In situ structural mechanism of epothilone-B-induced CNS axon regeneration

Abstract

Axons in the adult central nervous system (CNS) do not regenerate following injury, in contrast to neurons in the peripheral nervous system and neuronal growth during embryonic development. The molecular mechanisms that prevent regeneration of neurons in the CNS remain largely unknown^1,2. Here, to address the intracellular response to injury, we developed an in situ cryo-electron tomography and cryo-electron microscopy platform to mimic axonal damage and present the structural mechanism underlying thalamic axon regeneration induced by the drug epothilone B. We observed that stabilized microtubules extend beyond the injury site, generating membrane tension and driving membrane expansion. Cryo-electron microscopy reveals the in situ structure of microtubules at 3.19 Å resolution, which engage epothilone B within the microtubule lattice at the regenerating front. During repair, tubulin clusters are delivered and incorporated into polymerizing microtubules at the regenerating site. These microtubule shoots serve as scaffolds for various types of vesicles and endoplasmic reticulum, facilitating the supply of materials necessary for axon repair until membrane tension normalizes. We demonstrate the unexpected ability of neuronal cells to adjust to strain induced by epothilone B, which creates homeostatic imbalances and activates axons to regeneration mode.

Fig. 1. Integrative platform for axotomy and visualization of EpoB-mediated axon regeneration.

a, The experimental pipeline developed in this study. (1) Thalamus tissue was dissected from embryonic day 15.5 (E15.5) embryos and cultured on electron microscopy (EM) grids (top) and glass-bottom dishes (bottom). (2) Explants extended axons onto EM grids (top; scale bar, 200 µm) or glass-bottom dishes (bottom; scale bar, 200 µm), showing axon-dominant neurite extensions (middle; scale bar, 100 µm). (3) Upon maturation, a thin needle was used to cut the targeted axon (scale bar, 50 µm); the overall procedure is shown in Supplementary Video 1. (4) Cryo-EM or cryo-ET (top; scale bar, 200 nm) and light microscopy (bottom; scale bar, 100 µm) analyses were performed. b, Explant growth on electron microscopy grids. Explant images were acquired at different time points using a 10× objective and the longest neurite length from the explant centre was measured. n = 20 (DIV 1), n = 16 (DIV 2), n = 15 (DIV 3), n = 14 (DIV 4), n = 15 (DIV 5), n = 26 (DIV 6), n = 14 (DIV 7) and n = 12 (DIV 8). Data are mean ± s.e.m. c, Axon regeneration in the presence of EpoB recorded by differential interference contrast microscopy, showing the neurite tip at axotomy (Ax, t = 0 min, asterisk). Scale bars, 20 µm. Video of regenerating axons (Ax⁺EpoB⁺) and retracting or stalled axons (Ax⁺EpoB⁻) (control) are available in Supplementary Video 2. d, Quantification of axon reactions after axotomy. n = 29 (Ax⁺EpoB⁺) and n = 33 (Ax⁺EpoB⁻) axons. e, Snapshots of regenerating axons labelled with SiR-tubulin (Ax⁺EpoB⁺ and Ax⁺EpoB⁻). The mild microtubule-stabilizing effect of SiR-tubulin was negligible in the control (Ax⁺EpoB⁻) axons. Scale bar, 5 µm. f, Quantification of axon length over time. Start point is the axotomy site. Data are mean ± s.e.m. n = 29 (Ax⁺EpoB⁺) and n = 33 (Ax⁺EpoB⁻). Source data

Fig. 2. Cytoskeletal organization and membrane tension dynamics during EpoB-induced axon regeneration.

