Quantification of cristae architecture reveals time-dependent characteristics of individual mitochondria

A novel, machine-learning approach to quantify cristae density and architecture in live-cell super-resolution imaging characterizes the ultrastructural parameters of individual mitochondria over time.

Thank you for submitting your manuscript entitled "High-throughput quantification of mitochondrial cristae architecture and density in living cells" to Life Science Alliance. The manuscript was assessed by expert reviewers, whose comments are appended to this letter.
As you will see, your work received a mixed view from the reviewers. While reviewer #1 and #3 think that the method can get published here pending satisfactory revision, reviewer #2 thinks that the lack of benchmarking to existing methods makes it hard to appreciate the value provided. We think that you may be able to provide such benchmarking and would thus like to invite you to submit a revised version of your manuscript to us, addressing the reviewers' concerns. We would like to encourage you to also address the request of reviewer #2 to test your method for capturing dynamic changes (last point of the reviewer).
To upload the revised version of your manuscript, please log in to your account: https://lsa.msubmit.net/cgi-bin/main.plex You will be guided to complete the submission of your revised manuscript and to fill in all necessary information. Please get in touch in case you do not know or remember your login name.
We would be happy to discuss the individual revision points further with you should this be helpful.
While you are revising your manuscript, please also attend to the below editorial points to help expedite the publication of your manuscript. Please direct any editorial questions to the journal office.
The typical timeframe for revisions is three months. Please note that papers are generally considered through only one revision cycle, so strong support from the referees on the revised version is needed for acceptance.
When submitting the revision, please include a letter addressing the reviewers' comments point by point.
We hope that the comments below will prove constructive as your work progresses.
Thank you for this interesting contribution to Life Science Alliance. We are looking forward to receiving your revised manuscript.
--A letter addressing the reviewers' comments point by point.
--An editable version of the final text (.DOC or .DOCX) is needed for copyediting (no PDFs).
--High-resolution figure, supplementary figure and video files uploaded as individual files: See our detailed guidelines for preparing your production-ready images, http://www.life-sciencealliance.org/authors --Summary blurb (enter in submission system): A short text summarizing in a single sentence the study (max. 200 characters including spaces). This text is used in conjunction with the titles of papers, hence should be informative and complementary to the title and running title. It should describe the context and significance of the findings for a general readership; it should be written in the present tense and refer to the work in the third person. Author names should not be mentioned.

B. MANUSCRIPT ORGANIZATION AND FORMATTING:
Full guidelines are available on our Instructions for Authors page, http://www.life-sciencealliance.org/authors We encourage our authors to provide original source data, particularly uncropped/-processed electrophoretic blots and spreadsheets for the main figures of the manuscript. If you would like to add source data, we would welcome one PDF/Excel-file per figure for this information. These files will be linked online as supplementary "Source Data" files. ***IMPORTANT: It is Life Science Alliance policy that if requested, original data images must be made available. Failure to provide original images upon request will result in unavoidable delays in publication. Please ensure that you have access to all original microscopy and blot data images before submitting your revision.*** The report of Prof. Shirihai and colleagues describes the implementation of a semi-automated approach to quantify the architecture of the mitochondrial inner membrane.
The approach relies on the popular image processing package Fiji and several of its previously described plugins. Quantification of microscopic images is a challenging task and suitable software tools are urgently needed. Thereby this manuscript provides a valuable starting platform for researchers to analyze their own data sets. This protocol will certainly be valuable for the community.
I have only one serious concern that needs to be comprehensively addressed by the authors. It is claimed (page 5) that "the workflow was able to effectively segment cristae that were <50 nm apart from each other". This is a very bold and a very surprising statement, because the resolution of an AiryScan microscope is (at optimal conditions) around 120 nm. To my understanding it would be just impossible to resolve with such a microscope structures closer than 50 nm apart. This contradiction requires a detailed and careful explanation. The authors must provide several examples of a separation of structures closer than 50 nm and must provide the raw image data so that readers can verify this claim independently.
Minor concern Related to the issue phrased above, it appears unlikely that the authors always recorded single cristae. It appears much more likely that often two or more adjacent cristae are perceived as a single crista. This should be mentioned.
