In vitro DRG neurons recapitulate in vivo developmental stages.

(A) In vitro DRG neurons labeled with βIII-tubulin depicting different development stages. Scale bar, 10 µm. (B) Imaris segmentation of a pseudounipolar DRG neuron transduced with AAV-CMV-eGFP. Scale bar, 7 µm. (C) In vitro DRG neurons transduced with AAV-CMV-eGFP depicting stem axon formation. Scale bar, 10 µm. (D-E) Stem axon diameter (D) and length (E) of DRG neuron axons from the formation of the stem axon (Initial) to the final stage of pseudo-unipolarization (Final) (n= 13 neurons; paired t-test, diameter ****p<0.0001, length ***p=0.0004). (F) Stem axon and cell-body displacement during pseudo-unipolarization (n= 13 neurons; paired t-test, **p=0.0020). (G) In vitro diameter of DRG axons; n= 5-8 independent experiments, 5-10 neurons/experiment; paired t-test in bipolar neurons, ***p=0.0003; repeated measures (RM) one-way ANOVA in pseudo-unipolar neurons, stem-peripheral *p=0.0196, stem-central **p=0.0069, peripheral-central **p=0.0048; for comparisons amongst peripheral and central-like axons from bipolar and pseudo-unipolar neurons, a two-way ANOVA was used (peripheral: **p=0.0039; central: p=0.9829). (H) In vitro pseudo-unipolar DRG neuron transduced with the lentivirus CMV-EB3-GFP depicting different axon diameter. Scale bar, 5 µm. Data are represented as mean ± SEM.

In vitro DRG neurons recapitulate in vivo regenerative asymmetry.

(A-B) Live imaging of in vitro regeneration of a peripheral-like (A) and a central-like axon (B). Scale bar, 10 µm. The injury sites are marked by red lines; lesioned axon tips by black arrowheads and regenerating axons by green arrowheads. (C-D) Length (C) and duration (D) of axon retraction after laser axotomy (n= 10-14 axons, 3 independent experiments; unpaired t-test, length p=0.9541, duration *p=0.0228). (E) Number of regenerating and non-regenerating peripheral-like (P) and central-like (C) DRG axons. The chi-square test was used (5 independent experiments, **p=0.0082). (F) Regeneration length of peripheral and central-like axons following laser axotomy (n= 18-21 axons, 5 independent experiments; unpaired t-test, *p=0.0297). Data are represented as mean ± SEM.

DRG axons display asymmetric microtubule polymerization, attenuated by a conditioning lesion.

(A) In vitro pseudo-unipolar DRG neuron transduced with a Tom20-GFP lentivirus, labelling mitochondria. Scale bar, 5 µm. (B) Quantification of the anterograde mitochondria flux (n=4 independent experiments, 5 DRGs/experiment; paired t-test, *p=0.0143). (C) EB3-GFP comet density in in vitro DRG axons (n= 6-7 independent experiments, 5-10 neurons/experiment; paired t-test in bipolar axons, **p=0.0038; RM one-way ANOVA in pseudo-unipolar axons, stem-central *p=0.0221, peripheral-central *p=0.0171. (D) Kymographs of in vitro pseudo-unipolar DRG axons. (E) EB3-GFP comet velocity in in vitro pseudo-unipolar DRG axons (n= 6 independent experiments, 5-10 neurons/experiment; RM one-way ANOVA, stem-central *p=0.0443, peripheral-central *p=0.0183). (F) Representation of naive DRG neurons connected to the peripheral nerve (containing peripheral axons) and dorsal root (containing central axons). The dashed squares indicates the imaging locations. (G) Live imaging of DRG axons from Thy1-EB3-eGFP mice. Scale bar, 5 µm. (H) EB3-eGFP comet density in DRG axons (n=12-17 animals, 4 axons/animal; unpaired t-test, **p=0.0071). (I) High magnification electron microscopy images within individual DRG axons, depicting axonal microtubules (red arrowheads). Scale bar, 100 nm. (J) Total density of microtubules in DRG axons (n= 8 animals, 5-10 axons/animal; paired t-test, p=0.2299). (K) EB3-eGFP comet velocity in DRG axons (n= 11-15 animals, 4 axons/animal; unpaired t-test, *p=0.0129). (L) Representation of a sciatic nerve injury to DRG neurons (conditioning lesion). The dashed square indicates the imaging location, while the dashed line and scissor marks the lesion site. (M) EB3-GFP comet density in DRG explants from naive mice (n= 12-17 animals; 3-6 axons/animal, **p=0.0037) and mice with a peripheral conditioning lesion (CL) (n=9-10 animals, 3-5 axons/animal, p=0.1423). Two-way ANOVA; peripheral naive-peripheral CL, *p=0.0276; central naive-central CL, **p=0.0026. (N) Total density of axonal microtubules in DRG peripheral and central axons after peripheral CL (n= 8 animals, 5 axons/animal; paired t-test, p=0.4624). (O) EB3-GFP comet velocity in DRG explants from naive mice (n=11-15 animals, 3-6 axons/animal, **p=0.0048) and mice with peripheral CL (n=8-9 animals, 3-5 axons/animal, **p=0.0035). Two-way ANOVA, peripheral naive-peripheral CL, ***p=0.0003; central naive-central CL, **p=0.0038. Data are represented as mean ± SEM.

