Here, we provide protocols to perform three simple injury-induced axon degeneration (axon death) assays in Drosophila melanogaster to evaluate the morphological and functional preservation of severed axons and their synapses.
Axon degeneration is a shared feature in neurodegenerative disease and when nervous systems are challenged by mechanical or chemical forces. However, our understanding of the molecular mechanisms underlying axon degeneration remains limited. Injury-induced axon degeneration serves as a simple model to study how severed axons execute their own disassembly (axon death). Over recent years, an evolutionarily conserved axon death signaling cascade has been identified from flies to mammals, which is required for the separated axon to degenerate after injury. Conversely, attenuated axon death signaling results in morphological and functional preservation of severed axons and their synapses. Here, we present three simple and recently developed protocols that allow for the observation of axonal morphology, or axonal and synaptic function of severed axons that have been cut-off from the neuronal cell body, in the fruit fly Drosophila. Morphology can be observed in the wing, where a partial injury results in axon death side-by-side of uninjured control axons within the same nerve bundle. Alternatively, axonal morphology can also be observed in the brain, where the whole nerve bundle undergoes axon death triggered by antennal ablation. Functional preservation of severed axons and their synapses can be assessed by a simple optogenetic approach coupled with a post-synaptic grooming behavior. We present examples using a highwire loss-of-function mutation and by over-expressing dnmnat, both capable of delaying axon death for weeks to months. Importantly, these protocols can be used beyond injury; they facilitate the characterization of neuronal maintenance factors, axonal transport, and axonal mitochondria.
The morphological integrity of neurons is essential for sustained nervous system function throughout life. The vast majority of the neuronal volume is taken by axons1,2; thus life-long maintenance of particularly long axons is a major biological and bioenergetic challenge for the nervous system. Multiple axonal-intrinsic and glial-extrinsic support mechanisms have been identified, ensuring life-long axonal survival. Their impairment results in axon degeneration3, which is a common feature of nervous systems being challenged in disease, and by mechanical or chemical forces4,5. However, the underlying molecular mechanisms of axon degeneration remain poorly understood in any context, making the development of efficacious treatments to block axon loss challenging. The development of effective therapies against these neurological conditions is important, as they create an enormous burden in our society6.
Injury-induced axon degeneration serves as a simple model to study how severed axons execute their own disassembly. Discovered by and named after Augustus Waller in 1850, Wallerian degeneration (WD) is an umbrella term that comprises two distinct, molecularly separable processes7. First, after axonal injury, axons separated from their cell bodies actively execute their own self-destruction (axon death) through an evolutionarily conserved axon death signaling cascade within one day after injury8. Second, surrounding glia and specialized phagocytes engage and clear the resulting axonal debris within three to five days. The attenuation of axon death signaling results in severed axons that remain preserved for weeks9,10,11,12, while the attenuation of glial engulfment culminates in axonal debris which persists for weeks in vivo13,14,15.
Research in flies, mice, rats and zebrafish revealed several evolutionarily conserved and essential mediators of axon death signaling8. Axon death mutants contain severed axons and synapses that fail to undergo axon death; they remain morphologically and functionally preserved for weeks, in the absence of cell body support9,10,12,13,16,17,18,19,20,21,22,23. The discovery and characterization of these mediators led to the definition of a molecular pathway executing axon death. Importantly, axon death signaling is activated not only when the axon is cut, crushed or stretched24,25; it also seems to be a contributor in distinct animal models of neurological conditions (e.g., where axons degenerate in an injury-independent manner4, yet with a range of beneficial outcomes4,8). Therefore, understanding how axon death executes axon degeneration after injury might offer insights beyond a simple injury model; it could also provide targets for therapeutic intervention.
The fruit fly Drosophila melanogaster (Drosophila) has proven to be an invaluable system for axon death signaling. Research in the fly revealed four essential evolutionarily conserved axon death genes: highwire (hiw)11,14, dnmnat12,26, dsarm10 and axundead (axed)12. The modification of these mediators — loss-of-function mutations of hiw, dsarm and axed, and over-expression of dnmnat — potently blocks axon death for the life span of the fly. While severed wild type axons undergo axon death within 1 day, severed axons and their synapses lacking hiw, dsarm or axed remain not only morphologically, but also functionally preserved for weeks. Whether functional preservation can also be achieved through high levels of dnmnat remains to be determined.
