somas and apical dendrites than in basal dendrites and axons. The amplitude of these responses and metabolic activity were extremely sensitive to the firing frequency. Furthermore, while an EPSP alone caused no detectable Ca 2+ elevation in the dendritic mitochondria, the coincidence of EPSP with a backpropagating spike produced prominent, highly localized mitochondrial Ca 2+ hotspots. Our results indicate that mitochondria decode the spike firing frequency and the Hebbian temporal coincidences into the Ca 2+ signals, which are further translated into the metabolic output and most probably lead to long-term changes in synaptic efficacy.

Article activity feed

  1. Version published to 10.7554/elife.74606 on eLife
  2. eLife

    Evaluation Summary:

    This work describes how mitochondrial calcium in different regions of pyramidal neurons is controlled by action potentials and synaptic input. The authors show that calcium is controlled in a highly non-linear manner by calcium entry into cells (through voltage-dependent calcium channels) during sequences of action potentials. A particularly interesting finding is the high degree of localization of calcium rises in individual mitochondria in dendrites, and the requirement for both synaptic input and back-propagating action potentials to produce prominent rises of calcium in dendritic mitochondria. The work provides fundamental new information about how calcium entry during action potentials and synaptic input controls mitochondrial function.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 and Reviewer #2 agreed to share their name with the authors.)

  3. eLife

    Reviewer #1 (Public Review):

    The authors use electrical recordings and calcium imaging of adult neurons in cortical slices from mice to study how calcium transients in mitochondria respond to patterns of synaptic input and action potential firing that elicit long-term plasticity. Prior work has shown that mitochondria are required for cortical plasticity, making the question of how their calcium fluxes vary with the neurons inputs and outputs an important one. The authors show convincingly that features of the mitochondrial calcium transients can explain several previously identified features of "spike timing-dependent plasticity," such as its dependence on firing frequency and on repetition. The work opens up a series of further interesting questions about the molecular mechanisms involved in these mitochondrial calcium transients and about whether their subsequent influence on plasticity depends only on metabolism or on some additional function of mitochondria.

    This is a carefully carried out study that addresses an interesting and timely question: how do neuronal mitochondria "decode" coincident action potential firing and synaptic input? It has long been known that mitochondria have their own calcium transients, and that these are critical to their metabolic function. Recent work in cultured neurons showed that locally inhibiting dendritic mitochondria could impair long-term synaptic plasticity. The authors use viruses target the calcium indicator mitoGCaMP6m to neuronal mitochondria and simultaneously image their calcium transients and cytosolic calcium transients at two different wavelengths using two-photon imaging while also recording electrical activity and stimulating presynaptic axons.

    The experiments, carried out in layer 5 pyramidal neurons from ex vivo slices made from adult mice convincingly document highly nonlinear summation of mitochondrial calcium transients that are sensitive to the relative timing of synaptic input and neuronal firing in much the same way that long-term change in synaptic strength depends on these parameters. In addition, like long-term plasticity in these neurons, the summation is frequency dependent, showing a threshold above which it increases rapidly. The two also shows a similar dependence on the number of pre- and postsynaptic pairings. Together with prior work showing that mitochondrial function is required for long-term plasticity and that mitochondrial function is dependent on calcium fluxes, the present study suggests that the observed properties of mitochondrial calcium fluxes may be a defining feature of how correlated pre- and postsynaptic firing lead to long-term synaptic change.

    Although there are many remaining questions, such as exactly how these transients respond to synaptic input and how they subsequently influence synaptic strength, these are appropriately left for future studies.

  4. eLife

    Reviewer #2 (Public Review):

    This work describes how mitochondrial calcium in different regions of pyramidal neurons is controlled by action potentials and EPSPs. The authors show that calcium is controlled in a highly non-linear manner by calcium entry into cells (through voltage-dependent calcium channels) during sequences of action potentials. A particularly interesting finding is the high degree of localization of calcium rises in individual mitochondria in dendrites, and the requirement for both EPSPs and back-propagating action potentials to produce prominent rises of calcium in dendritic mitochondria. The work provides fundamental new information about how calcium entry during action potentials and EPSPs controls mitochondrial function.

