1. Cell Biology
  2. Neuroscience
Download icon

Identification of a critical sulfation in chondroitin that inhibits axonal regeneration

  1. Craig S Pearson
  2. Caitlin P Mencio
  3. Amanda C Barber
  4. Keith R Martin  Is a corresponding author
  5. Herbert M Geller  Is a corresponding author
  1. National Institutes of Health, United States
  2. University of Cambridge, United Kingdom
Research Article
  • Cited 0
  • Views 869
  • Annotations
Cite as: eLife 2018;7:e37139 doi: 10.7554/eLife.37139

Abstract

The failure of mammalian CNS neurons to regenerate their axons derives from a combination of intrinsic deficits and extrinsic factors. Following injury, chondroitin sulfate proteoglycans (CSPGs) within the glial scar inhibit axonal regeneration, an action mediated by the sulfated glycosaminoglycan (GAG) chains of CSPGs, especially those with 4-sulfated (4S) sugars. Arylsulfatase B (ARSB) selectively cleaves 4S groups from the non-reducing ends of GAG chains without disrupting other, growth-permissive motifs. We demonstrate that ARSB is effective in reducing the inhibitory actions of CSPGs both in in vitro models of the glial scar and after optic nerve crush (ONC) in adult mice. ARSB is clinically approved for replacement therapy in patients with mucopolysaccharidosis VI and therefore represents an attractive candidate for translation to the human CNS.

https://doi.org/10.7554/eLife.37139.001

Introduction

There is an urgent need for therapies to treat CNS injuries. Acute insult often results in axonal degeneration, and therefore many experimental strategies aim to stimulate regeneration of damaged axons. These efforts have been informed by two insights: (1) the extrinsic environment of the adult CNS is hostile to axonal growth due to the formation of a glial scar, and (2) adult CNS neurons have lost their intrinsic ability to express axon-growth-promoting factors. The inhibitory properties of the glial scar are primarily due to extracellular matrix molecules, particularly chondroitin sulfate proteoglycans (CSPGs) (Burnside and Bradbury, 2014). Digesting CSPG GAG chains with the bacterial enzyme chondroitinase ABC (ChABC) has been shown to promote axonal extension in several experimental models (Bradbury and Carter, 2011; Zhao and Fawcett, 2013). In parallel, many approaches have endeavored to restore the intrinsic growth capability of CNS axons, most prominently through the activation of cell growth programs such as the PTEN/mTOR pathway (Park et al., 2008). However, such approaches are not directly translatable to human patients: ChABC has failed to reach clinical trials, and manipulation of tumor suppressor genes is likely to prove clinically questionable (Barber et al., 2017).

Increasing evidence supports a critical role for GAG chain sulfation in CSPG signaling. Global removal of sulfate groups from GAG chains eliminates the inhibitory actions of CSPGs in culture (Smith-Thomas et al., 1995), and specific sulfation motifs dictate whether GAGs inhibit or permit axon growth. For instance, axons grow readily over surfaces coated with 6-sulfated (6S) CSPGs (Wang et al., 2008), and deleting the enzyme that adds 6S to CS GAGs impairs axonal regeneration in mice (Lin et al., 2011). In contrast, axons avoid 4-sulfated (4S) CSPGs, an effect abolished by treatment with 4-sulfatase (Wang et al., 2008). Sulfation at both positions (4,6S) has been shown to inhibit axonal growth in vitro and in vivo (Brown et al., 2012), and an increase in the ratio of 4S to 6S has been linked with age-related declines in plasticity (Foscarin et al., 2017; Miyata et al., 2012). Collectively, these observations suggest that reducing 4S while preserving 6S on intact GAG chains may enable growing axons to overcome CSPG-mediated inhibition more effectively than indiscriminate reductions in sulfation or destruction of GAGs.

We have previously shown that the deposition of CSPGs in the glial scar following brain and spinal cord injury is dominated by 4S GAGs (Wang et al., 2008). Here, we demonstrate the ability of arylsulfatase B (ARSB), a clinically approved enzyme that selectively removes 4S groups from the non-reducing ends of CSPGs (Litjens and Hopwood, 2001), to enhance axonal regeneration in vitro and in vivo. Treating CSPGs with ARSB or adding ARSB to TGF-β-treated astrocytes reverses their inhibition of neurite outgrowth by modifying CSPG sulfation. We then show that ARSB significantly enhances axon regeneration in vivo when the enzyme is delivered to the injured optic nerves of adult mice in combination with intravitreal injection of Zymosan and CPT-cAMP, an intrinsic growth stimulus. Importantly, this treatment is effective when administered several days after ONC, making it relevant for human conditions where interventions are rarely available immediately following injury. Crucially, ARSB (Naglazyme, Biomarin) is clinically approved for the treatment of mucopolysaccharidosis VI, a lysosomal storage disorder (Muñoz-Rojas et al., 2010; Harmatz et al., 2004; Harmatz et al., 2005); thus, its inclusion in future human therapeutic treatments is plausible. Taken together, these data establish a critical role for 4S at the non-reducing end of CS GAG chains in mediating the inhibitory actions of CSPGs. Moreover, they provide evidence for a promising translatable therapy that utilizes a highly selective human enzyme to modify chondroitin sulfation and enhance axon regeneration.

Results

ARSB reverses the inhibition of neurite growth caused by 4-sulfated CSPGs

The ability of ARSB to alter the inhibitory actions of CSPGs was first assessed in cell culture models of the glial scar (Wang et al., 2008). To assess whether neurite inhibition by CSPGs could be reduced through ARSB treatment, cultures of dissociated mouse hippocampal neurons were exposed to 5 µg/ml CSPGs with and without ARSB treatment for 48 hr. Cultures were stained for βIII-tubulin, and the lengths of neurites were measured. Neurons grown in the presence of CSPGs were significantly (p<0.0001) shorter than untreated neurons (neurite length [median]: 55.7 µm and 91.7 µm, respectively) (Figure 1c–d). Growth was not significantly altered by CSPGs that had been treated with ARSB (neurite length [median]: 93.2 µm), suggesting that ARSB treatment was sufficient to remove neurite outgrowth inhibiting characteristics of CSPGs (Figure 1c–d).

Figure 1 with 1 supplement see all
ARSB reverses neurite outgrowth inhibition caused by 4-sulfated CSPGs.

(a) Schematic diagram showing actions of ARSB and ChABC on GAG chains. (b) Western blot showing CS-56 signal in conditioned medium. (c) Micrographs showing hippocampal neurons treated with no treatment, CSPG (5 µg/ml), or CSPG +ARSB. Scale bar = 25 µm. (d) Plot showing lengths of longest neurite measured from β-III-tubulin stained neurons. Statistical significance was determined by one-way ANOVA, **p<0.005, ***p<0.001. (e) Micrographs showing co-cultures of CGNs grown on astrocytes and treated with TGF-β, TGF-β and ARSB, or no treatment. Scale bar = 25 µm. (f) Plot showing lengths of longest neurite measured from β-III-tubulin stained neurons. Statistical significance was determined by one-way ANOVA, *p<0.05, **p<0.005, ***p<0.001, ****p<0.0001.

https://doi.org/10.7554/eLife.37139.002

To test the actions of ARSB in a cellular model, monolayers of confluent mouse astrocytes were treated with TGF-β to stimulate elevated CSPG production (Wang et al., 2008). Mouse cerebellar granule neurons (CGNs) were then seeded onto these astrocytes and allowed to grow for 24 hr. Cultures were stained for GFAP and βIII-tubulin, and the lengths of CGN neurites were measured. Neurons growing on TGF-β-treated astrocytes exhibited significantly lower neurite outgrowth than those plated on untreated control astrocytes (p=0.0059, Mann-Whitney U test) (Figure 1e–f). However, incubating TGF-β-treated co-cultures with ARSB restored average neurite length to the levels observed in untreated controls (Figure 1e–f), significantly different from TGF-β-treatment alone (p<0.0001). This suggests that cleaving 4S from the non-reducing ends of GAG chains is sufficient to neutralize the inhibitory effects of CSPGs on neurons.

To demonstrate that ARSB acts on extracellular CSPGs, rather than being internalized into astrocytes and interfering with CSPG production or secretion, conditioned medium (CM) was collected from TGF-β-treated astrocytes and left untreated, treated with ARSB, or treated with ChABC. The isolated and treated CM was added to separately cultured CGNs. Application of CM from TGF-β-treated astrocytes significantly reduced neurite outgrowth while ARSB treatment reversed this effect to a degree equivalent to ChABC (Figure 1—figure supplement 1). Together, these findings demonstrate that the presence of CSPGs can inhibit neurite outgrowth, and that this inhibition is overcome by exposing the CSPGs to either ARSB or ChABC.

To further validate that ARSB does not interfere with CSPG secretion, the level of CSPGs in CM was measured by immunoblotting with the antibody CS-56, which reacts with 4S and 6S groups on GAG chains (Avnur and Geiger, 1984). The increase in CSPGs caused by TGF-β treatment (Wang et al., 2008) was not altered by treatment with ARSB, even after repeated additions (Figure 1b), indicating that its enhancement of neurite growth was derived from modifying the sulfation pattern rather than attenuating CSPG production or secretion. These data also demonstrate that CS-56 immunoreactivity is not altered by removal of 4S from the non-reducing end of CS GAG chains.

Optic nerve crush leads to astrogliosis and sustained elevation of CSPGs

Injury to the CNS is accompanied by astrogliosis, characterized by the accumulation of CSPGs in a glial scar (Burnside and Bradbury, 2014). Evidence suggests a similar effect in the optic nerve (Brown et al., 2012; Sellés-Navarro et al., 2001; Sengottuvel et al., 2011; Qu and Jakobs, 2013), but a comprehensive examination of this phenomenon has not been performed, especially regarding the production of 4S GAG chains following injury. A cohort of adult mice received optic nerve crush (ONC) or non-lesioned sham surgery (in which the nerve was exposed but not crushed) and optic nerves were collected at 1, 3, 5, 7, 14, and 21 days post crush (dpc). Optic nerve sections were stained with antibodies against GFAP (to identify reactive astrocytes) and Iba1 (to identify microglia and macrophages). By 7 dpc, GFAP+ reactive astrocytes had retracted from the lesion site to form a cavity, which was densely populated with Iba1+ microglia and macrophages (Figure 2—figure supplement 1). Astrocytes displayed reactive morphology, with elongated processes defining the lesion boundary. Some GFAP+ cells were also found within the lesion core. Microglia in and around the lesion displayed an activated morphology, with enlarged cell bodies and retracted processes, distinct from the striated morphology of cells found distal to the injury site and in non-lesioned sham control nerves. By 21 dpc, astrocytes had begun to repopulate the cavity and form a chronic scar.