a, Colocalization of microtubule (SiR-tubulin) and membrane (CellMask) signals at regenerating axon tips 1 h after axotomy and EpoB treatment. Scale bar, 10 µm. b, Actin organization (SPY555-FastAct) at the tip of axons under control (Ax⁻EpoB⁻) and regenerating (Ax⁺EpoB⁺) conditions. Scale bars, 3 µm. c–e, Membrane tension analysis by FLIM. c, Membrane tension is increased at tips relative to shafts (>15 µm downstream from the tip) 1 h after axotomy (n = 12). d, Tension normalizes at 4 h (n = 14). e, Control neurons show no difference between tip and shaft (n = 20). Data are mean ± s.e.m.; two-tailed Mann–Whitney test. f, Cryo-EM images of regenerating axons. The grid view gives an overview of the location (left; scale bar, 200 µm), with a magnified view showing an axotomy site (right; scale bar, 5 µm). Arrow indicates axonal growth direction and dashed line and scissors indicate the axotomy site. g, Distribution of microtubule lengths in cryo-EM snapshots 1 h after axotomy. Ax⁺EpoB⁺: n = 17, median = 19.1; Ax⁺EpoB⁻: n = 21, median = 0. The centre line shows the median; two-tailed Mann–Whitney test. h, Time course of axonal extension by cryo-EM snapshots, in the presence of EpoB after axotomy. n = 6 (15 min), n = 17 (1 h), n = 6 (3 h) and n = 12 (6 h) axons. Data are mean ± s.e.m. i, High-magnification montage of a regenerating axon 1 h after axotomy with EpoB. Top, cryo-EM montage; dashed line indicates the axotomy site. Scale bar, 400 nm. Bottom, segmentation depicting microtubules (green), membranes (vesicles and endoplasmic reticulum (ER)) (orange) and actin (magenta). Arrow depicts axonal growth direction. Source data

Fig. 3. In situ cryo-EM reconstruction of microtubule shoots in regenerating axon.

a, Large-scale axotomy of thalamus axons and regeneration in the presence of EpoB. Cryo-EM grid square 1 h after axotomy shows axons (white dotted lines) and the scission line (black dashed line). Beyond the cut, regenerating axons were visible. Magenta circles indicate grid holes with microtubule shoots used for SPA. Scale bar, 10 µm. b, Representative cryo-EM image of microtubule shoots taken for SPA. The polarities of the microtubules are indicated by + (plus end) and – (minus end). Scale bar, 50 nm. The corresponding tomogram is shown in Supplementary Video 6. c, Distributions of microtubule angles compared to reference microtubules. Angles close to 0° indicate polarities aligned in parallel, whereas angles close to 180° indicate an anti-parallel alignment. n = 1,663 microtubules, 4,123 segments. d, Distribution of microtubule protofilament (pf) numbers among observed microtubules (MTs). All 1,663 microtubules display the physiologically relevant 13-protofilament arrangement. e, 3D reconstruction of microtubule shoots at 3.19 Å resolution, showing the 13-protofilament arrangement. The reconstruction before focused refinement was used for display. Scale bar, 100 Å. f, 3D reconstruction of microtubule shoots, highlighting the inner lumen of microtubules. Left, B-factor sharpened map, with clearly distinguishable separation of α-tubulin (α-tub) and β-tubulin (β-tub) subunits within a dimer unit. Circles depict the S9–S10 loops and the magenta arrow indicates the EpoB binding pocket. Middle, higher volume threshold without sharpening, showing EpoB in the binding pocket. Right, magnified view of EpoB in the binding pocket. g, Left, atomic model of the microtubule shoots with an inter-dimer distance of 83.3 Å. For comparison, GMPCPP-stabilized microtubules adopt a stretched-dimer conformation (middle; inter-dimer distance 84.0 Å (Protein Data Bank (PDB): 6DPU)) and GDP-bound dynamic microtubules have a relaxed conformation (right; inter-dimer distance 81.8 Å (PDB: 6DPV)). EpoB-induced microtubule shoots adopt a stretched–stabilized lattice. Pink, α-tubulin nucleotide; salmon, β-tubulin nucleotide. Source data

Fig. 4. EpoB-induced tubulin cluster dynamics and microtubule maturation during axon regeneration.