Here, we will present three simple and recently developed protocols to study axon death (e.g., the morphology and function of severed axons and their synapses over time) in the absence of cell body support. We demonstrate how attenuated axon death results in severed axons which are morphologically preserved with a hiw loss-of-function mutation (hiw∆N) and how attenuated axon death results in severed axons and synapses that remain functionally preserved for at least 7 days with over-expression of dnmnat (dnmnatOE). These protocols allow for the observation of individual axonal and synaptic morphology either in the central, or peripheral nervous system (CNS and PNS, respectively)13,14, while the functional preservation of severed axons and their synapses in the CNS can be visualized by the use of a simple optogenetic setup combined with grooming as a behavioral readout12.
1. Observation of Axon Morphology During Axon Death in the PNS
Figure 1: Observation of axon morphology during axon death in the wing. (A) Schematic fly wing with two sparsely GFP-labeled sensory neurons, which are also separately indicated below. The site of injury and the field of observation are indicated. (B) Schematic setup for wing imaging. Injured and uninjured control wings (grey) are mounted in halocarbon oil 27 (red) on a glass slide (light blue) and covered with a cover slide (black). Please click here to view a larger version of this figure.
2. Observation of Axon and Synapse Morphology During Axon Death in the CNS
Figure 2: Observation of axon and synapse morphology during axon death in the brain. (A) Side view of a schematic fly head with GFP-labeled cell bodies, axons and synapses. (B) High-magnification front view of GPF-labeled olfactory receptor neurons and their axons and synapses. Cell bodies are housed in the 3rd antennal segment, and their axons project into the CNS. Axons form synapses in a glomerulus in the left olfactory lobe, cross the midline and form synapses in the glomerulus on the contralateral olfactory lobe. (C) Examples of fly heads with unilateral antennal ablation. Top: Uninjured control. Middle: Ablation of the 3rd antennal segment. Bottom: Ablation of the 2nd (and thus also 3rd) antennal segment. (D) Brain preparation. Top: Schematic dissected fly brain with indicated olfactory lobes and axonal projections in the field of view. Bottom: Schematic setup for brain imaging. Two clay rolls (green) are mounted onto a glass slide (light blue), they carry a cover slide sandwich, which contains fly brains (grey). Brains are mounted in antifade reagent(purple), surrounded by a lab tape (orange), and covered by two cover slides (black). Please click here to view a larger version of this figure.
3. Grooming Induced by Optogenetics as a Readout for Axon and Synapse Function
Figure 3: Optogenetic setup to induce grooming as a readout for axon and synapse function. (A) Illustration of assembled components required for optogenetics. Infrared (IR) LED spotlight, camera and red LED spotlight (from left to right, respectively). The components including a detailed description are listed in the Table of Materials. (B) Top view illustration of a behavior chamber including an IR emitter to indicate red LED spotlight activation. (C) Illustration of a single mount setup. A total of three mount setups are required for the two LED spotlights and the camera, respectively. Please click here to view a larger version of this figure.
Above, we described three methods to study the morphology and function of severed axons and their synapses. The first method allows for high-resolution observation of individual axons in the PNS. It requires clones generated by the MARCM technique14,31. Here, we performed crosses to generate wild type and highwire mutant MARCM clones (Figure 4A). A simple cut in the middle of the wing induces axon injury of neurons housed distal (e.g., at the outer side of the wing), while proximal neurons (e.g., between the cut site and the thorax) remain uninjured. This approach makes it feasible to observe axon death side-by-side of uninjured control axons in the same nerve bundle (Figure 1A, Figure 4B). Here, we used a genetic background resulting in low numbers of GFP-labeled clones (e.g., two in each experiment14). We present examples of 1 and 7 days after injury of wild type axons, to provide examples of control axons, axons undergoing axon death, and axonal fragments being cleared by surrounding glia, respectively. In addition, we repeated axonal injury in highwire mutants where we analyzed the outcome 7 days after injury.