CSPGs were detected using CS-56 and 2H6, an antibody that reacts predominantly with 4S (Yamamoto et al., 1995), and to a lesser degree, with 6S (Sugiura et al., 2012) and 2,6S (Matsushita et al., 2018). In non-lesioned sham control nerves, CSPGs were evenly distributed within the tissue (Figure 2a). Elevated CS-56 levels at the lesion site were first observed at 5 dpc and peaked around 7 dpc (Figure 2b). Levels remained high at 21 dpc. An increase in 2H6 staining was also observed, with levels reaching 2.5-fold those in non-lesioned sham controls (fold change [mean ± SE]: 2.53 ± 0.15) (Figure 2b). The axons of injured mouse RGCs visualized with fluorescently-tagged CTB, injected intravitreally 1 d prior to tissue harvest, failed to traverse the injury site and instead formed dystrophic endbulbs that appeared to be associated with areas of high CSPG deposition, which included areas of high 4S immunostaining (Figure 2c). CSPGs and 4S GAGs were associated with both GFAP+ and Iba1+ cells in nerves examined at 7 dpc (Figure 2—figure supplement 2). Taken together, these results illustrate that ONC in mice leads to astrogliosis and elevated expression of CSPGs, especially those with 4S GAGs, which is sustained for at least 21 days.

Figure 2 with 2 supplements see all
Optic nerve crush stimulates glial scar formation and sustained elevation of chondroitin sulfate proteoglycans.

(a) Micrographs show lesioned optic nerve tissue collected at 1, 3, 5, 7, 14 and 21 dpc and stained for CSPGs (CS-56), 4S GAGs (2H6), reactive astrocytes (GFAP), and microglia and macrophages (Iba1). Scale bar = 50 µm. (b) Fluorescence intensity of CS-56 and 2H6 immunostaining expressed as fold change vs. non-lesioned sham controls. Statistical significance versus sham was determined by Student’s t-test. *p<0.05, **p<0.005. Colored asterisks indicate significance for different groups (CS-56 = green, 2H6 = magenta). (c) Micrographs showing lesioned mouse optic nerve tissue at 7 dpc. Axons are visualized with CTB and form dystrophic endbulbs in areas of high CSPG and 4S GAG immunoreactivity. Scale bar = 100 µm, inset = 10 µm.

https://doi.org/10.7554/eLife.37139.004

Modifying CSPG sulfation enhances retinal ganglion cell axon regeneration

Given the in vitro evidence that 4S is critical to CSPG-mediated inhibition of neurite growth, we investigated whether cleaving 4S from the non-reducing ends of GAG chains at the ONC lesion site would enhance retinal ganglion cell (RGC) axon regeneration in the optic nerve. To accomplish this, an intrinsic pro-regenerative stimulus, Zymosan A and CPT-cAMP (Leon et al., 2000; Yin et al., 2003), was combined with direct application of ARSB to the lesioned nerve. ChABC was used as a control to evaluate the effects of digesting GAG chains entirely rather than selectively removing 4S groups.

Mice received ONC, followed 3 days later by an intravitreal injection of Zymosan A (12.5 µg/µL) supplemented with CPT-cAMP (50 mM), followed immediately by implantation of a gelfoam scaffold loaded with 5 μL of ARSB (1 mg/mL), ChABC (455 µg/mL), or control buffer. At 14 dpc, optic nerves were dissected, sectioned, and stained for GAP-43 to detect regenerating axons. In accordance with previous reports (Leaver et al., 2006), we found that GAP-43 selectively labels regenerating axons, as GAP-43 signal is absent from intact, non-lesioned optic nerves (data not shown). On its own, injection of Zymosan/CPT-cAMP induced significantly (p=0.0226) higher RGC axon regeneration than PBS controls at 14 dpc (axons at 0.25 mm distal to the lesion [mean ±SE]: 282 ± 83.4 and 42.3 ± 11.1, respectively) (Figure 3—figure supplement 1). Zymosan did not alter CSPG expression at the lesion site (Figure 3—figure supplement 1). When Zymosan was combined with enzyme delivery, both ARSB and ChABC significantly (p=0.0006 and p<0.0001, respectively) enhanced RGC axon regeneration compared with the buffer control (axons at 0.25 mm distal to the lesion [mean ±SE]: 472 ± 62, 535 ± 123, and 217 ± 53, respectively) (Figure 3). Interestingly, delivering ARSB or ChABC in the absence of Zymosan injection did not enhance basal RGC axon regeneration (Figure 3—figure supplement 2).

Figure 3 with 5 supplements see all
Selectively targeting inhibitory CSPGs enhances retinal ganglion cell axon regeneration.

(a) Experiment timeline and schematic diagram showing intravitreal injection of Zymosan/CPT-cAMP and implantation of gelfoam scaffolds containing ARSB, ChABC, or control buffer. (b) Micrographs showing GAP-43-labeled optic nerves from mice treated with Zymosan/CPT-cAMP and gelfoam scaffolds loaded with ARSB, ChABC, or control buffer. Arrows indicate lesion site. Scale bar = 200 µm. (c) Graph showing the number of regenerating axons at distances distal to the lesion site, displayed as mean ± SEM. Statistical significance was determined by two-way ANOVA with Bonferroni post-hoc test for multiple comparisons. *p<0.05, **p<0.005, ***p<0.001, ****p<0.0001. Colored asterisks indicate statistical significance for different groups (ARSB = magenta, ChABC = green). (d) Graph showing average length of longest GAP-43+ regenerating axon. Statistical significance was determined by Student’s t-test. *p<0.05.

https://doi.org/10.7554/eLife.37139.007

The products of the reaction catalyzed by ARSB are not readily detectable by immunohistochemistry or Western blot; therefore, to specifically validate the penetration of ARSB into the optic nerve fibers, mice received ONC surgery, and gelfoam scaffolds soaked in 200 µg/mL His-Tagged ARSB or control buffer were implanted behind the eyes at the ONC lesion site (Figure 3—figure supplement 3). Tissue collected at 1 dpc was analyzed by immunohistochemistry using anti-His antibody, and recovered scaffolds were tested for the presence of active ARSB. His-Tagged ARSB was detected in lesioned tissue using immunohistochemistry, and active enzyme was detected from recovered scaffolds (Figure 3—figure supplement 3). To further validate that the enzymes had successfully penetrated the optic nerve and modified CSPGs, we stained ChABC-treated samples with the antibody BE-123, which recognizes the ‘stubs’ produced on proteoglycans by ChABC digestion of the GAG chains. Western blot analysis of non-lesioned sham control tissue treated with ChABC revealed BE-123 signal exclusively in nerve segments exposed to ChABC-loaded scaffolds (Figure 3—figure supplement 4). ARSB treatment did not significantly alter immunoreactivity of CS-56 or 2H6 (Figure 3—figure supplement 5). Together, these observations establish that the enzymes released from the scaffold penetrate the tissue and digest GAG chains.

ARSB promotes axon regeneration with an extended therapeutic window

The duration of the regeneration enhancing effects of ARSB was assessed by measuring axon regeneration at early and late time points. At 7 dpc, only 4 days after implantation of the gelfoam scaffolds, a small but significant (p=0.0149) increase in the number of axons navigating through the lesion site was already detectable in the ARSB-treated group compared with the buffer control (axons at 0.50 mm distal to the lesion [mean ± SE]: 69.2 ± 12.3 and 16.0 ± 8.9, respectively) (Figure 4a–d). By 28 dpc, regenerating axons were found extending as far as 4.0 mm beyond the lesion site, to the optic chiasm entry point (Figure 4e–g). There was a significant (p=0.0002) increase in the number of axons in ARSB-treated animals versus buffer-treated controls (axons at 0.25 mm distal to the lesion [mean ± SE]: 568 ± 96.3 and 273 ± 63.0, respectively). The enhancing effect of ARSB treatment appeared to be concentrated at distances proximal to the lesion site (0.25–1.50 mm). At distances beyond 1.50 mm, there was relatively little difference between the ARSB-treated and buffer-treated groups (Figure 4—figure supplement 1). We isolated this effect by subtracting the number of regenerating axons in the Zymosan/buffer groups from those in the Zymosan/ARSB groups (Figure 4—figure supplement 1). ARSB strongly increased the number of axons regenerating through the lesion site but did not appear to substantially extend the distances of axons that were already regenerating.

Figure 4 with 1 supplement see all
ARSB enhances axon regeneration over an extended therapeutic window.

(a) Experiment timeline and schematic diagram showing intravitreal injection of Zymosan and CPT-cAMP and delivery of ARSB and control buffer to the lesioned optic nerve via implanted gelfoam scaffold. (b) Micrographs showing GAP-43-labeled optic nerves from mice treated with Zymosan/CPT-cAMP and gelfoam scaffolds loaded with ARSB or a control buffer. Arrows indicate lesion site. Scale bar = 200 µm. (c) Graph showing the number of regenerating axons at distances distal to the lesion site, displayed as mean ± SEM. Statistical significance was determined by two-way ANOVA with Bonferroni post-hoc test for multiple comparisons. *p<0.05. (d) Graph showing length of the longest regenerating axon, displayed as mean ± SEM. Statistical significance was determined by Student’s t-test. (e) Micrographs showing GAP-43-labeled optic nerves from mice treated with intravitreal injections of Zymosan and gelfoam scaffold loaded with ARSB or a control buffer. Arrows indicate lesion site. Scale bar = 200 µm. (f) Graph showing the number of regenerating axons at distances distal to the lesion site, displayed as mean ± SEM. Statistical significance was determined by two-way ANOVA with Bonferroni post-hoc test for multiple comparisons. **p<0.005, ***p<0.001, ****p<0.0001. (g) Graph showing length of the longest regenerating axon, displayed as mean ± SEM. Statistical significance was determined by Student’s t-test.

https://doi.org/10.7554/eLife.37139.013

ARSB does not alter the astrocytic scar or perineuronal nets

To determine whether treatment with ARSB alters glial cells at the lesion site, tissue from enzyme-treated nerves was stained with GFAP and Iba1. Neither ARSB nor ChABC disrupted formation of the astrocytic scar. The area delineated by GFAP+ astrocytes decreased over time but was not significantly different between treatment groups at any time point (Figure 5c–d). Correspondingly, the total GFAP immunoreactivity increased from 7 to 28 dpc as astrocytes repopulated the glial scar region, but no differences were observed between treatment groups (Figure 5e). Both ChABC and ARSB increased Iba1 immunoreactivity relative to the buffer control (fluorescence intensity [mean ± SE]: 21.7 ± 2.95, 12.9 ± 1.71, and 6.96 ± 1.79, respectively), but ChABC elicited significantly (p<0.05) higher Iba1 immunoreactivity than ARSB (Figure 5a–b).

