Dendritic spine remodeling following early and late Rac1 inhibition after spinal cord injury: evidence for a pain biomarker

Abstract

Neuropathic pain is a significant complication following spinal cord injury (SCI) with few effective treatments. Drug development for neuropathic pain often fails because preclinical studies do not always translate well to clinical conditions. Identification of biological characteristics predictive of disease state or drug responsiveness could facilitate more effective clinical translation. Emerging evidence indicates a strong correlation between dendritic spine dysgenesis and neuropathic pain. Because dendritic spines are located on dorsal horn neurons within the spinal cord nociceptive system, dendritic spine remodeling provides a unique opportunity to understand sensory dysfunction after SCI. In this study, we provide support for the postulate that dendritic spine profiles can serve as biomarkers for neuropathic pain. We show that dendritic spine profiles after SCI change to a dysgenic state that is characteristic of neuropathic pain in a Rac1-dependent manner. Suppression of the dysgenic state through inhibition of Rac1 activity is accompanied by attenuation of neuropathic pain. Both dendritic spine dysgenesis and neuropathic pain return when inhibition of Rac1 activity is lifted. These findings suggest the utility of dendritic spines as structural biomarkers for neuropathic pain.

Keywords: Rac1; biomarker; dendritic spines; pain; spinal cord injury.

Fig. 1.

Study design. Animals (n values) were weight-matched and randomly assigned to Sham (no injury) or spinal cord injury (SCI) groups. To assess the effect of early treatment on pain and dendritic spine remodeling, a subpopulation of animals received intrathecal catheters 7 days before SCI surgery. Within 24 h of SCI, these animals received an infusion regime of NSC23766, a Rac1-inhibitor, or Vehicle twice daily for 3 days. To determine whether drug withdrawal would allow post-SCI dendritic spine dysgenesis to return and permit a relapse of neuropathic pain, we implanted catheters 3 wk after SCI and infused NSC23766 or Vehicle twice daily for 3 days. Note that Sham animals received catheter implants in the 3rd wk after surgery (not shown in graphic). Testing for behavioral and electrophysiological signs of injury-induced pain was performed at the 4-wk time point post-SCI. Animals were killed for histological analysis of dendritic spine morphology following functional studies. In a subpopulation of animals, treatments were withdrawn and both functional and histological assessments were performed at the later 8-wk time point. These procedures produced 7 treatment groups: Sham, SCI + Early Vehicle (4 wk), SCI + Early anti-Rac1 (4 wk), SCI + Vehicle (4 wk), SCI + Vehicle (8 wk), SCI + Late anti-Rac1 (4 wk), and SCI + Withdrawn anti-Rac1 (8 wk).

Fig. 2.

SCI biomechanical data. All animals with SCI received a thoracic level (T9) contusive SCI with the Infinite Horizon (IH) Impactor device. Experimental applied impact force (value assigned into the IH system before executing the injury) was set at 170 kdyn (see dotted line). Analysis of biomechanical data provided automatically by the IH Impactor demonstrated no difference between any group for actual applied force (kdyn; A) or dorsal cord displacement (μm; B). C: Basso, Beattie, Bresnahan (BBB) testing demonstrated no significance differences in locomotor scores across SCI groups (P > 0.05). All naïve animals had baseline locomotor scores of 21. Testing was performed at the time point shown within parentheses. For clarity, scores were combined for SCI and vehicle-treated animals assigned to the 4- and 8-wk assessment timepoint.

Fig. 3.

Golgi-stained coronal sections of spinal cord tissue. A: representative image of the dorsal horn with a wide-dynamic range (WDR) neuron located in lamina V (highlighted green). B: as shown in a high-power field from A, inset, a WDR neuron exhibits dendritic branches that extend several hundred microns through the plane of section. C: dendritic spines appear on neuronal dendrites and vary in density and morphology, depending on the condition of the animal, e.g., injury, treatment. Note the qualitative differences in dendritic spine profiles between Sham and SCI animals with or without Rac1-inhibitor treatment. D: classification of thin and mushroom-shaped spines. Scale bar for A = 500 μm; B = 100 μm; C = 10 μm.

Fig. 4.

Reconstruction of spinal cord dorsal horn neurons. To profile dendritic spines, we reconstructed the entire branch structure of neurons sampled from the dorsal horn. A: contour traces were used to mark the location of sampled neurons (green dot, arrow) within the gray matter. Density and distribution measurements were performed on 3-dimensional reconstructions from Sham (B), SCI + early Vehicle (C), SCI + Vehicle (4 wk) (D), SCI + Vehicle (8 wk) (E), SCI + Early anti-Rac1 (F), SCI + anti-Rac1 (4 wk) (G), and SCI + Withdrawn anti-Rac1 (8 wk) (H). An ∼30- to 50-μm length of dendritic branch from neurons in B–H (gray shaded region) show thin-shaped (blue dots) and mushroom-shaped (red dots) dendritic spines. Scale bar in A = 500 μm; B–H = 50 μm.

Fig. 5.

