Systemic administration of epothilone B promotes axon regeneration after spinal cord injury

See allHide authors and affiliations

Science  17 Apr 2015:
Vol. 348, Issue 6232, pp. 347-352
DOI: 10.1126/science.aaa2958

Progress toward fixing a broken back?

Axon regeneration after a spinal cord injury requires interference with neuronal mechanisms to promote axon extension and early suppression of scar formation. Microtubule stabilization could provide, in principle, a basis for such intervention. Ruschel et al. used animal models of spinal cord injury, time-lapse imaging in vivo, primary neuronal cultures, and behavioral studies to tackle this challenge (see the Perspective by Tran and Silver). They showed that epothilone B, a U.S. Food and Drug Administration–approved microtubule-stabilizing drug that can cross the blood-brain barrier, does promote functional axon regeneration, even after injury.

Science, this issue p. 347; see also p. 285


After central nervous system (CNS) injury, inhibitory factors in the lesion scar and poor axon growth potential prevent axon regeneration. Microtubule stabilization reduces scarring and promotes axon growth. However, the cellular mechanisms of this dual effect remain unclear. Here, delayed systemic administration of a blood-brain barrier–permeable microtubule-stabilizing drug, epothilone B (epoB), decreased scarring after rodent spinal cord injury (SCI) by abrogating polarization and directed migration of scar-forming fibroblasts. Conversely, epothilone B reactivated neuronal polarization by inducing concerted microtubule polymerization into the axon tip, which propelled axon growth through an inhibitory environment. Together, these drug-elicited effects promoted axon regeneration and improved motor function after SCI. With recent clinical approval, epothilones hold promise for clinical use after CNS injury.

An ideal treatment to induce axon regeneration in the injured central nervous system (CNS) should reduce scarring (1) and growth-inhibitory factors at the lesion site (24), reactivate the axon growth potential (5), and be administrable as a medication after injury. Recently, a number of combinatorial approaches have led to axon regeneration (6, 7). These approaches, however, involve multiple drugs, enzymes, and interventions, rendering clinical translation difficult. Moderate microtubule stabilization by the anticancer drug Taxol promotes axon regeneration by reducing fibrotic scarring and increasing axon growth (8, 9). However, it remains elusive how microtubule stabilization induces such divergent effects. Moreover, Taxol cannot be used for clinical CNS intervention because it does not cross the blood-brain barrier (10).

We aimed to target microtubule stabilization in the injured CNS in a clinically feasible way and to decipher its distinct cellular actions. We used epothilones, a class of U.S. Food and Drug Administration (FDA)–approved blood-brain barrier–permeable microtubule-stabilizing drugs (11). Mass spectrometry confirmed that after intraperitoneal (i.p.) injection in adult rats, epoB was rapidly absorbed into the CNS and remained at comparable levels for 6 days (Fig. 1A). Rats i.p. injected with 0.75 mg of epoB per kilogram of body weight (BW) at day 1 and 15 after injury showed increased amounts of detyrosinated and acetylated tubulin in lesion site extracts 4 weeks after spinal cord dorsal hemisection (Fig. 1B), indicating increased microtubule stability (12). The dosage used presented no obvious adverse side effects, such as reduced animal weight or decreased white blood cell counts (fig. S1).

Fig. 1 EpoB reduces inhibitory fibrotic scarring after SCI by abrogating meningeal fibroblast polarization and migration.

(A) Mass spectrometric analysis of CNS tissue and blood after a single i.p. injection of epoB; n = 4 rats per time point. (B) Immunoblots (IB) of indicated proteins in lesion extracts; n = 3 rats. (C) Human spinal cord after injury (asterisk); laminin immunolabeling. (D) Immunolabeling for laminin, glial fibrillary acidic protein (GFAP), or chondroitin sulfates (CS-56) after rat spinal cord hemisection. (E) Laminin-immunopositive (+) area at the lesion; n = 7 to 8 rats per group. (F) Glycosaminoglycan amounts in spinal cord lesion extracts; n = 8 rats per group. (G) Rat meningeal fibroblasts (RMFs) in wound-healing assays. (H) Percentage of the area shown in (G) occupied with RMFs after 48 hours, n = 3 experiments. (I and J) Immunolabeling of tyrosinated (TyrTub) and detyrosinated tubulin (DetyrTub, arrowheads). (K) IB of indicated proteins in RMFs 24 hours after treatment. (L) Immunolabeling for fibronectin, detyrosinated and tyrosinated tubulin [4′,6-diamidino-2-phenylindole (DAPI); nuclear staining] in the rat meninges at the lesion. Bottom panel, magnification of fibroblasts (arrowheads) in top panel. dpi, days postinjury. Scale bars, 50 μm. Schemes in (D) and (L) indicate lesion and displayed region (red box). Values are plotted as means + SEM. *P < 0.05, ***P < 0.001 by Student’s t test.