a, Regenerating axons exhibit tubulin clusters in the presence of EpoB. Left, representative images of tubulin clusters (asterisks). Scale bars, 5 µm. Right, quantification of cluster densities per 100 µm of axon shaft and tips. Ax⁺EpoB⁺: n = 33 axons (178 particles); Ax⁺EpoB⁻: n = 28 axons (16 particles); Ax⁻EpoB⁺: n = 69 axons (149 particles); Ax⁻EpoB⁻: n = 67 axons (30 particles). Scatter plots show median with interquartile range. Two-tailed Mann–Whitney test, P < 0.0001. b, Temporal snapshots showing tubulin movements with or without EpoB and axotomy. Tubulin clusters in Ax⁺EpoB⁺ axons show pronounced anterograde movement. Axotomy sites are marked with scissors and a red dotted line. Representative videos are shown in Supplementary Video 8. c–e, Analysis of tubulin cluster movements within 30 min after axotomy. c, Distance of tubulin clusters from the injury site (distance travelled inward). Ax⁺EpoB⁺: n = 33, 12 axons; Ax⁺EpoB⁻: n = 31, 15 axons; P = 0.1764 (not significant). d, Distance of tubulin clusters from the regenerating tip (distance travelled outward) is decreased with EpoB (P < 0.0001). e, Net displacement of tubulin clusters (distance outward versus distance inward) shows a bias for anterograde transport with EpoB (P < 0.0001). Data are presented as dot plot with median and interquartile range. Two-tailed Mann–Whitney test. f, Spatial distribution of MIPs during regeneration. Stacked bar graphs show fractions of microtubules with sparse (grey, >50 nm spacing) versus dense (blue, <50 nm spacing) MIP occupancy. Fractions of sparsely populated filaments were 23% (control, n = 634), 29% (15 min, n = 256), 44% (1 h, n = 2025), 51% (3 h, n = 401), 51% (6 h, n = 336); and 34% (24 h, n = 912). Observed microtubules were located at pre-cut sites at 15 min after axotomy and at post-cut sites at 1 h after axotomy. g, Cryo-ET reconstructions showing MIPs at different stages of regeneration. A yellow arrowhead indicates a MIP-free microtubule tip. Scale bar, 50 nm. A representative reconstruction is presented in Supplementary Video 6. h, Tomographic snapshot of regenerating axon tips with MIP-free microtubules. Scale bar, 100 nm. Source data

Fig. 5. Spiral-like tubulin oligomers as precursors of polymerizing microtubules.

a, Snapshots of cryo-ET sections showing clusters of spiral-like tubulin oligomers (yellow arrowheads) at post-cut sites. Scale bars, 50 nm. The corresponding tomogram is shown in Supplementary Video 11. b, Segmentation of tubulin spirals from a tomographic reconstruction in 2 different views (80° rotated) to visualize the geometry of the spiral structures. Right image highlights oligomers with loose packing. Scale bar, 50 nm. c, Intermolecular spacing of microtubules, F-actin filaments and tubulin spirals. In box plots, the centre line represents the median, box edges delineate 25th and 75th percentiles, and whiskers extend from minimum to maximum values. n = 170 segments (6 microtubules), n = 97 segments (8 F-actin filaments), n = 182 segments (16 tubulin spirals). d, Size distribution of spiral-like oligomers. Dot plots show the distribution of diameter and width of tubulin oligomers with median and 25th and 75th percentiles as horizontal lines. Left, diameter: n = 163 oligomers, median: 37.8 nm, 25th percentile: 34.3 nm, 75th percentile: 40.7 nm. Right, width: n = 53 oligomers, median: 4.2 nm, 25th percentile: 3.2 nm, 75th percentile: 4.9 nm. e, Top, magnified images of tomographic sections, highlighting tubulin spirals attached to microtubule ends. Some spirals are unfolding into straighter conformations. Bottom, outline sketches of the polymerizing microtubules (green) with oligomers attached. Scale bar, 50 nm. f, Tomographic snapshot showing microtubule branching (green arrowhead) with a schematic on the right. Scale bar, 200 nm. The corresponding tomogram is shown in Supplementary Video 12. Source data

Fig. 6. Structural visualization and schematic sequence of EpoB-driven axon regeneration.

a, Cryo-EM image (grid view) of a regenerating axon in the presence of EpoB, 24 h after axotomy. The axotomy site is indicated by a black dotted line. The yellow shading highlights the path of the regenerating axon. Black boxes indicate positions at which cryo-ET data were collected and reconstructed. Sections are shown in b,c. Scale bar, 50 µm. b, Cryo-ET snapshots of a regenerating axon 24 h after axotomy in the presence of EpoB, from the boxed areas in a. Scale bar, 200 nm. c, Segmentation of the cryo-ET reconstructions shown in b. Some areas are surrounded by plasma membranes and some remain uncovered. Growth cones are established at the regenerating axon tip at this time. d, Summary of the molecular events in EpoB-induced axon regeneration. (1) The initial reaction after axon injury or axotomy is membrane closure, resembling the morphology of a retraction bulb. (2) In the presence of EpoB, stabilized microtubules shoot out from the injury site towards the plasma membrane, resulting in increased membrane tension. Using these stabilized microtubules as tracks, tubulin clusters and vesicles as membrane source are shuttled towards the growing axon. (3) Vesicles deliver materials for regeneration, including the necessary membrane support at the regenerating front, resulting in normalizing membrane tension (4). Ultimately, repair is complete, and a new growth cone is established at the axon tip that navigates the growth of the axon.

