Rolipram-loaded PgP nanoparticle reduces secondary injury and enhances motor function recovery in a rat moderate contusion SCI model

Abstract

Spinal cord injury (SCI) results in immediate axonal damage and cell death, as well as a prolonged secondary injury consist of a cascade of pathophysiological processes. One important aspect of secondary injury is activation of phosphodiesterase 4 (PDE4) that leads to reduce cAMP levels in the injured spinal cord. We have developed an amphiphilic copolymer, poly (lactide-co-glycolide)-graft-polyethylenimine (PgP) that can deliver Rolipram, the PDE4 inhibitor. The objective of this work was to investigate the effect of rolipram loaded PgP (Rm-PgP) on secondary injury and motor functional recovery in a rat moderate contusion SCI model. We observed that Rm-PgP can increase cAMP level at the lesion site, and reduce secondary injury such as the inflammatory response by macrophages/ microglia, astrogliosis by activated astrocytes and apoptosis as well as improve neuronal survival at 4 weeks post-injury (WPI). We also observed that Rm-PgP can improve motor functional recovery after SCI over 4 WPI.

Introduction

Traumatic spinal cord injury (SCI) is a major source of morbidity and mortality throughout the world. The global incidence of SCI is 10.5 cases per 100,000 persons, and there are over 700,000 new cases annually worldwide.¹ SCI frequently results in paralysis, hypoesthesia or autonomic disorder that persist throughout the patient’s lifetime with limited spontaneous recovery. For a patient with cervical trauma, the medical expenses in the first year post-injury are about one million dollars in the United States indicating that SCI poses a great financial burden to patients, their families, and the healthcare system.² Therefore, the treatment of SCI is an important issue in terms of medical and societal problems.

The initial traumatic insult results in primary damage to axonal tracts and neural cells. This is followed by the secondary injury response, a cascade of pathophysiological reactions that evolve over several weeks post-injury and substantially increase tissue destruction, cell death, and de-myelination.³ Compared to the unpredictable primary insult, the progressively developing nature of secondary injury offers an opportunity for therapeutic intervention. Furthermore, the area of secondary injury far exceeds the primary injury. Currently, systemic methyl prednisolone (MP) administration within 8 hours following an acute SCI is the only therapy, but questions raised about its efficacy and side effects have limited its use.^4–6 Therefore, there is a need for new therapeutics to reduce secondary in-jury and improve functional recovery after traumatic SCI.

One important secondary injury event is a significant decrease in intracellular cAMP level in the injured spinal cord.⁷ cAMP is a critical second messenger and intracellular metabolite that influences neuronal cell survival and neurite outgrowth during development and in response to injury.^8–10 Axonal outgrowth and glial reactivity can also be affected by the activation of the exchange protein activated by cAMP (EPAC), which acts as nucleotide exchange factors for the small GTPase Rap.^11–13 Indeed, restoring or elevating cAMP levels via administration of cAMP analogues such as dibutyl cAMP or phosphodiesterase 4 (PDE4) inhibitors has been investigated in several preclinical studies.^9,14,15 Following contusive SCI, systemic delivery of Rolipram (Rm), a PDE4 inhibitor, over 2 weeks prevented the injury-induced reduction in cAMP usually observed in the injured spinal cord and promoted axonal sparing and growth into Schwann cell grafts that spanned the injury site,^7,14 protected oligodendrocytes from secondary cell death, and enhanced functional recovery.^16–18 Despite these promising results in pre-clinical studies, continuous systemic administration of Rm is a challenge for clinical translation. All studies used osmotic minipumps for Rm administration, which inevitably increases the risk of infection.

Previously, we reported the synthesis and characterization of a cationic, amphiphilic graft copolymer, poly (lactide-co-glycolide)-graft-polyethylenimine (PgP) and demonstrated that rolipram can be loaded within its hydrophobic core and amount of Rm loaded into PgP was approximately 7 fold higher than its normal aqueous solubility.¹⁹ Local delivery of Rm by PgP (Rm-PgP) to the injury epicenter restored cAMP levels for up to 3 days post-injury (DPI) and mitigated secondary injury as determined by reductions in both apoptosis and the inflammatory response in a rat severe compression injury model.¹⁹ Although our previous study demonstrated the ability of Rm-PgP to mitigate many features of the secondary injury response, we used a 3 DPI time point (acute injury phase) in a rat compression SCI model and potential benefits to motor function recovery and neuropathic pain were not analyzed.

In this study, we extended our investigation of the therapeutic efficacy of Rm-PgP to motor function recovery and neuropathic pain development over 4 weeks (chronic injury phase) in a clinically relevant moderate, midthoracic contusion SCI model. We also evaluated the effect of Rm-PgP on cAMP restoration and the therapeutic efficacy of Rm-PgP on secondary injury such as the inflammatory response, apoptosis, and glial scar formation by histological analysis at 4 weeks post-injury (chronic injury phase).

