Overexpression of Reticulon 3 Enhances CNS Axon Regeneration and Functional Recovery after Traumatic Injury

Abstract

CNS neurons are generally incapable of regenerating their axons after injury due to several intrinsic and extrinsic factors, including the presence of axon growth inhibitory molecules. One such potent inhibitor of CNS axon regeneration is Reticulon (RTN) 4 or Nogo-A. Here, we focused on RTN3 as its contribution to CNS axon regeneration is currently unknown. We found that RTN3 expression correlated with an axon regenerative phenotype in dorsal root ganglion neurons (DRGN) after injury to the dorsal columns, a well-characterised model of spinal cord injury. Overexpression of RTN3 promoted disinhibited DRGN neurite outgrowth in vitro and dorsal column axon regeneration/sprouting and electrophysiological, sensory and locomotor functional recovery after injury in vivo. Knockdown of protrudin, however, ablated RTN3-enhanced neurite outgrowth/axon regeneration in vitro and in vivo. Moreover, overexpression of RTN3 in a second model of CNS injury, the optic nerve crush injury model, enhanced retinal ganglion cell (RGC) survival, disinhibited neurite outgrowth in vitro and survival and axon regeneration in vivo, an effect that was also dependent on protrudin. These results demonstrate that RTN3 enhances neurite outgrowth/axon regeneration in a protrudin-dependent manner after both spinal cord and optic nerve injury.

Keywords: Reticulon 3; axon regeneration; neurite outgrowth; optic nerve injury; protrudin; spinal cord injury.

Figure 1

RTN3 is upregulated in dorsal root ganglion neurons (DRGN) after preconditioning lesions. (A) The diagram shows the timeline of treatments and tissue harvesting. (B) qRT-PCR to confirm upregulated RTN3 mRNA in SN and pSN + DC injury models (n = 16 DRG/group (8 rats/group)). (C) Western blot to show levels of RTN3 in control, DC, SN and pSN + DC injury models (n = 16 DRG/group (8 rats/group)). (D) Densitometry to confirm upregulation of RTN3 in SN and pSN + DC injury models (n = 16 DRG/group (8 rats/group)). (E) Immunohistochemistry to show immunolocalisation of RTN3 (green) in DRGN (red). Data are means ± SEM. Scale bars in (E) = 50 µm. *** p = 0.0001, one-way ANOVA with Dunnett’s post hoc test.

Figure 2

Overexpression of RTN3 in naïve DRGN cultures promotes neurite outgrowth. (A) Timeline of experiment and analysis methods. (B) Western blot to show RTN3 plasmids overexpress RTN3 protein. (C) Densitometry to show that RTN3 was significantly upregulated in RTN3 transfected cultures. (D) Representative images to show DRGN neurite outgrowth after RTN3 overexpression. RTN3 overexpression increases (E) the mean DRGN neurite length, (F) % DRGN with neurites but does not affect (G) DRGN survival. LF2000 = Lipofectamine 2000, AU = arbitrary units, *** p = 0.0001, one-way ANOVA with Dunnett’s post hoc test, n = 9 wells/treatment. Scale bars in (D) = 100 µm.

Figure 3

RTN3 is required for disinhibited DRGN neurite outgrowth. (A) Experimental timeline for analysis of RTN3 knockdown in preconditioned DRGN. (B) qRT-PCR to show ~80% knockdown of RTN3 using siRTN3. (C) Western blot to show levels of RTN3 protein after shRNA-mediated knockdown in DRGN. β-actin is used as a loading control. (D) Densitometry to show that PEI-shRTN3 significantly knocks down RTN3 protein in DRGN cultures. (E) Representative images to show neurite outgrowth after LF2000, siEGFP and siRTN3 treatment. (F) Quantification to show the mean longest neurite length after shRNA-mediated knockdown of RTN3. (G) Quantification to show % DRGN with neurites after shRNA-mediated knockdown of RTN3. (H) Quantification to show that DRGN survival is unaffected after shRNA-mediated knockdown of RTN3. Data are means ± SEM. Scale bars in (E) = 100 µm. *** p = 0.0001, one-way ANOVA with Dunnett’s post hoc test. n = 9 wells/condition.

