Epac2 Elevation Reverses Inhibition by Chondroitin Sulfate Proteoglycans In Vitro and Transforms Postlesion Inhibitory Environment to Promote Axonal Outgrowth in an Ex Vivo Model of Spinal Cord Injury

Abstract

Millions of patients suffer from debilitating spinal cord injury (SCI) without effective treatments. Elevating cAMP promotes CNS neuron growth in the presence of growth-inhibiting molecules. cAMP's effects on neuron growth are partly mediated by Epac, comprising Epac1 and Epac2; the latter predominantly expresses in postnatal neural tissue. Here, we hypothesized that Epac2 activation would enhance axonal outgrowth after SCI. Using in vitro assays, we demonstrated, for the first time, that Epac2 activation using a specific soluble agonist (S-220) significantly enhanced neurite outgrowth of postnatal rat cortical neurons and markedly overcame the inhibition by chondroitin sulfate proteoglycans and mature astrocytes on neuron growth. We further investigated the novel potential of Epac2 activation in promoting axonal outgrowth by an ex vivo rat model of SCI mimicking post-SCI environment in vivo and by delivering S-220 via a self-assembling Fmoc-based hydrogel that has suitable properties for SCI repair. We demonstrated that S-220 significantly enhanced axonal outgrowth across the lesion gaps in the organotypic spinal cord slices, compared with controls. Furthermore, we elucidated, for the first time, that Epac2 activation profoundly modulated the lesion environment by reducing astrocyte/microglial activation and transforming astrocytes into elongated morphology that guided outgrowing axons. Finally, we showed that S-220, when delivered by the gel at 3 weeks after contusion SCI in male adult rats, resulted in significantly better locomotor performance for up to 4 weeks after treatment. Our data demonstrate a promising therapeutic potential of S-220 in SCI, via beneficial effects on neurons and glia after injury to facilitate axonal outgrowth.SIGNIFICANCE STATEMENT During development, neuronal cAMP levels decrease significantly compared with the embryonic stage when the nervous system is established. This has important consequences following spinal cord injury, as neurons fail to regrow. Elevating cAMP levels encourages injured CNS neurons to sprout and extend neurites. We have demonstrated that activating its downstream effector, Epac2, enhances neurite outgrowth in vitro, even in the presence of an inhibitory environment. Using a novel biomaterial-based drug delivery system in the form of a hydrogel to achieve local delivery of an Epac2 agonist, we further demonstrated that specific activation of Epac2 enhances axonal outgrowth and minimizes glial activation in an ex vivo model of spinal cord injury, suggesting a new strategy for spinal cord repair.

Keywords: Epac2; astrocyte; axonal regrowth; cAMP; organotypic; spinal cord injury.

Figure 1.

The effects of Epac2 modulation on cortical neurite outgrowth. A–C, Epac2 agonist S-220 promoted significant neurite outgrowth. A, Control. B, Treated with S-220. C, Quantification of total neurite length shows that S-220-treated neurons had significantly longer neurites. D–F, Epac2 antagonist ESI-05 significantly decreased cortical neurite outgrowth. D, Control. E, Treated with ESI-05. F, Quantification of total neurite length shows that ESI-05-treated neurons had significantly shorter neurites. G–I, siRNA knockdown of Epac2 significantly decreased cortical neurite outgrowth. G, Scrambled siRNA control. H, Epac2 siRNA-treated. I, Quantification of total neurite length shows that Epac2 siRNA-treated neurons had significantly shorter neurites. J–L, Overexpression of Epac protein using lentivirus promoted cortical neurite outgrowth. J, LV/GFP-transduced neurons. K, LV/Epac-YFP-transduced neurons. L, Quantification of total neurite length shows that LV/Epac-YFP-transduced neurons had significantly longer neurites. A, B, D, E, Cultures were grown for 24 h. G, H, Cultures were grown for 48 h. J, K, Cultures were grown for 7 d to allow enough time for gene expression. All cultures were immunostained for β-tubulin-III. Unpaired Student's t test: *p < 0.05; **p < 0.01; ***p < 0.001. n = 3/group. Data are mean ± SEM. Scale bars: A–E, J–K, 50 μm; G, H, 100 μm.

Figure 2.

