Two Algorithms for High-throughput and Multi-parametric Quantification of Drosophila Neuromuscular Junction Morphology

The overall goal of this procedure is to automatically quantify morphological features of the Drosophila neuromuscular junction using morphometric software macros. This method can help to identify regulators of synapse development, a key question in neurobiology. The main advantage of this technique is the automatic quantification of multiple NMJ features.

This allows for objective analysis of NMJ morphology in high through put. We decided to develop this methodology to be able to investigate NMJ morphology in a large amount of disease models. We thought of ways how to prevent in their personal differences and how to significantly speed up the quantification process.

Visual demonstration of this method is very helpful. Appropriate macro settings are crucial and some steps might be difficult to new users, especially when not familiar with VT software. For this protocol, generate image stacks of NMJs and save them as individual TIFF files, where channel one shows DLG1 staining, or a similar marker, and channel two shows BRP staining.

To begin, create C projections and hyper stacks of the NMJ image files. Open the plugin options and select Drosophila NMJ Morphometrics. Now, identify the unique file string that the microscope has assigned to the image series when storing them as TIFF.

This will be at the end of the image name. Copy and paste the string given to the lowest plane and channel number into the unique file string setting window. Then, select the sub macro convert to stack and choose the directory or folder where the images are located.

For each image file, two new files are made with the default names stack and flat stack, followed by the original image name. The original files can then be deleted to save storage space. Next, from the Drosophila NMJ Morphometrics interface select the define ROI sub macro, and choose the image file directory.

As the first projection opens select the freehand selections tool. Then use the mouse to define a region containing a complete NMJ terminal of interest. Once selected, click OK in the define terminal window.

Continue doing this until the NMJ terminals are defined in all of the projections. The macro advances the process automatically. For each image file one new file is made with the default names ROI followed by the original image name.

To quantify the NMJ features first go to the Drosophila NMJ Morphometrics interface and set the scale. For example, if one pixel in the image corresponds to 0.72 microns set scale pixels to one and scale distance to 0.072. Then select the sub macro analyze and if there are two channel images, also toggle weight.

Press OK and when prompted select the image file directory. The processing time can be several minutes per synapse. After the analysis, new image files for each analyzed synapse are stored in the parental folder and the quantitative measurements are the results.

txt file. Inspect all the images and exclude pictures with segmentation errors. For example, parts of the synaptic terminal might not be included in the yellow outline.

Parts of the background might be included in the synaptic terminal. A blue skeleton line may extend beyond the synaptic terminal. There may be too many active zones, or some active zones may remain undetected.

If more than 5%of images have segmentation errors explore different analysis algorithms to improve the image processing. The upcoming video sections describe how to define many of these macro analysis settings. To adjust the rolling ball radius value for the macro select three NMJ Z projections that are representative of the image data set.

Delete the result res image name and two active zones stack image name, previously created by the sub macro analyze. Open the stack image name files for each selected image and then from the tool bar select split channels to make an image of channel one and an image of channel two and save these files. Open the image belonging to channel one, which in this case corresponds to DLG one immuno labeling.

Now, run the filter subtract background, found under the process tab. Set the rolling ball radius to a value that increases the contrast between the synapse and background. Create a Z projection with a projection type as max intensity and save the image.

Then run the subtract background algorithm with the new rolling ball radius on all of the Z projections and save the results. To determine the best auto thresholds for the macro, open the saved C projections and run auto threshold with the try all option. From the resulting images, find the most suitable algorithm for the images and proceed using that threshold setting when running the macro for further images.

Defining the different auto thresholds, it’s critical for proper image segmentation by the macro. For this reason, to properly quantify eight of the signage parameters, so it’s important to be familiar with the 16 auto threshold options offered by the software. Press the plus button to zoom into the auto threshold’s result image.

The algorithm names are located underneath each result image. In this example, the best auto threshold algorithm for the setting NMJ outline threshold is Huang. For skeleton threshold, the best setting is Li, and the best setting for the active zone threshold is Huang.

To define the fine maxima noise tolerance value for the macro return to the original representative NMJ images. Open the BRP channel. Go to the plugins tab in the pop up menu and select maximum 3D.

After a short while, a new image will appear, then close the original image. Next, use the minimum 3D command. Then close the maximum of C2 stacks synapse one image and select the newly created image minimum of maximum C2 stack synapse one.

Now use the find maxima command. In the new window select preview point selection and set the noise tolerance to 50. The maxima points are then indicated with little crosses, which should only cover the synapse active zones.

If too many crosses are made, increase the noise tolerance value. If some of the active zones are not annotated decrease the noise tolerance value. Use the derived threshold value for the find maxima noise tolerance field of the macro.

Choosing the maxima noise tolerance value it’s important to properly quantify the number of active zones. Sometimes it’s necessary to try different values to define the proper one. Now adjust all the derived values in the threshold algorithms in the macro interface and run the sub macro analyze on the representative images that were originally used to define the macro settings.

A new file with the active zones indicated by white dots will be created. Open this file by dragging and dropping it into the tool bar and selecting Z project with a projection type as some slices. Thus, a projection file is made.

The next step is to adjust the threshold. In the new threshold window slide the upper bar to choose a threshold value where all the desired foci are read. These are the BRP positive spots.

Use this value as BRP puncta lower threshold. Now, rerun the analysis on the original representative NMJ images using the previously defined values, algorithms, and then new BRP puncta lower threshold value. The resulting images should meet the criteria for critical analysis.

Now use the define settings to run the macro on all NMJ images obtained under the same conditions. The Drosophila NMJ morphometrics macro was used to validate different known synaptic defects in three mutant genotypes. Ankyrin two mutants are known to have fused boutons and smaller NMJs.

Using the macro the area and perimeter of panuronil and quirin two RNAI knock down NMJs were measured and found to be significantly smaller than controls. The GTPase Rab3 is required for proper bruckpila distribution. When disrupted there are fewer active zones.

Using the macro panuronil Rab 3 knock down flies had an average of 138 active zones per NMJ terminal compared to 290 detected in controls. High wire is an important regulator of NMJ growth and when mutated NMJs have extended branching at their terminals. Using the macro panuronil high wire RNAI knock down lines showed significant increases in several skeleton derived parameters at their NMJs, including total length, longest branch length, number of branches, and number of branching points.

After watching this video you should have a good understanding how to operate and adjust the settings of the Drosophila NMJ morphometrics macro. Once mastered the analysis of 50 synapses can be done within one hour. This saves approximately 15 minutes of quantification time per synapse.

It’s important to acquire good quality images of the NMJ. The better the images, the better the macro will perform. This technique will help researchers in the neurobiology field to efficiently quantify morphological parameters of the Drosophila NMJ.