Graded spikes differentially signal neurotransmitter input in cerebrospinal fluid contacting neurons of the mouse spinal cord

  1. Emily Johnson
  2. Marilyn Clark
  3. Merve Oncul
  4. Claudia MacLean
  5. Jim Deuchars
  6. Susan A. Deuchars
  7. Jamie Johnston

  • Curated by eLife

    eLife logo

    Summary: The reviewers have found the topic of your study of high interest, with very intriguing findings on the different origins of calcium transients in CSFcNs.

    However, after a careful examination of your work , the reviewers have raised the following major concerns:

    1. To conclude on calcium spikes, the imaging data without electrophysiological calibration leaves too much unknown. A careful electrophysiological examination should reveal how calcium transients of different amplitude correlate with the electrical activity of the cell, calcium spikes and spontaneous PKD2L1 channel openings as described extensively in these cells, is absolutely mandatory to conclude.

    2. The manuscript shows a lack of consideration of the importance of the sensory functions and of the role of channel PKDL1 that are both well-established in CSFcNs in mice and other models. More work is necessary to relate to these critical aspects.

    3. The number of animals for juvenile and adult mice used by the authors should be clearly stated (the manuscript only refers to the total number) but also largely increased for the authors to reach robust conclusions.

    4. Overall, more rigor should be implemented throughout the entire manuscript, with a deep writing improvement and a careful inspection of figure panels (choice and fair / complete representation of the data) and more information on conditions used for experiments (promoter used, concentrations for pharmacological agents, selection of ROIs, ventral versus dorsal CSF-cNs, definition and proportion of silent cells, enrichment in the T and L type calcium channels, etc ... ).

    Reviewer #2 and Reviewer #4 opted to reveal their name to the authors in the decision letter after review.

Reviewed by

Read the full article See related articles

Discuss this preprint

Start a discussion What are Sciety discussions?

Listed in

Log in to save this article

Abstract

The action potential and its all-or-none nature is fundamental to neural communication. Canonically the action potential is initiated once voltage-gated Na + channels are activated, and their rapid kinetics of activation and inactivation give rise to the action potential’s all-or-none nature. Here we show that cerebrospinal fluid contacting neurons (CSFcNs) surrounding the central canal of the mouse spinal cord employ a different strategy. Rather than using voltage-gated Na + channels to generate binary spikes, CSFcNs use two different types of voltage-gated Ca 2+ channel, enabling spikes of different amplitude. T-type Ca 2+ channels generate small amplitude spikes, whereas large amplitude spikes require high voltage-activated Cd 2+ sensitive Ca 2+ channels. We show that these different amplitude spikes signal input from different transmitter systems; purinergic inputs evoke smaller T-type dependent spikes while cholinergic inputs evoke large T-type independent spikes. Different synaptic inputs to CSFcNs can therefore be signalled by the spike amplitude.

Article activity feed

  1. eLife

    Reviewer #3:

    -The authors claim in the first part of the results that the frequency of CSF-cN spontaneous activity is the same in juvenile and adult mice. In Fig.1G, 61 neurons from 7 animals are illustrated. The authors should state how many juvenile (P14-P24) and adult (P36-P47) mice have been included in the analysis (3 and 4 is different from 5 and 2) and how many neurons have been recorded in each animal. In the methods section, they indicate that acute slices were obtained from P14 to P55 mice. If the reviewer is correct, neurons from P55 mice are not included in Fig. 1G?

    -The immunohistochemical data have been obtained in P30-P52 mice. Are P14 CSF-cNs all VGaT positive?

    -The frequency of CSF-cN spontaneous activity could be the same but underlying mechanisms could completely differ with age. In Fig. 3, TTX fails to alter spontaneous Ca2+ spike expression in 3 animals. How old are these mice? Same questions for the results with Cd (2 animals, sample a little bit small...), ML218 4 animals (4 animals)...etc

    -The focal ejection of 40mM K+ triggers a depolarization of all CSF-cNs "including those previously silent". This is the first time page 9 that the authors mention the fact that some CSF-cNs are not spontaneously active. Is the proportion of silent CSF-cNs different with age? The effects of Cd have been tested in 1 animal. Same for the effect of MCA on Ach-evoked Ca2+ spikes. In my opinion, the sample size has to be increased.