Epac agonist activates Epac protein in DRG growth cones and induces growth cone attraction. A–D, Representative images of FRET SE before (A) and after (B) the addition of the Epac2 agonist S-220 and before (C) and after (D) the addition of the Epac2 antagonist ESI-05. E, FRET SE measured over time shows a significant activation and inactivation of Epac after the addition of the S-220 (blue) and ESI-05 (red), respectively, compared with control (black) (n = 4). F, G, Representative images of DRG growth cones turning toward a gradient of S-220 at time 0 (F) and 30 min (G). H, Cumulative frequency plots of turning angles show a clear switch toward the right in gradients of S-220 (blue) compared with control F12 (black), indicating greater attraction toward the direction of Epac2 activity. Average turning angles are shown above the abscissa (n = 10). Data are mean ± SEM. Scale bars: A–D, 10 μm; F, G, 50 μm. *p < 0.05; **p < 0.01; ***p < 0.001.

Figure 3.

The Epac2 agonist overcomes inhibitory environments for cortical neuron growth. A, β-tubulin-III-positive cortical neurons grew neurites on PDL-coated coverslips. B, Cortical neurons treated with CSPGs showed significantly shorter neurite lengths compared with control. C, Epac2 agonist S-220 attenuated the inhibitory effect of CSPGs on cortical neurite outgrowth. D, Quantification represents the percentage of change from control and shows a significant reduction in inhibition by CSPGs when S-220 was simultaneously applied. Cultures were grown for 48 h. E–G, S-220 also showed the effect in overcoming the astrocyte inhibition. Three neurite growth cone behaviors of DRG neurons cocultured with mature astrocytes were observed: retract (E, ▿), reflect (E, *), and crossover (F, #) using time-lapse live cell microscopy. Green color (E,F) = GFAP staining; Red color (E,F) = β-tubulin-III staining. G, Quantification showed a significant reduction in the retract/reflect behaviors of neurites and a significant increase in the crossover behavior of neurites in cells treated with the Epac 2 agonist compared with control. D, G, Mann–Whitney Rank Sum test: *p < 0.05; **p < 0.01. n = 3/group. Data are mean ± SEM. Scale bars: A–C, 50 μm; E, F, 25 μm.

Figure 4.

Epac2 agonist S-220 attenuates LPS-induced astrocyte and microglial activation in vitro. Representative fluorescent images of control (A), LPS-treated (B), and LPS + S-220-treated (C) astrocytes (immunostained for GFAP). Representative fluorescent images of control (D), LPS-treated (E), and LPS + S-220-treated (F) microglia immunostained for Iba-1 (green) and iNOS (red). G, Quantification of the mean fluorescence intensity of GFAP showed significant difference between control and LPS-treated astrocytes, and between LPS and LPS + S-220-treated astrocytes. H, Quantitative image analysis showing significant differences in iNOS-immunoreactive cell numbers among control, LPS-treated, and LPS + S-220-treated microglia. I, The Griess assay demonstrated a significant increase of nitrate concentration in the supernatant collected from LPS-treated microglial cultures compared with that of the control. Data are mean ± SEM. n = 3/group. One-way ANOVA with Bonferroni post hoc test: ***p < 0.001. Scale bars, 100 μm.

Figure 5.

The hydrogel degrades gradually, induces minimal immune response, and supports marked neurite outgrowth in 2-dimentional cultures. A, An image of a freshly prepared gel. B, The Griess assay showed low levels of nitrite in gel culture media compared with that of LPS-exposed primary microglia cultures. C, Gel stiffness showed a gradual decrease in gain modulus from the linear region of the raw data (0–100 Hz, rheology). D, In vitro cumulative percentage drug release versus time profile. Optimization of RGD concentration to promote neurite outgrowth of DRG explants in two dimensions. E, DRG explant in gel without RGD. F, DRG explant in 1.6 mm RGD gel. G, Quantification of maximal distance of neurite outgrowth showed significant differences at 1.6 and 3.2 mm RGD compared with control. H, Quantification of neurite densities of DRG explants at different concentrations of RGD. I, Cortical neurons on gel without RGD. J, Cortical neurons on gel with 1.6 mm RGD. K, Quantification of neurite length in gels with and without RGD, showing significantly enhanced neurite outgrowth on gels with 1.6 mm RGD. L, DRG explants grew extensive neurites in three dimensions within the gel with 1.6 mm RGD at 72 h after plating. M, Z-stack image of DRG neurons growing in three dimensions within the 1.6 mm RGD gel. N, Z-plane view of DRG neurons. DRG explants were immunostained for GAP-43 (E, F). DRG neurons and cortical neurons were immunostained for β-tubulin-III (I, J, M, N). Data are mean ± SEM. n = 3/group. Scale bars: E, F, 200 μm; I, J, L, 100 μm; M, 50 μm. One-way ANOVA with Bonferroni's post hoc (B, D, G, H), Kruskal–Wallis ANOVA on Ranks with Tukey post hoc (C), and unpaired Student's test (K): *p < 0.05; **p < 0.01; ***p < 0.001.

Figure 6.