Figure 4

Overexpression of RTN3 promotes disinhibited DRGN neurite outgrowth. (A) Timeline for experimental analysis of RTN3 overexpression in preconditioned DRGN. (B) Fold-change in mRNA after RTN3 overexpression in preconditioned DRGN. (C) Western blot to show overexpressed levels of RTN3 protein. β-actin is used as a protein loading control. (D) Densitometry to show significant RTN3 overexpression by PEI-RTN3. (E) Representative images to show neurite outgrowth after DC + PEI-EGFP, pSN + DC + PEI-EGFP and pSN + DC + PEI-RTN3 treatment. (F) Quantification to show the mean neurite length after DC + PEI-EGFP, pSN + DC + PEI-EGFP and pSN + DC + PEI-RTN3 treatment (G) Quantification to show % DRGN with neurites after DC + PEI-EGFP, pSN + DC + PEI-EGFP and pSN + DC + PEI-RTN3 treatment. Data are means ± SEM. *** p = 0.0001, one-way ANOVA with Dunnett’s post hoc test. Scale bars in (E) = 100 µm. n = 9 wells/treatment.

Figure 5

Knockdown of Protrudin and its interactors attenuates RTN3-induced DRGN neurite outgrowth. (A) Timeline for experimental analysis of Protrudin and its interactors in RTN3-overexpressed DRGN. Representative Western blots to confirm knockdown of (B) Protrudin, (C) FYCO1, (D) RAB7 and (E) SYT7. β-actin is used as a protein loading control. Representative images to show neurite outgrowth in (F) RTN3 overexpressed DRGN controls and after knockdown of (G) Protrudin, (H) FYCO1, (I) RAB7 and (J) SYT7 in RTN3-overexpressed cultures. (K) Quantification to show attenuation of RTN3-overexpressed DRGN neurite outgrowth by shProtrudin, shFYCO1, shRAB7 and shSYT7. (L) DRGN survival remains unaffected by knockdown of Protrudin, FYCO1, RAB7 and SYT7 in RTN3-overexpressed DRGN. n = 9 wells/condition Scale bars in (F–J) = 100 µm. Data are means ± SEM. *** p = 0.0001, one-way ANOVA with Dunnett’s post hoc test.

Figure 6

RTN3 overexpression promotes Protrudin-dependent DC axon regeneration. (A) Representative longitudinal sections of the spinal cord immunostained with GAP43 antibodies after treatment with pSN + DC + PEI-EGFP, pSN + DC + PEI-RTN3 + shEGFP and pSN + DC + PEI-RTN3 + shProtrudin (* = lesion site; C to R = caudal to rostral). FluoroRuby (FR) traced axons in the same sections as pSN + DC + PEI-RTN3 + shEGFP confirms overlap of GAP43⁺ axons (arrowheads), unequivocally demonstrating axon regeneration. Inset (i) shows high power images of GAP43⁺ axons in the boxed region (i). Inset (ii) and (iii) show high power images of FR⁺ axons in boxed regions (ii) and (iii). Note the overlap (arrowheads) between GAP43⁺ axons in inset (i) and FR+ axons in inset (ii). (B) Quantification of the % of GAP43⁺ axons at different distances caudal and rostral to the lesion site. Scale bars in (A) = 200 µm. n = 18 rats/group. * p = 0.05, ** p = 0.001, one-way ANOVA with Dunnett’s post hoc test.