Figure 3 – video 1. Mitochondria transport in in vitro DRG neurons.

Figure 3 – video 2. Microtubule dynamics in in vitro DRG neurons.

Figure 3 – video 3. Microtubule dynamics in in vivo DRG neurons.

DRG axons have a distinctive MAP signature that adapts upon conditioning lesion.

(A) Representation of a naive DRG neuron. Peripheral axons within the peripheral nerve (PN) are depicted in red and central axons within the dorsal root (DR) in blue. (B-E) Western blot analysis of the PN and DR (n= 5-7 animals) and respective quantification of (B) spastin (unpaired t-test, **p=0.0032), (C) katanin (paired t-test, **p=0.0063), (D) CRMP5 (unpaired t-test, ****p<0.0001), and (E) tau (Wilcoxon test, *p=0.0313) levels. (F-H) Immunofluorescence of (F) katanin and βIII-tubulin, (G) CRMP5 and βIII-tubulin, and (H) tau and βIII-tubulin in DRG axons (n= 4-5 animals) and respective quantification (katanin, paired t-test, **p=0.0021; CRMP5, paired t-test, ***p=0.009; tau, paired t-test, *p= 0.0114). Scale bar, 5 µm. (I) Representation of a DRG neuron with a priming peripheral lesion (peripheral lesion-PL). (J-M) Western blot of DRG peripheral nerve and dorsal root following PL (n= 6 animals) and respective quantification (n= 6 animals; paired t-test) showing (J) spastin (p=0.4085), (K) katanin (p=0.2661), (L) CRMP5 (p=0.1326) and (M) tau (**p=0.0052) levels. Data are represented as mean ± SEM.

Peripheral and central DRG axons show similar levels of tubulin post-translational modifications.

(A-D) Western blot analysis of DRG peripheral nerve (PN) and dorsal root (DR) and respective quantification of (A) tyrosinated tubulin (n= 6 animals; paired t-test, p=0.9075), (B) acetylated tubulin (n= 5 animals; Mann-Whitney U test, p=0.7302), (C) Δ2 tubulin (n= 5-6 animals; unpaired t-test, p=0.6639), and (D) polyglutamylated tubulin (n= 5-6 animals; unpaired t-test, p=0.2052) expression levels. Data are represented as mean ± SEM.

Potential Sorting Mechanism at the DRG T-Junction.

(A) Representative images of a peripheral DRG axon after performing RNAscope analysis to detect spastin and CRMP5 mRNA particles. Scale bar, 10 µm. (B-C) The density of (B) total spastin mRNA and (C) total CRMP5 mRNA in the DRG peripheral nerve and dorsal root (n= 7 animals, 3 non-consecutive DRG peripheral nerve and dorsal root sections were analyzed per animal). Paired t-test, spastin: p=0.2112, CRMP5 p=0.7671. (D) In vitro pseudo-unipolar DRG neurons transduced with a CMV-EB3-GFP lentivirus. Scale bar, 5 µm. (E-F) Sequential images of EB3-GFP comets, either stopping (E) or crossing (F) the DRG T-junction. This analysis spanned a 5 µm region from the end of the DRG stem axon to the start of either the peripheral-like (red line) or central-like (blue line) axonal branch. Scale bar, 5 µm. (G) Number of EB3-GFP comets that either stop or cross the DRG-T junction (n=8 independent experiments, 10 DRGs/experiment; RM two-way ANOVA, stop *p=0.0439, cross *p=0.0162, stop-cross peripheral p= 0.9169, stop-cross central **p=0.0032). (H) Immunofluorescence of polyglutamylated tubulin and βIII-tubulin in DRG peripheral and central like axonal branches. Scale bar, 10 µm, and close-up 5 µm. (I) Quantification of polyglutamylated tubulin fluorescence intensity normalized to βIII-tubulin (n=5 independent experiments, 8 pseudo-unipolar DRGs per experiment, paired t-test, p= 0.2957). (J) Immunofluorescence of acetylated tubulin and βIII-tubulin in DRG peripheral and central like axonal branches. Scale bar, 10 µm, and close-up 5 µm. (K) Quantification of acetylated tubulin fluorescence intensity normalized to βIII-tubulin (n=4 independent experiments, 8 pseudo-unipolar DRGs per experiment, paired t-test, p= 0.2946). (L) Immunofluorescence of Δ2 tubulin and βIII-tubulin in DRG peripheral and central like axonal branches. Scale bar, 10 µm, and close-up 5 µm. (M) Quantification of Δ2 tubulin fluorescence intensity (n=4 independent experiments, 8 DRGs/experiment; paired t-test, **p= 0.0012). Data are represented as mean ± SEM.