Figure 5 with 1 supplement see all
ARSB provokes muted immune response but does not alter astrocyte reactivity, glial scar size, or association of regenerating axons with astrocyte processes.

(a) Micrographs showing Iba1 immunostaining at the optic nerve crush site for samples treated with Zymosan/CPT-cAMP and ChABC, Zymosan/CPT-cAMP and ARSB, Zymosan/CPT-cAMP and a control buffer, no treatment, and non-lesioned controls. Scale bar = 50 µm. (b) Graph showing quantification of Iba1 fluorescence intensity measured as % area of thresholded insets centered at the lesion site. Statistical significance was determined by Student’s t-test. *p<0.05, ***p<0.001. (c) Micrographs showing GFAP and GAP-43 immunostaining at the optic nerve crush site for samples treated with Zymosan/CPT-cAMP and either ARSB, ChABC, or a control buffer and analyzed at 7, 14, and 28 dpc. Arrows indicate lesion site. Scale bar = 100 µm. (d) Graph showing quantification of glial scar size measured as the area delineated by GFAP+ astrocytes at the optic nerve crush site. Statistical significance was determined by Student’s t-test. (e) Graph showing quantification of GFAP immunoreactivity at the optic nerve crush site. Statistical significance was determined by Student’s t-test.

https://doi.org/10.7554/eLife.37139.015

In addition to their deposition in the glial scar, CSPGs are a major component of perineuronal nets (PNNs), structures that limit synaptic plasticity in the brain and spinal cord but are not present in the optic nerve. ChABC is known to disturb PNNs and alter plasticity in the visual cortex (Pizzorusso et al., 2002). To evaluate whether ARSB alters CSPG structure beyond the selective cleavage of 4S groups, we incubated post-fixed mouse brain tissue sections with ARSB (1 mg/mL), ChABC (≥20 µg/mL), and buffer control, and detected perineuronal nets (PNNs) with Wisteria floribunda agglutinin (WFA). ChABC completely eliminated WFA-stained PNNs (Figure 5—figure supplement 1). However, incubation with ARSB left PNNs intact, with no observable differences from PNNs in buffer-treated brain tissue (Figure 5—figure supplement 1).

Discussion

The glial scar is considered a major impediment to axonal regeneration. We show that the injured optic nerve develops a glial scar rich in CSPGs, including the axon-inhibiting 4S motif. The human enzyme ARSB selectively cleaves 4S groups from the non-reducing ends of GAG chains, reducing CSPG-mediated inhibition of neurite growth in vitro. We demonstrate that ARSB promotes neurite growth in culture without altering production or secretion of GAG chains. Furthermore, ARSB enhances the regeneration of RGC axons following optic nerve injury. The treatment is robustly effective even when administered 3 days after injury, an important consideration for translational therapies. Enhanced regeneration was evident as early as 7 days post ONC and remained significant at 28 days, illustrating an extended therapeutic window from a single treatment. ARSB is active in vivo, provokes less Iba1 immunoreactivity than ChABC, and preserves perineuronal structures that depend on intact GAG chains. Taken together, these findings demonstrate that the 4S motif at the non-reducing end of CS GAG chains plays a major role in mediating the inhibitory actions of CSPGs. Given the approval for ARSB as an enzyme replacement therapy in human patients, our evidence that ARSB enhances axon regeneration in the optic nerve means that future treatments could readily combine ARSB with clinically viable intrinsic approaches to achieve robust regeneration of damaged or degenerated axons in the CNS.

Sulfation dictates the effects of CSPGs on axon growth

Studies that link CSPGs to the failure of axon regeneration overwhelmingly fail to distinguish between differentially sulfated GAG chains, often showing instead that digestion of GAG chains with ChABC enhances neurite growth in vitro and axon regeneration in vivo (Bradbury and Carter, 2011). The importance of sulfation in governing CSPG function has been demonstrated using sodium chlorate, which broadly eliminates GAG sulfation (Smith-Thomas et al., 1995). Recent studies have characterized the behaviors of specific sulfation motifs, showing that both 4S and 4,6S inhibit neurite growth while 6S is growth-permissive (Wang et al., 2008; Brown et al., 2012). An age-related increase in the ratio of 4S to 6S was linked to declines in plasticity and memory (Foscarin et al., 2017; Miyata et al., 2012), and removal of 4S with ARSB improved motor function following spinal cord injury (Yoo et al., 2013). Blocking 4,6S with a custom antibody enhanced regeneration of RGC axons after ONC (Brown et al., 2012), which raises the question of whether 4S and 4,6S function similarly to inhibit axonal growth, and whether ARSB might convert 4,6S motifs to 6S.

The precise mechanism of how ARSB modifies the inhibitory actions of GAG chains is unknown. ARSB did not reduce the total amount of sulfated GAG in the culture medium as detected by the anti-CS antibodies, suggesting that its effects are mediated by altering GAG chain sulfation. ARSB, a lysosomal enzyme, maintains its highest activity at acidic pH, raising the question of whether it can cleave sulfate groups from secreted CSPGs, or whether lysosomal uptake is required. We observed that ARSB cleaves 4S from extracellular GAG chains in culture medium, suggesting that its activity at neutral pH is sufficient to perform its sulfatase function. This was validated by our discovery that ARSB promotes regeneration of optic nerve axons when administered exogenously. The prominent actions of ARSB are more remarkable considering that the average length of neuronal GAG chains is about 50 disaccharide units (Rauch et al., 1991). Removal of just the 4S at the non-reducing end leaves virtually the entire GAG chain intact, as demonstrated by the preservation of the immunoreactivity to CS-56, while the inhibitory activity is significantly diminished.

CSPG deposition is a key source of axon growth inhibition in the glial scar

The formation of a glial scar, including deposition of sulfated proteoglycans, is well documented in the brain and spinal cord (Bradbury et al., 2002; Bradbury and Carter, 2011; Yi et al., 2012; Burnside and Bradbury, 2014; Galtrey and Fawcett, 2007). Glial activation and macrophage recruitment have been observed in optic nerve lesions (Qu and Jakobs, 2013), and some studies have suggested that CSPGs are upregulated after ONC, but have not quantified this phenomenon or explored its time course (Brown et al., 2012; Sengottuvel et al., 2011; Sellés-Navarro et al., 2001). We found that CSPGs, and 4S GAGs in particular, were significantly elevated after ONC, reaching peak levels at 7 dpc. The sustained elevation of CSPGs at 21 dpc suggests that the optic nerve environment remains hostile to axon growth for extended periods after injury.

Cleaving 4S from the non-reducing ends of GAG chains with ARSB, or digesting GAG chains with ChABC, both enhanced axon regeneration without disrupting formation of the astrocytic scar. This supports a critical role for CSPG deposition, rather than scar formation per se, as the primary cause of axon growth inhibition. This is consistent with findings that ablating astrocytic scar formation without reducing CSPG levels does not lead to spontaneous regeneration of axons (Anderson et al., 2016; Silver, 2016). Intriguingly, blocking the transformation of reactive astrocytes into scar-forming astrocytes, which express elevated levels of CSPG-related transcripts, was found to significantly enhance axon regeneration (Hara et al., 2017).

Delayed application of ARSB promotes regeneration and preserves perineuronal structures

Most experimental therapies that stimulate RGC axon regeneration involve interventions at the time of injury or, in the case of many gene therapies, prior to injury (Buch et al., 2008). While such studies are valuable for identifying therapeutic targets and elucidating mechanisms of RGC axon regeneration, they are not readily translatable to human patients. We found that delivery of ARSB in conjunction with Zymosan/CPT-cAMP significantly enhanced RGC axon regeneration when administered 3 days after ONC, making a strong case for its clinical viability.

Delaying ARSB treatment may confer additional advantages: previous studies have argued that CSPG synthesis in the acute phase (0–2 dpi) may, in fact, promote recovery. Rolls et al., 2008 blocked CSPG synthesis immediately after spinal cord injury in mice and found that axon regeneration and functional recovery were impaired, whereas blocking synthesis in the subacute phase (2–7 dpi) enhanced regeneration and recovery. Other studies suggest that intact CSPGs recruit blood-borne monocytes and bias macrophages toward a resolving phenotype (Shechter et al., 2011), and that CSPGs regulate the spatial organization of microglia and macrophages and promote neurotrophic factor production by resident microglia (Rolls et al., 2008; Shechter et al., 2009). Stripping CSPGs of their GAG chains may impede these repair functions, whereas selectively modifying sulfation with ARSB could reduce GAG-mediated inhibition of neurons without disrupting their interactions with other cells. To demonstrate that ARSB preserves perineuronal structures composed of CSPGs, we treated mouse cortical tissue with ARSB and ChABC and showed that while ChABC eliminated perineuronal nets (PNNs), ARSB left PNNs intact.