Fibrotic scar tissue rich in fibronectin and laminin forms at the lesion site after spinal cord injury (SCI) in rodents (8) and humans (Fig. 1C and table S1). This scar tissue poses a key impediment to regenerating axons, because it contains axon growth–inhibitory factors, including chondroitin sulfate proteoglycans (CSPGs) (1, 8). Adult rats systemically treated after injury with 0.75 mg/kg BW epoB showed a significant reduction of fibronectin (Fig. 1B) and of laminin-positive fibrotic scar tissue even 4 weeks after dorsal hemisection (Fig. 1, D and E). We found a comparable decrease of fibrotic scarring when epoB was locally delivered to the injury site via an intrathecal catheter (fig. S2) (8). Reduction of fibrotic scar tissue by systemic epoB administration was associated with a decrease of CSPGs (Fig. 1, D and F), including neurocan (Fig. 1B) and NG2 (13), at the injury site (fig. S3). Astrogliosis and lesion area were similar between treated and control animals (fig. S1), indicating that neuroprotective glial sealing of the injury site (14) was not affected by the treatment.

Scar reduction upon epoB treatment resulted neither from decreased cell proliferation nor from increased apoptosis (fig. S4) but from a migratory defect of scar-forming meningeal fibroblasts (15). In wound-healing assays, epoB inhibited migration of meningeal fibroblasts (Fig. 1, G and H, and movies S1 and S2) by changing their microtubular network. Control cells polarized by forming a leading edge enriched in stable detyrosinated microtubules and a trailing edge containing dynamic, tyrosinated microtubules (Fig. 1I), both hallmarks of directed cell migration (16). In contrast, epoB-treated fibroblasts were round and nonpolar (Fig. 1J and fig. S5) with elevated amounts of detyrosinated microtubules (Fig. 1K) distributed throughout the cell (Fig. 1J). Similarly, systemic administration of epoB after dorsal hemisection prevented the polarization of meningeal fibroblasts at the lesion site into a bipolar, migratory shape (Fig. 1L), which reduced scar formation (Fig. 1, D and E).

In cocultures of meningeal fibroblasts and postnatal cortical neurons, epoB treatment (1 nM) perturbed fibroblast polarization while enhancing axon growth (fig. S5). Moreover, epoB restored axon growth when these neurons were confronted with the inhibitory molecules Nogo-A, CSPGs, or Semaphorin 3A (Fig. 2, A and B), which are abundant at the spinal cord lesion site (24, 17). In neurons expressing fluorescently tagged microtubule plus-end–binding protein 3 (EB3-mCherry), which labels polymerizing microtubules (18), epoB induced rapid and concerted microtubule polymerization into the neurite tips (Fig. 2, C and D, and movie S3), causing axon elongation despite inhibitory Nogo-A (Fig. 2E and movie S3). In accordance, low doses of the microtubule-destabilizing drug nocodazole abolished microtubule protrusion in neurites (Fig. 2, D and F) and abrogated the growth-promoting effect of epoB (Fig. 2B). EpoB also promoted axon growth of human cortical neurons under growth permissive as well as nonpermissive conditions (fig. S6). In meningeal fibroblasts, however, epoB prevented microtubule polymerization toward the cell edges (Fig. 2G), contrasting with the microtubule dynamics found in neurons. This dichotomy was due to neuron-specific expression of the microtubule-associated protein Tau (fig. S7), which regulates microtubule dynamics, bundling, and binding of microtubule-stabilizing agents (19, 20). In fibroblasts ectopically expressing Tau, epoB induced an accumulation of bundled microtubules (fig. S8) that polymerized toward the cell edge (Fig. 2, H and I, and movie S4), mimicking the effect observed in neurons. In turn, neurons depleted of Tau, by transfection with a plasmid encoding short hairpin RNA for tau (21), showed reduced microtubule polymerization into the distal neurite when exposed to epoB (fig. S9).