Materials and methods

Materials

Rolipram (Rm) was purchased from LC Laboratories (Woburn, MA, USA). Acetonitrile (HPLC grade), water (HPLC grade) and ethanol were purchased from Sigma (Sigma-Aldrich, MO, USA). Mouse/Rat cAMP Parameter Assay ELISA kit was purchased from R&D Systems (Minneapolis, MN, USA). The ApopTag Fluorescein In Situ Apoptosis Detection Kit were obtained from EMD Millipore (Darmstadt, Germany). Mouse anti-beta III tubulin (2G10) and goat anti-mouse Cy3 conjugated IgG antibodies were purchased from Abcam (Cambridge, MA, USA).

Methods

Preparation of rolipram-loaded PgP nanoparticles

PgP was synthesized using PLGA (4 kDa, 50:50, Durect Corporation, Cupertino, CA, USA) containing carboxylic end groups and branched PEI (bPEI, 25 kDa, Sigma-Aldrich, St. Louis, MO, USA) and characterized the structure via ¹H- NMR and molecular weight via gel permeation chromatography (GPC) in our laboratories as previously reported²⁰. Rm was loaded into PgP using solvent evaporation method as previously reported¹⁹. Briefly, Rm was dissolved in ethanol (10 mg/ml) and PgP was dissolved in water (1 mg/ml), respectively. 100 μl of Rm (10 mg/ml) solution was added into 1 ml PgP (1 mg/ml). The mixture of Rm and PgP was sealed and gently stirred for 4 hours at room temperature (RT), then ethanol was evaporated overnight. The Rm-PgP solution was filtered using a sterile 0.2 μm cellulose acetate syringe filter to remove unloaded Rm. The amount of Rm loaded in the PgP was measured by High Performance Liquid Chromatography (HPLC) using Rm standard curve. The HPLC system was composed of 1525 binary HPLC pump, 2998 photodiode array detector, and autoinjector (Waters, Milford, MA, USA) with a C18 column (Symmetry C18, 3.5 μm, 4.6x75mm, Waters), Showa Denko America, New York, NY). The mobile phase was water: acetonitrile (60:40) and flow rate was 0.6 ml/minute. The injection volume was 20 μl and run time was 6 minutes. Rolipram was detected by UV detection at 280 nm. The Rm-PgP was freshly prepared before each surgery and characterized by HPLC for all the experiments performed in the study.

Generation of a rat contusion SCI model and Rm-PgP treatment

All procedures involving animals were conducted under the supervision of the Clemson University Animal Research Committee (approved animal protocol no. AUP 2019–043). Sprague Dawley rats (male, 200–250 grams, 7–8 weeks, Charles River) were deeply anesthetized with ketamine/xylazine (80–100 mg/kg ketamine and 10–15 mg/kg xylazine). The T9-T10 spinous processes were identified and a 4-cm longitudinal incision over the dorsal mid-thoracic region made using a #10-blade scalpel. The T9 spinous process were removed using a rongeur and the ligamentum flavum removed to expose the intervertebral space. Infinite Horizon Impactor (IH-0400, PSI, LLC, KY, USA) was used and rat was positioned on the impactor stage and spinal cord stabilizing forceps were applied to the T8 and T10 processes of the vertebral column and the impactor tip was positioned on the exposed dorsal surface of the spinal cord. The impactor (IH-0400, PSI) was triggered to deliver a contusion injury with a force of 200 kDyne. Animals were randomly divided into 3 groups; 1) Sham groups: rats underwent the same T9 laminectomy without injury or intraspinal injection, 2) Untreated SCI group: Immediately after the impact, saline was injected intraspinally at the injury site using microinjector/pump (MICRO2T, WPI, Sarasota, FL) (Injection speed: 2 μl/min) with a 30-gauge Hamilton syringe, 3) Rm-PgP treated group: Immediately after the impact, Rm-PgP (0.5 μg Rm/μl, 10 μg Rm/rat) was injected intraspinally at the injury site. PgP without Rm group was not included because PgP alone did not show significant therapeutic benefit nor any toxic side effect in our previous in vitro study¹⁹.

Following injection, the paraspinal muscle was closed with 4–0 vicryl suture and the skin was closed with surgical wound clips. After surgery, animals were injected buprenorphine (0.03 mg/kg, Par Pharmaceutical) and cefazolin (40 mg/kg, Med Pharma) and kept warm until sternal. Bladders were expressed manually three times a day until return of bladder function. Scheme 1 shows the summary of experiments performed to evaluate the effect of Rm-PgP in this study.

Scheme 1.

Experimental design for Rm-PgP treatment in a rat moderate contusion SCI model.