ARSB is superior to ChABC in several other respects. ARSB has relatively lower immunogenicity and maintains its activity longer than ChABC in vitro (Yoo et al., 2013). In vivo studies indicate that ChABC injected into rat brains maintains activity for at least 10 days (Lin et al., 2008), and that low levels of ChABC activity suppress CSPG levels for weeks (Chau et al., 2004; Hyatt et al., 2010). While the durability of ARSB in vivo has not been characterized, ARSB is stable at physiological temperature and pH, meaning it should retain robust levels of activity for extended periods (Yoo et al., 2013). Crucially, ARSB is a human enzyme with preexisting approval for clinical use (Muñoz-Rojas et al., 2010; Harmatz et al., 2004; Harmatz et al., 2005).

Combining extrinsic and intrinsic stimuli enhances axon regeneration

Treating lesioned optic nerves with ARSB or ChABC alone failed to enhance regeneration, but combining them with Zymosan/CPT-cAMP promoted significantly greater regeneration than the intrinsic treatment alone. Most studies demonstrating long distance regeneration of RGC axons achieve their effects by modifying the intrinsic state of RGCs: knocking out the tumor suppressor PTEN (Park et al., 2008), delivering growth factors (Sieving et al., 2006), stimulating inflammatory pathways (Yin et al., 2003), enhancing the endogenous activity of RGCs (Lim et al., 2016), chelating neurotoxic ions in the retina (Li et al., 2017), and various combinations thereof. However, knowledge of how these regenerating axons traverse the glial scar and navigate the growth-inhibitory microenvironment is incomplete. Studies that have examined the three-dimensional growth patterns of regenerating RGC axons consistently find that axons induced to regenerate via intrinsic manipulations display highly irregular and aberrant growth patterns (Luo et al., 2013; Bray et al., 2017; Fischer et al., 2017). Understanding how axons respond to their extrinsic microenvironment, particularly GAG chains within the glial scar, will be vital to future efforts to stimulate robust long-distance regeneration of retinal neurons and successful innervation of visual targets in the brain.

Materials and methods

Key resources table
Reagent type (species)
or resource
DesignationSource or referenceIdentifiersAdditional information
AntibodyDAPIThermoFisher
Scientific
D3751;
RRID: AB_2307445
1/10,000
Antibodyanti-GAP-43 (Sheep)Benowitz et al., 1988;
PMID: 3339416
1/50,000;
Gift from Benowitz lab
Antibodyanti-GAP43
(Rabbit, polyclonal)
Abcamab7462;
RRID: AB_305932
1/500
AntibodyCS-56
(Mouse, monoclonal)
Millipore SigmaC8035;
RRID: AB_476879
1/500
Antibody2H6
(Mouse, monoclonal)
Amsbio370710-IEC1/500
AntibodyBE-123
(Mouse, monoclonal)
Millipore SigmaMAB2030;
RRID: AB_94510
1/500
AntibodyIba1
(Rabbit, polyclonal)
FUJIFILM Wako
Chemicals USA
019–19741;
RRID: AB_839504
1/500
AntibodyGFAP
(Rabbit, polyclonal)
Agilent (Dako)Z0334;
RRID: AB_10013382
1/500
AntibodyGFAP
(Chicken, polyclonal)
Abcamab74674;
RRID: AB_304558
1/500
Antibodyβ-III-tubulin
(Mouse, monoclonal)
Millipore SigmaT8660;
RRID: AB_477590
1/1000
Antibody6x His tagAbcamab1378391/500
AntibodyWFAMillipore SigmaL1516;
RRID: AB_2620171
1/500
AntibodyDonkey anti-sheep,
Alexa 488 (secondary)
ThermoFisher
Scientific
A-11015;
RRID: AB_2534082
1/1000
AntibodyDonkey anti-sheep,
Alexa 568 (secondary)
ThermoFisher
Scientific
A-21099;
RRID: AB_2535753
1/1000
AntibodyGoat anti-rabbit,
Oregon Green
488 (secondary)
ThermoFisher
Scientific
O-6381;
RRID: AB_2539800
1/1000
AntibodyGoat anti-rabbit,
Alexa 633 (secondary)
ThermoFisher
Scientific
A-21070;
RRID: AB_2535731
1/1000
AntibodyGoat anti-chicken,
Alexa 488 (secondary)
ThermoFisher
Scientific
A-11039;
RRID: AB_2534096
1/1000
AntibodyGoat anti-mouse,
Alexa 568 (secondary)
ThermoFisher
Scientific
A-11004;
RRID: AB_2534072
1/1000
AntibodyGoat anti-mouse
IgM mu chain,
Dylight 650
(secondary)
Abcamab98749;
RRID: AB_10672799
1/500
AntibodyTRITC-conjugated
streptavidin
(secondary)
Jackson
ImmunoResearch
016-020-081/1000
Recombinant
Protein
TGF-β (human)PeproTech100–21C
Recombinant
Protein
ARSB (human)BioMarin
Pharmaceuticals
Naglazyme, Provided
by BioMarin
Recombinant
Protein
ARSB (human)R and D Systems4415-SU-010
Recombinant
Protein
Cholera Toxin
Subunit B, Alexa 555
ThermoFisher
Scientific
CC22843
ProteinChABCMillipore SigmaC3667
ProteinCSPG (Chicken
Extracellular Chondroitin
Sulfate Proteoglycans)
Millipore SigmaCC117
Chemical
compound, drug
Zymosan AMillipore SigmaZ4250
Chemical
compound, drug
CPT-cAMPMillipore SigmaC3912
Chemical
compound, drug
PNCSAlfa AesarB23325
Chemical
compound, other
Can Get Signal
(Immunoenhancer)
CosmoBioTYB-NKB-101
Software,
algorithm
GraphPad Prism 7RRID:SCR_002798
Software,
algorithm
Adobe Illustrator
CC 2017
RRID:SCR_010279
Software,
algorithm
FijiRRID:SCR_002285
OtherGelfoamEthicon- Johnson
and Johnson
1972

Animals

All experiments and procedures were performed in accordance with protocols approved by the Institutional Animal Care and Use Committee (IACUC) at the National Institutes of Health. Female C57Bl/6 mice aged 6–8 weeks (Charles River) were housed in a pathogen free facility with free access to food and a standard 12 hr light/dark cycle. Sample sizes were determined by statistical power calculations from pilot experiments and the results of previous studies, as described below. Animals were randomly allocated into experimental groups. Animals were removed from the study if bleeding occurred during the optic nerve crush or scaffold implantation surgery.

Cell culture

Primary hippocampal neuron cultures were prepared from embryonic (e17-18) C57Bl/6 mouse brains. Hippocampi were dissected and dissociated into single cell suspensions. Dissociated cells were seeded onto coverslips coated with poly-L-lysine and cultured in 500 μL Neurobasal medium containing B27 supplement (Thermo Fisher) and 24 mM KCl. After allowing 2 hr for neuronal attachment, 500 μL of Neurobasal medium containing B27 supplement and 24 mM KCl that had been incubated for 4 hr with no treatment, 10 µg/ml CSPG (for final concentration of 5 µg/ml), or CSPG (10 µg/ml)+ARSB (2 µg/ml) (final concentrations 5 µg/ml and 1 µg/ml, respectively) was added. Cells were incubated for 48 hr at 37°C and 5% CO2 atmosphere and then fixed and stained for DAPI and βIII-tubulin.

Primary cortical astrocyte cultures were prepared from neonatal (1–3 days) C57Bl/6 mouse brains as described previously (Wang et al., 2008). Cerebral cortices were dissected and dissociated into single cell suspension. Dissociated cells were seeded into T-75 flasks and grown in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) at 37°C and 5% CO2 atmosphere until cells grew to confluence (10–14 days). Flasks were shaken for 20 hr (120 rpm, 37°C) to detach microglia, oligodendrocytes, and neurons from the more adherent astrocytes. After shaking, the medium was replaced. Media replacement was repeated 24 hr after the shaking period.

To harvest conditioned media from reactive astrocytes, purified astrocytes were plated into T-75 flasks in serum-containing medium. After reaching confluence, astrocytes were incubated with serum-free media overnight and treated with TGF-β (10 ng/mL), ARSB (1 µg/mL), TGF-β and ARSB, or neither (untreated controls), for 7 days. After harvesting, conditioned media was centrifuged at 800 rpm for 5 min to remove debris before being split into three aliquots of 2 mL each. Aliquots were treated with no enzyme, ARSB (1 μg/mL), or ChABC (1 μL/mL) for 4 hr prior to addition to neuronal cultures.

Cerebellar granule neurons (CGNs) were isolated at previously described (Wang et al., 2008). Dissociated cells were cultured in 500 μL Neurobasal medium containing B27 supplement and 24 mM KCl and plated on poly-L-lysine-coated coverslips in 24-well plates. After allowing 2 hr for neuronal attachment, 500 μL of treated conditioned medium was applied to each well in triplicate. Cells were incubated for 24 hr and then fixed and stained for DAPI and βIII-tubulin. In co-culture experiments, dissociated CGNs were plated at a density of 5 × 104 cells/well onto a confluent monolayer of astrocytes in 24-well plates that had been treated for 7 d with ARSB (1 µg/mL), TGF-β (10 ng/mL) or TGF-β and ARSB.

Neurite outgrowth analysis

After fixation and staining, at least 60 images were taken across two coverslips per condition. Files were analyzed by an experimenter blinded to the experimental conditions. Neurons were measured if they were isolated from other neurons and had distinct nuclei and at least one neurite longer than the diameter of the cell body. The longest neurite was measured for each neuron and at least 60 neurons were measured for each condition. Each experiment was performed in triplicate.

Preparation of zymosan/CPT-cAMP and enzyme delivery scaffolds

In accordance with established protocols (Yin et al., 2003; de Lima et al., 2012), Zymosan A (Sigma Z4250) was suspended in sterile PBS at a concentration of 12.5 μg/μL, incubated at 37°C for 10 min, and vortexed. CPT-cAMP (Sigma C3912) was added to achieve a final concentration of 50 mM CPT-cAMP. Aliquots were stored at 4°C for up to two weeks. ChABC (Amsbio 100332-1A) was reconstituted at 455 µg/mL in a buffer solution containing 100 mM Tris(hydroxymethyl)aminomethane hydrochloride (Tris-HCl) and 0.1% BSA in 1X phosphate buffered saline (PBS). ARSB (Naglazyme) was obtained in acidic PBS (pH 5.5) from Biomarin (San Rafael, CA). ARSB with His Tag was obtained from R and D Systems (4415-SU). Sterile gelfoam sponges were cut to roughly 2 mm3 and placed to soak in a sterile tube containing 5 μL of either ChABC, ARSB, or the control buffer. Tubes were stored on ice for up to 4 hr before surgical implantation.