In vitro evidence demonstrating the positive effects of combining the Fmoc hydrogel with Epac2 agonist application on neurite outgrowth. A–C, Assessing the effect of combining the Fmoc hydrogel with the Epac2 agonist S-220 in enhancing neurite outgrowth of DRG explants. A, Control explant growing in gel alone. B, Explant growing in gel in combination with S-220. C, Quantification of the maximal neurite length showed significantly longer neurites in the combination treatment group than the gel-only group. D–F, Assessing the effect of combining the gel with S-220 in enhancing neurite outgrowth of dissociated DRG neurons. D, DRG neurons growing in gel alone. E, DRG neurons growing in gel in combination with S-220. F, Quantification of the maximal neurite length showed significantly longer neurites in the combination treatment group than the gel-only group. DRG explants and neurons were immunostained for β-tubulin-III. Data are mean ± SEM. Scale bars: A, 250 μm; B, 500 μm; D, E, 100 μm. Unpaired Student's t test: *p < 0.05; **p < 0.01. n = 3/group.

Figure 7.

Timeline, study design, and schematic diagram of ex vivo preparation on organotypic spinal cord culture. A, Timeline of the experiments. B, Schematic diagram depicting the production and lesioning of organotypic spinal cord slice cultures. Yellow dotted lines indicate a spinal cord slice. Red dotted lines and the red arrow indicate a lesion gap. C, Quantification of live/dead assay did not show any difference before and after lesion (n = 10–15/group). D, Representative live/dead-stained fluorescence micrograph of a slice before lesioning and 2 d after lesion. E–I, Using the ex vivo model and β-tubulin-III staining, we found that, at 7 d after injury, treatment with either S-220 or hydrogel alone produced similar degrees of neurite outgrowth into the lesion gaps, but significantly more than that of the nontreated slices. E, Nontreated slice. F, S-220-treated slice. H, Hydrogel-treated slice. I, Combination-treated slice. White dashed lines indicate lesion margins. The agonist was added to culture media. G, The axonal profile analysis showed that, when the gel was combined with S-220, there was significantly more axon growth in the lesion gap than those following treatments with S-220 or hydrogel alone. C, Data are mean ± SEM. G, Data are mean ± SEM (box limits). Bars above and below each box represent 5% and 95% confidence limits. Circles represent individual biological replicates (n = 4–6). One-way ANOVA, Bonferroni's post hoc test: *p < 0.05; **p < 0.01; ***p < 0.001. Scale bars: D, 500 μm; E–I, 200 μm.

Figure 8.

S-220 incorporated into the hydrogel promotes neurite outgrowth and suppresses astrocyte activation. A, B, Representative images showing significant difference in neurite outgrowth between gel-only (A) and gel + 5 μm S-220 (B). White dashed lines indicate lesion margins. C, Neurite growth profiles across the lesion gap, showing a progressive increase with gel-only and gel + 5 μm S-220. D, Quantification showing the numbers of β-tubulin-III⁺ processes per square millimeter between control and different concentrations of S-220 delivered in the gel, with 5 μm having the greatest effect. E–G, Representative images exhibit astrocyte activation in the lesion using GFAP staining as an astrocyte marker. E, Control. F, Gel-only. G, Gel + 5 μm S-220. H, Quantification of the GFAP immunoreactivity intensity showed a significant reduction of mean gray value (optical density) in gel + 5 μm S-220 compared with the control and the gel-only treatment. I, Representative image of the relationship between GFAP (red) and β-tubulin-III (green) immunoreactive processes in a control injury condition. J, Higher-magnification image showing the collapse of growth cones (white arrowheads) when they meet activated astrocytes. K, Representative image of the relationship between GFAP and β-tubulin-III-immunoreactive processes in a lesion with combined treatment with gel + 5 μm S-220. L, Higher-magnification image showing the alignment of the astrocytes and β-tubulin-III-immunoreactive processes (white arrowheads). M, N, Levels of astrocyte reactivity were estimated by the overlapping of GFAP (red) and nestin (green). M, Representative image of GFAP/nestin reactivity in lesion sites of nontreated slices. N, Representative image of GFAP/nestin overlapping in slices treated with S-220 delivered by the hydrogel. O, Quantification of GFAP/nestin pixel overlapping. D, H, O, Data are mean ± SEM (box limits). Bars above and below each box represent 5% and 95% confidence limits. Circles represent individual biological replicates (n = 4–6). One-way ANOVA with Bonferroni's post hoc test: **p < 0.01; ***p < 0.001. Scale bars: A, B, E–G, M, N, 100 μm; I, K, 50 μm; J, L, 25 μm.

Figure 9.

