Characterization of Neuromuscular Junctions in Mice by Combined Confocal and Super-Resolution Microscopy

Care and manipulation of mice were performed according to national and European legislation on animal experimentation and approved by the institutional ethical committee. Males and females of Smn^2B/- (C57Bl/6J background) and ColQ^Dex2/Dex2 (B6D2F1/J background) mice at 3- and 6-weeks of age, respectively, were used in the study.

1. Euthanasia of mice and dissection of muscles: tibialis anterior  and gastrocnemius

1. Proceed to mouse anesthesia by intraperitoneal injection of a ketamine (87.5 mg/kg)/xylazine (12.5 mg/kg) mixed solution (0.1 mL/20 g of body weight) prior to euthanasia by cervical dislocation.
   NOTE: Since SMA and ColQ-CMS affect individuals independently of their sex, male and female mice were used in the present protocol.
2. Remove hindlimbs hair using a small electric shaver and rinse the legs with 70% ethanol.
   NOTE: The dissection procedure will differ for each muscle. For dissection of tibialis anterior (TA), follow steps 1.2.1-1.2.3, and for gastrocnemius (GA) (medial and lateral parts), follow steps 1.2.4-1.2.6. Handle the muscles gently to prevent tissue damage and crushing or stretching them during dissection.
   1. Place the mouse in the supine position.
   2. Make a skin incision of 5 mm with sharp-blunt scissors along the antero-external part of the distal hindlimb, parallel to the tibia, to expose the muscle. Use extra-thin scissors to remove the fascia.
   3. Cut the distal tendon first (close to the paw) and then the proximal tendon (close to the knee) using extra-thin scissors and a curved thin forceps. Handle the muscle carefully to avoid damage to myofibers and nerves.
      NOTE: The proximal tendon must be sectioned as close as possible to the bone to harvest the entire muscle.
   4. Place the mouse in the prone position, use sharp-blunt scissors to make a skin incision from the upper part of the distal hindlimb posterior compartment down to the paw, and remove the skin.
   5. Grasp the Achilles tendon with medium serrated forceps, cut it with an extra-thin scissor and gently separate the GA from the surrounding tissue back to its proximal insertion.
   6. On the proximal side, insert the medium serrated forceps into the pocket formed between the biceps femoris (BF) and the GA. Separate the two muscles to cut the GA tendon as close as possible to the bone insertion with an extra-thin scissor.
3. For tissue fixation, place each muscle into a 2 mL microcentrifuge tube containing 1 mL of 4% w/v paraformaldehyde (PFA) solution diluted in phosphate buffer saline (PBS without Ca²⁺Mg²⁺) and keep at 4 °C for 18-24 h.
   CAUTION: Paraformaldehyde and formaldehyde are toxic and must be handled in a chemical fume hood with adequate protective equipment.
4. Next day, wash the fixed muscles 3x for 5 min with PBS in 12-well plates by shaking gently at room temperature (RT) inside a chemical fume hood.
   NOTE: The protocol can be stopped at this step and continued within a month. In this case, add PBS supplemented with 0.01% sodium azide to store samples at 4 °C.
5. Tease each muscle in small fiber bundles of about 1 mm wide using two fine serrated forceps.
   NOTE: It is crucial to manipulate muscles very gently with the forceps, without excessive force, to prevent tissue damage during teasing.
   1. Dissociate the TA muscle into 3 or 4 bundles depending on its size.
   2. For GA, separate the medial and lateral parts of the muscle and then dissociate each part into 4-5 bundles depending on their size.

