Time-lapse imaging of cortical projection neuron migration in mice using mosaic analysis with double markers

Summary Mosaic analysis with double markers (MADM) technology enables the sparse labeling of genetically defined neurons. We present a protocol for time-lapse imaging of cortical projection neuron migration in mice using MADM. We describe steps for the isolation, culturing, and 4D imaging of neuronal dynamics in MADM-labeled brain tissue. While this protocol is compatible with other single-cell labeling methods, the MADM approach provides a genetic platform for the functional assessment of cell-autonomous candidate gene function and the relative contribution of non-cell-autonomous effects. For complete details on the use and execution of this protocol, please refer to Hansen et al. (2022),1 Contreras et al. (2021),2 and Amberg and Hippenmeyer (2021).3


Background
This protocol describes 4D time-lapse imaging of Mosaic Analysis with Double Markers (MADM)labeled brain tissue.MADM is a genetic technology that allows the introduction of a homozygous mutation of choice with unprecedented single cell resolution in a defined genetic lineage with concomitant fluorescent labeling. 2,4As MADM technology allows for the analysis of sparse genetic mosaic versus global/whole tissue ablation of a candidate gene with single cell resolution it provides a unique genetic tool to investigate cell-autonomous gene functions and the relative contribution of non-cell-autonomous effects, which can be quantitatively analyzed by time-lapse imaging.Other techniques for sparse labeling, such as viral-induced, in utero electroporation, mT/mG mice, are readily available and could be utilized in conjunction with this protocol.Several potential sparse labeling techniques have been reviewed elsewhere [5][6][7] In this protocol, we present an experimental pipeline for the isolation, culturing and 4D imaging of MADM-labeled brain tissue in situ that allows for the analysis of neuronal migration dynamics.Please refer to 1 for further information.]8 See Contreras et al. and Amberg et al. for details on choosing and breeding the MADM line of choice. 2,3epare dissection equipment (on the day of experiment) Timing: 15 min Clean all dissection tools (2 curved forceps, large and small scissors) with 70% ethanol.Prepare 24-well plates filled with ice-cold PBS to collect embryos (store plate on ice).Prepare 24 1.5 mL Eppendorf tubes to collect tissue for genotyping.Prepare respective genotyping reagents.
Prepare slice-culture dishes (on the day of experiment) Timing: 15 min Fill each well (35 mm glass-bottom dish or 6-well glass-bottom plate) with 1.5 mL of culture medium (see culture medium).Place a cell culture insert in each well/dish.Use high edge filter-inserts as they are heavier and therefore not so prone to move.Place in incubator (37 C, 5% CO2) until needed.

Vibratome sectioning preparations
Keep the vibratome cutting chamber and cutting stage at -20 C before use.
Note: For MADM to work, the presence of a Cre recombinase, active in stem or progenitor cells of your tissue of interest, is required.In other words, any Cre-driver can be used, provided Cre is expressed in dividing stem and progenitor cells, to target cell types and lineages of interest.

Institutional permissions
Institutional and governmental permission and oversight information for the animal study should be obtained.In this study, experimental procedures were discussed and approved by the institutional ethics and animal welfare committees at IST Austria in accordance with good scientific practice guidelines and national legislation (license number: BMWF-66.018/0007-II/3b/2012 and BMWFW- The embryonic age used in this protocol was E14 and E16; however, other developmental stages can be used as desired.

Retrieval of tissue and culturing of acute brain-tissue slices
Timing: 2-5 h (depending on the genotyping protocol) Here, we will describe the process of retrieving the mouse brain tissue.Depending on the gene of interest and the Cre-driver that is used in the experimental MADM paradigm, use the appropriate genotyping protocol to identify the embryos with the correct genotype.Time is of the essence when dealing with living tissue, so minimizing the time between dissection and imaging is highly recommended to ensure tissue viability.
1. Sacrifice pregnant dam by cervical dislocation and remove embryos by c-section.2. Retrieve embryos and isolate brain (Figure 1A). 3. Place each embryo in a 24-well plate containing ice cold 1X PBS and place a small piece of embryonic tissue in a 1.5 mL Eppendorf tube for genotyping.4. Decapitate the embryo and dissect out the brain in cold ACSF.  5. Place the retrieved brain into ice-cold pre-oxygenated ACSF in another 24-well plate until genotyping has finished.