Effect of Rm-PgP on cAMP level in injured spinal cord

To evaluate the effect of Rm-PgP treatment on cAMP level, animals (n=4/group) were decapitated under deep anesthesia by isoflurane gas at 1, 2, and 3 DPI. The 0.5 cm long spinal cords (0.25 cm-long piece from the epicenter) were harvested and rapidly frozen in liquid nitrogen and stored at −80 °C until assayed. The concentration of cAMP in the injured tissues were measured by ELISA using a Mouse/Rat cAMP Parameter Assay Kit. Briefly, the tissue samples were weighed individually, then homogenized in 0.1N HCl solution at a 1: 5 (weight of tissue: volume of HCl) ratio. The homogenized solutions were then centrifuged at 10,000 g for 10 minutes at 4 °C. The supernatant was collected, neutralized with 0.1N NaOH and diluted 2-fold with Diluent RD5–55. Streptavidin-coated plates were washed thoroughly using wash buffer and then incubated with biotinylated mouse monoclonal antibodies to cAMP for 1 hour at RT. After incubation, solution was removed, washed using wash buffer. cAMP conjugated to horseradish peroxidase was added to all the wells and then cAMP standards, controls, and samples were added into appropriate wells and incubated for 2 hrs at RT on a shaker (500 rpm, Southwest Science, Trenton, NJ, USA). After incubation, the solutions were removed, washed thoroughly 4 times with wash buffer. The chromogenic substrate solution was added into each well and incubated for 30 minutes in the dark and then stop solution was added. The absorbance was measured at 450 nm and 570 nm on a plate reader (BioTek, Winooski, Vermont, USA) within 30 min. The absorbance at 570 nm was subtracted from that at 450 nm to correct the measurement from background. The data was presented with mean ± standard deviation (STD).

Effect of Rm-PgP on Motor function after SCI

To evaluate the effect of cAMP restoration by Rm-PgP on motor function after spinal cord injury, Basso, Beattie, Bresnahan (BBB) test were performed in an arena (90 cm x90 cm, 30 cm height) up to 4 weeks post-injury (WPI)²¹. Briefly, the rats (n=6/group) were acclimated in the behavior test room for 1 hour before testing. Animals were individually observed for a period of 4 min by two investigators blinded to the treatment condition. The movements of the hip, knee and ankle joints were mainly observed at the early stage of recovery. As recovery occurs, rats may progress into an intermediate stage exhibiting stepping and may reach final stage of recovery coordinating limb movements with the tail up and no trunk instability. Each hindlimb was scored separately, and left and right scores were averaged at each timepoint to generate 1 score per rat. The data was presented with mean ± STD.

Effect of Rm-PgP on Neuropathic pain

The effect cAMP restoration by Rm-PgP treatment on neuropathic pain was evaluated by von Frey test^22,23. For mechanical allodynia, hindpaw withdrawal thresholds were determined using a series of von Frey filaments, with calibrated bending forces ranging from 0.08 to 200 g (Bioseb, Pinellas Park, FL, USA) up to 4 WPI. Rats (n=6/group) were placed individually in wire mesh-bottom cages and acclimated for 10 min before assessment began. von Frey filaments were applied to the hindpaws using the Up-Down method beginning with the 5.18 filament²⁴. The withdrawal threshold was defined as the minimum gauge von Frey filament that elicited a withdrawal reflex in more than 50% of its applications. Withdrawal thresholds were determined for both hindpaws and average thresholds were calculated. The data was presented with mean ± STD.

Tissue preparation for histological analysis

At 4 weeks post-injury (WPI) after behavior test, the rats were deeply anesthetized by euthasol (Intraperitoneal injection, 100 mg/kg, Virbac) and sacrificed by cardiac perfusion with 0.9 % saline followed by 4 % paraformaldehyde in PBS (pH 7.4). 10 mm long block of spinal cord was excised centered at the injury epicenter and post-fixed in 4% paraformaldehyde solution for 48 hours at 4 C° followed by sucrose solution gradient (10 %, 20 %, and 30 % sucrose). The fixed tissue was then embedded in optical cutting temperature OCT compound (Fisher HealthCare, Houston, TX) at −20 C°. Spinal cords were sagittally sectioned at 10 μm thickness and mounted onto positively charged glass slides in 10 series (each section was separated by 100 μm) and stored at −20 °C.

Effect of Rm-PgP on lesion size and spared tissue volume

Spinal cord tissue sections (5 sections/rat, 6 rats/group) were selected at 5 different positions as shown in Figure 3A. Midline of spinal cord sections was identified and five equidistant sagittal sections spanning mediolaterally across the spinal cord (left (-) 0.4 mm, left (-) 0.2 mm, midline, right (+) 0.2 mm, right (+) 0.4 mm) were selected. The sections were stained with hematoxyline and eosin (H&E), imaged, and the lesioned area was measured by ImageJ. For each sample, the percentage of spared tissue of each section with a length of 10 mm was calculated by subtraction of measured lesion area from the total area of the section and then normalized to total area of sham.

Figure 3.

Effect of Rm-PgP on lesion size. (A) 5 tissue sections per rat were selected as shown in the schematic. (B) Representative images of midline section from untreated and Rm-PgP treated SCI rats stained with H & E. (C) The area of spared tissue was measured in each section using Image J and the % average spared area at each position was calculated by subtraction of measured lesion area from the total area of the section and then normalized to total area of sham. (D) Lesion length was measured in the midline tissue section.