2. Immunostaining

1. Proceed with muscle fiber permeabilization: Transfer muscle bundles into 24-well plates containing 1% (v/v) Triton X-100 in PBS and keep them under gentle agitation (50 rpm) for 1 h at RT or 5 h at 4 °C.
   NOTE: Split the muscle bundles between two plates to proceed with separate immunostainings and to minimize the risk of antibody confusion. Do not split them into more than two wells (1 well/plate); otherwise, the number (N) of NMJs that are representative of their general status in the analyzed muscle may be insufficient.
2. Wash the samples 3x for 5 min with PBS at RT and incubate them with a blocking solution composed of 4% bovine serum albumin (BSA) in PBS/Triton X-100 1% for 4 h at 4 °C, under gentle agitation (50 rpm).
   NOTE: Do not use an aspiration pump during the washing steps, but rather aspirate the solution manually with a 200 µL pipette and small-sized tips (the reference is indicated in the Table of Materials).
3. Incubate the samples overnight (O/N) at 4 °C under gentle agitation (50 rpm) with the blocking solution indicated in step 2.2 containing primary monoclonal antibodies against either neurofilament M (NF-M, 2H3, dilution 1/200) or synaptic vesicle glycoprotein 2 (SV2, dilution 1/200) to label presynaptic axon terminals or active zones, respectively.
4. Next day, wash the muscle bundles 3x for 5 min in PBS under agitation (50 rpm).
   1. For confocal imaging: Incubate the muscle bundles with secondary anti-mouse antibodies conjugated with a red-emitting fluorophore (F594) (dilution 1/500) and α-bungarotoxin conjugated with a green-emitting fluorophore (α-BTX-F488) (dilution 1/1000) in PBS for 2 h at RT under agitation (50 rpm).
   2. For STED imaging: Incubate the muscle bundles with secondary anti-mouse antibodies conjugated with a green-emitting fluorophore (F488) (dilution 1/500) and α-bungarotoxin conjugated with a far-red-emitting fluorophore characterized by high photostability (α-BTX-F633) (dilution 1/1000) in PBS for 2 h at RT under agitation (50 rpm).
      NOTE: Do not expose the samples to light during incubation to avoid photobleaching.
5. Wash the labeled muscle bundles 3x for 5 min with PBS under agitation (50 rpm) and place them on a slide with a mounting medium.
   NOTE: Place a maximum of 4 to 5 muscle bundles per slide to allow sealing.
6. Add a grade #1.5 (or #1.5H) glass coverslip (0.17 mm thickness) on the top, and place cylindrical magnets on both sides of the slide to apply pressure and flatten the muscles.
7. Keep the slides protected from light O/N at 4°C. Seal the slides permanently with nail polish.

3. Image acquisition

1. Acquisitions by a confocal microscope
   NOTE: Images were collected with an inverted laser-scanning confocal microscope using a 63x magnitude oil immersion objective (HCX Plan Apo CS, 1.4 numerical aperture (NA)).
   1. For blinded analysis, let a person not involved in the analysis code each slide with a given number. Remain blinded to the experimental groups until the quantification of NMJ parameters is complete for all samples.
   2. Launch the microscope software in Configuration Mode > machine.xlhw (Supplemental Figure 1).
   3. Place the slide on the microscope stage and find the observation plane within the sample by looking under DAPI wide-field fluorescence illumination with the DAPI filter set.
   4. Click on Open Project > New Projects and create a folder to store image acquisitions (Supplemental Figure 1).
      NOTE: Create a new project for each NMJ to limit folder size and prevent computer memory issues.
   5. To manage acquisition parameters, click on the Acquisition tab window and set the confocal pinhole to 1.0 Airy unit and laser power to optimize the gain and offset levels for the green/F488 (α-BTX) fluorescence using a 488 nm laser at the endplate that has to be imaged (Live mode ON).
   6. Next, optimize the red/F594 (NF-M or SV2) fluorescence acquisition using a laser adapted to F594 observation. In this study, a 552 nm laser was used (Live mode ON). Set the spectrum of dye emission with the following ranges for each laser: laser 405 (DAPI) from 414 to 483 nm, laser 488 (F488-α-BTX) from 506-531 nm, and laser 552 (NF-M/SV2) from 622-650 nm.
   7. Collect image stacks of neuromuscular junctions in each experimental group with the same settings: image size 1024 x 1024 pixels (73.7 x 73.7 µm) at 400 Hz sampling rate, Bidirectional X ON, Zoom factor 2.5, Z-step size 0.5 µm in Z-Wide mode.
      NOTE: For each NMJ, the number of slices is set to acquire the whole junction. The acquisition settings described above fulfill the Nyquist-Shanon sampling theorem. However, the user can click on the Optimize Format button, present on all recent confocal operating software, to ensure pixel size and Z-step meet the ideal Nyquist sampling rate. This action will avoid over or under-sampled images, which will cause a loss of accuracy in volume measurements.
   8. Save the original file (.lif) or Z-stack images (.tif) in a folder with a name that includes the code name of the slide, the staining type, and the endplate number.
      NOTE: Collect sequentially (not simultaneously) the scans using the 488 nm and 552 nm lasers (F488 and F594) to avoid crosstalk of the F488 fluorescence into the F594 channel and vice versa (bleed-through). NB: the beam path can be configured with the Dye Assistant in the microscope software.
   9. Change to the next coded slide and repeat steps 3.1.3-3.1.8 for each NMJ.
   10. At the end of the session, click on Open in 3D Viewer and choose an NMJ representative of an experimental group to visualize the 3D labeling.
       NOTE: This view mode will help to verify that the acquisition parameters were correct.
   11. Close the microscope software, clean the objectives with lens tissues, and then turn off the system.
2. Acquisitions by STED microscopy
   NOTE: Images were collected with an inverted laser-scanning confocal microscope equipped with Gated STED at 775 nm using a 100x oil immersion objective (HC PL APO CS2 1.4 NA).
   1. For blinded analysis, let a person not involved in the analysis code each slide with a given number. Remain blinded to the experimental groups until the quantification of NMJ parameters is complete for all samples.
   2. Launch the microscope software in Configuration Mode > machine.xlhw and STED ON (Supplemental Figure 2).
   3. Click on Open project > New Projects to create a folder to store image acquisitions.
      NOTE: Generate a new folder for each slide to limit folder size and prevent computer memory issues.
   4. Place the slide on the microscope stage and view it under wide-field fluorescence illumination using the 488 nm laser to find the observation plane within the sample.
   5. Search for a NMJ labeled with neurofilament M (NF-M) or SV2 stainings using the 488 nm laser with a spectral detection from 506-531 nm.
   6. When a NMJ has been identified, click on Activate STED and start acquiring images in a region that contains several junctional folds (Supplemental Figure 3) using the 635 nm laser with a spectral detection from 640-750 nm.
      NOTE: Be careful to the saturation look-up table during image acquisition and click the Quick LUT button to avoid overexposure (grey values >255; for 8 bit).
   7. Collect the images of each experimental group with the same settings: image size 2048 x 2048 pixels (38.75 x 38.75 µm) at a 400 Hz sampling rate.
      NOTE: The depletion laser (STED) power is set to 65%.
   8. Save the images with a file name that includes the code of the slide.
      NOTE: It is possible to click on Optimized XY Format: Set Format に to obtain the best acquisition setting with STED imaging.
   9. Change to the next coded slide and repeat steps 3.2.3-3.2.8. Repeat this procedure for all slides.
   10. At the end of the STED microscopy session, transfer the image files to another computer and save the original files (.lif) in an external drive or server.
   11. Turn off the microscope software, clean the objectives with lens tissues, and then turn off the system.