Fig. 2 EpoB promotes microtubule protrusion and axon elongation in neurons while dampening microtubule dynamics in scar-forming fibroblasts.

(A) Beta-3 tubulin (Tuj-1) immunolabeling of neurons on inhibitory substrates (CSPGs, chondroitin sulfate proteoglycans; Sema 3A, Semaphorin 3A). (B) Neurite length of cortical neurons after 48 hours under indicated conditions; n = 3 to 4 experiments. (C) EB3-mCherry time-lapse projections in Nogo-A exposed neuron before and after epoB treatment (asterisks, stable landmarks). Bottom panels, high magnification of boxed areas in top panels. (D) EB3-mCherry fluorescence intensity in neurites under indicated conditions; n = 9 to 16 neurons (from three experiments). (E) Neurite growth on Nogo-A. Black arrowhead, time of indicated treatment. n = 12 to 15 neurons (from three experiments). (F and G) EB3-mCherry time-lapse projections of nocodazole-treated neuron (F) and epoB-treated meningeal fibroblast (G). Bottom panels, high magnification of boxed areas in top panels. (H) EB3-mCherry time-lapse projections before and after epoB treatment in cultured meningeal fibroblasts with (arrowhead) or without Tau-expression. Bottom panels, magnification of boxed areas in top panels. (I) EB3-mCherry fluorescence intensity in fibroblast periphery under indicated conditions; n = 20 cells per condition (from four experiments). Scale bars, 25 μm. Values are plotted as means [+ SEM in (B) and (E)]. *P < 0.05, **P < 0.01 by Student’s t test. n.s., not significant.

Injured axons in the rodent and human CNS form dystrophic retraction bulbs (Fig. 3, A to D, and table S2), a consequence of microtubule depolymerization and disorganization (Fig. 3, A and B) (22, 23). Because epoB induced microtubule polymerization and axon growth in cultured neurons, we assessed its ability to promote axon regeneration after SCI. In vivo imaging of adult transgenic mice, expressing green fluorescent protein (GFP) in spinal cord dorsal column axons (23, 24), revealed that transected axons of animals injected with 1.5 mg/kg BW epoB exhibited significantly fewer retraction bulbs (Fig. 3, C and D), reduced axonal dieback, and increased regenerative growth (Fig. 3, C and E). Moreover, in adult mice, systemic and postinjury treatment with epoB promoted axon regeneration after complete dorsal column transection (Fig. 3, F and G).

Fig. 3 EpoB reduces dystrophy and promotes regeneration of injured spinal cord axons.

(A) Electron microscope images of human SCI. (Top) Undamaged axon containing microtubules (black arrowheads). (Bottom) Retraction bulb (indicated by white arrowheads) without microtubules. (Middle) Magnification of boxed area in bottom panel. Scale bars, 500 nm. (B) Beta-3 tubulin (Tuj-1) immunolabeling of retraction bulbs in chronic human SCI. Scale bar, 10 μm. (C) Lesioned GFP-positive spinal cord axons in mice forming retraction bulbs (yellow arrowheads), dying back (red arrowheads), or regenerating (green arrowheads). Boxed area in top panels, displayed region in panels below. Scale bars, 100 μm. (D and E) Percentage of injured axons forming retraction bulbs (D) and distance between injured axons and injury site (E); n = 8 mice per group. Values are plotted as means + SEM. (F) Microruby-traced mouse dorsal column axons after injury (white arrowheads), laminin and GFAP immunolabeling (dashed line, lesion border). Scale bar, 100 μm. (G) Average distance between caudal lesion margin and injured axons in individual animals (circles) and group means (vertical bars) ± SEM. *P < 0.05, **P < 0.01 by Student’s t test.