Effect of Rm-PgP on secondary injury

The effects of Rm-PgP treatment on inflammatory cell activation, glial scar formation, neuronal cell survival by immunohistochemistry (IHC) and apoptosis by TUNEL assay were evaluated (5 sections/rat, n=6 (rats/group). For IHC, sections were selected as shown in Fig 4A, 5A, and 6A. Sagittal sections from sham, SCI untreated, and Rm-PgP treated groups were washed with 0.1M PBS and incubated using blocking solution (5 % bovine growth serum and 0.05 % TritonX 100 in 0.1M PBS) for 1 hour at RT. Sections were rinsed with PBS and incubated overnight at 4 °C with either mouse monoclonal anti-macrophage/monocyte CD68 clone-ED1 (ED1) antibodies (1:200; Cat.# MAB1435, Millipore, Billerica, MA) for activated microglia and macrophages, rabbit polyclonal anti- glial fibrillary acid protein (GFAP) (1:200; Cat.# ab7260, Abcam, Cambridge, MA) for activated astrocytes, or mouse monoclonal anti-NeuN antibodies (1:200; Cat.# MAB377, Millipore) for neuronal nuclei. Sections were washed with PBS and then incubated with goat anti-mouse Cy3-conjugated secondary antibodies (1:200; Cat.# 115–165-003, Jackson ImmunoResearch, West Grove, PA, USA) or AlexaFluor^® 488-conjugated goat anti-rabbit secondary antibodies (1:500; Cat.# A-11008, Thermo Fisher Scientific, Hampton, NH, USA) for 1 hour at RT. Sections were washed with PBS and then coverslipped with VectaShield mounting media containing DAPI (Vector Laboratories, Burlingame, CA, USA). Images were captured using an Axiovert 40 CFL microscope (Carl Zeiss, Oberkochen, Germany). The effect of Rm-PgP treatment on apoptosis was evaluated by terminal deoxynucleotidyl transferase-mediated dUTP nicked end labeling (TUNEL) assay using an ApopTag^® Plus Fluorescein in situ Apoptosis Detection Kit (S7111, EMD Millipore, Temecula, CA, USA). Sections were selected as shown in Fig 7A. Images of the entire stained sections were captured at 50× using an AZ100 fluorescence microscope (Nikon, Tokyo, Japan) and the images were stitched together using Photoshop software. For ED1+ cell and TUNEL + cells, five equidistant sagittal sections spanning mediolaterally across the spinal cord (l-0.4 mm, −0.2 mm, midline, +0.2 mm, +0.4 mm) were analyzed to determine whether Rm-PgP alters cell distribution mediolaterally within the spinal cord. The number of positively labeled cells was normalized to total spinal cord area in each section. To determine if the number of labeled cells differs in the rostrocaudal direction with Rm-PgP treatment, high magnification images at five equidistant rostrocaudal locations centered at the epicenter in each of the 5 sagittal sections (epicenter (blue), 0.2 mm (yellow) and 0.4 mm (green) in the rostral direction as well as 0.2 mm (red) and 0.4 mm (black) in the caudal direction were analyzed. Cell counts at each of the rostrocaudal locations in 5 sections were summed for each animal For activated astrocytes, the fluorescence intensity of GFAP stained sections was measured and normalized to intensity from sections from sham animals.

Figure 4.

Effect of Rm-PgP treatment on inflammatory response after moderate contusion SCI. (A) Sections were selected as shown in scheme (Total 30 sections: 5 sections/ rat, 7 rats/ group). Five equidistant sagittal sections spanning mediolaterally across the spinal cord (left (-) 0.4 mm, left (-) 0.2 mm, midline, right (+) 0.2 mm, right (+) 0.4 mm) are denoted by the gray planes and are utilized to determine whether Rm-PgP alters cell distribution mediolaterally within the spinal cord. Colored, transverse planes represent equidistant locations in each of the 5 sagittal sections where high magnification images were utilized to determine if the number of labeled cells differs in the rostro-caudal direction with Rm-PgP treatment. Injury epicenter (blue), R 2 mm (yellow) and R 4 mm (green) in the rostral direction, and C 2 mm (red) and C 4 mm (black) in the caudal direction. (B) Number of ED1+ cells normalized to total spinal cord area of each of the 5 sagittal sections spanning the width of the spinal cord, (C) Number of ED1+ cells normalized to the area of spinal cord at the 5 different rostrocaudal positions (R 4 mm, R 2 mm, epicenter, C 2 mm, C 4 mm) on each of the 5 sagittal sections examined. (D) The number of total ED1+ cells within each rostro-caudal position were summed across the 5 sagittal sections examined. Left bar: Untreated and Right bar: Rm-PgP treated SCI group. *p<0.05 and **p<0.01 (E) Representative images of ED1+ microglia/ macrophages (red) from sagittal sections (Top: untreated SCI, bottom: Rm-PgP treated SCI). Blue: DAPI countstained cells 50X original magnification. The white boxes in each image represent higher magnification (200 X).