Note:
The genotyping protocol we used typically took 3 hours to complete.If using MADM, one can also check for fluorescent signal using e.g. a fluorescent stereo-microscope to verify correct genotype and/or appropriate fluorescent signal before proceeding to the next step.If a time-consuming genotyping protocol is required, please see problem and solution 1 (see Troubleshooting Problem 1).
6. Embedding the brain (Figure 1B). 7. Carefully remove any excess ACSF and place embryonic brain into an embedding mold.8. Pour 4% low-melting agarose solution into the embedding mold to cover the brain.You may need to gently swirl the brain around in the agarose to dilute any ACSF that was transferred along with the brain to ensure the agarose can bind to the tissue.9. Orient the brain as needed and keep on ice to let the agarose harden.

Vibratome sectioning
10. Remove the embedding mold from the agarose-embedded brain (Figure 1C).11.Dry the agarose-block with embedded brain on the wider bottom by placing it on tissue paper.12. Place a few drops of super glue on the vibratome specimen disc (Figure 1D) and position the dried agarose block on the glue as needed and let dry for 10 s (coronal sectioning, olfactory bulbs facing upwards).13.Trim the agarose block with a blade in a pyramidal shape, base of the block wider than the top, making sure the olfactory bulbs are directed to the top of the pyramid shape (Figure 1E).Make sure to keep a few millimeters of agarose around the brain.14.Place the specimen disc with the trimmed agarose block containing the brain in the buffer tray.15.Immediately pour ice-cooled oxygenated ACSF in the buffer tray to completely cover the embedded brain and oxygenate the chamber with carbogen (Figure 1C).16.Fill the vibratome ice-chamber with ice surrounding the buffer tray (Figure 1G).17.Section the embedded brain coronally at 300 mm (Vibratome setting: 0.4 mm/s, amplitude 1.00 mm).18.As soon as the slice has been cut within the brain region of interest, grab the slice carefully with a forceps (grabbing the surrounding agarose, if possible) and transfer it directly to the millicell insert in the prepared 35 mm glass-bottom culture dish (Figure 1H) (see Troubleshooting Problem 2).19.Add culture media on the filter-insert so that the brain slice is covered but make sure the slices does not float and stays positioned on the filter-insert (Figure 1H).Slices can be moved in the desired position by slowly pushing on the edge of the slice while on the filter-insert (see Troubleshooting Problem 3).20.Place the glass-bottom dish with positioned brain slices in an incubator (37 C, 5% CO2) until ready for the microscope (Figure 1I).We advise readers to commence imaging of the tissue promptly as the disintegration and decrease of viability of the brain slices can occur (see Troubleshooting Problem 4 & 5).Note: Prenatal brains can stay on ice for some time, whereas postnatal and adult brain tissue should be sectioned immediately.Several brain slices from the same animal can be placed on the same filter-insert to increase the number of slices, which can be imaged.
Note: Buffer trays can be kept in the freezer at -20 degrees to ensure a cold environment for longer time when in use.

