Time-Lapse Imaging of Neuronal Arborization using Sparse Adeno-Associated Virus Labeling of Genetically Targeted Retinal Cell Populations

Single-cell labeling and visualization of dendritic or axonal arbors is required to study neuronal morphogenesis. By labeling developing arbors in a minimally-invasive manner, retinal tissue remains healthy and suitable for prolonged confocal live imaging. The main advantage of this method is the ability to visualize individual arbors with multi-color fluorescent proteins within four days post-injection.

This makes studying neonatal arbor morphologies possible. This injection imaging and analysis protocol can be used to study neuronal morphologies across the central nervous system. Appropriate Cre selection and injection location must be determined to target the cell population of interest.

Begin by selecting a Cre mouse line to label the retinal cell populations of interest. Obtain recombinase-dependent AAV virus encoding fluorescent proteins. For optimal labeling of fine processes, select vectors that express modified fluorescent proteins targeting the plasma membrane.

For the AAV aliquot on ice, and prepare a one to four AAV dilution using sterile saline or PBS. To visualize the injections, color the solution blue by adding approximately one microliter of 0.02%Fast Green FCF dye solution for every 15 microliters of AAV dilution. Sterilize the injection area with 70%ethanol.

Backfill the micropipette with the AAV dilution using a microsyringe. Under a stereo microscope, break the micropipette tip with a 30-gauge needle to unseal the tip. After anesthetizing the neonatal mice, tattoo their paw pads with tattoo ink using a 30-gauge needle to identify the animals.

Collect tail clippings for DNA isolation and genotyping. Swab the skin overlying the eyes with 70%ethanol. Use a 30-gauge needle to open the fused eyelid while applying light pressure with the fingers to open the eye.

Poke a small hole through the cornea at the cornea sclera junction. Insert the glass micropipette into the hole and press the micro injector foot pedal two to four times to inject the AAV into the intravitreal space. Slowly remove the micropipette and confirm AAV injection by visualizing the blue dye in the pupil.

After preparing the retinal aCSF as described in the text manuscript, oxygenate the retinal aCSF by bubbling with carbogen for a minimum of 15 minutes, then adjust the pH to 7.4. Embed a 60 millimeter diameter Petri dish into an ice tray and fill it with oxygenated retinal aCSF. Place the filled Petri dish under a stereo microscope.

Next, cut the eyelid flap of the euthanized mouse to expose the eye. Use blunt forceps to enucleate the eyes and transfer them into the cold retinal aCSF placed under the stereo microscope. Under the stereo microscope, stabilize the eye by clasping the optic nerve using Dumont number five forceps and poke a hole in the center of the cornea with a 30 gauge needle, then insert one tip of the micro scissors into the hole to make an incision from the hole to the end of the cornea.

Repeat to make four slices in the cardinal directions, creating four flaps. Grasp and pull the two adjacent flaps apart, gently peeling the sclera from the retina for all the cornea flaps. Then remove the lens from the retinal cup using the forceps.

Finally, use micro scissors to make four radial incisions from the edge of the retina towards the optic nerve, creating four equal petals. If using large diameter gray MCE membrane filters for mounting, cut the disc into quadrants roughly one centimeter across, then place the MCE disc onto the center of a larger white filter paper. Using two size three by zero paint brushes, flip one retinal cup onto a paint brush with the retinal ganglion cell side down.

Gently lift the retina out of the aCSF, making sure the water tension does not tear the retina. While still holding the paint brush with the retina, move the Petri dish containing aCSF out of the way and place the white filter paper with the MCE membrane under the stereoscope. Using a transfer pipette, place a droplet of aCSF in the center of the MCE filter paper.

Float the retina into the aCSF droplet created by the surface tension and charged MCE membrane. Use paint brushes to position the retina within the droplet with the retinal ganglion cell side up, then unfold the four petals. Once positioned, create a water bridge between the paintbrush and white filter paper to break the surface tension of the droplet.

Assemble the live imaging incubation chamber and fill it with oxygenated aCSF. Turn on the pump and temperature controller, making sure that the temperature does not rise above 34 degrees Celsius. It can take up to an hour for the chamber temperature to stabilize.

It is recommended to set up the incubation chamber before beginning retinal dissection. Stable chamber temperature helps reduce sample drift. To transfer the retinal flat mound into the perfusion chamber, stop the pump and remove the aCSF that is in the chamber.

Then place the MCE disc with the retinal flat mount into the incubation chamber. Pre-wet a sample weight to break the surface tension, then place it onto the flat mount ensuring that the sample weight is placed on the MCE paper to reduce sample drift. Refill the chamber with the warmed aCSF and circulate aCSF at approximately one milliliter per minute.

Position the nosepiece with the 25 times water dipping objective into the imaging chamber. Screen for labeled cells of interest using epifluorescent light, and adjust the imaging volume to capture dendritic features of interest. Using a lookup table that identifies both oversaturated and undersaturated pixels, adjust the laser power such that no pixels are oversaturated.

Acquire the 3D imaging volume and repeat acquisition at the desired frame rate. Imaging volume can be adjusted between each time point to compensate for sample drift. Import the image series, splitting the images by time.

Save all time points manually or by running a macro plugin. Create a theoretical point spread function using the ImageJ plugin by clicking Defraction PSF 3D, lens numerical aperture, fluorophore emission wavelength, pixel size, and Z-spacing are required to create an accurate PSF. Select plugins, macros, and run to perform batch parallel iterative deconvolution for all time points using the provided macro.

Import all deconvolved time points as a stack by clicking file, import, then image sequence. Convert the stack into a hyperstack by clicking image, hyperstacks, then stacks to hyperstack. Add the number of Z-slices and timeframes, then correct the 3D drift using the correct 3D drift plugin.

Save the 3D drift corrected hyperstack and convert the image back to a regular stack by clicking image, hyperstacks, hyperstack to stack, then split the time points by clicking image, stacks, tools, and stacks splitter. For the number of substacks to split, enter the number of time points. Use batch processing to create a maximum projection for all time points.

Import the time lapse image sequence by clicking file, import, then image sequence. And use conventional ImageJ tools for the desired analysis of deconvolved and post processed two dimensional video, A high resolution 3D video of developing starburst cell dendrites was acquired, deconvolved, and corrected for 3D drift. Z-plane maximum projections were produced to make 2D videos for analysis, showing an area of dendritic refinement and an area of dendritic outgrowth.

3D deconvolution of each time point increased the resolution of fine filopodia projections of single P6 starburst amacrine cell before and after deconvolution with ImageJ. Prolonged AAV infection periods do not necessarily lead to increased fluorescent signal as shown by similar levels of protein expression at five and 11 days post-injection. Reducing sample drift during imaging is important.

Drift is minimized by maintaining a consistent imaging temperature and by appropriately positioning the weight on the sample. If drift occurs, manual acquisition of the time series allows for adjustment of the imaging volume between frames. This ensures the features of interest remain in frame.

This protocol produces high resolution, deconvolved, and drift-corrected 3D videos of developing neurites. Such videos lead to a better understanding of neuron wiring and are required to advance computational analysis methods of 3D morphogenesis. Better automatic tracing and tracking softwares are required to achieve high throughput analysis of neurite development.