Figure 5.

Effect of Rm-PgP on Astrogliosis. A) Scheme of the 3 different sagittal section selected, B) Average fluorescence intensity of sections stained against GFAP normalized to that of sham group at 3 sagittal sections, *p<0.05 and **p<0.001. C) Representative images of GFAP+ cells in green. Dotted line indicates lesion area. 50X original magnification. Left: SCI untreated and Right: Rm-PgP treated group. The white boxes in each image represent higher magnification (200 X).

Figure 6.

Effect of Rm-PgP on neuronal cell survival. A) Scheme of the 3 different sagittal section selected, B) Normalized average number of NeuN+ cells at each position (n=6 sections/position). C) Normalized total number of NeuN+ cells (total 12 sections/group) (*p < 0.05). D) Representative images of NeuN + cells in red and DAPI + cells in blue. Left: untreated SCI, Right: Rm-PgP treated group. 50X original magnification. Scale bar=0.1 mm.

Figure 7.

Effect of Rm-PgP on apoptosis. A) Scheme of the 5 different sagittal section position, B) Number of TUNEL+ cells normalized to total spinal cord area at 5 positions of sagittal sections, C) Number of TUNEL+ cells normalized to the area of spinal cord at the 5 different positions from rostral (R) to caudal (C) (R 4 mm, R 2 mm, epicenter, C 2 mm, C 4 mm) at the 5 different sagittal section position. Left bar: untreated SCI and Right bar: Rm-PgP treated group. *p<0.05 and **p<0.005, D) The number of total ED1+ cells within each rostro-caudal position were summed across the 5 sagittal sections examined. E) Representative images of TUNEL+ apoptotic cells in green (Top: untreated SCI, bottom: Rm-PgP). 50X original magnification.

Statistical analysis

All the data were evaluated using a two-tailed student t-test and presented as mean ± standard deviation (STD). The results showed significant when a p value < 0.05 on a 2-tailed test.

Results

Preparation of rolipram loaded PgP nanoparticles

Rm-PgP was freshly prepared for each individual experiment and the amount of Rm loading was measured by HPLC system. The amount of Rm loaded in PgP (1 mg PgP/ml water) was calculated based on the standard curve generated by dissolving Rm in ethanol. The Rm loading efficiency was approximately 98 ± 4 % (n= 3/individual experiment, total of 7 experiments).

Effect of Rm-PgP on cAMP levels

To evalaute the effect of Rm-PgP on cAMP level in lesion site, animals were sacrificed. at 1, 2, and 3 days post-injection and spinal cords (0.5 cm long-piece centered around the SCI epicenter) were harvested, homogenized, and cAMP level was evaluated by ELISA. We observed that the cAMP level in the untreated SCI group was significantly lower than that in sham group at 1, 2 and 3 DPI (P<0.05), while the cAMP level in Rm-PgP treated SCI group was significantly different only at 1 DPI. We also observed that the cAMP level in Rm-PgP treated SCI group was higher than that in the untreated SCI group at all time points and was significantly higher than untreated SCI group at 1 DPI (p< 0.01) (Figure 1).

Figure 1.

Effect of Rm-PgP on cAMP levels at the SCI lesion site. cAMP level in injured spinal cord (0.5 cm length) was measured by ELISA at 1, 2, and 3 DPI. cAMP level was normalized by cAMP level of the sham group at each time point. Data presented mean ± standard deviation (n=4). (^#p < 0.05 compared to Sham, ^##p < 0.01 compared to Sham, **p < 0.01 compared to untreated SCI).

Effect of Rm-PgP on motor function recovery and neuropathic pain

The effect of Rm-PgP treatment on motor functional recovery was evaluated by Basso, Beattie, Bresnahan (BBB) scoring system at 1, 3, 7, 14, 21, and 28 DPI. Untreated SCI animals and sham animals were used as controls (n=6 rats/group). The sham group retained normal locomotion throughout the 4 weeks study, displaying consistent plantar stepping and normal forelimb-hindlimb coordination. Both untreated SCI and Rm-PgP treated groups showed significant impairments in hindlimb function compared to sham that persisted through the duration of the experiment. Figure 2A shows that BBB scores of the Rm-PgP treated animals were higher at 7 DPI and significantly higher than those of untreated SCI animals at 2, 3 and 4 week time point. Functionally, untreated SCI rats recovered isolated hindlimb joint movements and some regained plantar placement with weight support. 5 rats (83.3 %) out of 6 rats treated with Rm-PgP regained the ability to take frequent plantar weight-supported steps (BBB score >9), compared to only 1 rat (16.7 %) out of 6 rats of untreated SCI rats could step at 28 DPI (Table 1). These results suggest that Rm-PgP treatment can significantly improve overground locomotor recovery.

Figure 2.