Optic nerve crush and implantation of enzyme scaffolds

Optic nerve crush was performed as described previously (Park et al., 2008). Mice were anesthetized using 1–2% isoflurane; the depth of anesthesia was confirmed by lack of response to a toe pinch. The optic nerve was exposed intraorbitally, and curved forceps were inserted beneath the external ocular muscle, avoiding the ophthalmic artery and retrobulbar sinus. The nerve was crushed approximately 1 mm behind the eye for 10 s. Immediately after the crush, eyes were monitored fundoscopically for signs of ischemia, and mice were observed for bleeding in the hours following surgery. Mice received a subcutaneous injection of 1 mg/kg buprenorphrine as an analgesic and topical application of ophthalmic ointment to prevent corneal drying.

For implantation of enzyme scaffolds, the optic nerve was exposed by gently reopening the conjunctiva and inserting curved forceps behind the eye. Carefully avoiding the ophthalmic artery and retrobulbar sinus, the enzyme- or buffer-soaked gelfoam scaffold was placed in direct contact with the optic nerve at the site of the crush lesion, approximately 1 mm behind the eye. Retinal blood flow was assessed fundoscopically, and mice received a subcutaneous injection of 1 mg/kg buprenorphrine and topical application of ophthalmic ointment.

Intravitreal injection

Intravitreal injections of Zymosan or a PBS control were administered immediately following implantation of the gelfoam scaffold, and injections of CTB were administered 1 d prior to tissue harvest. 2 μL of the injecting solution was drawn into a sterile 5 μL Hamilton syringe with a 33-gauge removable needle. In the case of Zymosan injections, the syringe was inspected to ensure that the needle was not blocked by Zymosan particles. The solution was then slowly injected through the superior nasal sclera at a 45° angle, avoiding the lens, external ocular muscle, and blood vessels. A sterile 33-gauge needle was used to puncture the cornea and drain the anterior chamber before removing the injecting needle, to reduce intraocular pressure and prevent reflux of the injected solution. Different needles were used for Zymosan and PBS injections to prevent contamination, and the syringe was rinsed thoroughly with ethanol followed by sterile PBS between injections.

Western blot

Tissue preparation

Mice were anesthetized using 1–2% isoflurane and exsanguinated, followed by cervical dislocation. Optic nerves were severed between the globe and the optic and cut into four equally sized segments of approximately 1.0–1.5 mm each. Nerve segments were immediately placed in sterile 1.5 mL Eppendorf tubes containing cold 40 µL lysis buffer (cOmplete Lysis-M, EDTA-free, Roche). Tissue was mechanically homogenized using a sterile pestle and centrifuged to separate un-homogenized tissue. Protein concentration in the supernatant was determined using the BCA assay (ThermoFisher). Samples were frozen and stored at −80°C.

Immunoblotting

Proteins were separated by SDS-PAGE under reducing conditions and transferred to a 0.45 µm PVDF membrane. Membranes were blocked with PBS containing 0.2% Tween-20% and 5% skim milk for 1 hr at room temperature. To detect ChABC-digested CSPGs, membranes were incubated with the primary mouse monoclonal antibody BE-123 (Millipore MAB2030) diluted in an immunoenhancing reagent (Can Get Signal, Toyobo) and 5% skim milk for 2 hr at 4°C, then washed and incubated with an HRP-conjugated anti-mouse IgG secondary antibody for 30 min at room temperature. Signals were visualized with myECLTH Imager (ThermoFisher, Waltham, MA, USA).

Enzyme activity

Activity of ChABC and ARSB was assessed immediately before surgery. ChABC activity was measured by spectrophotometrically detecting the production of disaccharides cleaved from the glycosaminoglycan chains of CSPGs, as has been previously described (Suzuki et al., 1968).

ARSB activity was measured by detecting the cleavage of a sulfate group from p-nitrocatechol sulfate (PNCS), which yields a product with an absorbance peak at 510 nm (Porter et al., 1969; Knaust et al., 1998). To measure enzyme activity after in vivo implantation, scaffolds were recovered from freshly dissected optic nerves, placed in 1.5 mL Eppendorf tubes, and stored on ice. 250 μL of 50 mM 2-(N-morpholino)ethanesulfonic acid (MES) at pH 6.5 was added to the tube containing the recovered scaffold. After approximately 1 hr, three aliquots of 75 μL were removed from this solution and combined with 75 μL 4-PNCS in a 96-well microplate. Samples were incubated at 37°C for 24 hr, after which the reaction was quenched by adding 150 μL of 0.2 N NaOH. Absorbance was measured at 510 nm. Recovered scaffolds loaded with enzyme buffer served as controls.

Enzyme treatment of brain sections

Free-floating 30 µm sections of mouse brain were incubated with either ChABC (Sigma C3667,≥20 µg/mL), ARSB (Naglazyme, Biomarin, 1 mg/mL), or control buffer (50 mM Tris, 60 mM sodium acetate, and 0.02% BSA, pH 8.0) in a 24-well plate. ChABC and ARSB were assayed to confirm activity before being added to the wells. Brain sections were incubated with enzyme and control solutions at 37°C for 8 hr.

Immunohistochemistry

Tissue preparation

Mice were anesthetized using 1–2% isoflurane and transcardially perfused with PBS followed by 4% paraformaldehyde (PFA). Optic nerves were dissected, laid flat on 13 mm filter paper (Millipore AABG01300), and immersed in 4% PFA. The tissue was post-fixed overnight, then immersed in 30% sucrose for at least 24 hr for cryoprotection. Nerves were embedded in Tissue-Tek OCT and snap-frozen for cryosectioning. 14 µm longitudinal sections were obtained on charged Superfrost microscope slides using a Leica CM3050 cryostat. Slides were dried and stored at −80°C.

For analysis of perineuronal nets, fresh brain tissue was dissected from a C57Bl/6 mouse and immediately immersed in 4% PFA. Tissue was post-fixed for 24 hr, cryoprotected in 30% sucrose for 24 hr, embedded in Tissue-Tek OCT, and snap-frozen for sectioning.

Immunostaining

For antibodies detecting CSPGs and glial cell activation, slides with optic nerve sections were incubated for 1 hr in blocking solution (PBS containing 3% goat serum and 0.2% Triton X-100), then incubated overnight at 4°C in primary antibodies diluted in the blocking solution. Slides were washed three times for 5 min with PBS, incubated for 2 hr with secondary antibodies, washed, and mounted onto glass coverslips with Fluoromount medium (Sigma).

The GAP-43 antibody (Benowitz et al., 1988) was incubated as previously described (Leon et al., 2000). Briefly, slides were rinsed in TBS (50 mM Tris buffer containing 8.766 g/L NaCl) and then washed with methanol for 10 min. Slides were blocked in TBS containing 10% donkey serum for 1 hr. The GAP-43 antibody was diluted 1:50,000 in a solution of TBS2T (50 mM Tris buffer, 17.532 g/L NaCl, and 0.1% Tween) containing 5% donkey serum and 2% BSA. Slides were incubated with primary antibody overnight on a rocking platform. Slides were then washed with TBS2T for 1 hr, with TBS2T plus 5% donkey serum and 2% BSA for 1 hr, and with TBS2T for 1 hr, all on a rocking platform. The secondary antibody was diluted 1:1000 in TBS2T plus 5% donkey serum and 2% BSA. Slides were incubated with the secondary antibody solution for 2 hr, followed by 30 min washes with TBS2T, TBS2T, and TBS. Slides were mounted using Fluormount and glass cover slips, and stored at 4°C for imaging.

For detection of perineuronal nets, immediately after incubation with enzymes, free-floating brain sections were washed with 1 mL of PBS containing 0.02% Triton-X100 three times for 30 min. Sections were incubated with 250 µL biotinylated Wisteria floribunda agglutinin (WFA) overnight at 4°C on a rocking platform. Sections were then washed with 1 mL PBS/0.02% Triton three times for 5 min, incubated with 250 µL TRITC-conjugated streptavidin for 1 hr at room temperature, washed with 1 mL PBS three times for 5 min, stained with DAPI, and mounted using Fluoromount and glass cover slips. Slides were stored at 4°C prior to imaging.

Microscopy and Image Processing

Tissue was imaged using a Zeiss 780 confocal microscope with 40X and 63X objectives. Z-stacks were maximally projected onto a single plane using Zeiss image processing software. For images used in fluorescence quantification, image capture settings were held constant, and samples from within each group were imaged at the same time. Fluorescence intensity was measured using ImageJ, with identical settings for all samples within each analysis.

Quantification of regenerating axons

In ImageJ, vertical lines were drawn through each nerve section at 0.25 mm intervals starting from the lesion site, and the number of GAP-43+ axons crossing each line was manually counted. Four sections were counted for each nerve. The number of regenerating axons per nerve was then calculated at each distance using a previously developed formula (Lim et al., 2016; Bei et al., 2016), with the total number of axons equal to πr2 (r being the maximum recorded radius of the optic nerve section) times the average number of counted axons, divided by the thickness of the section (14 µm). Axon counting was verified by a separate observer blind to the experimental conditions. For quantification of longest axon, the same images were used. GAP-43+ axons were identified, and the length of the longest detectable axon was measured from the lesion site using ImageJ.

Statistics

Sample size for axon regeneration experiments was determined based on preliminary data from a pilot experiment. The number of regenerating axons counted at 0.50 mm distal from the lesion site was obtained from groups of mice treated with either Zymosan + ARSB (n = 4) or Zymosan + Buffer (n = 5). The control group had a mean of 104 ± 53 axons at 0.50 mm, while the ARSB-treated group had a mean of 260 ± 84 axons. Based on these numbers, we assumed a standard deviation of 75, to be equal for each group, and estimated using a two-sided two sample t-test that n = 9 mice per group would be required to achieve 80% power (at the 0.025 level) to compare ARSB treatment to a buffer control.