Uninjured control wings harbor two wild-type clones, thus two GFP-labeled wild-type axons (Figure 4B, wild type, uninjured control). One day after cutting the middle of the wing by the use of micro scissors, axon death is induced in GFP-labeled axons where cell bodies are distal to the cut site, while axons from proximally housed cell bodies serve as an internal control within the same nerve bundle (Figure 4B, wild type, 1 day post injury). Note the axonal debris trace in the upper part indicated by the arrow. 7 days after axonal injury, GFP-labeled axonal debris is cleared by surrounding glia, while GFP labeled uninjured control axons remain in the nerve bundle (Figure 4B, wild type, 7 days post injury, arrow). In contrast, highwire mutant axons that have been severed for 7 days remain morphologically preserved, consistent with previous findings11,14 (Figure 4B, highwire, 7 days post injury, arrow). These results demonstrate the powerful visual resolution of the Drosophila wing. Axon death can be observed side-by-side of uninjured controls in the same nerve bundle. While wild-type axons undergo axon death within 1 day after injury and the resulting debris is cleared within 7 days, axon death deficient highwire mutants remain morphologically preserved for 7 days.
Figure 4: Approach to study axon death of GFP-labeled sensory neuron axons in the wing. (A) Schematic crosses to generate wild type and highwire clones in the wing (P0 and F1 generation, respectively). Virgin females are on the left, males on the right. See Table of Materials for genotype details. (B) Examples of control and injured GFP-labeled axons. The field of view is indicated in (Figure 1A). From left to right: uninjured wild type control axons, wild type axons 1-day post injury, wild type axons 7 days post injury, highwire mutant axons 7 days post injury, respectively. Arrows indicate severed axons, Scale bar = 5 µm. Please click here to view a larger version of this figure.
The second method describes how to visualize whole axon bundles projecting into the CNS where they form synapses, which belong to neurons housed in both left and right antennae (Figure 2A-C). Here, we performed crosses to generate wild type and highwire mutant MARCM clones (Figure 5A). Uninjured, GFP-labeled axons and their synapses can be visualized over the course of days to weeks, in the absence of injury (Figure 5B, Wild type, uninjured control). Alternatively, animals can be subjected to 3rd antennal segment ablation, and severed GFP-labeled axons and their synapses can be observed during a time course over hours to days. We focused on 7 days after antennal ablation, because at this time point, axons and their synapses have undergone axon death, and resulting debris is cleared by surrounding glia. If unilateral ablation of the right antenna is performed, then the right axon bundle is severed and will disassemble and the resulting debris is fully cleared 7 days after injury (Figure 5B, wild type, unilateral ablation, 7 days post injury, arrows), consistent with previous findings13. Alternatively, both the right and the left antennae can be ablated, which will sever both axon bundles, and 7 days after injury, axons and their synapses disappeared (Figure 5B, wild type, bilateral ablation, 7 days post injury, arrow). In contrast, unilateral ablation of the right antennae in highwire mutants results in severed axons that remain preserved 7 days post injury, consistent with previous findings11,14 (Figure 5B, highwire, unilateral ablation, 7 days post injury, arrow). These results demonstrate that severed wild-type axons undergo axon death and the resulting debris is cleared within 7 days, while axon death deficient highwire mutants fail to undergo axon death and remain morphologically preserved for 7 days.
Figure 5: Approach to study axon death of GFP-labeled sensory neuron axons in the brain. (A) Schematic crosses to generate wild type and highwire clones in the brain (P0 and F1 generation, respectively). Virgin females are on the left, males on the right. See Table of Materials for genotype details. (B) Examples of control and injured GFP-labeled axons. From left to right: uninjured wild type controls, wild type 7 days post unilateral antennal ablation, wild type 7 days post bilateral antennal ablation, and highwire mutants 7 days post unilateral antennal ablation, respectively. Arrows indicate severed axon bundles, Scale bar = 10 µm. Please click here to view a larger version of this figure.