Extended Data Fig. 1. EpoB dose response and cytoskeletal organization in regenerating axons.

a) Representative images of regenerating thalamus axons treated with different doses of EpoB shown at different time points. Magenta dotted lines indicate the cut site. Scale bar: 10 µm. Corresponding movies are shown in Supplementary Video 3. b) Quantification of regeneration in response to different doses of EpoB. The stacked bar graph shows mean as a horizontal line ± s.e.m. n = 7 (0.1 nM), n = 13 (1 nM), n = 14 (10 nM), n = 16 (100 nM), n = 17 (1 µM), n = 16 (5 µM) and n = 13 (10 µM) axons. c) Representative images of regenerating thalamus axons treated with taxol shown at different time points. Magenta dotted lines indicate the cut site. Scale bar 20 µm. The corresponding movie is shown in Supplementary Video 4. d) Immunofluorescence image of thalamus axons after axotomy and treated with EpoB for 1 h, fixed and stained with tubulin antibody 12G10 (cyan) and CellBrite membrane labeling dye (yellow). The axons were not permeabilized to maintain membrane integrity. Scale bar: 5 µm. e) Control thalamus explants stained with SPY555-actin (magenta) and SPY650-tubulin (cyan). Scale bar: 5 µm. f) Regeneration test of thalamus explants after axotomy at different time points. The axons were stained with SPY650-tubulin (Cyan, microtubules), SPY555-actin (magenta, plasma membrane) and MemGlow 488 (plasma membrane, Yellow). Scale bar: 5 µm. The corresponding videos are shown in Supplementary Video 5. Source data

Extended Data Fig. 2. Time course of ultrastructural changes in regenerating axons with and without EpoB.

a) Axon injury site in the presence of EpoB, 15 min after axotomy. Top left: Cryo-EM snapshot with a white dotted line indicating the axotomy site and white boxes indicating positions of magnified insets. The black arrow indicates the axonal growth direction. Scale bar: 2 µm. Top right: Magnified images of square insets. Scale bar: 200 nm. Bottom left: Segmentation of axon borders from image above, which appears swollen, showing a retraction bulb-like morphology. Bottom right: Segmentation of axon borders and visible cellular components from magnified insets above with microtubules (green), actin (magenta), membranes (orange) and vesicles (brown circles). b) A regenerating axon in the presence of EpoB 3 h after axotomy. Top: cryo-EM image of a regenerating axon beyond the axotomy site (black dotted line). The black arrow indicates the axonal growth direction. Scale bar 1 µm. Zoomed-in images of the boxes i-v are shown on the right. Scale bar: 100 nm. Bottom: Schematic depicting the border and visible cellular components at this magnification. Microtubules are shown in green and other membranous structures are in light yellow or orange. c) A regenerating axon in the presence of EpoB 6 h after axotomy. Top: cryo-EM image of a regenerating axon beyond axotomy site (black dotted line). The black arrow indicates axonal growth direction. Scale bar 1 µm. Zoomed-in images of the boxes i-v are shown on the right. Scale bar: 100 nm. Bottom: Schematic depicting the border and visible cellular components at this magnification. Microtubules are shown in green, actin in magenta and other membranous structures are in light yellow or orange. The tip of microtubule-shoots is topped by actin bundles. d) A tip of an axon in the absence of EpoB, 1 h after axotomy. Top: Low magnification cryo-EM image. The black arrow indicates axonal growth direction. Scale bar: 1 µm. Bottom: Schematic of visible structures. Actin is shown in magenta and membranous structures are in light yellow or orange.

Extended Data Fig. 3. Quality control of cryo-EM images showing membrane rupture under high tension.

a) Cryo-EM montage of a thalamus axon incubated in hypotonic media conditions before vitrification. The top panel shows a high-magnification montage of an axon. Under hypotonic conditions, membrane tension increases and compromises membrane integrity during the vitrification process. Scale bar: 400 nm. Dotted white lines indicate the positions of zoomed-in insets shown below the montage. A light orange shade highlights the trace of the axon. Scale bar: 200 nm. b) High-magnification cryo-EM montage of a control axon (neurons grown in normal culture media). Scale bar: 400 nm. Zoomed-in insets at different locations were shown below the montage. A light orange shade is used to highlight the trace of the axon. Scale bar: 200 nm. c) Live imaging of axons after axotomy and treated with EpoB. Fluorescently-labelled tubulin antibody (cyan) was added. No microtubule staining, showing the plasma membrane was sealed immediately after axotomy. Scale bar: 20 µm.

Extended Data Fig. 4. SPA processing workflow for high-resolution reconstruction of microtubule shoots 1 h after axotomy.

