All-Optical Electrophysiology in hiPSC-Derived Neurons With Synthetic Voltage Sensors

  1. Francesca Puppo
  2. Sanaz Sadegh
  3. Cleber A. Trujillo
  4. Martin Thunemann
  5. Evan P. Campbell
  6. Matthieu Vandenberghe
  7. Xiwei Shan
  8. Ibrahim A. Akkouh
  9. Evan W. Miller
  10. Brenda L. Bloodgood
  11. Gabriel A. Silva
  12. Anders M. Dale
  13. Gaute T. Einevoll
  14. Srdjan Djurovic
  15. Ole A. Andreassen
  16. Alysson R. Muotri
  17. Anna Devor

  • Curated by eLife

    eLife logo

    Evaluation Summary:

    This manuscript will be interesting for people performing all optical electrophysiology. It describes a new combination of previously available genetic tools to allow simultaneous optogenetic manipulation and optical electrophysiology. The manuscript does not provide a major conceptual advance but provides good evidence that this assay can be employed for large-scale screening in hIPSC-derived neurons.

    (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.The reviewers remained anonymous to the authors.)

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Voltage imaging and “all-optical electrophysiology” in human induced pluripotent stem cell (hiPSC)-derived neurons have opened unprecedented opportunities for high-throughput phenotyping of activity in neurons possessing unique genetic backgrounds of individual patients. While prior all-optical electrophysiology studies relied on genetically encoded voltage indicators, here, we demonstrate an alternative protocol using a synthetic voltage sensor and genetically encoded optogenetic actuator that generate robust and reproducible results. We demonstrate the functionality of this method by measuring spontaneous and evoked activity in three independent hiPSC-derived neuronal cell lines with distinct genetic backgrounds.

Article activity feed

  1. eLife

    Evaluation Summary:

    This manuscript will be interesting for people performing all optical electrophysiology. It describes a new combination of previously available genetic tools to allow simultaneous optogenetic manipulation and optical electrophysiology. The manuscript does not provide a major conceptual advance but provides good evidence that this assay can be employed for large-scale screening in hIPSC-derived neurons.

    (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.The reviewers remained anonymous to the authors.)

  2. eLife

    Reviewer #1 (Public Review):

    In this manuscript, entitled All-optical electrophysiology in hiPSC-derived neurons with synthetic voltage sensors" the authors present an alternative method for expressing activity-dependent sensors for hiPSC-derived neurons to overcome the harmful effects of expressing genetically-encoded voltage dyes. They used a red-shifted synthetic voltage sensor - BeRST-1 to measure spontaneous and evoked spiking activity, with and without the optogenetic actuator, CheRiff to stimulate the neurons. The data recorded from the iPSC are with good signal to noise ratio and convincing but do not advance the field substantially as the paper applies an existing tool to iPSC, which is important, but describe neither a unique technological development nor a novel biological finding. As a methodological paper that introduces or develops new a method, I expected to see a thorough characterization of the technique and the advantages of using this technique with iPSC-derived neurons. For example:
    Is the dye sensitive enough to measure dendritic voltage changes? To differentiate pre and post-synaptic activity? What is the maximal action potentials frequency it can separate single action potentials?

  3. eLife

    Reviewer #2 (Public Review):

    In this manuscript, Puppo et al. explore the possibility of all-optical electrophysiology in human IPSC derived neurons using a synthetic voltage sensor, with the long term aim to upscale electrophysiological analysis of human neurons. The authors build on previous studies describing the utility of the near infrared sensor BeRST-1, as an advantage compared to genetically encoded voltage indicators to reach high sensitivity, and low cytotoxicity. The authors report that the BeRST-1 signal correlates well with Ca2+ signals, and reliably responds to the optogenetic actuator CheRiff. Additionally, the author demonstrate that this assay is sensitive to modulation by high K+ or GABA antagonistss, and temperature. The data provide good evidence that this assay can be employed for large-scale screening in hIPSC-derived neurons. The impact of successful all-optical electrophysiology is substantial, but the voltage sensor and OG actuator are both previously published, and the data presented here in hIPSCs is often anecdotal. Furthermore, the paper lacks a proper discussion of the results. Hence, the manuscript in its current form does not make a convincing case, given the quality of the data and the fact that the main elements of the assay were previously published.

    Major points

    1. All experiments on primary rat neurons (half the paper) have been published before (albeit with a different OG actuator) in a study by Huang et al. (2015). Figure 2 shows that the same method can be applied to hIPSC derived neurons. This is an important, but expected result, given other comparative studies in rodent and hIPSC derived neurons.

    2. A considerable part of the data is anecdotal and should be supplemented with more experiments to allow quantitative conclusions. Voltage responses to optical stimulation (Fig. 1G and 2H-K) are not quantified. In Fig 1 H, the quantification on "instantaneous firing rate" and "burst/s" are not convincing (high variability, low number of observations). Typical examples in Fig. 2J and K suggest that there is actually quite some variation in the shape of the recorded APs. This is important and should be quantified.