We then tested whether the treatment also promoted axon regrowth of descending axons important for locomotion. In adult rats injected after injury with 0.75 mg/kg BW epoB, we found a threefold increase of serotonergic fibers caudal to a dorsal hemisection (Fig. 4, A and B). Increased serotonergic innervation strongly correlates with recovery of motor function after SCI (2527). Therefore, we asked whether the treatment improves walking of adult rats that underwent a moderate, mid-thoracic spinal cord contusion, a clinically relevant SCI model (28). After contusion injury, epoB administration (0.75 mg/kg BW) reduced fibrotic scarring at the injury site (fig. S10) and promoted serotonergic axon regrowth in the caudal spinal cord (Fig. 4, C and D). Moreover, epoB treatment increased stride length and gait regularity and reduced external rotation of the hind paws (fig. S11), indicating improved walking balance and coordination. Accordingly, epoB-treated animals showed a 50% reduction of foot misplacements on the horizontal ladder compared to injured controls (Fig. 4E and movies S5 and S6). These functional improvements were abrogated by pharmacological ablation of serotonergic innervation (Fig. 4, D and E, and movies S7 and S8) (25).

Fig. 4 EpoB promotes regrowth of raphespinal axons and improves walking after spinal cord contusion injury.

(A) Serotonin (5HT) immunolabeling (dashed line, lesion border) and (B) number of 5HT-labeled (+) fibers caudal to a spinal dorsal hemisection; n = 7 to 8 rats per group. (C) Coronal sections of the lumbar spinal cord after contusion injury. (Left panel) Coimmunostaining of 5HT, synaptophysin (Syn), and choline acetyltransferase (ChAT). (Right panels) Magnification of each marker in boxed area (left panel) visualizing serotonergic innervation of motor neurons (arrowheads). (D) Total length of 5HT-immunopositive fibers in the ventral horn (5,7-DHT, 5,7-dihydroxytryptamine); n = 4 (uninjured), 6 (7 dpi), 11 to 12 rats (56 and 70 dpi) per group. (E) Number of footfalls on the horizontal ladder; n = 10 to 11 rats per group. dpi, days postinjury. Scale bars, 50 μm. Schemes in (A) and (C) indicate lesion and displayed region (red box). Values are plotted as means + SEM. *P < 0.05; n.s., not significant by Student’s t test.

The finding that the stabilization of microtubules inhibits cell division established the usage of systemic microtubule-stabilizing agents as a therapeutic standard for the treatment of cancer (29). Here, at low doses, systemic administration of the microtubule-stabilizing agent epoB promoted functional recovery after SCI. Our approach differs from other experimental regenerative paradigms (17) by pharmacologically focusing on a single molecular target, the microtubules, yet overcoming multiple pathological obstacles. This is possible due to divergent effects of pharmacological microtubule stabilization on microtubule dynamics and, hence, the polarization of neurons and meningeal fibroblasts. This dual effect, and the efficacy after systemic and postinjury administration, give epothilones a promising translational perspective for treatment of the injured CNS.

Supplementary Materials

Materials and Methods

Figs. S1 to S11

Tables S1 and S2

Movies S1 to S8

References (3037)

References and Notes

  1. Acknowledgments: Materials and methods and other supporting materials are available on Science Online. We thank L. Meyn, K. Weisheit, D. Fleischer, and N. Thielen for technical assistance and animal care and C. Hill for teaching the spinal cord contusion injury model. We also thank C. Laskowski, C. H. Coles, A. Kania, M. Hübener, W. Jackson, and G. Tavosanis for critically reading and discussing the manuscript. We are grateful for support from the Human Spinal Cord Tissue Bank and the electron microscopy core at the Miami Project, as well as to B. Kakulas (University of Western Australia and Royal Perth) for providing anonymized postmortem sections after human spinal cord injury. This work was supported by NIH, International Foundation for Research in Paraplegia, Wings for Life, and Deutsche Forschungsgemeinschaft. H. Witte, A. Ertürk, F.H., and F.B. filed a patent on the use of microtubule-stabilizing compounds for the treatment of lesions of CNS axons (European Patent no. 1858498; European patent application EP 11 00 9155.0; U.S. patent application 11/908,118). The authors declare no competing financial interests.
View Abstract

Stay Connected to Science

Navigate This Article