    This paper complements very nicely a collection of recent papers that have described some key mechanisms and consequences of calcium entry into neuronal mitochondria, including the demonstration by Rangaraju et al (2019) that mitochondria serve as highly-localized energy reserves for morphological synaptic plasticity, the paper by Ashrafi et al (2020) showing that MICU3 allows neuronal mitochondria to achieve calcium entry with much smaller increases in cytoplasmic calcium than non-neuronal cells, the paper by Garg et al (2021) showing how MICU proteins regulate the uniporter channels, and the paper by Diaz-Garcia et al (2021) showing that Ca2+ uptake into the mitochondria is responsible for controlling buildup of mitochondrial NADH, probably through Ca2+ activation of dehydrogenases in the TCA cycle. The current manuscript explores how the time course of mitochondrial calcium is controlled in a highly non-linear manner by calcium entry into cells during sequences of action potentials. A particularly interesting finding is the high degree of localization of calcium rises in individual mitochondria in dendrites, and the requirement for both EPSPs and back-propagating action potentials to produce prominent rises of calcium in dendritic mitochondria. This observation complements the 2019 Rangaraju Cell paper very nicely in giving a picture of a mechanism with a central role for dendritic mitochondria in spike-timing dependent plasticity.

    The work in the paper is done to a very high technical standard, and all the results were convincing. The work is clearly presented and the paper is clearly and concisely written.

  5. eLife

    Reviewer #3 (Public Review):

    In this study the authors combine brain slice electrophysiology with two-photon calcium imaging of fluorescent dye and genetically encoded calcium reporters (in cytoplasm and mitochondria, respectively) to further investigate, the relationship between neuronal activity and mitochondrial calcium handling. They find that, from a certain calcium threshold, neuronal activity levels closely correlate with both mitochondrial calcium handling and energetic capacity (determined using NAD(P)H imaging). The authors also demonstrate that in proximal dendrites mitochondrial calcium transients upon single synapse stimulation require a co-incident back-propagating action potential. They go on to hypothesise this could be important for some forms of synaptic plasticity.

    By combining brain slice electrophysiology with two-photon calcium imaging of fluorescent dye and GECI reporters (in cytoplasm and mitochondria, respectively)the authors aimed to further investigate, the relationship between neuronal activity, mitochondrial calcium handling and metabolic rate.

    Quite a lot has been done to look at the impact of neuronal activity (and frequency dependency thereof) on mitochondrial calcium handling in axons using combinations of electrophysiological and imaging techniques, (e.g. Kwon et al., 2016; Gazit et al., 2016; Vaccaro et al ., 2017; Lewis et al., 2018; Devaraju et al., 2017; Styr et al., 2019; Ashrafi et al., 2020 - several of these key studies not cited). Several of these earlier studies also go on to address the physiological relevance of mitochondrial calcium handling for neuronal function, synaptic properties and plasticity.

    In contrast less is know about these relationships in somato-dendrites (but see also Divakaruni et al., 2018). In the current work the authors aim to address this gap in the knowledge. Importantly a strength of the study is that the work is performed in acute brain slices using an elegant combination of electrophysiology and two-photon imaging. The work is interesting and appears well performed throughout. However, a weakness of the current study as it stands is that it is mostly correlative and descriptive in nature. No attempt is made to better understand mechanisms of some of the reported observations (e.g. the spatial variation in activity-dependent mitochondrial calcium handling), or to causally test (genetically or pharmacologically) the importance of the observed phenomena on neuronal function or plasticity. Thus while the title of the study implies a relationship between frequency dependent mitochondrial calcium handling and metabolic rate this is not robustly tested. Similarly the final sentence of the abstract is a significant over-interpretation of the presented results. No evidence is provided that the proposed co-incidence detection by mitochondria can be read out as metabolic or plasticity changes. As such the mostly observational nature of the findings currently dampens their potential impact.

  6. Version published to 10.1101/2021.06.07.447340v1 on bioRxiv