Mitochondria and Caspases Tune Nmnat-Mediated Stabilization to Promote Axon Regeneration

Abstract

Axon injury can lead to several cell survival responses including increased stability and axon regeneration. Using an accessible Drosophila model system, we investigated the regulation of injury responses and their relationship. Axon injury stabilizes the rest of the cell, including the entire dendrite arbor. After axon injury we found mitochondrial fission in dendrites was upregulated, and that reducing fission increased stabilization or neuroprotection (NP). Thus axon injury seems to both turn on NP, but also dampen it by activating mitochondrial fission. We also identified caspases as negative regulators of axon injury-mediated NP, so mitochondrial fission could control NP through caspase activation. In addition to negative regulators of NP, we found that nicotinamide mononucleotide adenylyltransferase (Nmnat) is absolutely required for this type of NP. Increased microtubule dynamics, which has previously been associated with NP, required Nmnat. Indeed Nmnat overexpression was sufficient to induce NP and increase microtubule dynamics in the absence of axon injury. DLK, JNK and fos were also required for NP. Because NP occurs before axon regeneration, and NP seems to be actively downregulated, we tested whether excessive NP might inhibit regeneration. Indeed both Nmnat overexpression and caspase reduction reduced regeneration. In addition, overexpression of fos or JNK extended the timecourse of NP and dampened regeneration in a Nmnat-dependent manner. These data suggest that NP and regeneration are conflicting responses to axon injury, and that therapeutic strategies that boost NP may reduce regeneration.

Fig 1. Reducing mitochondria in dendrites increases axotomy-induced neuroprotection.

(A-A”) A schematic of the axotomy-induced neuroprotection/NP assay is shown. (A) Without pre-axon injury, dendrites degenerate within 18h after injury. (A’) An axon injury 8h prior to dendrite injury induces NP so that dendrite degeneration is delayed; this timeline is the standard one used to assay NP throughout. (A”) When 48h elapses between the axon injury and dendrite severing, very little NP is observed and most dendrites are gone 18h after they are removed. Laser-induced injury is indicated by red lightning bolts. The ddaE neuron is drawn in green and other neurons labeled by 221-Gal4, which was used in most experiments to drive expression, are drawn in grey. (B and C) The NP assay as illustrated in Fig 1A’ was performed in wide-type (yw indicates control neurons that do not express an RNAi hairpin, and Rtnl2 indicates neurons that express a control RNAi hairpin) and Miro RNAi neurons. Neurons were labeled with EB1-GFP under the control of 221-Gal4. In control conditions half the neurons have a dendrite that remains at 18h, and (top row) and half have a fully degenerated dendrite (bottom row in B). Red arrows are the site of axon injury; purple arrows mark the site of dendrite injury. Green lines indicate stabilized dendrites. The scale bar is 20 μm. (D) Quantification of NP is shown. The number of neurons analyzed for each genotype is indicated above the bars. A Fisher’s exact test was used to determine statistical significance. Rtnl2 RNAi was used as a control as it targets a non-essential gene for which we have never observed phenotypes. * p<0.05. (E) Dendrites in ddaE neurons expressing EB1-GFP and control or Miro RNA hairpins were severed without prior axon injury. The presence of intact dendrites (no breaks in continuity) was scored at 4h, 7h and 11h after dendrites were severed. The numbers above the bars are the numbers of cells analyzed; one cell per animal.

Fig 2. Axotomy-induced mitochondrial fission inhibits neuroprotection.

(A and A’) Representative images of dendritic mitochondria in control (A) and Drp1 RNAi (A’) neurons before and 8h post axon injury (hpa) are shown. mito-GFP and mCD8-RFP were coexpressed in ddaE neurons under the control of 221-Gal4 in order to visualize mitochondria and the cell membrane, respectively. Orange, blue and magenta arrows indicate long (>1.5 μm), medium (1–1.5 μm) and short (<1 μm) mitochondria respectively. Scale bars are 10 μm. (B-D’) The length, length distribution, and total number of mitochondria in control (B-D) and Drp1 RNAi neurons (B’-D’) before and after axon injury were measured. Statistical significance was determined using a Fisher’s exact test (C and C’), or a t test (B, B’, D and D’). * p<0.05, ** p<0.01, *** p<0.001, N.S. not significant. Error bars represent SD. (B, B’, D and D’) The numbers of neurons analyzed are indicated above the bars. (C) 227 and 229 mitochondria from 9 neurons were analyzed for uninjured and 8hpa, respectively. (C’) 173 and 169 mitochondria from 8 neurons were analyzed for uninjured and 8hpa, respectively. (E) The NP assay was performed in Drp1 RNAi neurons labeled with EB1-GFP. Red arrows indicate the site of axon injury and purple arrows, dendrite injury. Green lines mark stabilized dendrites. Scale bar, 20 μm. (F) Quantification of NP is shown with control data from Fig 1D for comparison. * p<0.05 and **p<0.01, determined by Fisher’s exact test. The numbers of neurons are indicated above the bars. For the RNAi experiment, Rtnl2 data is shown as the matched control and yw (no RNAi hairpin) control data was used for the overexpression comparison. (G) Dendrite injury was performed in Drp1 RNAi neurons without pre-axotomy. Dendrite degeneration assayed at the indicated times. Control data from Fig 1E is included for comparison. The numbers of cells analyzed for each condition are shown above the bars.