Early treatment effects on dendritic spine density, spatial distribution, and shape. We analyzed reconstructed neurons and assessed changes in dendritic spine density (A, B, and C), distribution (D, E, and F), and shape, including spine length and head diameter (G and H). Total spine density (A), which included all spine shape categories, thin-shaped spines (B), and mushroom-shaped spines (C) resulted in significant increased after SCI and treated early (within 24 h postinjury) with vehicle or Rac1-inhibitor compared with Sham (*P < 0.05). There was no statistical difference between the 2 SCI groups (P > 0.05). An analysis of the distribution of total spines (D), thin-shaped spines (E), and mushroom-shaped spines (F) revealed significant and preferential increases in density within the most proximal regions from the cell body. Compared with Sham, SCI with early vehicle treatment resulted in density increases in the 50- to 150-μm regions for total spines, 50- to 200-μm region for thin spines, and in the 50-μm region for mushroom spines (*P < 0.05). Early treatment with the Rac1-inhibitor had similar increases (note this was further restricted to the closest regions): 50–100 μm for total and thin spines, and 50 μm for mushroom spines (#P < 0.05). Both spine length (G) and spine head diameter (H) increased following SCI and early vehicle treatment. Early Rac1-inhibitor treatment failed to restore close-to-normal spine head diameter, as demonstrated by a significant increase compared with Sham (*P < 0.05). A comparison across SCI groups demonstrated a decrease in spine head diameter in animals treated early with the Rac1 inhibitor compared with vehicle-treated SCI animals (#P < 0.05). All graphs are means ± SE.

Fig. 6.

Neuropathic allodynia develops following SCI. All animals with SCI treated early with vehicle or the Rac1-inhibitior continued to develop significant neuropathic allodynia. Von Frey filament testing demonstrated a significant decrease in hindpaw withdrawal thresholds compared with Sham (*P < 0.05). There were significant differences between the SCI groups (P > 0.05). Graph is mean ± SE.

Fig. 7.

Electrophysiological signs of neuropathic pain after SCI with early treatments. A: all SCI groups treated early with vehicle or Rac1-inhibitor exhibited significant increases in single unit spike firing activity compared with Sham following brush, press, and pinch stimuli applied to cutaneous receptive fields (*P < 0.05). Similarly, in testing for punctate hyperalgesia (B) with graded Von Frey filament application, all SCI animals had significantly increased unit-firing activity compared with Sham (*P < 0.05, SCI + Early Vehicle vs. Sham; #P < 0.05, SCI + early anti-Rac1 vs. Sham). There was no significant difference between either SCI group (P > 0.05). An assessment of cutaneous receptive field size (C) demonstrated a significant increase in the receptive fields of both SCI groups compared with Sham (P < 0.05). Although early treatment with the Rac1-inhibitor did not return receptive field area back to normal, Sham levels, drug treatment decreased cutaneous receptive field area compared with SCI animals treated only with vehicle (P < 0.05). D: representative maps show the qualitative changes in receptive field area for a subset of sampled neurons. All graphs are mean ± SE.

Fig. 8.

Treatment withdrawal permits the reappearance of abnormal dendritic spine profiles. To determine whether withdrawal of the Rac1-inhibitor would allow post-SCI dendritic spine dysgenesis to return we analyzed reconstructed sampled WDR neurons. As with above, we profiled dendritic spine density (A, B, and C), distribution (D, E, and F), and shape, including spine length and head diameter (G and H). The density of total spines (A), thin (B)-, and mushroom (C)-shaped spines increased in SCI + Vehicle animals at the 4-wk time point compared with Sham (*P < 0.05). At 8-wk post SCI, mushroom spine density increased in SCI animals with vehicle treatment compared with Sham (*P < 0.05). There was no significant difference between total or thin-shaped spines compared with Sham in SCI animals treated with vehicle and analyzed at the 8-wk time point (P > 0.05). Assessed at the 4-wk time point and within 24 h following the end of treatment administration, the Rac1-inhibitor decreased spine density in total and mushroom-shaped spine categories compared with SCI and vehicle treatment (#P < 0.05). In addition, there was a significant decrease in mushroom-shaped spine density with Rac1-inhibitor treatment at 4 wk when compared with SCI and vehicle treatment at the 8-wk time point. There was no difference in thin-shaped spine density between these 2 groups (P > 0.05). Withdrawal of Rac1-inhibitor treatment and assessment at the 8-wk time point demonstrated no significant difference in total or thin-shaped spines compared with Sham (P > 0.05). In contrast, mushroom-shaped spines increased significantly compared with Sham following withdrawal of the Rac1-inhibitor treatment (*P < 0.05). Total and thin-shaped spines continued to have decreased spine density at the 8-wk time point (in the absence of Rac1-inhibitor treatment; SCI + Withdrawn anti-Rac1) when compared with SCI and vehicle treatment at the 4-wk time point (#P < 0.05). There was no significant difference in mushroom shaped spines between SCI animals treated with the Vehicle at the 4- or 8-wk time point compared with SCI + withdrawn anti-Rac1 group at the 8-wk time point (P > 0.05). An assessment of distribution of total spines (D) and thin-shaped spines (E) demonstrated a preferential increase in density in the most proximal dendritic branch regions, at 50–150 μm, in SCI animals treated with vehicle at the 4-wk time point compared with Sham (*P < 0.05). An analysis of mushroom-shaped spine distribution (F) showed a significant increase in density at the 50 and 100 μm dendritic branch regions in the SCI and vehicle treatment group at both the 4- and 8-wk time point compared with Sham (*P < 0.05 and #P < 0.05). There were no differences in distribution for all spine categories in comparisons across any other treatment group (P > 0.05). Spine length (G) and head diameter (H) increased following SCI and vehicle treatment at both the 4- and 8-wk time point (*P < 0.05). Treatment with the Rac1-inhibitor in SCI animals significantly decreased spine length compared with SCI and vehicle treatment when assessed at the 4-wk time point and decreased head diameter compared with SCI and vehicle at the 8-wk time point (#P < 0.05). There was no significant difference between animals with Rac1-inhibitor treatment at 4-wk and Sham animals (P > 0.05). Withdrawal of the Rac1-inhibitor and measurements at 8-wk demonstrated an increase in spine head diameter compared with Sham (*P < 0.05, arrows denote drug treatment followed by withdrawal in the same animals). There was no difference between SCI animals that had drug withdrawn compared with SCI and vehicle treatment (P > 0.05), suggesting that in the absence of drug treatment spine shapes returned to SCI-induced levels. All graphs are mean ± SE.