Spastin deficiency is sufficient to abolish the asymmetry of DRG axons.

(A) Density of EB3-GFP comets in wild-type DRG axons in vitro (n= 4 independent experiments, 10 cells/experiment; RM one-way ANOVA, stem-peripheral p= 0.0536, stem-central **p= 0.0082, peripheral-central **p= 0.0027). (B) Density of EB3-GFP comets in spastin knock-out DRG axons in vitro (n= 4 independent experiments, 10 cells/experiment; RM one-way ANOVA, stem-peripheral *p= 0.0168, stem-central *p= 0.0250, peripheral-central p= 0.8762). (C-D) EB3-eGFP comet density in (C) wild-type and (D) spastin knock-out mice (n= 10-14 animals; 3 axons/animal; unpaired t-test; wild-type, *p=0.0388; knock-out, p=0.9792). (E-F) EB3-eGFP comet velocity in (E) wild-type and (F) spastin knock-out mice (n= 5-7 animals; 3 axons/animal; unpaired t-test; wild-type, *p=0.0405; knock-out, p=0.0823). (G) Representation of the conditioning lesion (CL). A dorsal column hemisection is preceded by a sciatic nerve transection 1 week before. Lesion sites are indicated with dashed red lines and DRG axons in green. (H-J) Longitudinal spinal cord sections of (H) wild-type mice with spinal cord lesion or (I) CL and (J) spastin knockout mice with CL. Dorsal column tract axons were traced with cholera toxin-B (white). The lesion border is highlighted by a yellow line. Regenerating axons are highlighted by red arrowheads. C-caudal; R-rostral; D-dorsal; V-ventral. Scale bar, 100 µm. (K) Number of regenerating axons in wild-type mice with spinal cord injury (n= 5 animals) and CL (n= 6 animals), and spastin knockout with CL (n= 7 animals); 6 sections per animal. One-way ANOVA; wild-type SCI-CL, ***p=0.0005; wild-type-knockout CL, **p=0.0020; wild-type SCI-knockout CL, p=0.3335. (L) Representative in vitro wild-type and spastin knockout adult DRG neurons labeled with βIII-tubulin. Scale bar, 30 µm. (M) Quantification of the number of primary neurites in adult wild-type and spastin knockout DRG neurons. n= 4-5 independent experiments for wild-type and spastin knockout; unpaired t-test; *p= 0.0205. Data are represented as mean ± SEM.

Signs of axon degeneration are not observed in 15 weeks old spastin knockout animals.

(A) Representative electron microscopy images (left) and respective quantification (right) of the total density of axonal microtubules in wild-type and spastin knockout peripheral and central DRG axons (n= 6-7 animals, up to 10 axons/animal). Two-way ANOVA followed by Fisher’s LSD post hoc test, p=0.6030 wild-type peripheral-central axons and p=0.3615 spastin knockout peripheral-central axons. (B) Representative images (left) of the dorsal column tract central axons of naive 15 weeks old wild-type and spastin knockout animals traced with cholera toxin-B (black) injection (zoom-ins of boxed regions are provided; Scale bars: 200 µm and 50 µm in zoom-ins) and respective quantification (n=3 animals; up to 6 spinal cord sections analyzed per animal; unpaired t-test, p=0.5918) (right). (C) Representative images of sciatic nerve semi-thin sections of naive 15 weeks old wild-type and spastin knockout animals (left; scale bar, 10 µm) and respective quantification of myelinated axons (n=4-5 animals, all myelinated axons in a complete nerve cross-section were quantified; unpaired t-test, p=0.2455) (right). (D) Representative electron microscopy images of the sciatic nerve of naive 15 weeks old wild-type and spastin knockout animals (left; scale bar, 2 µm) and respective quantification of unmyelinated axons (n=5 animals, unmyelinated axons inside Remak bundles were quantified; paired t-test, p=0.6315) (right). Data are represented as mean ± SEM.

Microtubules in DRG axons are distinctively regulated both in vivo and in vitro.

(A) In vitro DRG cultures replicate in vivo asymmetry in microtubule polymerization (established already at the bipolar stage) and reveal a potential cargo filtering mechanism operating at the DRG T-junction, where a higher continuum of microtubules from the stem axon to the central-like axonal branch is found, together with increased levels of Δ2 tubulin in the central-like axonal branch. (B) Under physiological conditions, peripheral DRG axons show a decreased density of growing microtubules when compared to central axons due to an asymmetric MAP signature. (C) After a peripheral conditioning lesion, the DRG axon MAP signature is remodeled, leading to a decrease in the density of growing microtubules and regeneration of both peripheral and central axons. (D) Perturbing the DRG MAP signature by knocking out spastin, abolishes DRG axon asymmetry and central axon regeneration following conditioning lesion.