The 3D reconstruction was performed using FREALIGN and CryoSPARC. Using CryoSPARC, microtubules were segmented into 399,396 overlapping boxes (703 Å) with an 82.5 Å step size, corresponding to the tubulin heterodimer length. After alignment and 2D classification in CryoSPARC, 129,110 particles with clear filament features were retained for 3D classification and 3D reconstruction. All analyzed microtubules (n = 1663) exhibited a 13-protofilament architecture. Seam detection was conducted in FREALIGN v9.11 using previously established protocols. Final reconstructions were performed using 114,673 particles with the identification of clear seam position using CryoSPARC. The reconstruction was performed in C1 without imposing pseudo-helical symmetry, preserving the α/β-tubulin distinction. After per-particle defocus refinement and focused refinement, the analysis yielded a C1 reconstruction at 3.19 Å resolution (gold-standard FSC at 0.143) visualizing a clear EpoB engagement. GSFSC, Local resolution map, orientation distribution and map-to-model FSC are included in the figure.

Extended Data Fig. 5. EpoB enhances microtubule dynamics and oligomer transport during regeneration.

a) Live imaging of EpoB-induced regenerating thalamus axon tip transduced with EB3-mScarlet encoding lentiviruses. EB3 signals were found at the regenerating axon tip, suggesting active growth of plus-end-out microtubules. The axotomy site is shown with a white dotted line. Corresponding movies are available in Supplementary Video 7. b) Scatter dot plots showing counts of tubulin tracker puncta in thalamus control axons in the presence of different concentrations of EpoB. Mean is indicated as horizontal line. One-Way Anova was used for statistical testing, and p-values are 0.0259 (0.1 nM vs 1 nM), <0.0001 (0.1 nM vs 10 nM) and 0.0020 (1 nM vs 10 nM). N = 19 (0.1 nM EpoB), N = 24 (1 nM EpoB) and N = 27 (10 nM EpoB) puncta. c) Tubulin Tracker binding affinity to tubulin monomers, polymerized microtubules and tubulin spirals. Tubulin Tracker alone does not show any fluorescence, but it can bind and fluoresce in the presence of tubulin spirals and microtubules. The data is shown as bar graphs overlaid with data points, in which the fold change in fluorescence between different experimental conditions is plotted. Tubulin spirals were induced by p150 protein. d) Distances traveled by tubulin clusters away from the axotomy site within 30 min of observation. The violin plots represent median and interquartile range. Ax−EpoB+: n = 39 particles (16 axons), and Ax−EpoB−: n = 66 particles (29 axons). Significance was tested using the two-tailed Mann-Whitney test. P-values are reported at the corresponding comparisons on the graphs. e) Distances traveled by tubulin clusters toward regenerating sites within 30 min of observation. The violin plots represent median with interquartile range. Ax−EpoB+: n = 39 particles (16 axons), and Ax−EpoB−: n = 66 particles (29 axons) from at least three biologically independent experiments. Significance was tested using the two-tailed Mann-Whitney test. p-values are reported at the corresponding comparisons on the graphs. f) Net distance (outward vs inward) traveled by tubulin clusters within 30 min of observation. The violin plots represent median with interquartile range. Ax−EpoB+: n = 39 particles (16 axons), and Ax−EpoB−: n = 66 particles (29 axons) from at least three independent experiments. Significance was tested using the two-tailed Mann-Whitney test. p-values are reported at the corresponding comparisons on the graphs. g) Distances traveled by tubulin clusters away from the axotomy site within 10 min of observation. The violin plots represent median with interquartile range. Ax−EpoB+Monas−: n = 60 particles (6 axons), and Ax−EpoB+Monas+: n = 32 particles (7 axons) from at least three independent experiments. Significance has been tested using the two-tailed Mann-Whitney test. p-values are reported at the corresponding comparisons on the graphs. h) Distances traveled by tubulin clusters towards the axotomy site within 10 min of observation. The violin plots represent median with interquartile range. Ax−EpoB+Monas−: n = 60 particles (6 axons), and Ax−EpoB+Monas+: n = 32 particles (7 axons) from at least three independent experiments. Significance has been tested using two-tailed Mann-Whitney test. p-values are reported at the corresponding comparisons on the graphs. i) Distribution of distances between neighboring MIPs at different times after axotomy in the presence of EpoB. The dot plots show median with interquartile range. n = 634 (control), 256 (t = 15 min), 2025 (t = 1 h), 401 (t = 3 h), 336 (t = 6 h), 912 (t = 24 h) particles. Median: 18.5 nm (control) 17.6 nm (15 min) 24.2 nm (1 h), 22.6 nm (3 h), 24.3 nm (6 h), 18.3 nm (24 h). At t = 15 min, the observed microtubules are located at the pre-cut site as no regeneration of microtubules are observed at this time point. At t > 1 h, all the microtubules are located at the post-cut sites. One-Way Anova was used for statistical testing and p-values are >0.9999 (control vs 0.25 h), <0.0001 (control vs 1 h), <0.0001 (control vs 3 h), <0.0001 (control vs 6 h) and >0.9999 (control vs 24 h). Source data