The third method allows for the observation of functional preservation of severed axons and their synapses in the CNS. It relies on the manipulation of a subset of JO neurons housed in the 2nd antennal segment which are sufficient to induce antennal grooming. Expression of a red-shifted channelrhodopsin (CsChrimson) in JO neurons, combined with dietary supplementation of all trans-retinal, is sufficient to elicit a simple post-synaptic grooming behavior upon red light exposure12,30. Here, we performed crosses to generate wild type JO neurons, and JO neurons over-expressing dnmnat (dnmnatOE) (Figure 6A). Wild type flies or flies containing JO neurons with attenuated axon death (dnmnatOE), both harbor a potent grooming behavior before injury. However, 7 days post injury (e.g., bilateral ablation of the 2nd antennal segment), grooming fails to be elicited by optogenetics in wild type flies due to injury-induced axon and synapse degeneration, while animals with attenuated axon death continue to groom (Figure 6B, Movie 1,2). Attenuated axon death is therefore capable of functionally preserving severed axons and their synapses for 7 days.
Figure 6: Approach to visualize axonal and synaptic function after axotomy. (A) Schematic crosses to generate wild type and dnmnat over-expressing JO sensory neurons (P0 and F1 generation, respectively). Virgin females are on the left, males on the right. See Table of Materials for genotype details. (B) Individual ethograms of grooming behavior induced by optogenetics. Top: individual ethograms of wild type flies before and 7 days after injury (blue). Bottom: individual ethograms of flies over-expressing dnmnat (dnmnatOE) in JO neurons before and 7 days after injury (red). Each bin indicates at least 1 grooming behavior within 1 s. The black line indicates the sum of all bins. (C) Quantification of grooming behavior. Data is shown as average ± standard deviation, p > 0.001 (one-way ANOVA, multiple comparison with Tukey's post hoc test). Please click here to view a larger version of this figure.
Movie 1: Representative wild type grooming behavior elicited by optogenetics before and 7 days after antennal ablation. Please click here to download this video.
Movie 2: Representative grooming behavior elicited by optogenetics in flies over-expressing dnmnat in JO neurons before and 7 days after antennal ablation. Please click here to download this video.
The protocols described here allow for the robust and reproducible observation of morphology as well as function of axons and their synapses separated from their cell bodies in Drosophila. The wing assay facilitates the observation of axon death side-by-side of uninjured control axons in the PNS14, while the antennal assay facilitates the observation of whole nerve bundles of GFP-labeled axons and their synapses, to assess both morphology and function in the brain (CNS)12. There are critical steps and certain advantages for each approach to study morphology that have to be taken into consideration when designing experiments.
To observe axon morphology in the PNS in the wing, experiments can be readily performed, because of the transparency of the wing: it allows to bypass dissection and immunohistochemistry. However, due to the lack of fixation, the wings have to be imaged immediately after mounting14. Currently, two distinct Gal4 drivers are frequently used, either ok371Gal4 or dpr1Gal4, and both references offer distinct approaches to quantify degeneration14,26. Sparse labeling of a few neurons is recommended, by using "Mosaic Analysis with a Repressible Cell Marker (MARCM)"14,31, as the resolution of axonal morphology is unprecedented. In contrast, the observation of synapses is not possible in wings, they are located in the ventral nerve cord inside the thorax of the flies. Furthermore, additional axonal markers cannot be visualized by immunohistochemistry: the waxy cuticle makes it impossible for the diffusion of fixatives and antibodies into the underlying tissue.
To observe axon and synapse morphology in the CNS, brain dissections have to be performed. They offer the advantage of visualizing additional axonal and synaptic markers by the use of immunohistochemistry, and synapses can be observed alongside axons in the same field of view10,13. A large collection of characterized olfactory receptor neuron (ORN) Gal4 drivers is readily available32, and frequently, OR22aGal4 is the driver of choice. For antennal ablation, cell bodies of OR22a neurons are housed in the 3rd segment (Figure 2B). A fluorescence intensity-based quantification is used to quantify the degeneration of either axons or synapses13. Conversely, experiments are time consuming due to brain dissection and antibody staining.
To visualize axonal and synaptic function after axotomy, optogenetics is used to trigger antennal grooming: it serves as a readout for functional preservation of severed axons and their synapses12. The grooming circuit and corresponding sensory, inter- and motorneuron Gal4 drivers have been thoroughly described29,30. GMR60E02Gal4 labels a subset of Johnston's organ (JO) sensory neurons, which are required and sufficient for grooming29,30. For antennal ablation, cell bodies of JO neurons are housed in the 2nd antennal segment (Figure 2B). An optogenetic setup can readily be built from scratch, or an existing setup adjusted. Importantly, experiments have to be performed in a dark room, and flies thus visualized with an infrared (IR) LED spotlight. When using CsChrimson as a channel, it is crucial to supply the food with all trans-retinal and a red LED spotlight to activate JO neurons29. Alternatively, blue light-sensitive channels and a blue LED spotlight, or the TrpA1 channel and temperature can be used for neuronal activation29,33. The quantification of grooming behavior has already been described12,29.