Effect of Rm-PgP on motor function and neuropathic pain after contusive SCI. A) Motor function by BBB scoring system, and B) Paw withdrawal threshold by von Frey study. Data presented mean ± standard deviation (n=6/group). *p<0.05 compared to untreated SCI.

Table I.

Number of stepping rats in Rm-PgP treated groups compared to untreated SCI group

Groups	Time (DPI)
	1	3	7	14	21	28
SCI (n=6)	0	0	0	0	1	1
Rm-PgP (n=6)	0	0	0	3	4	5

The effect of Rm-PgP treatment on neuropathic pain-like behavior was evaluated by von Frey test at 7, 14, 21, and 28 DPI. Untreated SCI animals and sham animals were used as controls (n=6 rats/group). The sham group maintained normal paw withdrawal thresholds pre and post-operatively, responding to a tactile stimulus of 60g force. Both untreated and Rm-PgP treated groups exhibited a significant decrease in paw withdrawal thresholds that persisted over 28 DPI. We observed that slight improvement paw withdrawal thresholds at 14 and 28 DPI in Rm-PgP treatment group even though it was not significantly different compared to untreated SCI group (Fig 2B).

Effect of Rm-PgP on lesion size

At 28 DPI, animals were sacrificed by cardiac perfusion to evaluate the effect of Rm-PgP treatment on secondary injury by histological analysis. A 1 cm length of spinal cord centered around the lesion epicenter was sectioned, and 5 sagittal sections from 5 different positions across the spinal cord at 0.2 mm intervals were selected for lesion analysis (Fig. 3A). Representative images of H & E-stained sections for untreated and Rm-PgP treated SCI groups were shown in Figure 3B. We observed that average spared tissue area in Rm-PgP treated SCI group is slightly higher than that in the untreated SCI group at all positions, even though it is not statistically significant (Fig 3C). We also observed that the average rostrocaudal lesion length (mm) in Rm-PgP treated group was 23% shorter than that in the saline SCI group even though it was not significantly different (Fig. 3D).

Effect of Rm-PgP on secondary injuries

Inflammatory response

The effects of Rm-PgP treatment on the inflammatory response was analyzed by immunostaining for the inflammatory cell marker, ED1 at 28 DPI (n=7 rats/group). Five sagittal sections from 5 different positions (left (-) 0.4 mm, left (-) 0.2 mm, midline, right (+) 0.2 mm, right (+) 0.4 mm) were selected and stained for activated microglia and infiltrated macrophages using ED1 antibody (Fig 4A). The number of ED1+ cells in the spinal cord at each position was counted and normalized to the total area of spinal cord. We observed significantly lower number of ED1 + cells, such as activated microglia and/or infiltrated macrophages, in the spinal cords from animals treated with Rm-PgP relative to the spinal cords from the untreated SCI animal group at all 4 positions except left (-) 0.4 mm position (Fig. 4B). We also compared ED1+ cells at the 5 different positions from rostral (R) to caudal (C) with 2 mm distance (R 4 mm, R 2 mm, epicenter, C 2 mm, C 4 mm) at each 5 different sagittal section positions (− 0.4 mm, − 0.2 mm, midline, + 0.2 mm, + 0.4 mm). We observed significantly lower number of ED1 + cells in the spinal cords from animals treated with Rm-PgP relative to the spinal cords from the untreated SCI animal group at all 4 positions except rostral 4 mm position (Fig. 4C). Figure 4D shows the total number of ED1+ cells normalized to spinal cord area at 5 different positions on 5 different sagittal sections and the total number of ED1+ cells was significantly lower in Rm-PgP treated group than in untreated SCI group. Figure 5E shows the representative images of ED1+ cells in midline spinal cord section in untreated SCI animal (top) and Rm-PgP treated animal (bottom).

Astrogliosis

The effects of Rm-PgP treatment on the astrogliosis was analyzed. Three sagittal sections from 3 different positions (left (-) 0.2 mm, midline, right (+) 0.2 mm) per each animal (n=6 rats/group) were selected and stained for reactive astrocytes using antibody targeting GFAP (Fig. 5A). The fluorescence intensity of GFAP stained sections was measured and normalized to intensity from sections from sham animals. We observed that the fluorescence intensity from sections in Rm-PgP treated SCI animal group was significantly lower compared to those in untreated SCI animal group at all 3 positions (Fig 5B). Figure 5C shows the representative images of GFAP stained sections from untreated SCI group (left) and Rm-PgP treated group (right).

Neuronal cell survival

The effect of Rm-PgP treatment on neuronal survival was evaluated by immunostaining for the neuronal nuclear marker, NeuN using sagittal sections from 3 positions (6 sections from each position, n=6 rats/group) shown in Figure 6A. We observed that the average number of NeuN+ cells at all 3 positions in the Rm-PgP treated group was higher than in the untreated SCI group even though it was not significantly different (Fig. 6B). However, total number of NeuN+ cells in the Rm-PgP treated group was significantly higher than in the untreated SCI group (Fig. 6C). Figure 6D shows the representative images of NeuN stained sections from untreated SCI group (left) and Rm-PgP treated group (right).

