A small molecule M1 promotes optic nerve regeneration to restore target-specific neural activity and visual function

Abstract

Axon regeneration is an energy-demanding process that requires active mitochondrial transport. In contrast to the central nervous system (CNS), axonal mitochondrial transport in regenerating axons of the peripheral nervous system (PNS) increases within hours and sustains for weeks after injury. Yet, little is known about targeting mitochondria in nervous system repair. Here, we report the induction of sustained axon regeneration, neural activities in the superior colliculus (SC), and visual function recovery after optic nerve crush (ONC) by M1, a small molecule that promotes mitochondrial fusion and transport. We demonstrated that M1 enhanced mitochondrial dynamics in cultured neurons and accelerated in vivo axon regeneration in the PNS. Ex vivo time-lapse imaging and kymograph analysis showed that M1 greatly increased mitochondrial length, axonal mitochondrial motility, and transport velocity in peripheral axons of the sciatic nerves. Following ONC, M1 increased the number of axons regenerating through the optic chiasm into multiple subcortical areas and promoted the recovery of local field potentials in the SC after optogenetic stimulation of retinal ganglion cells, resulting in complete recovery of the pupillary light reflex, and restoration of the response to looming visual stimuli was detected. M1 increased the gene expression of mitochondrial fusion proteins and major axonal transport machinery in both the PNS and CNS neurons without inducing inflammatory responses. The knockdown of two key mitochondrial genes, Opa1 or Mfn2, abolished the growth-promoting effects of M1 after ONC, suggesting that maintaining a highly dynamic mitochondrial population in axons is required for successful CNS axon regeneration.

Keywords: axon regeneration; mitochondrial dynamics; optic nerve crush; peripheral nerve injury; visual function recovery.

Fig. 1.

M1 increases axonal mitochondrial length and expression of key mitochondrial fusion proteins in cultured adult DRG neurons. (A) Representative fluorescence micrographs of distal axonal segments of DRG neurons showing that there was a marked increase in mitochondrial clustering and size after M1 treatment. Yellow arrowheads indicate individual mitochondria; white arrowheads indicate clustered mitochondria. The graph depicts mitochondrial length; there was a significant increase in mitochondrial size in M1-treated (2.5 µM) neurons compared with vehicle-treated controls (0.1% dimethyl sulphoxide; DMSO). A total of 3,291 (vehicle treatment) and 2,988 (M1 treatment) mitochondria from three independent cell cultures were evaluated. Each dot represents one mitochondrion. Data points are shown in gray; the median is indicated as a black line. (B) M1 (2.5 µM) induced a significant increase in the cumulative frequency of larger mitochondria in DRG neurons. (C and D) DRG neurons were costained with phalloidin to detect filopodia (red) and with βIII-tubulin (green) to identify axons. In vehicle-treated DRG neurons, OPA1 (blue; C) or MFN2 (blue; D) was evenly distributed along the axons; however, both proteins were heavily localized to the growth cones (filopodia) of regenerating axons in neurons treated with M1 (2.5 µM). White arrowheads indicate OPA1 or MFN2 localization in filopodia. Photomicrographs of individual fluorescence filters are shown in SI Appendix, Fig. S1. (Scale bars: 10 µm in A, C, and D.) (E) The growth cone area was outlined and defined as a region of interest (ROI). OPA1 and MFN2 fluorescence intensity (arbitrary units; A.U.) was measured within the ROI using ImageJ software. OPA1 and MFN2 immunoreactivity was markedly increased in the growth cones of M1-treated (2.5 µM) DRG neurons. Each dot represents one growth cone. (F) The mRNA expression levels of both Opa1 and Mfn2 were markedly increased compared with vehicle-treated controls (0.1% DMSO). Each dot represents the mRNA expression from one independent cell culture experiment. (G) M1 treatment up-regulated OPA1 and MFN2 protein expression in the mitochondrial fraction of DRG neurons. OPA1 and MFN2 expression in mitochondria was normalized to that of cytochrome c oxidase subunit IV (COX IV). Each dot represents the protein expression from one independent cell culture experiment. Adult DRG neurons were prepared from C57BL/6 mice (8 to 12 wk). Data are presented as means ± SEM from six to eight independent experiments in A and B and as means ± SEM of triplicates in E–G. The Mann–Whitney U test (A), the two-sample Kolmogorov–Smirnov test (B), and the Student’s t test (E–G) were used. *P < 0.05; **P < 0.01; ***P < 0.001.

Fig. 2.