Extended Data Fig. 6. Workflows for MIP analysis and stress-fiber subtomogram averaging.

a) Processing workflow for the MIP analysis. i) Zoomed-in view of a MT in a 2D slice of a tomogram, scale bar: 50 nm. ii) The centerline is traced by Amira. iii) The lumen is segmented by applying distance transform on the centerline. iv) PySeg analysis extracting the local minima and their connections in the MT lumen and MT shell. The color bar represents the topological persistence, the metric used to rank the minima relevance. The information of the shell is used to determine the threshold for the lumen. v) Thresholded local minima by persistence in the MT lumen. vi) MIP obtained by applying mean shift clustering to v. b) Processing workflow for stress-fiber subtomogram averaging. Tilt-series alignment and tomographic reconstructions were performed within RELION 5.0. A total of 1555 segments of filaments were manually selected from the area of interest using IMOD, and the corresponding coordinates were imported into RELION for subtomogram extraction in the form of a 2D stack. For subtomogram averaging, a cylinder reference was used, followed by 3D classification to create an initial model, and further refinement was performed using a combination of classification and refinement schemes without applying the actin helical symmetry. The nominal resolution was 27.7 Å. CTF correction was applied only during subtomographic analysis, following 3D reconstruction using RELION 5.0.

Extended Data Fig. 7. Cellular and vesicular components at regenerating axon sites.

a) Representative cryo-EM snapshot of cellular components at regenerating sites after axotomy in the presence of EpoB. Membranous and vesicular entities accumulate at regeneration sites. Vesicles are highlighted in light blue and microtubules (MT) in green. Scale bar: 100 nm. b) Vesicle size distribution as dot plot. n = 110 vesicles, median: 43.9 nm, 25th percentile: 32.7 nm, 75^th percentile: 76.6 nm. c) Representative sections of a cryo-ET reconstructions of a regeneration site showing endoplasmic reticulum (ER) membranes (highlighted in light yellow), presynaptic vesicles (highlighted in light blue), and clusters of materials (highlighted in light brown) near microtubules. Scale bar: 100 nm. d) Cryo-ET section showing actin stress fiber formation (highlighted in light pink). Scale bar: 100 nm. Subtomogram average of actin fibers at 27.7 Å resolution for validation. e) Various cryo-EM snapshots of regenerating axons showing vesicles coated with proteins (yellow arrowheads) and vesicles with components in inner compartments (magenta arrowheads). Scale bar: 100 nm. Multivesicular body (MVB). Source data

Extended Data Fig. 8. Dynamic progression of EpoB-induced axon regeneration over hours.

a) Live images of an EpoB-induced, regenerating axon tip shown at different time points during repair. For about 2 h, the regenerating front displays a flat rapidly moving axon tip. After two hours, a typical actin-rich finger-like growth cone appears. Scale bar: 10 µm. Zoomed-in images of the tip are shown in the insets. Scale bar: 5 µm. b) Live images of EpoB-induced regenerating axons recorded continuously for 6 h. Straightened axon snapshots at different time points are shown for comparison with the axotomy site indicating with a dotted white line. Scale bar: 10 µm. c) Quantification of the cumulative length of axons during EpoB-induced regeneration over 6 h. n = 2 axons from two biologically independent experiments. The bar graph is overlaid with corresponding data points. d) Snapshots of EpoB-induced regenerating axons captured at 0, 24 and 48 h. Images were straightened for comparison. EpoB-induced axonal growth continues for at least 48 h beyond axotomy. Scale bar: 10 µm. e) Quantification of the cumulative length of axons during EpoB-induced regeneration. The bar graph shows mean ± s.e.m. Significance was tested using One way ANOVA. P-values are 0.0030 (0 h vs 24 h) and <0.0001 (0 h vs 48 h). n = 5 axons from three biologically independent experiments. Source data