All statistical tests were performed using GraphPad Prism 7.0 (GraphPad Software, La Jolla, CA). Neurite lengths in culture were compared using Kruskal-Wallis Analysis of Variance and Mann-Whitney U tests. Axon regeneration was assessed using two-way ANOVA and Bonferroni post-hoc analysis. Asterisks indicate significance levels as specified in the corresponding figure legends.

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20
  21. 21
  22. 22
  23. 23
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
  29. 29
  30. 30
  31. 31
  32. 32
  33. 33
  34. 34
  35. 35
  36. 36
  37. 37
    Isolation and characterization of developmentally regulated chondroitin sulfate and chondroitin/keratan sulfate proteoglycans of brain identified with monoclonal antibodies
    1. U Rauch
    2. P Gao
    3. A Janetzko
    4. A Flaccus
    5. L Hilgenberg
    6. H Tekotte
    7. RK Margolis
    8. RU Margolis
    (1991)
    The Journal of Biological Chemistry 266:14785–14801.
  38. 38
  39. 39
  40. 40
  41. 41
  42. 42
  43. 43
  44. 44
  45. 45
    Increased axon regeneration in astrocytes grown in the presence of proteoglycan synthesis inhibitors
    1. LC Smith-Thomas
    2. J Stevens
    3. J Fok-Seang
    4. A Faissner
    5. JH Rogers
    6. JW Fawcett
    (1995)
    Journal of Cell Science 108:1307–1315.
  46. 46
  47. 47
    Formation of three types sulfates of disulfated by chondroitinase disaccharides from chondroitin digestion
    1. S Suzuki
    2. H Saito
    3. T Yamagata
    4. K Anno
    5. N Seno
    6. Y Kawai
    7. T Furuhashi
    (1968)
    The Journal of Biological Chemistry 243:1543–1550.
  48. 48
  49. 49
    Immunohistochemical localization of chondroitin sulfate in the forestomach of the sheep
    1. Y Yamamoto
    2. Y Atoji
    3. A Oohira
    4. Y Suzuki
    (1995)
    European Journal of Histochemistry : EJH 39:265–272.
  50. 50
  51. 51
  52. 52
  53. 53

Decision letter

  1. Joseph G Gleeson
    Reviewing Editor; Howard Hughes Medical Institute, The Rockefeller University, United States

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

Thank you for submitting your work entitled "Identification of a critical sulfation in chondroitin that inhibits axonal regeneration" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by a Reviewing Editor and a Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: Hudson H Freeze (Reviewer #3).

Our decision has been reached after consultation between the reviewers. All three reviewers appreciate the quality of the science and the potential impact of chondroitin sulfating in axon regeneration. They appreciated the quality of the cell culture and animal model data, as well as the potential tor translatability to axon regeneration. However, they identified a number of flaws that will preclude publication in its current form. These center around the choice of the experimental system, the need to prime the system with zymosan, and size of the effect observed, and how these results are quantitated. Addressing these issues may be within the scope of what is possible, after consulting with reviewers, we came to the conclusion that this would be beyond what could be expected with the turnaround time that we seek at eLife for acceptance of papers for publication.

Reviewer #1:

The authors have taken advantage of the properties of Arylsulfatase B (ARSB) to specifically degrade C4sulphate groups on GAG chains on inhibitory proteoglycans to promote regeneration in vitro but also in vivo using an optic nerve crush model. While the use of ARSB to promote regeneration is a novel and exciting idea given that this enzyme is already approved clinically, the paper contains a number of issues that are of concern.

1) In the Introduction, the authors mention that myelin is thought to be part of the "glial scar." Actually, myelin inhibitory proteins are located broadly and not thought to be specifically related to the scar.

2) Also, in the Introduction the authors champion their own work (which is reasonable) suggesting that C4S proteoglycans are those that are the most potently inhibitory. However, they do not even mention the strongly opposing and well published view from the Hsieh-Wilson lab that CS-E (rather than A) is critical. I think it would be best to mention up front that there is some controversy. Then, later, the authors can discuss how their current results help to provide evidence in support of their idea. However, it would have been welcomed if the authors had compared ARSB efficacy with CS-E specific degradation in their various models.

3) Figure 1 has a major problem in that the lengths of the control neurons with axons on untreated, TGF-b and TGF-b ARSB cultures are very similar, with very tiny differences (only around 10 um) between the groups. The same problem is apparent in the conditioned media experiments. However, the authors suggest that such small differences are statistically significant. I see no obvious differences between any of the neurons in Figure 1E suggesting that the differences really are incredibly slight. It would have been much preferable if an adult neuron (such as DRGs) would have been used and the cultures had been allowed to mature until long axons (in the controls) had been elaborated.

4) In Figure 2 the absolute levels of C4S in the optic nerve lesion is very small compared to the CS56 staining especially at day 7 which is supposedly at peak of expression. Indeed, the immunohistochemistry overall is not especially convincing. This suggests that C6S GAG moieties are in high abundance in the lesion and apparently in much greater abundance than C4S. Fold changes in C4S are not as interesting as are absolute amounts.

5) In Figure 3 there is a critical flaw. In the control optic crush animals there should be zero regeneration of axons into and past the lesion. None-the-less, the authors report regeneration of numerous axons in the controls as far as 1mm beyond the lesion. This is clear evidence of axonal sparing after lesion which confounds the interpretation of all of the data. In addition, the differences that are shown between control and ARSB treated optic nerves is again, incredibly small.

In its present form this manuscript should not be acceptable for publication.

Reviewer #2:

In their submission entitled "Identification of a critical sulfation in chondroitin that inhibits axonal regeneration" Pearson et al. examine the role of terminal 4S groups on GAG chains of CSPGs in imparting growth inhibiting properties of CSPGs, using the enzyme arylsulfatase B. They first show in astrocyte and neuronal coculture, or neuronal culture exposed to astrocyte conditioned medium, that ARSB performs similarly to chondroitonase ABC in restoring axonal growth potential after TGF-β induced inhibition. They then characterize an optic nerve crush model to show that 4S CSPGs are a part of the scar that forms after injury, which is present for at least 4 weeks post crush. Then the authors combine a zymosan+cAMP combinatorial treatment to induce RGC regeneration with ChABC or ARSB and show that RGC axon regeneration is improved similarly between the two treatments and can act on a delay of up to three days and does not alter astrocytic distributions.

While the finding that, in this model system, targeting 4S terminal GAG chains for cleavage can mediate effects similar to that of ChABC with an already clinically approved drug is quite interesting, overall the chosen application seems inappropriate. In the introductory sections, the authors make a strong point that ARSB is already approved for clinical use, but then they apply it to a system where ChABC itself does nothing without a combinatorial therapy that would not be seen in clinics. Without the application of this suggested therapy to an injury model where beneficial results would be expected sans untranslatable combinatorial therapies (eg spinal cord injury) a major thrust of this study is lost. I would highly recommend considering a different system for this application.

Additionally, although the authors were able to demonstrate that their delivery of ChABC was able to penetrate the optic nerve to some extent and breakdown CSPGs to some extent, they do not examine the efficiency very clearly. How much of the intact CSPGs remain should be demonstrated by immunostaining, and while it is appreciably difficult to clearly show ARSB function in vivo, it needs to be shown that the treatment worked as intended. Is there no biochemical approach to examining the GAG terminal chains?

Reviewer #3:

This is an important paper. It is well done, thoughtful about controls and relatively well versed in glycobiology. The results of these complicated animal studies, especially, are clearly presented and the overall results are unexpected. The rationales for carrying on further work using ARSB are clear and these initial results are hopeful.

I have no objection to the general direction, but I wonder about the need for the zymosan injection to prime the system in some way. What would be the equivalent in a patient? or what other insults would occur in a therapeutic setting? Clearly, they would hope this approach is something general and not only restricted to ONC. I think it is a well worthwhile study.

All that said, I do have some suggestions for improvement. First, I don't accept their claim of a 2.5 fold increase in 4S (antibody 2H6) as shown in Figure 2—figure supplement 1. The level at Day "0" is already ~1.75 fold (not baseline of 1.0) and it is shown as an increase to 2.5 fold. That's only a 1.4 fold increase, which is about the same as the CS-56 antibody. So, there is no selectivity. It seems misleading to state that. If I have misinterpreted the data, I'd like to see their justification for the claim based on these results.

In Figure 1, there are so few tubulin staining cells. I'd like to see more pictures, perhaps a small composite. How many of these cells were actually counted? How many were in the "longest" category. Also, it would be helpful to see the distribution of lengths, not only the (few?) longest ones. I can't tell based on the presentation. So, I have some skepticism.

[Editors’ note: what now follows is the decision letter after the authors submitted for further consideration.]

Thank you for submitting your article "Identification of a critical sulfation in chondroitin that inhibits axonal regeneration" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by Joseph Gleeson as the Reviewing Editor and Marianne Bronner as the Senior Editor. Two of the three reviewers were the same and one reviewer is new for this submission. The following individual involved in review of your submission has agreed to reveal his identity: Hudson H Freeze (Reviewer #3).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

This manuscript leverages off of previous findings that implicated 4-sulfation as a mechanism of CSPG inhibition of axonal growth. The three reviewers agree that the major finding presented are that the human enzyme arylsulfatase B (ARSB), which cleaves 4-sulfate groups from glycosaminoglycan (GAG) chains of chondroitin sulfate proteoglycans (CSPGs), abrogates the inhibitory actions of chondroitin sulfate in vitro and promotes regeneration of retinal ganglion cell axons in the optic nerve. The reviewers from the previous submission appreciate the positive changes and the correction to the manuscripts of the previous errors. Reviewers commented that the work is consistent with the report of Yoo et al., 2013 using ARSB together with ChABC that suggested a comparable impact of these enzymes on locomotor function recovery after spinal cord injury. The reviewers appreciated the quality of the cell culture and animal model data and the potential impact of ARSB in axon regeneration. They mention that the study is important and well thought out, with immediate therapeutic implications and possibly applications. However, there were several concerns expressed by the reviewers that require attention before the manuscript can be further considered at eLife.