M1 promotes axonal mitochondrial transport and axon regeneration in cultured adult DRG neurons. (A) Representative kymographs of vehicle- (0.1% DMSO) or M1-treated (2.5 µM) adult DRG neurons showing the movement of individual mitochondria (white lines) recorded at one frame per 30 s for 2.5 h. Mitochondrial trafficking was faster in M1-treated than vehicle-treated DRG neurons. Vertical lines represent stationary mitochondria; oblique lines indicate motile mitochondria. A total of 1,198 (vehicle treatment) and 906 (M1 treatment) mitochondrial events were included in the kymograph analysis. (Vertical scale bar: 20 min; horizontal scale bar: 5 μm.) (B) M1 treatment increased the motile pool of mitochondria significantly when compared with vehicle control. (C) M1 treatment did not alter the mitochondrial density in the 50-µm distal-most axonal segments. (D) The moving frequency (indicated by the time each mitochondrion spent in motion) was markedly elevated in M1-treated DRG neurons when compared with vehicle-treated controls. (E) The average mitochondrial velocity was increased dramatically after M1 treatment and calculated from the motile pool of mitochondria only. (F) M1 treatment up-regulated the mRNA expression of major axonal transport machinery genes (Kif5a, Miro1, Milton, Dync1h1, and Dctn1) in DRG neurons. Each dot represents the mRNA expression from one independent cell culture experiment. (G) Adult DRG neurons were treated with M1 at various concentrations (0.5 to 10 µM), with DMSO (0.1%) serving as the vehicle control. After 17 h of incubation, the neurons were fixed and immunostained with anti–βIII-tubulin antibody for the neurite outgrowth assay. (H) Magnified view of the red boxes in G. (Scale bars: 500 µm in G and H.) (I) Compared with vehicle-treated controls, the administration of M1 at the 2.5 µM concentration increased maximal neurite outgrowth. Each dot represents the average total neurite outgrowth from one independent cell culture experiment. Adult DRG neurons were prepared from C57BL/6 mice (8 to 12 wk). Data are presented as means ± SEM from six to eight independent experiments in B–E and as means ± SEM of triplicates in F and I. The Mann–Whitney U test (B–E), the Student’s t test (F), and one-way ANOVA with Bonferroni’s post hoc test (I) were used. n.s., not significant. *P < 0.05; ***P < 0.001.

Fig. 3.

Peripheral axon regeneration and axonal mitochondrial dynamics are enhanced substantially in M1-treated mice. (A) Adult C57BL/6 mice were subjected to SNC, following which 20 µg of M1 was applied to the lesion site immediately after the crush for 3 consecutive days. (B) M1 treatment markedly increased the distal extent of axonal regrowth 3 d after injury as determined by the sciatic nerve pinch test (n = 12 for vehicle treatment; n = 13 for M1 treatment). Each dot represents the length of axonal regrowth from one mouse. (C) The extent of sensory axonal regrowth was assessed by immunostaining of SCG10 3 d after SNC. SCG10 immunoreactivity was prominent even in the far-distal region of the crushed nerve in the M1-treated mice (red arrowheads). The red dotted line indicates the crush site. (Scale bar: 500 µm.) (D) The SCG10 fluorescence intensity was measured in the longitudinal section of the sciatic nerve and normalized to the crush site referred to as the regeneration index (n = 4 per treatment group). Each dot represents the normalized SCG10 fluorescence intensity from one mouse. (E) Ex vivo time-lapse imaging of AAV2/9-mitoGFP–labeled mitochondria in sciatic nerves. M1 (20 µg) was applied directly to the sciatic notch, and time-lapse images were captured for kymograph and mitochondrial length analyses 1 d after M1 administration. (F) Representative photomicrographs of mitoGFP-labeled mitochondria in sciatic nerves. M1 induced a 22.4% increase in mitochondrial length in the sciatic nerves. A total of 2,390 (vehicle treatment) and 2,053 (M1 treatment) mitochondria from six sciatic nerves per treatment group and three independent experiments were evaluated. (Scale bar: 5 µm.) (G) There was a significant increase in the cumulative frequency of larger mitochondria 24 h after M1 treatment. (H) Representative kymographs of mitochondrial movement in sciatic nerves of vehicle- and M1-treated mice. The movement of individual mitochondria (white lines) was recorded at one frame per 5 s for 10 min. Vertical lines represent stationary mitochondria, and oblique lines indicate motile mitochondria. M1 markedly accelerated mitochondrial transport in the peripheral axons of sciatic nerves. A total of 1,115 (vehicle treatment) and 1,487 (M1 treatment) mitochondrial events from six sciatic nerves and three independent experiments were included for kymograph analysis. (Vertical scale bar: 200 s; horizontal scale bar: 5 μm.) (I) M1 treatment led to a marked increase in motile pool of mitochondria. (J) M1 did not affect the mitochondrial density in the 100-µm-long axonal segments in sciatic nerves. (K) The moving frequency (indicated by the time each mitochondrion spent in motion) was significantly increased in M1-treated sciatic nerves. (L) M1 treatment promoted the average mitochondrial velocity significantly when compared with vehicle controls. Average mitochondrial velocity was calculated from the motile pool of mitochondria only. Data are presented as means ± SEM. The Student’s t test (B), two-way ANOVA followed by Bonferroni’s post hoc test (D), the Mann–Whitney U test (F and I–L), and the two-sample Kolmogorov–Smirnov test (G) were used. n.s., not significant. *P < 0.05; **P < 0.01; ***P < 0.001.