Fig. 9.

SCI-induced pain relapse after treatment withdrawal. SCI and vehicle treatment at the 4- and 8-wk time point resulted in a significant decrease in hindpaw withdrawal threshold compared with Sham (*P < 0.05), indicative of injury-induced tactile allodynia. Compared with SCI and vehicle-treated animals at the 4- and 8-wk time point Rac1-inhibitor treatment significantly increased withdrawal threshold (#P < 0.05). There was no significant difference in thresholds between Rac1-inhibitor treatment in SCI animals, tested at the 4-wk time point, and Sham levels (P > 0.05). Graph is mean ± SE.

Fig. 10.

Electrophysiological signs of SCI-induced neuropathic pain return after treatment withdrawal. A: all SCI groups treated with vehicle at the 4- and 8-wk time point demonstrated increased spike firing activity in response to brush, press, or pinch stimulation applied to units' cutaneous receptive fields, compared with Sham (*P < 0.05). Compared with SCI and vehicle treatment at the 4-wk time, Rac1-inhibitor treatment in SCI animals resulted in a significant decrease in evoked firing activity in response to brush stimuli (*P < 0.05); treatment with the Rac1-inhibitor did not significantly affect evoking firing spike responses to press or pinch stimuli compared with SCI and vehicle-treated animals at the 4-wk time point (P > 0.05). Withdrawal of the Rac1-inhibitor from SCI animals resulted in an increase in firing activity evoked by brush, press, and pinch stimuli at the 8-wk time point compared with Sham (*P < 0.05, arrows denote drug treatment followed by withdrawal in the same animals). In testing for punctate hyperalgesia (B), SCI animals with vehicle treatment at the 4- or 8-wk time point exhibited significant increases in firing activity in response to all or a majority of the graded Von Frey filament intensities (0.04 to 26 g) compared with Sham [*P < 0.05, SCI + Vehicle (4 wk) vs. Sham; #P < 0.05, SCI + Vehicle (8 wk) vs. Sham]. Treatment with the Rac1 inhibitor in SCI animals, assessed at the 4-wk time point, decreased peripherally evoked firing activity at several filament intensities compared with SCI and vehicle treatment at the 4- or 8-wk time point [+P < 0.05, SCI + Vehicle (4 wk) vs. SCI + anti-Rac1 (4 wk); &P < 0.05, SCI + Vehicle (8 wk) vs. SCI + anti-Rac1 (4 wk)]. Withdrawal of the Rac1-inhibitor and assessment at the 8-wk time point resulted in a significant increase in spike firing at activity at graded intensities at 0.04 and 6.0 g compared with ongoing drug treatment at the 4-wk time point (@P < 0.05). In addition, drug withdrawal from SCI animals (and then assessed at the 8-wk time point) also resulted in a significant increase in evoked firing activity compared with Sham at nearly all graded intensities (§P < 0.05). C: cutaneous receptive fields increased in receptive areas following SCI with vehicle treatment at the 4- and 8-wk time point compared with Sham (P < 0.05). The receptive field size in SCI animals treated with the Rac1-inhibitor was similar to Sham levels (P > 0.05). Withdrawal of the Rac1-inhibitor from SCI animals by the 8-wk time point resulted in a significant increase in receptive field area compared with both Sham and Rac1-inhibitor treatment at 4-wk (P < 0.05). D: representative maps show the qualitative changes in receptive field area for subset of sampled neurons. All graphs are mean ± SE.