Essential revisions:

1) Considering the known variation associated with zymosan treatment, it is questionable how much meaningful impact ARSB has on axon regeneration. In addition, in the image in Figure 3B, "labeled axons" in the distal optic nerve appear to be bundled, possibly spared axons.

2) 1- CS-56 staining in conditioned media is not altered by ARSB treatment. However, 2H6 antibody has more affinity for 4S groups versus 6S. It would be interesting to document how ARSB treatment impacts the elevation of chondroitin sulfate proteoglycan after optic nerve crush as documented in Figure 2.

3) The authors don't provide significant data to further incriminate 4S versus 4,6S role in the inhibitory effect of CSPGs as well as any new data to specifically illustrate the fundamental role of 4S. A Title reflecting the effect of ARSB enzyme in their specific ONC paradigm will be more appropriate.

4) The representation of the data related to the neurite outgrowth does not allow assessing easily the inhibitory effect of CSPGs. Several experiments were performed in Figure 1 and Figure 1—figure supplement 1; however different representations of the data are used. Additionally, the authors should provide standard deviations to assess the reproducibility between triplicates.

5) The authors should provide information on the amount of enzymes used in subsection “Modifying CSPG sulfation enhances retinal ganglion cell axon regener”. One is expressed as U/ml and others as ug/ml. Please include both, even though the assays are based on rather artificial substrate measurements.

6) In subsection “Sulfation dictates the effects of CSPGs on axon growth” says that ARSB cleaves extracellular GAG chains, but only an assay using an artificial substrate is presented. Besides, the enzyme should not cleave GAG chains- it cleaves terminal 4-sulfate. That statement should be corrected or explained.

7) No data was shown on the effects of ARSB on 2H9 staining/reactivity. Presumably there is not any, since it only removes 1 sulfate per CS chain, but this should be stated explicitly.

8) The size of the CS chains in these cells/settings is not known. Presumably its 50-100 monosaccharides, but again ARSB should only remove a single 4-sulfate. This remarkably small change (1/20, 1/50, 1/100) is worth stressing, and citing relevant literature.

https://doi.org/10.7554/eLife.37139.019

Author response

[Editors’ note: the author responses to the first round of peer review follow.]

This current submission has been extensively revised in response to the reviewers’ comments on our original submission. We have readily corrected two key errors (noted in the bullets below), and have additionally revised the manuscript to answer each of the other major and minor criticisms. We hope that this document summarizing our changes will persuade you to consider our revised manuscript for publication.

Having carefully read each of the reviewers’ comments, we noted that two of the major flaws highlighted by reviewers were, in fact, due to our figure presentation, rather than experimental design. Briefly, these included:

1) An omission in Figure 3 led Reviewer #1 to suggest that our optic nerve samples were detecting spared axons in negative control samples. In fact, these were regenerating axons from a Zymosan-treated control. However, we mistakenly omitted the “Zymosan” label from one panel of our graph.

2) An omission in Figure 2–Figure Supplement 1 led Reviewer #3 to question the validity of our data showing that CSPGs are elevated after optic nerve crush. We had failed to include the baseline “Day 0” condition where fold change = 1.0. We have fixed both of these errors.

Of the remaining criticisms, we address the most significant first:

1) “Spared” axons in controls

Reviewer #1:

In Figure 3 there is a critical flaw. In the control optic crush animals there should be zero regeneration of axons into and past the lesion. None-the-less, the authors report regeneration of numerous axons in the controls as far as 1mm beyond the lesion. This is clear evidence of axonal sparing after lesion which confounds the interpretation of all of the data. In addition, the differences that are shown between control and ARSB treated optic nerves is again, incredibly small.

The appearance of axonal sparing was due to our mislabeling of the graph in panel (c) of Figure 3, where we failed to specify that all animals in this experiment were treated with Zymosan/CPT-cAMP. This information was conveyed in the text and figure legend, but mistakenly omitted from the graph labels. We have corrected the figure. Our data confirming that no axons extended beyond the lesion in untreated control animals can be found in Figure 3—figure supplement 1. We therefore argue that this “critical flaw” was merely a labeling omission and does not in any way confound the interpretation of the data.

Regarding the differences shown between control and ARSB treated optic nerves, we believe our reported differences are substantial and noteworthy. The enhancement of axon regeneration in Zymosan + ARSB vs. Zymosan + buffer was on the order of ~250 axons at 0.50 mm from the lesion after only 14 d (Figure 3C). The inclusion of ARSB essentially doubles the effect of Zymosan alone and is comparable to the effects achieved in many landmark publications in the field of optic nerve regeneration. For instance, Lim et al., (2016) also used a combination stimulus and, in several experiments, reported effects of a similar magnitude to ours (e.g. in Lim et al., Figure 3, treatment group 3 = ~75 axons and group 4 = ~275 axons at 0.50 mm).

2) Use of intrinsic stimulus

Reviewer #2:

While the finding that, in this model system, targeting 4S terminal GAG chains for cleavage can mediate effects similar to that of ChABC with an already clinically approved drug is quite interesting, overall the chosen application seems inappropriate. In the introductory sections, the authors make a strong point that ARSB is already approved for clinical use, but then they apply it to a system where ChABC itself does nothing without a combinatorial therapy that would not be seen in clinics. Without the application of this suggested therapy to an injury model where beneficial results would be expected sans untranslatable combinatorial therapies (eg spinal cord injury) a major thrust of this study is lost. I would highly recommend considering a different system for this application.

Reviewer #3:

I have no objection to the general direction, but I wonder about the need for the zymosan injection to prime the system in some way. What would be the equivalent in a patient? or what other insults would occur in a therapeutic setting? Clearly, they would hope this approach is something general and not only restricted to ONC. I think it is a well worthwhile study.

We appreciate that both reviewers acknowledge the interest and significance of the work and welcome the opportunity to defend our choice of the optic nerve system and our use of a combined intrinsic/extrinsic regeneration stimulus. The optic nerve has many advantages for the study of axonal regeneration as compared to spinal cord injury models, which have enabled this work. The surgical accessibility of the optic nerve enables a high degree of reproducibility and allowed us to deliver ARSB directly to the lesion site. All axons in the optic nerve come from retinal ganglion cells, making the measurement of regeneration unambiguous.

Virtually all experimental therapies with robust effects in the optic nerve or spinal cord require combinatorial treatments to stimulate significant axon regeneration. While we appreciate that Zymosan itself could not be implemented clinically, ARSB in combination with other stimuli is a distinct possibility. Reporting the efficacy of ARSB in enhancing intrinsic pro-regenerative stimuli will certainly encourage other groups to integrate ARSB into new experimental therapies. We ourselves are engaged in ongoing research to this effect. We believe that holding back our findings with ARSB due to its failure to achieve an effect independently of intrinsic stimulation is not justified, given the state of the field and the potential impact of our findings.

3) CSPG immunohistochemistry

Reviewer #1:

In Figure 2 the absolute levels of C4S in the optic nerve lesion is very small compared to the CS56 staining especially at day 7 which is supposedly at peak of expression. Indeed, the immunohistochemistry overall is not especially convincing. This suggests that C6S GAG moieties are in high abundance in the lesion and apparently in much greater abundance than C4S. Fold changes in C4S are not as interesting as are absolute amounts.

We believe that the impression of low C4S levels may have been due to the false-coloring of our figures – the magenta coloring for 2H6 is more difficult to see on a black background than the green coloring of CS-56 when comparing them side by side. We have thus modified Figure 2 to show all images in grayscale, where features are more readily detected by eye. For the sake of full transparency, we have also modified the Figure 2—figure supplement 1 to include both raw fluorescence intensity values and fold change. These graphs are essentially identical, showing increase in both total GAG (using CS-56) and 4S GAG (using 2H6) as compared with the uninjured control in both analyses.

We appreciate reviewer #1’s suggestion that we compare absolute levels of CSPG and C4S, as this would indeed resolve the question of whether C4S or C6S moieties are in higher abundance within the lesion. However, it is not possible to compare absolute levels of two different antigens using immunohistochemistry, as each antibody has a unique epitope and affinity for the antigen, and the images are collected in different confocal channels with distinct gain and contrast settings. To quantify absolute amounts of GAG would require biochemical methods, which is precluded by the small amount of tissue in each optic nerve, as isolating the optic nerve lesion site produces only 6-8 µg of protein per sample and would require a prohibitive number of mice. We therefore chose to report fold change of total GAG and 4S GAG as compared to control.

Reviewer #2:

In their submission entitled "Identification of a critical sulfation in chondroitin that inhibits axonal regeneration" Pearson et al. examine the role of terminal 4S groups on GAG chains of CSPGs in imparting growth inhibiting properties of CSPGs, using the enzyme arylsulfatase B. They first show in astrocyte and neuronal coculture, or neuronal culture exposed to astrocyte conditioned medium, that ARSB performs similarly to chondroitonase ABC in restoring axonal growth potential after TGF-β induced inhibition. They then characterize an optic nerve crush model to show that 4S CSPGs are a part of the scar that forms after injury, which is present for at least 4 weeks post crush. Then the authors combine a zymosan+cAMP combinatorial treatment to induce RGC regeneration with ChABC or ARSB and show that RGC axon regeneration is improved similarly between the two treatments and can act on a delay of up to three days and does not alter astrocytic distributions.

We agree and have therefore modified the text to simply describe an increase in 4S and removed any claims regarding disproportionate expression. Our previous publications have documented a relative increase in 4S expression after injury as evaluated by biochemical methods. However, as stated above, these biochemical measurements are impossible to perform at the site of an optic nerve lesion given the extremely small amount of tissue.

Reviewer #3:

First, I don't accept their claim of a 2.5 fold increase in 4S (antibody 2H6) as shown in Figure 2—figure supplement 1. The level at Day "0" is already ~1.75 fold (not baseline of 1.0) and it is shown as an increase to 2.5 fold. That's only a 1.4 fold increase, which is about the same as the CS-56 antibody. So, there is no selectivity. It seems misleading to state that. If I have misinterpreted the data, I'd like to see their justification for the claim based on these results.

This impression arose from our failure to include the baseline condition (1.0) at Day 0, instead starting our data from Day 1 (where the fold change was indeed ~1.75). We are grateful for the correction and have amended the figure to include “Day 0” baseline data for clarity.