Time-lapse imaging of live brain slices
Timing: 10-24 h depending on the desired imaging time Here we will describe a setup using an inverted Zeiss LSM800 fitted with a Plan-Apochromat 10x/ 0.45, WD = 2.1 mm objective and equipped with a heating chamber and an ibidi stage-top incubator chamber & gas mixer.Time-lapse images were collected on a PC running Zeiss ZEN Blue software.See Methods videos S1, S2, S3, and S4 for desired outcome examples related to step 31.
21. Start the microscope incubator with gas-mixer and set to 37 C, 5% CO2 and let it reach the set values.22. Mount the FoilCover (PeCon) lid on the glass-bottom-dish to allow gas exchange but reduce the evaporation of the medium fluid (Figure 1J).23.Place the glass-bottom dish with FoilCover lid holding the prepared brain slices in the microscope stage (Figure 1K).24.Let the sample acclimatize for $45 min to reduce the risk of sample drift.25.Select the needed laser lines and filters.For MADM tissue, record in two distinct channels with excitation/emission wavelengths at 488/509 GFP and 554/581 nm for tdTomato.26.Select appropriate objective (we recommend 10x objective for neuronal migration studies), scanning speed and resolution appropriate to the needs of your experiments.27.Adjust laser intensity and gain for each channel.28.Select the region of interest for each sample and adjust the microscope accordingly in xyz directions (Figure 1L).29.Record a time-lapse for each sample (Figure 1M) e.g., for 15 h with a framerate of 15 min recorded unidirectionally at 7 Z-positions with 5 mm spacing between z-planes.30.Save recorded time-lapse images as *.czi files.31.Export time-lapse images as *.tiff files for analysis.
Note: When samples for imaging are placed on filter-inserts, i.e. placed away from the bottom of the glass-bottom dish, it is important to make sure to use a microscope objective with a long working distance e.g.use a Plan-Apochromat 10x/0.45,WD = 2.1 mm objective.
Note: For higher framerates one could apply an inverted spinning-disc microscope setup if available.
Note: Depending on the scanning time of the microscope it is important to note that there is a limit how many regions/samples one can image in one session because the desired framerate is a limiting factor.E.g. a framerate of 15 min (one image every 15 min) with a scan time of e.g. 5 min per imaging block/region allows maximum three regions to be imaged in the session.
CRITICAL: It is important to record the exact framerate in between acquired time-lapse images.This can be done in retrospect by determining the amount of time in between recorded images: e.g.right click on first created image from one time-lapse file and note the time the file was created and do the same for the next consecutive image of the same time-lapse file.The time in between the two images determines the framerate (e.g. 15 min).The framerate is important to know as it determines the units in which the dynamics parameters (Velocity in m/s etc.) are calculated in the further analysis.
Note: Brain slices can be imaged without the PeCon FoilCover, but it will increase the risk of the media drying out and thereby causing unwanted drift of the sample.

Correction of non-linear local drift in time-lapse images
Timing: 1-5 h depending on the processing power of the computer Here we describe the application of 'Undrift', a recently developed tool 1   ii.It is possible to change to change the spatial smoothing of Undrift (standard is set to 31 pixels).
To use undrift with spatial smoothing of e.g., 51 pixels use command.
iii.The output will generate three new files: filename_drift_visu.tiff (visualization of the image drift in a checkboard style, with fields corresponding to the smooth_xy pixel value), filename_optflow_field.tiff (Optical flow field estimation of the image) and filename_undrift.tiff(final corrected image file).iv.Check if the time-lapse images (filename_undrift.tiff)have been corrected.If not, run Undrift again with different smooth_xy values until optimal drift correction has been achieved.v. Apply Undrift with the same parameters to all time-lapse images planned to be used in the same future dataset.35.Analyze the corrected time-lapse images with the software of choice.

Note:
The time-lapse images have to be assembled as an image stack i.e. all individual frames in a single *.tiff file to be used with Undrift.
Note: To use Undrift, it is required to install Anaconda, an open-source Python distribution platform.
Note: Most commands described in this protocol can be copied directly into the Anaconda prompt.