    3. The authors should provide more insight in the data, for instance by using boxplots, overlaid with the individual data points as in fig.1-suppl. 1B-E. Furthermore, they should argue why they used unpaired T-tests only. That seems not appropriate? It is unclear why the authors did not use a paired T-test for the (paired) experiments in fig.1H, and why the n-number for the washout group is 25, while n=9 for the base and KCL group. The authors should report whether data are normally distributed in all experiments.

    4. The authors claim that one of the major advantages of using BeRST-1 over genetically encoded voltage indicators is the improved neuronal viability, but they provide no data to justify this conclusion. According to the authors expression of CheRiff does not cause viability issues. The authors provide data on viability of the hIPSC-derived neurons.

    5. On p9, the authors claim that "the heterogeneity of firing properties within a cell line could be in part due to the difference in neuronal cell types". However, at the end of the paragraph, it is stated that the majority (~90%) of cells are glutamatergic and expression of CheRiff-GFP is driven by CaMKII promoter. Alternative explanations are not considered, such as different NPCs maturation states? To substantiate this, a basic characterization of the hIPSC-derived neurons is necessary (a few established markers for IPSC, NPC and mature neurons).

    6. Some parts of the results are incomplete. E.g. experiments on iPSC neurons in Fig.1-suppl. 2(B-D) are not described in the text, while the quantification of the data of primary neurons in the same figure is mentioned but not shown. Furthermore, P9: "Photoactivation of CheRiff robustly induced depolarization and spiking irrespective of the level of spontaneous activity (Fig. 2 I-J)". On what analysis is this based?

    7. How much was the increase in BeRST-1 signal upon 450 nm laser pulse? The data should be shown

    8. The paper lacks a discussion section, but ends with a conclusions paragraph that seems to emphasize only the benefits of the all-optical approach. A proper discussion of the results, and the opportunities, but also the limitations of the all optical approach compared to conventional electrophysiological techniques, such as MEA recordings and patch-clamp, should be added. Furthermore, conclusions should be clear about which physiological parameters can be measured with this technique (Spike width, ISI, ISR, bursts rate) and which not (resting membrane potential, synaptic potentials, synaptic currents, calibrated current-clamp experiments to measure excitability (rheobase)).

    9. The use of OGB1 to select active neurons for voltage imaging leads to a bias towards higher neuronal activity in the data. P7: "Co-labeling with OGB1 was critical for quick and efficient evaluation of the level of spiking activity and for choosing FOVs for subsequent voltage imaging." and P8: "we used OGB1 for quick evaluation of the level of activity in human neurons (Fig. 2B) taking advantage of larger FOVs achievable with Ca2+ imaging due to a tradeoff between FOV and imaging speed." This should be discussed and a measure for the selected subset should be provided.

    10. It is not clear how sensitive the method is to the voltage sensor loading of the cells. The concentration of 5uM that is used is based on previous experiments in rat primary neurons, but is this also optimal for hIPSC derived neurons? Along these lines, how homogenous is the loading of different cells in the culture, and does this affect the measurements?

    Minor
    • Vertical scale bar is missing in Fig 2G
    • Legend Fig2E: values for mean and SD not given for IFR and Burst rate
    • Resolution of images in 2A,B is too low to judge the morphological integrity of the cells
    • Fig1 suppl 2: What temperature is "heated"?
    • Zoom in on the individual bursts in Fig. 1H. Difficult to see in 1H how well individual spikes can be distinguished
    • How is instantaneous firing rate computed, during bursts only or average firing frequency during total recording?
    • No information about the number of cells stimulated in the network. Why is there no recurrent activity in the network after single evoked APs?
    • On page 7 the authors state that "the algorithm was used to extract several key parameters" while referring to Fig 1 E-F which represent the segmentation of the neurons and the characterization of the bursts by the number of action potentials. Can the author provide more detailed examples of such parameters in the figure? This will benefit other groups that are considering using this system, but are not yet familiar with all the possibilities.

  4. eLife

    Reviewer #3 (Public Review):

    All optical electrophysiology, which combines an optogenetic actuator with an optical voltage sensor, offers the potential for versatile and higher throughput neurophysiology studies. One possible challenge with using a genetically encoded voltage sensor is that the resulting large delivery cassette is challenging to package in a viral vector at high titer, and in general these constructs can compromise cell health in some cases. This manuscript combines viral vector delivery of the optogenetic actuator with a previously developed organic voltage sensitive dye for all optical electrophysiology. The work does represent an advance, particularly the combination of optogenetics with two organic dyes (for voltage and calcium), though with some caveats.

    One concern is that the toxicity of vectors with large payloads or use of multiple vectors is somewhat overstated. Toxicity due to lentiviral transduction of primary cells or differentiated neurons can be challenging. However, when studying human pluripotent stem cell (hPSC) derived neurons (the approach of this manuscript), it's very straightforward to use lenti generate a stable hPSC line carrying multiple expression cassettes, then do the differentiation into neurons.

    In addition, voltage dyes have been used extensively in hPSC-derived neurons, in work for example out of E. Miller's lab (which generated the voltage dye used in this work). That prior work didn't use optogenetics, but it does reduce the novelty of this study somewhat.

  5. Version published to 10.3389/fncel.2021.671549
  6. Version published to 10.1101/2021.01.18.427081v1 on bioRxiv