Fig 3. Caspases inhibit axotomy-induced neuroprotection.

(A) Left, images of control and Dronc RNAi neurons in which dendrites were severed without axon pre-cut are shown. Neurons were labeled with EB1-GFP. Purple arrows indicate the site of dendrite injury. The scale bar is 20 μm. Right, presence of intact dendrites was scored at different time points. Control data from Fig 1E is included for comparison. The numbers of neurons analyzed for each condition are shown above the bars. (B) The NP assay was performed in Dronc RNAi neurons labeled with EB1-GFP. Red arrows indicate site of axon injury; purple arrows, dendrite injury; green lines, stabilized dendrites. The scale bar is 20 μm. (C) Quantification of NP is shown. The numbers on the bars indicate the numbers of neurons analyzed. Control data (Rtnl2 for RNAi and yw for other genotypes) from Figs 1D and 2F is shown for comparison. A Fisher’s exact test was used to determine statistical significance. * p<0.05. N.S. not significant. (D) Left, images of mitochondria in the dendrites of Dronc RNAi neurons before and 8h post axon injury are shown. Orange, blue and magenta arrows indicate long, medium and short mitochondria respectively. The scale bar is 10 μm. Right, quantification of mito-GFP length is shown. The numbers of neurons analyzed are indicated above the bars. ** p<0.01, determined with a t test. Error bars are SD.

Fig 4. Nmnat is required for stabilization of dendrites in response to axotomy.

(A) Left, dendrites of Nmnat RNAi neurons were severed without axon pre-cut. Neurons were labeled with EB1-GFP. The scale bar is 20 μm. A purple arrow marks the site of dendrite injury. Right, quantification of dendrite degeneration at various time points is shown with control data from Fig 1E for comparison. * The numbers above the bars are the number of cells analyzed for each condition. (B) Left, the NP assay was performed in Nmnat RNAi neurons labeled with EB1-GFP. A red arrows shows the site of axon injury, a purple arrow, dendrite injury. The scale bar is 20 μm. Right, quantification of NP is shown, with control data from Figs 1D and 2F. The numbers of neurons analyzed are indicated on the bars. ** p<0.01, determined by Fisher’s exact test. (C) The NP assay was performed in neurons expressing Dronc RNAi in conjunction with a control RNAi or Nmnat RNAi in cells labeled with EB1-GFP. Images of neurons 18hpd are shown. The green line indicates a stabilized dendrite. The scale bar is 20 μm. Right, quantification of NP is shown. The numbers of neurons analyzed are indicated above the bars. A Fisher’s exact test was used to determine statistical significance. ** p<0.01, *** p<0.001. (D) Left, kymographs of EB1-GFP in the dendrites of control and Nmnat RNAi neurons before and 8h after axon injury are shown; the cell body is off to the right in each image. The trajectory of EB1 comets appears white. The X- and Y- axes represent distance and time, respectively. Right, microtubule dynamics is quantified. The numbers of neurons analyzed are indicated above the bars. *** p<0.001, determined by unpaired t test. Error bars are SD. (E) A proposed model that summarizes the finding so far is shown. Steps that cannot be definitely resolved with the data are indicated by question marks.

Fig 5. Nmnat is sufficient to delay dendrite degeneration and increase microtubule dynamics.

(A) Dendrites were severed in neurons expressing GFP-Nmnat-B-delta N without prior axotomy. An example of a cell immediately after dendrite injury and then 18h later is shown with the injury site indicated by a purple arrow and persistent dendrite with a green line. The scale bar is 20 μm. Quantitation of dendrite degeneration is shown at the right. In control (yw) and Nmnat overexpressing neurons, the number of intact dendrites was scored 18h after dendrites were severed. A Fisher’s exact test was used to calculate significance with *** indicating p<0.001. The number of cells analyzed for each genotype is shown above the bars. (B) Movies of EB1-GFP were acquired in the trunk of the ddaE comb dendrite. Neurons expressed a control protein, Kaede, GFP-Nmnat-B-deltaN or Wlds. Kymographs from a portion of the dendrite are shown with the cell body to the right. Quantitation of comet number in the different genetic backgrounds is shown at the right. The central line shows the mean and the error bars are the SD. Numbers of cells analyzed are shown above the plots. Significance was calculated with an unpaired t test and * indicates p<0.05.