Reviewer #2 (Comments to the Authors (Required)): New methods for quantifying cristae in an unbiased way represent a new challenge in the field of mitochondrial research with considerable potential in the study and characterization of these organelles. Segawa et al reported a novel approach to efficiently characterize several features (size, shape) of mitochondrial cristae. I have the concern that more than developing a new method, the authors simply show that an already existing method can be used to quantify cristae from high resolution microscopy images obtained with airyscan. Another important issue is how general is the claimed advantage of this protocol when applied to other high resolution microscopy techniques for living cells like STED or SIM. Overall, their protocol efficiently selects and quantifies important aspects of mitochondrial cristae. However, in my opinion the advantage of this strategy with respect a conventional thresholding protocol is not clear. The authors argue that their method overcomes the limitations of conventional segmentation, but these limitations could also be overcome by a more restrictive thresholding algorithm. They should compare their approach with conventional segmentation under different algorithms of thresholding and demonstrate that their approach is still significantly more capable of detecting mitochondrial cristae shape size density etc. Besides comparing their method to conventional thresholding, how do their results compare with cristae data obtained from EM on the same type of cells? They should also include this comparison for the cardiolipin deficient cells vs. wild type. Moreover, the number of cells analysed in figures 5 and 7 should be increased in order to validate the approach for high -throughput analysis. In addition, the analysis is done in living cells but with simple snaps, would be good to detect changes in cristae evolution over the time upon CCCP treatment or any alternative treatment.
Reviewer #3 (Comments to the Authors (Required)): This is a very nice tool to analyze the latest Cristae fluorescence images semi-automatically after a training session. However, some information and steps are still missing in the workflow. Is the requirement an 8bit image? Does the scale have to be entered beforehand? What to enter in Circularity? To the authors: Please go through the workflow step by step again, maybe have someone with less experience run the macro in the lab. He/she will point to open questions. If it is taken care of this, everything speaks for a quick publication.
Reviewer #1 (Comments to the Authors (Required)): The report of Prof. Shirihai and colleagues describes the implementation of a semi-automated approach to quantify the architecture of the mitochondrial inner membrane. The approach relies on the popular image processing package Fiji and several of its previously described plugins. Quantification of microscopic images is a challenging task and suitable software tools are urgently needed. Thereby this manuscript provides a valuable starting platform for researchers to analyze their own data sets. This protocol will certainly be valuable for the community.
We appreciate these encouraging comments and hope that our study will serve as an effective platform for quantifying the inner mitochondrial membrane.
I have only one serious concern that needs to be comprehensively addressed by the authors. It is claimed (page 5) that "the workflow was able to effectively segment cristae that were <50 nm apart from each other". This is a very bold and a very surprising statement, because the resolution of an AiryScan microscope is (at optimal conditions) around 120 nm. To my understanding it would be just impossible to resolve with such a microscope structures closer than 50 nm apart. This contradiction requires a detailed and careful explanation. The authors must provide several examples of a separation of structures closer than 50 nm and must provide the raw image data so that readers can verify this claim independently.
Thank you for bringing this to our attention! This has been corrected in the revised version. We see that our choice of words was poor in relation to this panel. We should not have implied that we were resolving cristae that were 50 nm apart, because this would be beyond the resolution of the Airyscan. For this panel, we simply wanted to note that our segmentation protocol could distinguish between objects, which have edges that are very close together. In other words, we did not want to claim that we were resolving two cristae that were in reality only 50 nm apart, since, to do this, we would have needed to measure the distance between the two peaks of the objects, in terms of their maximal fluorescence intensities. To address this point, we have removed this image and associated scale bar from the manuscript, because it was misleading. Furthermore, in the revised version, we refrained from deducing a resolution based on the images. Please also see the response below associated with the minor concern, because it is related to this point.
Minor concern Related to the issue phrased above, it appears unlikely that the authors always recorded single cristae. It appears much more likely that often two or more adjacent cristae are perceived as a single crista. This should be mentioned.