  2. eLife

    Reviewer #2:

    The present study investigates how CSF-contacting neurons (CSFcNs) of the mouse spinal cord integrate and translate different synaptic inputs using distinct calcium-dependent spike mechanisms. Indeed two different types of voltage-gated calcium channels can be activated, resulting in the generation of spikes with different amplitudes. T-type Ca2+ channels would be involved in the generation of low amplitude spikes while HVA-Ca2+ channels participate in the generation of large amplitude spikes. Then these distinct spikes allow signaling different neurotransmitter systems. Consequently, the data provided here argue in favor of CSF-contacting neurons acting as a sensory system that uses Ca2+ channels-dependent spike activity with graded amplitude corresponding to the activation of different neurotransmitter receptors. This study is based on two-photon calcium imaging performed on spinal cord slices preparations obtained from young and adult mice. My comments are as follows:

    1. All data are based on calcium imaging. Therefore, traces correspond to calcium-dependent fluorescent changes in the cells of interest. Can the author provide at least one sample showing that these calcium events are indeed linked to the generation of spikes; i.e., electrophysiological recordings? In addition, is there any electrophysiological evidence for the existence of calcium-dependent conductances in the CSFcNs? In the same vein, the authors conclude that spontaneous activity of CSFcNs depends upon calcium- but not sodium-spikes as TTX has apparently no effect. But, are the authors sure that in their experimental conditions individual sodium spikes could be detected given the genetic encoded probe used, the kinetic of such spikes and the frequency of the sampling during image acquisition? Note that this does not preclude the conclusion that CSFcNs express calcium-dependent spikes. See also comment 4 below.

    2. Using the activation of different calcium channels to trigger spikes of different amplitude to code distinct signaling pathways associated with distinct neurotransmitter systems is a very attractive mechanism. I was wondering whether the authors ever observed the two processes in one single cell, meaning: did they ever try to apply Ach and ATP on the same cell? To my point of view, this would be an extremely elegant way to show that spikes of variable amplitudes imply the activation of distinct calcium-dependent conductances and are linked to different neurotransmitter signaling in one neuron. This should be possible as they said that 100% of the examined cells responded to Ach, suggesting that the only limitation would be to find a cell that also expresses purinergic receptors (should be highly feasible). In addition, this would strongly demonstrate how much this coding mechanism is valuable if this is present in a single cell, otherwise one could consider that the coding system just depends upon each cell, the neurotransmitter and its associated receptor signaling that by definition can involve distinct calcium-dependent channels. Then it would rather be a mechanism specific to each receptor than a sophisticated coding system.

    3. As a general comment on figures, I would suggest to the authors to provide samples that are more illustrative of the results they claim on. For example on Figure 3 they state that TTX has no effect on spike amplitude and frequency, but the two traces shown (in blue and green) rather indicate a decrease in spike frequency and even an increase in spike amplitude after a few minutes of recording (green trace). [See also comment 4 below]. Another example is in Figure 6 in which one important data is the distinct amplitude of spikes triggered by either Ach or ATP. While this is properly illustrated in panels C, D and E, in contrast the samples chosen for panels A and B show events with the exact same amplitude. Please choose other traces. By the way, panel C is not necessary because the same info are included in panels D and E. I would suggest removing panel C. Finally, in Figure 7 it is stated in the text that in some cells ATP induced first a decrease in fluorescence followed by a large Ca2+ spike, while this specific spike looks much smaller than all the other ones illustrated in the study (Fig 7G). Also, the spike triggered by UTP looks different than the one triggered by ATP. Is it a typical response?

    4. Several experimental details must be provided. First, the justification for the choice of VGAT promoter to drive the GCaMP6f indicator into PKD2L1 neurons is missing. Second, drug concentrations are not justified. This is important as the authors argue that Ach and ATP trigger Ca2+ spikes with different amplitudes, but isn't there the possibility that this is dose-dependent? Did the authors try different concentrations? Third, on TTX experiments (Fig 3), after how long under TTX exposure were measurements performed? While this is a crucial parameter, this is not indicated in the paper. Given the traces provided different conclusions could be reached depending on this timing.

    5. It remains unclear to me why only some of the data (for example Fig 7) make a distinction between dorsal and ventral CSF-contacting neurons. In the zebrafish it is established that ventral and dorsal CSFC neurons have different developmental origins and distinct types of projections related to different functions. Then, if these neurons are suspected to play different roles depending on their ventro-dorsal position also in mice, the entire study should take this into account.