4. Image analysis- confocal microscopy

NOTE: All images were processed with computers using Microsoft Windows 10 professional operating system.

1. Launch ImageJ and custom macro to calculate postsynaptic NMJ endplate volume, Maximum Intensity Projection (MIP) area, and relative tortuosity.
   1. Process NMJ image stacks using NIH ImageJ freeware²³, the iGeodesic plugin and the custom macro to obtain NMJ parameter measurements. Launch ImageJ software.
      NOTE: The latest version of ImageJ is freely available and can be downloaded²⁴. In order to open proprietary file formats, Bio-Formats Package²⁵ plugin has to be downloaded²⁶ . This step is not necessary in case the operator uses Fiji because the plugin is already installed in the software. The iGeodesic plugin²⁷ to compute tortuosity is also available online²⁸; verify the availability of this plugin in the ImageJ/Fiji version that will be used. The custom-made Macros are also available online²⁹.
   2. Drag and drop the Macro_NMJ_VOL_Marinelloetal.ijm (custom made, Supplementary Coding File 1) to the ImageJ window; the macro will be open in a second window. In this new window, click on Macros > Run Macro.
      NOTE: The macro can process both proprietary and TIFF files. Files must fulfill the following criteria: for proprietary file formats, save only one junction (i.e., the stack of images) per file, ordered in a folder; for TIFF images, files must be saved in a folder containing subfolders, each named JunctionX (X corresponds to a NMJ number) with the image stacks of a given junction (RGB TIFF) (Supplemental Figure 4).
   3. Select the native folder containing the Junction subfolders that have to be analyzed and click on Select.
   4. In the new pop-up menu called Saving Folder, select the storage folder and click on Select.
   5. In the new pop-up menu called Image Type, select the format of the Z-stack acquisitions.
   6. Select the RGB channel corresponding to the staining of interest and indicate XY pixel size and Z-step (z). The macro will automatically perform the analysis.
      NOTE: In case proprietary file formats are selected, the macro directly reads pixel size and Z-step (z). However, the user still has to indicate the channel of interest (C1, C2, or C3). The macro will provide a DataSheet (.csv) for each junction parameter (endplate volume, MIP area, and tortuosity) in the saving folder. The macro also generates three .TIF files, which correspond to the perimeter of α-BTX staining Drawing_MaxprojX.tif, DrawingJunctionX.tif, and MIP MaxprojX.tif. These TIFF files are generated to verify the quality of the acquisitions and to ensure that image processing has been performed correctly.
      Postsynaptic NMJ volume (V): The macro will separate images from a single NMJ and keep the α-bungarotoxin F488 channel corresponding to the postsynaptic endplate. The stack is segmented using Otsu threshold³⁰ on the intermediate slice of the stack. The resulting binary image is 1-pixel dilated, and the Analyze Particles feature is used to measure the endplate area of each detected object. To obtain the postsynaptic NMJ volume, the macro sums all measured endplate areas of the stack and multiplies it by Z-step value (z) in µm.
      