We appreciate this criticism, which we have addressed in the revised manuscript. We further elaborated on this point on pgs. 14 and 15 in the Discussion: "It is important to note that, while stateof-the-art high-and superresolution live-cell imaging technologies are providing exciting new windows onto the complex biology of mitochondrial membranes, they remain, to date, unable to supersede EM as a way to probe the spatial dimensions of the organelle. For example, the resolution of electron 1st Authors' Response to Reviewers April 27, 2020 tomography (ET) systems typically ranges from 5 to 20 nm. Therefore, given average ET measurements of crista length and width to be 240 nm and 20 nm, respectively, the cross-sectional area of a crista is 4,800 nm 2 (or 0.0048 µm 2 ) (Mannella et al., 2013). The lateral resolution of SIM (including the GE DeltaVision OMX SR system used in this study) is approximately 100 nm. Our measurements of mean crista area taken from SIM micrographs were 0.0489 µm 2 . This value is about one order of magnitude larger than estimates obtained from ET micrographs, a disparity consistent with an approximately 10fold difference in the resolution of the two imaging technologies. Therefore, researchers should bear in mind that measurements of cristae area from live-cell, superresolution images are likely to be overestimated by roughly a factor of 10. Given that the mean crista-to-crista distance in HeLa cells is 120 nm (Stephan et al., 2019), it is likely that our Airyscan and SIM images did not fully capture all of the cristae within a particular frame -especially in cases where cristae were positioned more closely together than could be resolved by the imaging systems employed in this study." Reviewer #2 (Comments to the Authors (Required)): New methods for quantifying cristae in an unbiased way represent a new challenge in the field of mitochondrial research with considerable potential in the study and characterization of these organelles. Segawa et al reported a novel approach to efficiently characterize several features (size, shape) of mitochondrial cristae. I have the concern that more than developing a new method, the authors simply show that an already existing method can be used to quantify cristae from high resolution microscopy images obtained with airyscan. Another important issue is how general is the claimed advantage of this protocol when applied to other high resolution microscopy techniques for living cells like STED or SIM.
We thank the referee for bringing up this important point. To address whether our machine-learning segmentation of mitochondrial cristae would also be useful for superresolution images obtained with other microscopy techniques, we include new SIM images in the revised version. Using these images, we present a validation that our machine-learning protocol is still significantly better than conventional (manual) thresholding, when applied to SIM images of mitochondria. Please see the new Overall, their protocol efficiently selects and quantifies important aspects of mitochondrial cristae. However, in my opinion the advantage of this strategy with respect a conventional thresholding protocol is not clear. The authors argue that their method overcomes the limitations of conventional segmentation, but these limitations could also be overcome by a more restrictive thresholding algorithm. They should compare their approach with conventional segmentation under different algorithms of thresholding and demonstrate that their approach is still significantly more capable of detecting mitochondrial cristae shape size density etc.
To address this point in the revised manuscript, we provide new data, where we examined more restrictive thresholding algorithms available in ImageJ (i.e., MaxEntropy and Shanbhag) in addition to our conventional (manual) thresholding. We determined that MaxEntropy performed about as well as conventional (manual) thresholding, and Shanbhag tended to underestimate the cristae density, and tended to select only smaller regions within cristae area. Please see the new Figs 2-3 for details about the performance of these other benchmarks. Altogether, our machine-learning (TWS) approach provided better segmentation than any other algorithms that we tested.
Besides comparing their method to conventional thresholding, how do their results compare with cristae data obtained from EM on the same type of cells?
We thank the reviewer for this important comment. In the revised version of this manuscript, we addressed how our measurements of superresolution images of cristae compare to those performed on electron micrographs in the Discussion. Please see the text associated with the response to reviewer #1.
They should also include this comparison for the cardiolipin deficient cells vs. wild type.
We thank the reviewer for bringing up this point. We provide a comparison to EM images of PTPMT1deficient cells on pg. 9 of the results section, where we note that EM images show swollen cristae, resembling what we observed using Airyscan microscopy. Numerical comparison to images from other studies was not feasible, since previous studies did not perform any quantifications of architectural parameters, and the images provided in these publications are insufficient for us to perform a thorough comparative analysis.
Moreover, the number of cells analysed in figures 5 and 7 should be increased in order to validate the approach for high -throughput analysis.
We thank the reviewer for requesting clarification on this point. In the graphs, each point represents an independent experiment. In each experiment, we analyzed approximately 15 different cells, including hundreds of mitochondria, and many thousands of cristae.
In addition, the analysis is done in living cells but with simple snaps, would be good to detect changes in cristae evolution over the time upon CCCP treatment or any alternative treatment.
We thank the referee for this making this wonderful suggestion as to detecting cristae dynamics over time, which led us to include the following exciting new data. In the revised manuscript, we addressed this point in two ways: 1) we performed a series of SIM experiments using FCCP on HeLa cells, where we tracked acute changes in cristae parameters over a period of 1 hr (please see new Fig. 8 and new Fig  S4); and, 2) we examined changes in cristae architecture at high temporal resolution using SIM, examining changes in cristae structure independently of mitochondrial fusion and fission as well as in the context of classical mitochondrial dynamics (please see new Figs 9-12).
Reviewer #3 (Comments to the Authors (Required)): This is a very nice tool to analyze the latest Cristae fluorescence images semi-automatically after a training session.