Fig. 4.

Intravitreal injection of M1 induces robust axon regeneration 2 wk after ONC. (A) Schematic diagram illustrating the M1 (1-µg) treatment paradigm for ONC. Pten was knocked down by the intravitreal injection of an AAV2 vector expressing Cre recombinase (AAV2-Cre) in Pten^fl/fl mice 2 wk before ONC. AAV2-eGFP served as the control. M1 (1 µg) or vehicle was intravitreally injected into the injured eyes at days 0 and 7 after ONC. Axon regeneration was analyzed by CTB-555 tracing. (B and C) Compared with that in Pten^fl/fl controls, mice treated with M1 displayed long-distance axon regeneration 14 d after ONC, with a significant number of regenerating axons reaching 2 mm from the crush site. Pten deletion further increased M1-induced axon regeneration. Red dotted lines indicate the crush sites. Each dot represents the average number of regenerating axons from one mouse. Magnified views of the yellow boxes in B, Left are shown in B, Right. (D) Over 32% of RGC survival was observed in Pten^−/−, M1 treatment, and Pten^−/− + M1 treatment; however, only ∼20% of RGCs survived in Pten^fl/fl control mice 2 wk after ONC injury. Each dot represents the average number of RBPMS-positive RGCs from one mouse. Data are presented as means ± SEM (n = 5 mice for the vehicle-treated group; n = 6 mice for the Pten^−/−- and Pten^−/− + M1–treated groups; n = 7 mice for the M1-treated group). One-way ANOVA followed by Bonferroni’s post hoc test was used in C and D. *P < 0.05. (Scale bars: 200 µm in B; 20 µm in D.)

Fig. 5.

M1 induces sustained axon regeneration that reaches the optic chiasm 4 wk after ONC. (A) Schematic diagram illustrating the M1 (1-µg) treatment paradigm for ONC. Pten was knocked down by the intravitreal injection of an AAV2 vector expressing Cre recombinase (AAV2-Cre) in Pten^fl/fl mice 2 wk before ONC. AAV2-eGFP served as the control. M1 (1 µg) or vehicle was intravitreally injected into the injured eye at days 0, 7, 14, and 21 after ONC. Axon regeneration was analyzed by CTB tracing (CTB-555). (B and C) Sustained long-distance axon regeneration was observed in M1-treated mice 4 wk after ONC; some CTB-555–labeled axons regenerated along the entire length of the optic nerve to reach the optic chiasm in M1-treated mice (yellow boxes in B). The number of regenerating axons in the combination treatment group (Pten^−/− + M1) was comparable with that of M1-treated mice, indicating that M1 alone is sufficient to induce sustained axon regeneration. Red dotted lines indicate the crush sites. Each dot represents the average number of regenerating axons from one mouse. Magnified views of the yellow boxes in B, Left are shown in B, Right. (D) M1 treatment, Pten deletion, or a combination of Pten deletion and M1 administration led to a similar significant increase in RGC survival. M1 treatment did not further enhance RGC survival in Pten^−/− mice. Each dot represents the average number of RBPMS-positive RGCs from one mouse. Data are presented as means ± SEM (n = 5 mice for the vehicle-treated group; n = 6 mice for the Pten^−/−-, M1-, and Pten^−/− + M1–treated groups). One-way ANOVA followed by Bonferroni’s post hoc test was used in C and D. *P < 0.05. (Scale bars: 200 µm in B; 20 µm in D.)

Fig. 6.

Regenerating RGC axons reinnervate multiple subcortical visual targets 6 wk after M1 treatment. (A) Schematic diagram illustrating the M1 (1-µg) treatment paradigm for ONC. Pten was knocked down by the intravitreal injection of an AAV2 vector expressing Cre recombinase (AAV2-Cre) in Pten^fl/fl mice 2 wk before ONC. AAV2-eGFP served as the control. M1 (1 µg) or vehicle was intravitreally injected into the injured eyes at days 0, 7, 14, and 21 after ONC. Axon regeneration was analyzed by CTB tracing (CTB-555). (B–G) Representative confocal micrographs of coronal brain sections showed that a considerable number of CTB-555–positive regenerating RGC axons (red arrowheads) was observed in the (B) hypothalamic SCN, (C) optic tract (OT), (D) thalamic vLGN, (E) dLGN, (F) OPN, and (G) SC of both M1- and Pten^−/− + M1–treated mice at 6 wk post-ONC. (Scale bars: 50 µm.) (H) Quantification of the area fraction of CTB-positive axons revealed that both M1-treated and Pten^−/− + M1–treated mice exhibited a significant number of regenerating axons that reinnervated into different subcortical visual targets at 6 wk post-ONC. Each dot represents the area fraction of CTB-positive axons from one mouse. Data are presented as means ± SEM (n = 3 mice for the vehicle-treated group; n = 4 mice for the M1- and Pten^−/− + M1–treated groups). One-way ANOVA followed by Bonferroni’s post hoc test was used. *P < 0.05; **P < 0.01.