Note:
The amount of spatial smoothing that Undrift will use for the correction vector field can be adapted to the respective images.E.g.If the time-lapse images have a resolution of 512 3 512 pixels and you want to apply $10% smoothing of file pixel size, apply 51 pixels smoothing in Undrift: >undrift C:\time-lapse-images\filename_time-lapse.tiff --smooth_xy 51 For $5% smoothing from file a pixel file size of 512 3 512 pixels xy, use 25 pixels).
The smooth function only accepts integers so decimal values are not accepted.
Note: For help with undrift use command:>undrift --help Note: If you used another path than C:\ to clone the repository, the command is: Note: Undrift can be applied recursively to many time-lapse images in the same folder e.g. using Windows: >for /r %i in (*.tif) do undrift --smooth_xy 51 "%i" Note: In case of extreme drift in the images, it might not be suitable to use such a time-lapse as a correction using Undrift might not be sufficient and therefore proper analysis might not be possible.
Note: At the edges of images with local drift, Undrift processing will remove a part of the image to correct the drift.Note: If correction of drift is necessary for any recorded time-lapse video, it is important to apply the drift correction in all time-lapse images used for later analyses to avoid any possible correction bias.

Analysis of neuronal trajectories
Timing: the automatic tracking is fast (1-5 min) whereas manual curation and correction of tracks can take longer (30 min to 2 h per video).
Here we will describe an example of how one can analyze migrating neurons in time-lapse images by using ImageJ (FIJI) and the TrackMate plugin semiautomatically.For further details refer to. 1 36.Open ImageJ 9  Note: Depending on the goal of the analysis, it can be important to distinguish doublelabeled cells (yellow) from purely red and green labeled cells.Yellow cells can be filtered by setting filters on detected spots in TrackMate for e.g.Median intensity to be above or below a certain value for each analyzed image channel to filter out only red or green labeled cells.

EXPECTED OUTCOMES
The expected outcome should be similar to that observed in Methods videos S1, S2, S3, and S4, a time-lapse of migrating neurons.Depending on the quality of the recorded images, one can use the Undrift tool as described above, to attain videos that can be analyzed with minimal distortion (e.g., tissue drift).
One can proceed with the image analysis to track the migrating neurons over time when the timelapse of neurons has been obtained.Once neurons have been tracked they can be analyzed on different parameters such as e.g., speed, directionality, track distance, acceleration, etc.In Hansen et al. 2022, we analyzed several parameters of the neuronal migration dynamics of projection neurons in the developing mouse cortex.The neuronal migration tracking data from the time-lapse imaging experiments also served as the basis for a quantitative modeling approach to neuronal migration. 1

QUANTIFICATION AND STATISTICAL ANALYSIS
For the tracking data, a range of parameters can be analyzed for describing the dynamics of neuronal migration as found below: Distance of one cell between two frames as: Total distance traveled: Net distance traveled: Net time traveled: Mean straight line speed: Directionality (meandering) index: The tools to analyze the tracked neurons could e.g., be R Studio or Excel and the final statistical analysis can e.g., be carried out with GraphPad Prism or similar software.

LIMITATIONS
Our protocol describes time-lapse imaging of passively cultured brain tissue that allows the analysis of the neuronal dynamics during mouse brain development.Tissue can be cultured and imaged for many hours (>24 hours, See Methods video S4: MADM-5 Emx1-labeled brain tissue at time point E16 for an example), however, over extended periods (>12 hours) the tissue can start to disintegrate and be less viable.Depending on the goal of the experiment, one can adjust the time needed to record the biological process of interest.

TROUBLESHOOTING Problem 1
The genotyping process is long and the tissue needs to be stored longer before imaging can start (related to step 5).

Potential solution
If a time-consuming genotyping protocol is required, we recommend oxygenating the ice cold ACSF in which the brains are stored to ensure their viability.Alternatively, one can continue with the protocol until step 19 before genotyping and store the cultured slices in the incubator until the genotyping has finished.Of note, long storage of slices before imaging most likely reduces the viability of the tissue.In extreme cases, one may decide to image immediately and genotype in retrospect.

Figure 1 .
Figure 1.Time-lapse imaging of MADM-labeled tissue (A) Retrieve embryos from the pregnant mouse at the desired developmental stage and genotype to select the brains with MADM labeling.Extracted brains should be stored in ice-cold ACSF until further use.(B) Embed the MADM-labeled brain in an embedding mold and fill with low-melting agarose and position it with the olfactory bulbs facing down.