Apoptosis

The effect of cAMP restoration by Rm-PgP treatment on apoptosis was evaluated by TUNEL assay. The total number of TUNEL+ cells in the spinal cord at each position was counted and normalized to the total area of spinal cord. Five sagittal sections from 5 different positions (left (-) 0.4 mm, left (-) 0.2 mm, midline, right (+) 0.2 mm, right (+) 0.4 mm) were selected (Fig 7A). We observed that the number of TUNEL+ cells in Rm-PgP treated animals was significantly lower compared to the untreated animal group at both left and right 0.4 mm positions, while the number of TUNEL + cells was not significantly different in midline and both left and right 0.2 mm position, (Fig. 7B). Figure 7C shows the number of TUNEL+ cells at 5 different positions from rostral to caudal (2 mm distance, R 4 mm, R 2 mm, epicenter, C 2 mm, C 4 mm)) and the number of TUNEL+ cells in Rm-PgP treated animals was lower than that in the untreated animal group at all positions even though it was not significantly different. Figure 7D shows the total number of TUNEL+ cells normalized to spinal cord area at 5 different positions on 5 different sagittal sections and the total number of TUNEL+ cells was significantly lower in Rm-PgP treated group than that in untreated SCI group. Figure 5E shows the representative images of TUNEL+ cells (green) in untreated SCI animal (top) and Rm-PgP treated animal (bottom).

Discussion

The inability of injured axons to regrow in the adult mammalian central nervous system is due to both the presence of extrinsic inhibitory molecules and age-related changes in intrinsic neuronal growth capacity. One of the consequences of SCI is a rapid decline in cAMP levels,⁷ which is associated with the onset of growth inhibition by myelin, impaired regenerative responses, and increased glial reactivity.^8–10 Cyclic AMP activity can be enhanced in many way-either by increasing adenyl cyclase activity, administering a membrane-permeable cAMP analogue (dibutyl-cAMP), or by decreasing phosphodiesterase 4 (PDE4) activity. Restoration of cAMP after SCI, often induced by Rm, has been shown to be an effective strategy, either alone or in combination with other therapeutic interventions, to improve axonal growth and functional recovery in preclinical SCI models.¹⁴ Although not completely elucidated, cAMP signaling appears to act through independent activation of PKA and EPAC2. PKA signaling through CREB leads to transcriptional activation of multiple targets including arginase I that support axonal growth, while EPAC2 activates the growth-promoting b-Raf pathway.²⁵

Previously, we demonstrated that Rm’s aqueous solubility can be increased ~6.8 times when loaded within the hydrophobic core of the PgP micelle delivery system. Live animal imaging of PgP loaded with a hydrophobic fluorescent dye demonstrated its retention in and around the SCI lesion epicenter for at least 7 days, likely allowing sustained, local release. Consequently, locally injected Rm-PgP restored cAMP to sham animal level, reducing both the inflammatory response and apoptosis at 3 DPI (acute phase) in a severe compression SCI model.¹⁹ Here, we extend this work to investigate the therapeutic effect of Rm-PgP administered by intraspinal injection into the midthoracic spinal cord contusion lesion epicenter on motor functional recovery and development of neuropathic pain over 4 weeks post-injury (chronic injury phase), as well as secondary injury responses at the 4 week endpoint. We observed that the cAMP level in the Rm-PgP treated SCI group was higher than that in the untreated SCI group at all time points and was significantly higher than the untreated SCI group at 1 DPI (p< 0.01) (Figure 1). This is consistent with the trend we observed in our previous study.¹⁹ Moreover, acute Rm-PgP treatment improved locomotor function after SCI. We observed that 100% of SCI rats treated with intraspinal Rm-PgP (10 mg Rm) at the time of SCI regained the ability to take frequent plantar weight-supported steps by 28 DPI, compared to 30% of untreated SCI rats. Several studies reported that Rm showed regenerative effects^14,26 and functional improvements^7,17 after SCI. When comparing studies, it is important to note that Rm is sometimes studied as the sole therapeutic modality, but also frequently included as one component of combinatorial therapies. Costa et al., showed that Rm administered subcutaneously by osmotic mini-pump (3.18 mg/kg/day, dissolved in dimethylsulfoxide [DMSO]), for 2 weeks to rats with contusive SCI improved motor performance by BBB scoring system up to 8 weeks compared to control group received DMSO only.²⁵ Their behavioral results align with our results, despite differences in route of administration and total Rm delivered. Costa et al. administered Rm (795 μg/ rat (250 g)) for 2 weeks systemically, while we administered only one single Rm (10 μg/ rat (250 g) intraspinally at the injury epicenter. These results suggest that acute, local delivery of Rm by PgP can be an effective therapeutic intervention following SCI.