  3. eLife

    Reviewer #1:

    The authors provide interesting evidence on the properties of CSF-contacting neurons, referred to as 'CSFcNs' in their manuscript, using 2 photon calcium imaging in mice.

    Their work relies on calcium imaging using 2 photon microscopy in slices of the mouse spinal cord. The authors observed calcium transients with two different amplitudes and propose that these transients reflect the activation of different voltage dependent calcium channels (T and L).

    Although the work is of interest, there are throughout the manuscript numerous issues: -shortcuts and oversimplified assumptions (calcium transients do not equal spikes!) (see title of Figure 2, 3) -the massive ignorance of the relevant literature for this small field on CSF-cNs in mice. In particular, but not only, the authors should know and refer to the work of Orts Dell'Immagine, Wanaverbecq, Trouslard who have shown since 2012 that CSF-cN in mice are chemosensory cells whose spontaneous activity is driven by the channel PKD2L1.

    Major comments

    1. The authors assume that calcium transients equal to firing (Figure 2) or calcium spikes (Figure 3) but these are far from being the same. No deconvolution algorithm can use calcium transients to infer spiking with better than 70% accuracy.

      In the recordings of the Wanaverbecq group, spontaneous firing in slices was 0.4Hz in control and 0.1Hz in PKD2L1 KO. The authors find here calcium transients occurring at 0.16Hz (n = 63 cells), suggesting that some of the sparse firing activity is missed by the authors.

      Since calcium transients reflect spiking but not in a linear manner, a calibration is necessary via cell attached or loose patch recordings in order to infer on CSF-cN spiking, or perforated patch to validate the evidence for calcium spikes.

    2. This assumption of calcium = firing does not hold in cells that have an input resistance of GOhms and whose activity has been shown to be driven by the opening of the channel PKD2L1 (Orts Del Immagine et al Neuropharmacology 2016). In particular, observations of the TTX insensitive calcium transients may be due to the PKD2L1 channel.

      => The authors need to combine recordings with perforated Patch Clamp together with the 2P calcium imaging in order to tackle the question of the role of the channel openings in the generation of the different calcium transients observed in WT or KO for PKD2L1.

      From introduction to discussion, the authors should properly cite the work of the Wanaverbecq group as well as other groups in the field, whose contributions were relevant and ignored.

    3. Activation leading to calcium spikes (K+, ATPergic, Cholinergic inputs, ...) was done without blockage of the neurotransmission in the slices and could therefore originate from indirect sources, including activation of metabotropic receptors presynaptically.

      The authors need to solve these issues.

    4. In Figure 7, there are diverse responses that the authors should better illustrate. Many cells appear to not respond for multiple stimuli tested : what is the rational criterion to define that a cell responded or not? Can the authors quantify the proportion of cells responding? Did the author take into account the high level of spontaneous activity? Can the negative dip in response possibly from a motion artifact in panel G and H?

  4. eLife

    Summary: The reviewers have found the topic of your study of high interest, with very intriguing findings on the different origins of calcium transients in CSFcNs.

    However, after a careful examination of your work , the reviewers have raised the following major concerns:

    1. To conclude on calcium spikes, the imaging data without electrophysiological calibration leaves too much unknown. A careful electrophysiological examination should reveal how calcium transients of different amplitude correlate with the electrical activity of the cell, calcium spikes and spontaneous PKD2L1 channel openings as described extensively in these cells, is absolutely mandatory to conclude.

    2. The manuscript shows a lack of consideration of the importance of the sensory functions and of the role of channel PKDL1 that are both well-established in CSFcNs in mice and other models. More work is necessary to relate to these critical aspects.

    3. The number of animals for juvenile and adult mice used by the authors should be clearly stated (the manuscript only refers to the total number) but also largely increased for the authors to reach robust conclusions.

    4. Overall, more rigor should be implemented throughout the entire manuscript, with a deep writing improvement and a careful inspection of figure panels (choice and fair / complete representation of the data) and more information on conditions used for experiments (promoter used, concentrations for pharmacological agents, selection of ROIs, ventral versus dorsal CSF-cNs, definition and proportion of silent cells, enrichment in the T and L type calcium channels, etc ... ).

    Reviewer #2 and Reviewer #4 opted to reveal their name to the authors in the decision letter after review.

  5. Version published to 10.1101/2020.09.18.303347 on bioRxiv