      Maximum intensity projection (MIP) endplate area: After the stack is thresholded, the maximum intensity projection (MIP) is obtained using the Z-project ImageJ feature. The Analyze particles feature is then used to quantify the MIP endplate area.
      NMJ MIP tortuosity (T): The NMJ tortuosity, which reflects the degree of complexity of the postsynaptic motor endplate including folds and perforations³¹, is calculated based on each MIP using the following formula, where d_Obj(AB) is the distance between A and B along the perimeter of the object, and d_Euc(AB) is the Euclidian distance between A and B (straight line).
      
   7. Set the highest tortuosity value in the wild-type group of each experimental condition to 100%, and normalize all the other values of the experimental condition to this value in order to obtain the relative NMJ tortuosity.
2. Launch ImageJ and custom macro to quantify presynaptic neurofilament accumulation and synaptic vesicle glycoprotein 2 staining.
   NOTE: Neurofilament accumulation (here, NF-M) and/or altered distribution of synaptic vesicles (here, SV2) are markers of abnormal axonal transport and/or impaired vesicle trafficking and were previously observed in NMJs of various SMA mouse models^32,33,34.
   1. Drag and drop the Macro_NMJ_ACCU_Marinelloetal.ijm (custom made, Supplementary Coding File 2) to the ImageJ window; the macro will open in a second window. In this new window, click on Macros > Run Macro.
      NOTE: The macro can process both proprietary file formats and TIFF files. Files must fulfill the criteria indicated in the NOTE below step 4.1.2.
   2. Select the native folder containing the Junction subfolders that have to be analyzed and click on Select.
   3. In the new pop-up menu called Saving Folder, select the storage folder and click on Select.
   4. In the new pop-up menu called Image Type, select the format of the Z-stack acquisitions.
   5. In the Staining Infos pop-up, indicate the presynaptic and postsynaptic label and color and click on OK. For instance, Presynaptic label: SV2 or NF, Presynaptic color: R, Postsynaptic label: BTX, Postsynaptic color: G.
      NOTE: For proprietary file formats, labels and corresponding channels (C1, C2, or C3) have to be indicated.
   6. In the Pixel Size pop-up, indicate XY pixel size 0.072 µm and Z step 0.5 µm (z) and click on OK. The macro will automatically perform the analysis.
      NOTE: This parameter corresponds to the image size 1024 x 1024 pixels (73.7 x 73.7 µm) selected before confocal microscope acquisitions, and it is correlated to the objective and zoom settings. If the proprietary file formats are selected, the macro directly reads pixel size and Z-step (z). The macro will store, in the saving folder, a DataSheet (.csv) of presynaptic and postsynaptic volumes, a multipage TIFF image of the current detection for each (pre- and postsynaptic) labeling. As indicated above, these TIFF files are generated to check the quality of the acquisitions and to ensure that image processing has been performed correctly.
      The macro calculates the volume of axonal neurofilament M staining (NF volume) from the NF-M-F594 channel that colocalizes with α-bungarotoxin-F488 labeling and the volume of NMJ synaptic vesicle glycoprotein 2 staining (SV2 volume) from the SV2-F594 channel that colocalizes with α-BTX-F488 labeling. The NF-M accumulation is quantified by calculating the ratio between NF volume and postsynaptic endplate (α-BTX) volume and NMJ axon terminal occupancy by the ratio of SV2 and α-BTX volumes, as shown below.
      
      

5. Image analysis- STED microscopy

NOTE: Image processing was performed with the offline software of the STED microscope manufacturer.