SCI-induced pain-like behavior as measured by the von Frey test was not resolved with Rm-PgP treatment, which slightly improved paw withdrawal thresholds at 14 and 28 DPI compared to untreated SCI group, although the changes were not significantly different. In contrast, prolonged, systemic Rm administration has been an effective treatment in peripheral nerve injury models of pain. Zhang et al. administered Rm dissolved in DMSO intrathecally over 14 days after partial sciatic nerve ligation (PSNL) injury in mice and observed reduced mechanical hypersensitivity of the paw.²⁷ They also reported that repeated intrathecal treatment with Rm reduced both PSNL-induced downregulation of cAMP and Connexin 43 (Cx43) and upregulation of proinflammatory cytokines tumor necrosis factor-α (TNF-α) and interleukin-1β. In our study, we administered only one single Rm dose (10 μg/ rat (250 g) intraspinally at the injury epicenter and we did not investigate the effect of repeated Rm-PgP treatment on neuropathic pain development. Histological analysis showed that average spared tissue area in Rm-PgP treated SCI group was slightly higher than that in the untreated SCI group at all positions, even though it was not statistically significant (Fig 3C) and average rostro-caudal lesion length in Rm-PgP treated group was 23% shorter than that in the saline SCI group even though it was not significantly different (Fig. 3D). Costa et al. similarly reported no significant differences in lesion volume and length between SCI groups receiving Rm dissolved in DMSO and DMSO only, although they did observe that spared white matter was significantly higher in the Rm treatment group.²⁸ One challenge in interpretation of our results is the possibility that some axons that were spared by the initial trauma may have been damaged by the intraspinal injection. We also did not perform white matter staining even though white matter sparing is an important parameter for the correlation with functional outcome and neuroprotection. In our future study, we will perform the white matter staining by eriochrome cyanine staining to evaluate the effect of Rm on white matter sparing specifically,^24,29,30 as well as investigate alternative administration routes such as intrathecal injection.

The effect of Rm-PgP on secondary injury including the inflammatory response, astrogliosis, neuronal survival, and apoptosis was further evaluated by histological analysis. Several studies reported that Rm can influence neuronal and oligodendrocyte cell function through intracellular cAMP signaling pathway.^14,15,17 We observed that a single intraspinal injection of Rm-PgP at the lesion site immediately after injury reduced the inflammatory response (Fig. 4), astrogliosis (Fig. 5) and apoptosis (Fig. 7) and increased neuronal survival (Fig. 6). These data are consistent with our previous study, where Rm-PgP reduced both the inflammatory response and apoptosis at 3 days after severe compression SCI.¹⁹

Since cAMP is ubiquitously expressed and the current PgP system is not cell type specific, it is likely that the robust and meaningful recovery of frequent hindlimb stepping with Rm-PgP treatment after SCI is mediated by the interactions between several different cell types. That microglial/macrophage activation was significantly reduced in the SCI site with Rm-PgP treatment, suggests that Rm was affecting these cells. Indeed, cAMP regulates essential actions in microglia and macrophages to resolve inflammation. These include their polarization to pro-resolving phenotypes expressing arginase-1, CD206, Ym-1 and Interleukin-10,^31–33 promoting biosynthesis of pro-resolving molecules (i.e., lipoxinA4 or annexinA1) and anti-inflammatory cytokines,³⁴ and reducing proinflammatory cytokine production.^35,36 Our current results, combined with these insights from the literature, suggest that acute, local injection of Rm-PgP may be driving meaningful functional recovery by altering microglial activation and astrogliosis and indirectly affecting neuronal populations in the injury site via these glial cells. Table II summarizes our key findings on the effect of Rm-PgP treatment compared to untreated SCI group.

Table II.

Summary of our key findings on the effect of Rm-PgP treatment compared to untreated SCI animal groups at 4 weeks post-injury

Results	Rm-PgP treatment
Locomotion	↑
Mechanical Sensitivity	↓
Neuronal Sparing	↑
Apoptosis	↓
Microglia/Macrophage Activation	↓
Astrocyte Reactivity	↓

In conclusion, we demonstrated that the acute local administration of Rm by PgP nanocarrier (Rm-PgP) can restore cAMP level at the lesion site, and restored cAMP can improve functional recovery after moderate, contusive SCI. We also observed that one single intraspinal injection of Rm-PgP can reduce secondary injury such as the inflammatory response by infiltrated macrophages/ activated microglia, astrogliosis by activated astrocytes and apoptosis as well as improve neuronal survival in the chronic phase of injury (4 WPI). These data are consistent with our previously published study that Rm-PgP can reduce secondary injury in the acute recovery phase (3 DPI) in a rat severe compression injury model. One limitation of these studies is that we evaluated therapeutic effects of immediate (< 30 min) local treatment with Rm-PgP following SCI, however, in clinical settings treatment is often more delayed. In future studies, we will evaluate the effects of treatment window by delayed treatment of Rm-PgP via intrathecal catheter on motor function, neuropathic pain, and secondary injury in a rat moderate thoracic contusion model.