Fig. 7.

Long-distance regenerated axons elicit neural activity in target brain regions to restore the PLR and responses to looming visual stimuli. (A) Schematic diagram illustrating the M1 (1-µg) treatment paradigm for ONC and LFP recording in the SC upon optogenetic stimulation of RGCs. M1 (1 µg) or vehicle was intravitreally injected into the injured eye at days 0, 7, 14, and 21 after ONC. Adult mice were administered AAV2-ChR2-mCherry by intravitreal injection at day 28 post-ONC. LFPs were evoked in the SC by the focal laser stimulation of ChR2-expressing RGCs at day 42 post-ONC. (B) Eye-evoked LFP amplitudes were determined using Spike2 software and a customized MATLAB program. M1 administration led to a marked increase in eye-evoked LFP amplitudes compared with vehicle-treated controls. The combination of Pten^−/− + M1 treatment further augmented the M1-induced increase in LFP amplitudes after the optogenetic stimulation of injured RGCs; in contrast, only minimal eye-evoked LFPs were recorded from the SCs of vehicle-treated mice 6 wk after ONC (n = 4 mice for the uninjured and vehicle-treated groups; n = 8 mice for M1- and Pten^−/− + M1–treated groups). Each dot represents the average eye-evoked LFP amplitude from one mouse. (C) Representative eye-evoked LFPs for all the treatment groups. Red lines indicate optogenetic stimulation in the injured eye. (D) Schematic diagram illustrating the PLR test. Mice were dark adapted for 1 h before the PLR test. The relative pupil constriction was calculated as the percentage change in pupil area between the baseline reading after 1 h of dark adaptation and the reading obtained after 30 s of light stimulation (30 lx) at 470 nm. (E) Representative images of the PLR from vehicle-treated control, M1-treated, and Pten^−/− + M1–treated mice. (F) The vehicle-treated pupils of control mice failed to fully constrict upon light stimulus. Pupil constriction was restored to baseline levels in both M1- and Pten^−/− + M1–treated mice (n = 4 mice for the uninjured group; n = 6 mice for the vehicle-, M1-, and Pten^−/− + M1–treated groups). Each dot represents the pupil constriction from one mouse. (G) Uninjured mice with normal vision immediately froze and ran under the shelter following looming visual stimulation. In contrast, none of the vehicle-treated lesioned mice responded to the looming stimulus. Half of the M1-treated mice responded to the looming stimulus by hiding in the shelter (n = 6 mice per treatment group). Each dot represents the looming response from one mouse. Data are presented as means ± SEM. One-way ANOVA with Bonferroni’s post hoc test was used. n.s., not significant. *P < 0.05.

Fig. 8.

The knockdown of genes encoding mitochondrial fusion proteins abolished M1-induced optic nerve regeneration. (A) In vivo AAV2-mediated silencing of the Opa1 or Mfn2 genes in retinae of M1-treated mice by the intravitreal injection of AAV2 vectors expressing shRNAs 2 wk before ONC. AAV2-scramble (AAV2-scr-shRNA) was used as a control. M1 (1 µg) or vehicle was intravitreally injected into the injured eye at days 0 and 7 after ONC. Three days before tissue harvesting, mice were intravitreally injected with 2 µg of CTB-555 to trace regenerating axons. (B and C) M1 treatment induced robust axon regeneration 14 d after ONC; however, this effect was abolished following the knockdown of either Opa1 or Mfn2, suggesting that the regeneration-promoting effects of M1 in the optic nerve were mediated through the activation of mitochondrial fusion. Red dotted lines indicates the crush sites. Each dot represents the average number of regenerating axons from one mouse. (D) The protective effect of M1 on RGC survival was also abolished after Mfn2 or Opa1 knockdown. Each dot represents the average number of RBPMS-positive RGCs from one mouse. Data are presented as means ± SEM (n = 5 per group). One-way ANOVA followed by Bonferroni’s post hoc test was used in C and D. *P < 0.05 compared with AAV2-scr-shRNA + vehicle controls; ^#P < 0.05 compared with the AAV2-scr-shRNA + M1 treatment group. (Scale bars: 200 µm in B; 20 µm in D.)