Fig 6. Excess Nmnat suppresses axon regeneration.

(A) Images of neurons of the indicated genotypes 0h and 96h after proximal axotomy are shown. The neurons are labeled with EB1-GFP. Red arrows indicate the site of axon injury. A green star marks the tip of the dendrite that converts to an axon after injury; all neurites that grew more than 50 microns have a star. Green lines mark cut off axons that have not degenerated. Scale bars are 20 μm. The average amount of tip growth (growth beyond the length increase due to growth of the whole animal) of the regenerating axon 96h after injury is shown in the graph. The numbers of neurons analyzed are indicated above the bars. Error bars, SD. See methods for statistics. Error bars represent SD. * p<0.05, ** p<0.01. (B) Axon regeneration of ddaE neurons co-expressing Dronc RNAi with a control RNAi or Nmnat RNAi was assayed. Green stars mark tips of the converted dendrites. The scale bar is 20 μm. Statistical significance was determined with an unpaired t test. * p<0.05. Error bar show SD. The numbers of neurons analyzed are indicated above the bars.

Fig 7. DLK/fos signaling coordinates neuroprotection and axon regeneration.

(A) Left, the NP assay was performed in fosDN-expressing neurons labeled with EB1-GFP. Red arrows point to the site of axon injury, purple arrows, dendrite injury. The scale bar is 50 μm. Right, quantification of NP is shown. The numbers of neurons analyzed are indicated above the bars, and control data is from Fig 2F. Statistical significance was determined with a Fisher’s exact test. * p<0.05. (B) Left, kymographs of EB1-GFP comets in fosDN-expressing neurons before and 8h after axon injury are shown. The graph at the right shows the quantification of microtubule dynamics in the ddaE comb dendrite 8h after axon injury. The numbers of neurons analyzed are indicated above the bars; control data is from Fig 4D. ** p<0.01, determined with an unpaired t test. Error bars are SD. (C) Left, mitochondria were imaged in the dendrites of fosDN-expressing neurons before and 8h post axon injury. mito-GFP (mitochondria) and mCD8-RFP (cell membrane) were expressed under the control of 221-Gal4. Orange and blue arrows indicate long (>1.5 μm) and medium (1–1.5 μm) mitochondria, respectively. The scale bar is 10 μm. Right, quantification of the average length of mitochondria in fosDN neurons is shown. Control data from Fig 2B is included for comparison. The numbers of neurons analyzed are indicated above the bars. A t test was used to test for significance. Error bars are SD. (D) Left, the NP assay was performed in WT and fos-overexpressing (fosWT) neurons as in (A), except that 48h elapsed between axon injury and dendrite injury rather than the 8h used elsewhere. Red arrows indicate the site of axon injury and purple arrows, dendrite injury. The green line marks a stabilized dendrite. The scale bar is 50 μm. Right, quantification of NP with the 48h gap between axon and dendrite injury is shown. The numbers of neurons analyzed are indicated above the bars. ** p<0.01, determined by Fisher’s exact test. (E) Left, images of neurons that co-express fosWT with a control RNAi or Nmnat RNAi at 96h post axon injury are shown. The green star marks the tip of the converted dendrite. The scale bar is 20 μm. Right, the average amount of regeneration was quantified and shown in the graph. The numbers of neurons analyzed were indicated above the bars. * p<0.05, determined by an unpaired t test. Error bars are SD.

Fig 8. Effects of JNK (bsk) overexpression on NP and axon regeneration, and a summary model.

(A) A 48h NP assay was performed in bsk-overexpressing neurons. The green line indicates a stabilized dendrite. Statistical significance was determined by a Fisher’s exact test and the numbers above the bars indicate numbers of neurons analyzed. *** p<0.001. Control data is from Fig 7D. (B) Axon regeneration assays were performed in neurons overexpressing bsk paired with either mCD8-RFP or Nmnat RNAi. UAS-mCD8-RFP was expressed as a control for Nmnat RNAi to keep the number of UAS-controlled transgenes constant and rule out Gal4 dilution effects. Red arrows mark sites of axon injury. The green star indicates the tip of the converted dendrite. Statistical significance was determined by a Mann-Whitney test. Error bars represent SD. *** p<0.001. (C) A summary model of the results is shown. Axon injury activates the DLK/bsk/fos response pathway. The AP-1 transcription factor fos turns on early injury responses that include Nmnat-mediated NP (indicated by darker cell outline in middle image), microtubule dynamics (short green lines in middle image) and mitochondrial fission. NP is mediated by Nmnat, which, if unchecked dampens subsequent regeneration. Caspases and mitochondrial fission counteract NP. (D) Reduction of caspases, increased bsk or fos, or increased Nmnat result in excess or longer than normal NP. Unbalanced NP dampens regeneration.