When these assays are used to specifically study axon death, it is important to note that the phenotype of morphological or functional preservation should be robust over time. There are cases where axon death leads to a consistent yet less pronounced phenotype in morphological preservation34,35, and whether such a phenotype translates into functional preservation remains to be determined.
Axon death phenotypes have also been observed in neurons during development of Drosophila larvae, where nerves were crushed rather than injured11,23. Here, we specifically focused on adult Drosophila neurons which completed development. In this context, the use of RNA interference36, or tissue-specific CRISPR/Cas937 can readily be implemented. Importantly, the above techniques can be used in an axon death independent context: they facilitate the characterization of neuronal maintenance factors38, axonal transport39, age-dependent axonal mitochondria changes40, and morphology of axonal mitochondria41.
The authors have nothing to disclose.
We would like to thank the entire Neukomm lab for contributions. This work was supported by a Swiss National Science Foundation (SNSF) Assistant Professor award (grant 176855), the International Foundation for Research in Paraplegia (IRP, grant P180), SNSF Spark (grant 190919) and by support from the University of Lausanne and the Department of Fundamental Neurosciences (État de Vaud) to LJN.
Tweezers (high precision, ultra fine) | EMS | 78520-5 | Antennal ablation |
MicroPoint Scissors (5-mm cutting edge) | EMS | 72933-04 | Wing injury |
1.5 mL microcentrifuge tube | Eppendorf | 30120086.0000 | |
35mm tissue culture dish | Sarstedt | 83.3900 | |
Cover Slips, Thickness 1 | Thermo Scientific™ | BB02400600A113MNT0 | |
Superfrost Microscope Slides | Thermo Scientific™ | AA00008032E00MNT10 | |
High-Sensitivity USB 2.0 CMOS Camera, 1280 x 1024, Global Shutter | Thorlabs | DCC1240M | Camera setup |
SM1 Retaining Ring for Ø1" Lens Tubes and Mounts | Thorlabs | SM1RR | |
25mm 1/1.2" C mount Lens | Tamron | M112FM25 | |
Adapter with External M27 x 0.5 Threads and Internal SM1 Threads | Thorlabs | SM1A36 | |
Aspheric Condenser Lens, Ø25 mm, f=20.1 mm, NA=0.60, ARC: 650-1050 nm | Thorlabs | ACL2520U-B | |
Ø25.0 mm Premium Longpass Filter, Cut-On Wavelength: 700 nm | Thorlabs | FELH0700 | |
SM1 (1.035"-40) Coupler, External Threads, 0.5" Long | Thorlabs | SM1T2 | |
SM1 Lens Tube Without External Threads, 1" Long, Two Retaining Rings Included | Thorlabs | SM1M10 | |
850 nm, 900 mW (Min) Mounted LED, 1200 mA | Thorlabs | M850L3 | IR LED spotlight |
SM1 (1.035"-40) Coupler, External Threads, 0.5" Long | Thorlabs | SM1T2 | |
SM1 Lens Tube Without External Threads, 2" Long, Two Retaining Rings Included | Thorlabs | SM1M20 | |
Aspheric Condenser Lens, Ø25 mm, f=20.1 mm, NA=0.60, ARC: 650-1050 nm | Thorlabs | ACL2520U-B | |
Ø25.0 mm Premium Longpass Filter, Cut-On Wavelength: 850 nm | Thorlabs | FELH0850 | |