Figure 1 .
Figure 1.Continued (C) Cool on ice to stiffen the agarose where after the embedding mold can be removed.(D) Mount the embedded brain on the specimen disc with glue.(E) Trim the agarose block into a pyramidal shape.(F) Place the sample in the buffer tray and fill with ice-cold oxygenated ACSF until covered.(G) Place the prepared buffer tray with the mounted specimen in the ice-cooling tray of the vibratome and start slicing the brain.(H) Place the freshly sliced brain tissue on the prepared filter-inserts mounted in a glass-bottom dish.(I) Store the sample in an incubator until imaging.(J) Place the FoilCover lid on the glass-bottom dish containing the sample before placing it in the microscope with an incubator.(K) Place the sample in the microscope and make sure that the climate within the microscope incubator is as desired.(L) Locate the desired region of interest and start imaging with the desired settings.(M) Record the time-lapse over desired length of time.
iii.Clone repository to <path> (e.g., the path where you want the python repository to be stored).For installing in the directory you are in use the command: iv.Move to the directory where Undrift were installed:>cd undrift v. Install required packages by running the command: b.Use Undrift.i. Basic command to apply Undrift to time-lapse images: >git clonehttps://github.com/hippenmeyerlab/undrift >pip install -r requirements.txt-e .>undrift <time-lapse file> E.g. >undrift C:\time-lapse-images\filename_time-lapse.tiff >undrift C:\time-lapse-images\filename_time-lapse.tiff --smooth_xy 51 >cd <path>/undrift) to correct non-linear local drift in recorded time-lapse images available at http://github.com/hippenmeyerlab/undrift (https:// doi.org/10.5281/zenodo.10183022).Undrift is based on a python script that can be executed by using the Anaconda python platform.Unlike other drift correction options, such as current ImageJ plugins, Undrift facilitates the automatic application of local drift correction to time-lapse images containing two fluorophores.The text highlighted in gray are commands to be typed in the Anaconda terminal.See Methods video S1: MADM-11 Emx1-labeled brain tissue at time point E16, not yet Undrifted and Methods video S2: MADM-11 Emx1-labeled brain tissue at time point E16, Undrifted for an example of before and after Undrift application related to Step 34.Run Anaconda and install git by running the command: >conda install git.
and load the desired time-lapse file to be analyzed (File / Open or simple drag and drop your file into ImageJ).37. Go to Image / Properties and set the Frame interval to the respective framerate at which the time-lapse was recorded 15 min) and set the Pixel width and Pixel height to the corresponding values in micrometer (e.g., pixel width and height 1.2479 mm and Voxel depth 5 mm).38.Save the file while keeping it open in ImageJ (File / Save / Replace).39.Start the plugin TrackMate 10 (ImageJ (FIJI) / Plugins / Tracking / TrackMate ).40.Check that the Calibration and Crop settings are set correctly for the loaded file.41.Choose the LoG detector and press next.42.Choose the channel to be analyzed (Segment in channel) and set Estimated blob diameter: 10.0 mm, threshold: 2.0, Median filter: enabled, Sub-pixel localization: enabled.Press next.43.After detection, press next.44.Choose Auto in Initial thresholding and press next.45.In select a view, Choose Hyperstack displayer and press next.46.In the next window one can apply filters on spots if needed and press next.47.In ''Select a tracker'' choose Linear motion LAP tracker and set values (Initial search radius: 15, Search radius: 15, Max frame gap: 2).48.Curate each track manually to ensure that the tracking was successful.Either correct the tracking manually or discard track using the function ''TrackScheme''.49.Save the TrackMate file by Clicking ''Save''.50.Extract all parameters by clicking ''Analysis'' and three windows with analysis data will pop-up.Save the data as a .csvfile by pressing File / Save in each pop-up window.51.Repeat the TrackMate analysis for the other channel in the image by restarting the process.52.Analyze extracted data as desired by e.g., calculating velocity, directionality etc. (See expected outcomes).