S-220 incorporated into the gel suppresses microglia activation in lesioned spinal cord slice. A, Iba-1-immunoreactive cells inside the injury site in control lesion slices. White asterisks indicate activated morphology. B, Iba-1-immunoreactive cells inside the injury site in combination-treated slices. Yellow arrowheads indicate resting morphology. C, Noninjured microglia. Yellow arrowheads indicate resting morphology. D, Cell body perimeter was quantified and used as an indication of activation. Cell body size decreased more significantly in the presence of the hydrogel-only or hydrogel + 5 μm S-220, respectively. E, Representative image of the relationship between GFAP- and Iba-1-immunoreactive cells in an injured control. F, Representative image of the relationship between GFAP- and Iba-1-immunoreactive cells in injured slices treated with a combination of hydrogel + 5 μm S-220. D, Data are mean ± SEM (box limits). Bars above and below each box represent 5% and 95% confidence limits. Circles represent individual biological replicates (n = 4). One-way ANOVA with Bonferroni's post hoc test: **p < 0.01; ***p < 0.001. Scale bars: A–C, 25 μm; E, F, 50 μm.

Figure 10.

S-220 delivered by Fmoc hydrogel into the lesion of contusion injured spinal cord enhances locomotor functional recovery. Both S-220-treated and nontreated contusion-injured animals showed similar levels of locomotor function at 3 weeks after injury as assessed on the BBB open field task. However, by 2 weeks after treatment (indicated by the black arrow), S-220-treated animals showed significantly better locomotor function (3 points on the BBB scale) compared with the nontreated animals; a difference was maintained until 4 weeks after treatment (the last time point assessed). Data were analyzed with two-way ANOVA followed by Bonferroni's post hoc test: *p < 0.05. n = 4 or 5. Data are mean ± SEM.

Figure 11.

A schematic diagram summarizing the potential mechanisms of Epac2 activation by S-220 in the in vitro and ex vivo studies. Blue dotted rectangle and blue dotted arrows represent the mechanisms for in vitro studies. Red dotted rectangle and red dotted arrows represent the mechanisms for ex vivo studies. Our findings from the in vitro studies with individual monocultures of neurons, astrocytes, and microglia demonstrate that S-220 has direct and individual effects on the following: (a) neurons, promoting their growth possibly via the subsequent activation of Rap 1/B-Raf/ERK signaling pathway (Murray and Shewan, 2008; Murray et al., 2009; Wei et al., 2016); (b) astrocytes, reducing their activation by LPS possibly via the IL-6/STAT3 signaling pathway (Takanaga et al., 2004); and (c) microglia, reducing their activation by LPS possibly via arginase I signaling pathway (Ghosh et al., 2016). S-220 binds to the cyclic nucleotide binding (CNB) domains of Epac2 protein within these cells, triggering downstream signaling pathways (Tucker, 2014). Our findings from the ex vivo studies, however, suggest that S-220 manifests its effects on these cells in a more complex and intermingled manner as follows: (1) elongated astrocytes following S-220 treatment may resemble radial glial progenitors, releasing axon-growth supporting molecules (Garcia et al., 2004); (2) elongated astrocytes following S-220 treatment may provide direct guidance for axonal growth (Anderson et al., 2016; Robichaux and Cheng, 2018); reduced astrocyte activation by S-220 may lead to upregulate the gene for NMDA receptor subunit NR2C, which is a key part in the tripartite synapse regulating astrocyte–neuron communication (Paco et al., 2016); (4) reduced astrocyte activation by S-220 may downregulate the genes responsible for the production of proteoglycans, including CSPGs, thereby reducing their inhibitory effects on axonal outgrowth (Paco et al., 2016); (5) reduced astrocyte activation by S-220 may involve STAT3/nestin, and subsequently may regulate other inflammatory cell behavior, such as that of microglia at the spinal cord lesion site (Wanner et al., 2013); and (6) reduced microglial activation via astrocyte regulation may lead to reduced production of proinflammatory mediators (Steininger et al., 2011), thereby promoting axonal outgrowth. It is also highly likely that S-220 may have direct and individual effects on these three types of cells in the ex vivo model, including transcriptional changes in arginase I, IL-6, secretory leukocyte protease inhibitor, and metallothionein in neurons (d) (Hannila and Filbin, 2008; Siddiq and Hannila, 2015), which have been shown to promote axonal outgrowth and overcome inhibitory molecules. Red and blue small circles with diagonal across represent “stop/inhibit”. Green arrows represent “promote/enhance”; ERK = extracellular signal-regulated kinases; STAT3 = signal transducer and activator of transcription 3; IL-6 = interleukin 6; NMDA = N-methyl-D-aspartate.