SM1 Retaining Ring for Ø1" Lens Tubes and Mounts | Thorlabs | SM1RR | |
660 nm, 940 mW (Min) Mounted LED, 1200 mA | Thorlabs | M660L4 | Red LED spotlight |
Aspheric Condenser Lens, Ø25 mm, f=20.1 mm, NA=0.60, ARC: 650-1050 nm | Thorlabs | ACL2520U-B | |
SM1 (1.035"-40) Coupler, External Threads, 0.5" Long | Thorlabs | SM1T2 | |
SM1 Lens Tube Without External Threads, 2" Long, Two Retaining Rings Included | Thorlabs | SM1M20 | |
15 V, 2.4 A Power Supply Unit with 3.5 mm Jack Connector for One K- or T-Cube | Thorlabs | KPS101 | LED control |
T-Cube LED Driver, 1200 mA Max Drive Current | Thorlabs | LEDD1B | |
150 mm x 300 mm x 12.7 mm Aluminum Breadboard, M6 Double-Density Taps | Thorlabs | MB1530/M | Mount base |
Ø12.7 mm Universal Post Holder, Spring-Loaded Locking Thumbscrew, L = 75 mm | Thorlabs | UPH75/M | Mount, 3x (IR LED, red LED, cam) |
Ø1.20" Slip Ring for SM1 Lens Tubes and C-Mount Extension Tubes, M4 Tap | Thorlabs | SM1RC/M | |
Ø12.7 mm Optical Post, SS, M4 Setscrew, M6 Tap, L = 150 mm | Thorlabs | TR150/M | |
Ø12.7 mm Optical Post, SS, M4 Setscrew, M6 Tap, L = 40 mm | Thorlabs | TR40/M | |
Right-Angle Clamp for Ø1/2" Posts, 5 mm Hex | Thorlabs | RA90/M | |
M6 x 1.0 Stainless Steel Cap Screw, 16 mm Long, Pack of 25 | Thorlabs | SH6MS16 | screws for mount onto base |
USB-6001 14-Bit 20 kS/s Multifunction I/O and NI-DAQmx | National Instruments | 782604-01 | Red LED spotlight controller |
20k Ohm 1 Gang Linear Panel Mount Potentiometer | TT Electronics/BI | P230-2EC22BR20K | fintuner for indicator |
IR (860nm) emitter, 100 mA radial | Osram | 475-1365-ND | Red light indicator |
cable | – | – | Misc |
All-trans retinal | Sigma | R2625 | |
Ethanol absolute | Vwr | 20821.296 | |
Halocarbon Oil 27 | Sigma | H8773 | |
Mowiol | Merk | 81381 | |
Paraformaldehyde | Sigma | F8775 | |
Phosphate buffered saline (PBS) | Sigma | P5493 | |
Sylgard 184 silicone elastomer base | Dow Corning Corp | 4019862 | |
Sylgard 184 silicone elastomer curing agent | Dow Corning Corp | 4019862 | |
Triton X-100 | Sigma | T8787 | |
Chicken anti-GFP antibodies | Rockland | 600-901-215 | Antibodies |
Goat Dylight anti-Chicken | Abcam | ab96947 | |
FM7a, B | BDSC | RRID:BDSC_785 | X chromosome |
FRT19A[hs-neo] | BDSC | RRID:BDSC_1709 | |
hiw[ΔN] | BDSC | RRID:BDSC_51637 | |
hs-FLP[12] | BDSC | RRID:BDSC_1929 | |
tub-Gal80[LL1] | BDSC | RRID:BDSC_5132 | |
w[1118] | BDSC | RRID:BDSC_3605 | |
20xUAS-IVS-CsChrimson::mVenus | BDSC | RRID:BDSC_55135 | 2nd chromosome |
5xUAS-Gal4[12B] | Kyoto | RRID:Kyoto_108492 | |
5xUAS-HA::dnmnat | BDSC | RRID:BDSC_39702 | |
5xUAS-mCD8::GFP[LL5] | BDSC | RRID:BDSC_5134 | |
ase-FLP[2d] | Freeman laboratory | Neukomm et al., 2014 (PNAS) | |
CyO | BDSC | RRID:BDSC_2555 | |
dpr1-Gal4 | BDSC | RRID:BDSC_25083 | |
OR22a-Gal4 | BDSC | RRID:BDSC_9952 | |
ey-FLP[6] | BDSC | RRID:BDSC_5577 | 3rd chromosome |
GMR60E02-Gal4 | BDSC | RRID:BDSC_39250 | |
TM3,Sb,e | BDSC | RRID:BDSC_3644 |