We thank the referee for these generous comments! However, some information and steps are still missing in the workflow. Is the requirement an 8bit image?
We thank the referee for bringing up this point. In our study, we analyzed Airyscan images, which have a bit depth of 16, and SIM images, which have a bit depth of 15. It is likely that using images with higher bit depths will improve the ability to segment cristae. We added this information to the revised manuscript, in the new Supplementary Methods section.
Does the scale have to be entered beforehand?
To address this point, we modified the Supplementary Methods in the revised manuscript. To train the Trainable Weka Segmentation plugin to segment objects (e.g., cristae), it is not necessary to first enter the scale of the image. This information is retained in the metadata of each image and will be automatically applied during the machine-learning training process.
What to enter in Circularity?
Thank you for bringing up this point. Your point illustrated for us that we should clarify the specific steps required to make use of the software. We addressed this point in the revised manuscript, by addition of text to the Supplementary Methods section. Circularity is a shape descriptor that ImageJ provides as a standard readout or measurement of regions of interest (ROIs). To obtain values of circularity associated with specific ROIs, it is only necessary to make sure the "Shape descriptors" box is checked in the "Set Measurements" window of the "Analyze" tab. Typically, this box is checked by default.
To the authors: Please go through the workflow step by step again, maybe have someone with less experience run the macro in the lab. He/she will point to open questions.
We thank the referee for this suggestion, which we have gladly followed. We added additional explanations to the workflow (Supplementary Methods: pgs 29-32), following various recommendations of new users.
If it is taken care of this, everything speaks for a quick publication.
Thank you very much! Thank you for submitting your revised manuscript entitled "Quantifying cristae architecture reveals time-dependent characteristics of individual mitochondria". As you will see, the reviewers appreciate the introduced changes and now support publication, pending minor revision. We would be happy to publish your paper in Life Science Alliance pending addressing the following: -Please address the remaining reviewer concerns -Please check one more time whether the author order in our system matches the one in the manuscript file -Please incorporate the "Detailed, Step-by-step Workflow of TWS of mitochondrial cristae" in the main manuscript, Methods section, in order to make it easier for readers to apply your method If you are planning a press release on your work, please inform us immediately to allow informing our production team and scheduling a release date.
To upload the final version of your manuscript, please log in to your account: https://lsa.msubmit.net/cgi-bin/main.plex You will be guided to complete the submission of your revised manuscript and to fill in all necessary information. Please get in touch in case you do not know or remember your login name.
To avoid unnecessary delays in the acceptance and publication of your paper, please read the following information carefully.

A. FINAL FILES:
These items are required for acceptance.
--An editable version of the final text (.DOC or .DOCX) is needed for copyediting (no PDFs).
--High-resolution figure, supplementary figure and video files uploaded as individual files: See our detailed guidelines for preparing your production-ready images, http://www.life-sciencealliance.org/authors --Summary blurb (enter in submission system): A short text summarizing in a single sentence the study (max. 200 characters including spaces). This text is used in conjunction with the titles of papers, hence should be informative and complementary to the title. It should describe the context and significance of the findings for a general readership; it should be written in the present tense and refer to the work in the third person. Author names should not be mentioned.

B. MANUSCRIPT ORGANIZATION AND FORMATTING:
Full guidelines are available on our Instructions for Authors page, http://www.life-sciencealliance.org/authors We encourage our authors to provide original source data, particularly uncropped/-processed electrophoretic blots and spreadsheets for the main figures of the manuscript. If you would like to add source data, we would welcome one PDF/Excel-file per figure for this information. These files will be linked online as supplementary "Source Data" files. **Submission of a paper that does not conform to Life Science Alliance guidelines will delay the acceptance of your manuscript.** **It is Life Science Alliance policy that if requested, original data images must be made available to the editors. Failure to provide original images upon request will result in unavoidable delays in publication. Please ensure that you have access to all original data images prior to final submission.** **The license to publish form must be signed before your manuscript can be sent to production. A link to the electronic license to publish form will be sent to the corresponding author only. Please take a moment to check your funder requirements.** **Reviews, decision letters, and point-by-point responses associated with peer-review at Life Science Alliance will be published online, alongside the manuscript. If you do want to opt out of having the reviewer reports and your point-by-point responses displayed, please let us know immediately.** Thank you for your attention to these final processing requirements.
Thank you for this interesting contribution, we look forward to publishing your paper in Life Science Alliance.