Figure 7

RTN3 overexpression promotes electrophysiological, sensory and locomotor functional recovery after DC injury. (A) Representative Spike 2 processed CAP traces in sham, DC + PEI-EGFP, pSN + DC + PEI-EGFP, pSN + DC + PEI-RTN3, pSN + DC + PEI-RTN3 + shEGFP and pSN + DC + PEI-RTN3 + shProtrudin-treated animals. (B) Negative CAP amplitudes at different stimulation intensities in sham, DC + PEI-EGFP, pSN + DC + PEI-EGFP, pSN + DC + PEI-RTN3, pSN + DC + PEI-RTN3 + shEGFP and pSN + DC + PEI-RTN3 + shProtrudin-treated animals. (C) CAP areas at all stimulation intensities in sham, DC + PEI-EGFP, pSN + DC + PEI-EGFP, pSN + DC + PEI-RTN3, pSN + DC + PEI-RTN3 + shEGFP and pSN + DC + PEI-RTN3 + shProtrudin-treated animals. (D) Mean tape sensing and removal times in sham, DC + PEI-EGFP, pSN + DC + PEI-EGFP, pSN + DC + PEI-RTN3, pSN + DC + PEI-RTN3 + shEGFP and pSN + DC + PEI-RTN3 + shProtrudin-treated animals. (E) Mean error ratio to show the number of slips over the total number of steps in sham, DC + PEI-EGFP, pSN + DC + PEI-EGFP, pSN + DC + PEI-RTN3, pSN + DC + PEI-RTN3 + shEGFP and pSN + DC + PEI-RTN3 + shProtrudin-treated animals. Data are means ± SEM. n = 18 rats/group. *** p = 0.0001, one-way ANOVA with Dunnett’s post hoc test. # p = 0.0012, linear mixed models (LMM); ## p = 0.00012, generalised linear mixed models (GLMM).

Figure 8

RTN3 overexpression promotes RGC axon regeneration and functional recovery after ONC, which is also protrudin-dependent. (A) Western blot and (B) quantification to show that RTN3 plasmids significantly upregulate RTN3 protein in the retina after intravitreal injection in vivo. (C) Immunohistochemistry to localise RTN3 protein to the RGCs (green) in the ganglion cell layer (GCL) and the inner plexiform and inner nuclear layer (INL). ONL = outer nuclear layer. (D) High power images to show RTN3 protein localised to βIII-tubulin⁺ RGCs (arrows) in the GCL. (E) GAP43 immunohistochemistry and (F) quantification to show that in control optic nerve, few if any GAP43⁺ axons pass beyond the lesion site (*) whilst in RTN3 overexpressed eyes, significant GAP43⁺ axons are present beyond the lesion site, an effect which is obliterated when Protrudin is knocked out at the same time as RTN3 overexpression. (G) Representative images to show FluorGold (FG) backfilled RGC in retinal wholemounts after RTN3 overexpression. (H) Quantification of the number of FG⁺ RGC in retinal wholemounts shows that overexpression of RTN3 is significantly neuroprotective, an effect that is independent of protrudin. (I) Representative ERG traces and (J) quantification of the pSTR amplitude shows significant improvements in RTN3 overexpressed eyes which are ablated after protrudin knockdown. Arrows show the peak of the pSTR. Data are means ± SEM. n = 12 eyes/optic nerves/treatment. Scale bars in (C) = 100 µm, scale bars in (D) = 25 µm, scale bars in (E) = 200 µm, scale bars in (G) = 50 µm. *** p = 0.0001, one-way ANOVA with Dunnett’s post hoc test.

Figure 9

(A) Endoplasmic reticulum (ER)-localised protrudin forms contact sites with late endosomes (LEs) after detection of Rab7 and phosphoinositide 3 phosphate PI(3)P (yellow) and stopping the movement of LE (STOP). Kinesin-1 bound to protrudin is then handed over to the LE protein FYCO1, which mediates plus-end (+)-directed LE movement (MOVE) along microtubules (MT) to the periphery. Synaptotagmin 7 (SYT7)-mediated LE fusion with the plasma membrane then enables protrusion formation and neurite outgrowth, as previously proposed [14,57]. (B) Overexpression of RTN3 allows the handover of kinesin-1 to occur more rapidly and hence greater numbers of LEs available for protrusion (neurites/axons) at the plasma membrane.