4) Small differences in neurite length

Reviewer #1:

Figure 1 has a major problem in that the lengths of the control neurons with axons on untreated, TGF-b and TGF-b ARSB cultures are very similar, with very tiny differences (only around 10 um) between the groups. The same problem is apparent in the conditioned media experiments. However, the authors suggest that such small differences are statistically significant. I see no obvious differences between any of the neurons in Figure 1E suggesting that the differences really are incredibly slight. It would have been much preferable if an adult neuron (such as DRGs) would have been used and the cultures had been allowed to mature until long axons (in the controls) had been elaborated.

Reviewer #3:

In Figure 1, there are so few tubulin staining cells. I'd like to see more pictures, perhaps a small composite. How many of these cells were actually counted? How many were in the "longest" category. Also, it would be helpful to see the distribution of lengths, not only the (few?) longest ones. I can't tell based on the presentation. So, I have some skepticism.

5) Efficiency of enzyme delivery

Reviewer #2:

Additionally, although the authors were able to demonstrate that their delivery of ChABC was able to penetrate the optic nerve to some extent and breakdown CSPGs to some extent, they do not examine the efficiency very clearly. How much of the intact CSPGs remain should be demonstrated by immunostaining. And while it is appreciably difficult to clearly show ARSB function in vivo, it needs to be shown that the treatment worked as intended. Is there no biochemical approach to examining the GAG terminal chains?

We appreciate the suggestion and agree that quantitatively demonstrating ARSB function in vivo would be ideal. However, absolute levels of GAG can only be obtained biochemically, which is limited by the amount of tissue in the injured optic nerve. To answer the question would take prohibitively high numbers of animals per condition, which is inconsistent with current pressures to reduce unnecessary use of animals in research.

Using immunohistochemistry, we found a significant increase in BE-123 “stub” immunoreactivity, indicative of ChABC activity (Figure 3—figure supplement 2), and a small but not significant decrease in CS-56 immunoreactivity at 14 days post crush. We also recovered the implanted gelfoam scaffolds from enzyme-treated mouse optic nerves and detected significant residual ARSB activity from ARSB-loaded scaffolds when compared with buffer-loaded scaffolds, which showed no ARSB activity. Taken together with the pro-regenerative effects seen in enzyme-treated mice, we believe these observations adequately support our conclusion that ChABC and ARSB were active and functioning in vivo.

[Editors' note: the author responses to the re-review follow.]

Essential revisions:

1) Considering the known variation associated with zymosan treatment, it is questionable how much meaningful impact ARSB has on axon regeneration. In addition, in the image in Figure 3B, "labeled axons" in the distal optic nerve appear to be bundled, possibly spared axons.

We agree that analyzing optic nerve samples for the presence of spared axons is essential to ensure the validity of the results, including careful adherence to such guidelines as those recently published by Fischer et al.,(2017). We carefully reexamined the GAP-43-labeled axons in each image quantified from the sample shown in Figure 3B, and confirmed that no signs of bundled, excessively straight, obviously spared axons were present in any of the other sections analyzed. We also stained optic nerve samples for GFAP and found a clear injury site demarcated by retracted astrocyte processes. Further, we observed that, unlike CTB, which will appear in non-lesioned or spared axons if they are present, GAP-43 stains only regenerating axons and does not stain intact, non-lesioned axons. We have added a sentence to this effect in the manuscript and cited previously published work that reports the same findings. Collectively these observations give us confidence that our analysis and results include only regenerating axons.

2) 1- CS-56 staining in conditioned media is not altered by ARSB treatment. However, 2H6 antibody has more affinity for 4S groups versus 6S. It would be interesting to document how ARSB treatment impacts the elevation of chondroitin sulfate proteoglycan after optic nerve crush as documented in Figure 2.

We appreciate this suggestion and have added a new supplementary figure (Figure 3—Figure supplement 5) to address the question. We show that treating lesioned mouse optic nerves with ARSB at 3 dpc does not significantly alter immunoreactivity of CS-56 or 2H6 at 7 dpc when compared with the buffer-treated control.

3) The authors don't provide significant data to further incriminate 4S versus 4,6S role in the inhibitory effect of CSPGs as well as any new data to specifically illustrate the fundamental role of 4S. A Title reflecting the effect of ARSB enzyme in their specific ONC paradigm will be more appropriate.

The only known action of ARSB is to remove the sulfate group from the 4-position of GalNAc. This confirms the essential role of 4S (even if it is in the presence of 6S) as an important signal to neurons. Whether ARSB attacks 4,6S is not at all known, and we found a very low level of 4,6S at the non-reducing end of CSPGs by our HPLC methods. Based on this we feel that the title to the manuscript is appropriate.

4) The representation of the data related to the neurite outgrowth does not allow assessing easily the inhibitory effect of CSPGs. Several experiments were performed in Figure 1 and Figure 1—figure supplement 1; however different representations of the data are used. Additionally, the authors should provide standard deviations to assess the reproducibility between triplicates.

We have changed all data representations to dot plots, which we believe to be a more accurate picture of the data presented. Each experiment showed a significant action of both CSPGs and ARSB in reversing the CSPG actions on neurite outgrowth. We have included a Table of in vitro experimental statistics below containing the mean, standard deviation, standard error, and p-values (for both overall ANOVA and posthoc comparison to untreated samples) for each experiment in the triplicate of the in vitro experiments.

5) The authors should provide information on the amount of enzymes used in subsection “Modifying CSPG sulfation enhances retinal ganglion cell axon regener”. One is expressed as U/ml and others as ug/ml. Please include both, even though the assays are based on rather artificial substrate measurements.

We agree and have standardized our reporting of enzyme amounts, using mg/mL or µg/mL units for both enzymes.

6) In subsection “Sulfation dictates the effects of CSPGs on axon growth” says that ARSB cleaves extracellular GAG chains, but only an assay using an artificial substrate is presented. Besides, the enzyme should not cleave GAG chains--it cleaves terminal 4-sulfate. That statement should be corrected or explained.

We agree and have amended the sentence to more clearly state that ARSB cleaves terminal 4S groups from the non-reducing ends of GAG chains.

7) No data was shown on the effects of ARSB on 2H9 staining/reactivity. Presumably there is not any, since it only removes 1 sulfate per CS chain, but this should be stated explicitly.

We have included this data, showing no significant change in 2H6 immunoreactivity following ARSB treatment, in a new supplementary figure (Figure 3—figure supplement 5).

8) The size of the CS chains in these cells/settings is not known. Presumably its 50-100 monosaccharides, but again ARSB should only remove a single 4-sulfate. This remarkably small change (1/20, 1/50, 1/100) is worth stressing, and citing relevant literature.

We appreciate this suggestion and have added to our Discussion section emphasizing the fact that ARSB only removes a single, terminal sulfate from the non-reducing end of GAG chains, and yet produces a pro-regenerative effect of the same magnitude as that of ChABC, which digests entire GAG chains. We have cited literature on the average length of GAG chains for neurocan and phosphacan that shows an average chain length of approximately 50 disaccharides, which implies a small change in sulfation produces a big change in inhibitory activity.

https://doi.org/10.7554/eLife.37139.020

Article and author information

Author details

  1. Craig S Pearson

    1. Laboratory of Developmental Neurobiology, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, United States
    2. Department of Clinical Neurosciences, University of Cambridge, Cambridge, United Kingdom
    Contribution
    Conceptualization, Data curation, Formal analysis, Investigation, Writing—original draft
    Competing interests
    No competing interests declared
    ORCID icon 0000-0002-1906-6347
  2. Caitlin P Mencio

    Laboratory of Developmental Neurobiology, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, United States
    Contribution
    Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Writing—review and editing
    Competing interests
    No competing interests declared
  3. Amanda C Barber

    Department of Clinical Neurosciences, University of Cambridge, Cambridge, United Kingdom
    Contribution
    Conceptualization, Writing—review and editing
    Competing interests
    No competing interests declared
  4. Keith R Martin

    Department of Clinical Neurosciences, University of Cambridge, Cambridge, United Kingdom
    Contribution
    Conceptualization, Resources, Supervision, Funding acquisition, Writing—review and editing
    For correspondence
    krgm2@cam.ac.uk
    Competing interests
    No competing interests declared
  5. Herbert M Geller

    Laboratory of Developmental Neurobiology, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, United States
    Contribution
    Conceptualization, Resources, Data curation, Supervision, Funding acquisition, Methodology, Project administration, Writing—review and editing
    For correspondence
    gellerh@nhlbi.nih.gov
    Competing interests
    No competing interests declared
    ORCID icon 0000-0002-7048-6144

Funding

National Institutes of Health (1ZIAHL006135)

  • Craig S Pearson
  • Caitlin P Mencio
  • Herbert M Geller

Cambridge Eye Trust

  • Amanda C Barber
  • Keith R Martin

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

This research was supported by the Division of Intramural Research of the National Heart, Lung, and Blood Institute, the Cambridge Eye Trust, Fight for Sight UK, and the Jukes Glaucoma Research Fund. Help with statistical analyses was provided by Dr. Nancy Geller. We also appreciate the input and suggestions of Dr. Yasuhiro Katagiri. GAP-43 antibody was generously provided by Drs. Larry Benowitz and Yuqin Yin.

Ethics

Animal experimentation: This study was performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All of the animals were handled according to approved institutional animal care and use committee (IACUC) protocols (#H-0186R3) of the National Heart, Lung, and Blood Institute, NIH. Mice were anesthetized using 1-2% isoflurane, and every effort was made to minimize suffering.

Reviewing Editor

  1. Joseph G Gleeson, Reviewing Editor, Howard Hughes Medical Institute, The Rockefeller University, United States

Publication history

  1. Received: March 29, 2018
  2. Accepted: May 14, 2018
  3. Accepted Manuscript published: May 15, 2018 (version 1)
  4. Version of Record published: May 30, 2018 (version 2)

Copyright

This is an open-access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.

Metrics

  • 869
    Page views
  • 160
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Cell Biology
    2. Developmental Biology
    Hannah J Green et al.
    Research Article Updated
    1. Cell Biology
    2. Stem Cells and Regenerative Medicine
    Aurore Lhonore et al.
    Research Article