1. Launch the microscope software.
2. Open the project by clicking on the Open Project button. Select the project file (.lif) and open it. The images are displayed on the screen along with their names.
3. In the Process window: Click on Noise Reduction > Median. At the bottom of the middle window, set Radius to 5.00 and Iteration to 1.00, and untick 3D Filtering.
4. Then select the Open Projects tab on the top left of the window and choose an image.
5. Click on Apply to validate parameters. A new image called "nameofimage_median001" is created.
   NOTE: It is possible to click on プレビュー prior to Apply to monitor the effect of the median filter, which will enhance the image contrast and smooth the line profiles used for quantification.
6. Apply the filter to all images as indicated in steps 5.4-5.5.
7. In the Open Projects tabs, click on the floppy drive icon to save all the projects, including the newly created filtered images.
   NOTE: The next step will be done using the filtered image named "nameofimage_median001".
8. Calculate the distance between AChR stripes
   NOTE: Changes in the morphology of postjunctional folds are often observed in neuromuscular disorders as a sign of NMJ pathology (immaturity or degeneration). The distance (d) between AChR stripes, which are detected by α-bungarotoxin staining, is calculated by generating intensity profiles and quantifying the distance between each maximum intensity peak by drawing a line profile (Supplemental Figure 5).
   1. Using the microscope software, select the Quantify menu on the top of the central window.
   2. Click on the Tools tab in the top left. Select Intensity in the top left panel and click on the Line profile icon. Set Oversampling to 1, and tick Sort Channels.
   3. Click on the Open Projects tabs and select the filtered image to be analyzed.
      NOTE: It is possible to zoom in on the image by scrolling with the computer mouse. The dynamic range of the image can be modified using the bar on the left side next to the displayed image, which facilitates the visualization of the stripes.
   4. Then, click on the Draw Line icon in the top menu of the right window and trace a line crossing perpendicularly several stripes/junctional folds.
      NOTE: The intensity profile is displayed in the central window.
   5. Click on the top of the first peak and move the mouse pointer while keeping the left mouse button pressed until the next maximum peak is reached.
      NOTE: The information is shown in the intensity profile, while the distance between the two peaks is displayed under the chart with "dx" denomination.
   6. Click right on the mouse while being in the image of the right window and select Save ROIs. Open the saved ROIs (Regions of Interest) by clicking on Load ROIs.
   7. Click on the arrow icon in the top left of the right window, click on the ROI and delete it by clicking on the bin icon.
   8. Repeat this operation as many times as needed from different intensity profiles to obtain the predicted number of AChR stripe distances that will represent the global value in the analyzed muscle.
      NOTE: The optimal N value can be calculated in advance based on the estimated difference between groups, α risk, power, and one- or two-tailed test. In the current experimental design, a one-tailed Mann-Whitney test (α risk = 10%; power = 80%) was applied, and the N value was estimated to be at least five AChR stripe distances per NMJ in order to compare the two groups of animals.
9. AChR stripe width
   NOTE: The stripe width (w) corresponds to the full-width half-maximum (FWHM) of the intensity profile, which is the distance between the points where the α-BTX signal fluorescence value is half of its maximum intensity (Supplemental Figure 5).
   1. Using the microscope software, select the Quantify menu in the central window.
   2. Click on the Tools tab on the top left side. Select Intensity in the top left panel and click on the Determine FWHM icon. Tick Sort Channels.
      NOTE: To optimize the peak detection by the software, Set Threshold and 幅 were set to 50 and 3, respectively. Adapt these values for each experiment and ask for advice from an experienced imaging scientist.
   3. Click on the Open projects tabs and select the filtered image to be analyzed.
      NOTE: It is possible to zoom in on the displayed image in the right window by scrolling with the computer mouse. As indicated above (NOTE following step 5.8.3), the dynamic range of the image can be modified for optimal stripe visualization.
   4. Next, click on the Draw Rectangle icon in the top menu of the right window. Select a stripe that is either horizontal or vertical and draw a rectangle perpendicularly to the stripe. A profile appears in the central window.
   5. Click on either Vertical or Horizontal of the Average Projection menu located in the left panel, depending on whether the stripe orientation is vertical or horizontal.
   6. Click on 統計 in the central window and read the FWHM value.
   7. Make a right-click with the computer mouse on the image displayed in the right window and select Save ROIs.
      NOTE: Open the saved ROIs by clicking on Load ROIs.
   8. Click on the arrow icon in the top left of the right window, click on the ROI and delete it by clicking on the bin icon.
   9. Repeat this operation as many times as needed from different rectangle ROIs until obtaining the predicted number of AChR stripe widths, which will be representative of the global value in the analyzed muscle.

6. Experimental design and statistical tests

1. Perform statistical analyses using specific software.
   NOTE: Data were collected from N ≥ 3 biological replicates and at least 20 NMJs per genotype for confocal microscope imaging, and N ≥ 5 biological replicates and N = 5 NMJs per genotype for STED imaging, in each experimental group. Significance was assessed by unpaired Mann-Whitney test (nonparametric), and p-values are indicated in the corresponding figure legends.