Therapeutic efficacy of rolipram delivered by PgP nanocarrier on secondary injury and motor function in a rat TBI model

Abstract

Aim: To develop poly(lactide-co-glycolide)-graft-polyethylenimine (PgP) as a nanocarrier for the delivery of rolipram (Rm) and evaluate the therapeutic efficacy of Rm-loaded PgP (Rm-PgP) on secondary injury and motor function in a rat traumatic brain injury (TBI) model. Materials & methods: Rm-PgP was injected in the injured brain lesion immediately after TBI using a microinjection pump. Secondary injury pathologies such as inflammatory response, apoptosis and astrogliosis were assessed by histological analysis and functional recovery was assessed by assorted motor function tests. Results: Rm-PgP restored cyclic adenosine monophosphate level in the injured brain close to the sham level and Rm-PgP treatment reduced lesion volume, neuroinflammation and apoptosis and improved motor function at 7 days post-TBI. Conclusion: One single injection of Rm-PgP can be effective for acute mild TBI treatment.

Traumatic brain injury (TBI) remains a leading cause of death and disability in the United States with approximately 2.5 million new cases occurring annually [1]. The heterogeneous effects of TBI are debilitating and the extent of these effects depends on the location and severity of the injury [2]. The progressive secondary injury phase can last for months following the initial trauma and is driven by biochemical and cellular cascades that provide an exploitable therapeutic window [3–5]. Current research aims to develop therapeutics that can reduce neuronal and glial cell death, reduce neuroinflammation and stimulate plasticity and repair following CNS injury [6–9]. However, therapeutics with promising preclinical results continue to fall short of desired functional improvement upon translation to clinical application [10,11]. Therefore, there is a need to develop therapeutic strategies capable of eliciting neuroprotection, repair, regeneration and functional recovery following TBI. The secondary messenger cAMP plays important roles in neural growth cone development, the regenerative capacity of adult neurons and the expression levels of proinflammatory cytokines. After TBI, the cAMP level decreases due to degradation by activated phosphodiesterase 4 (PDE4) enzymes and decreased synthesis by adenylyl cyclase [12,13]. Therefore, restoration of the cAMP level is a promising target for the treatment of TBI. Rolipram (Rm), a PDE4 inhibitor, can restore cAMP level, suppress proinflammatory cytokine expression [14,15], support antiapoptotic pathways [16–18] and potentiate the effects of neurotrophic molecules [18] after TBI. One challenge for the clinical use of Rm is its low water solubility [19]. As a result, Rm has to be dissolved in organic solvents such as DMSO or ethanol and then diluted in water, however, DMSO or ethanol can exhibit neurotoxic effects [20–22]. Therefore, the development of an efficient delivery carrier for Rm is necessary to avoid the need of organic solvents in systemic administration. Several types of nanoparticles (NPs) such as lipid-based NPs (LNPs) and polymer-based NPs have been developed for loading and delivery of Rm [23–25]. Lamprecht et al. prepared Rm-loaded NPs using poly(ε-caprolactone) by two preparation methods: pressure homogenization–emulsification with a microfluidizer or a modified spontaneous emulsification solvent diffusion method. They reported that the pressure homogenization–emulsification method showed higher Rm loading efficiency and improved controlled release compared with the spontaneous emulsification solvent diffusion method [23]. Glueckert et al. used LNPs to deliver Rm to tyrosine kinase positive cells in the inner ear [24]. When comparing polymer NPs and LNPs, each approach has both strengths and limitations. LNPs have low cytotoxicity, however, they can be challenged by low drug encapsulation efficiency and require chemical modification to improve circulation time [25]. Polymer-based NPs are a more versatile class of nanocarriers synthesized from combinations of natural and synthetic polymers that provide control of particle stability, circulation and drug encapsulation and can also be easily modified to include targeting antibodies and ligands [25]. In our group, we have developed a cationic, amphiphilic block copolymer, poly(lactide-co-glycolide)-graft-polyethylenimine (PLGA-g-PEI: PgP), which spontaneously assembles to form micellar NPs with a hydrophobic PLGA core capable of loading hydrophobic drugs and cationic branched PEI shell that can electrostatically complex therapeutic nucleic acids [26]. The novelty of our PgP as a nanocarrier is that PgP is capable of combinatorial delivery of hydrophobic drugs and therapeutic nucleic acids for CNS injury [27–30]. We have demonstrated that Rm can be loaded in the core of PgP nanocarriers and that local delivery of Rm-loaded PgP (Rm-PgP) to the injured spinal cord reduced neuroinflammation and apoptosis in a rat compression spinal cord injury model [29]. We have also previously reported the ability of PgP nanocarriers to efficiently deliver nucleic acids such as plasmid DNA (pDNA) and siRNA both in vitro [27] and in the injured spinal cord in vivo [25] and in the injured brain in vivo [30]. Here, we investigated the therapeutic efficacy of Rm-PgP on secondary injury and motor function recovery in a rat mild controlled cortical impact (CCI) TBI model in vivo at 7 days postinjury (DPI). We first evaluated the effect of Rm-PgP treatment on cAMP level in injured brain at 1, 3 and 7 DPI. Then, we evaluated the effect of cAMP restoration by Rm-PgP treatment on secondary injury such as lesion volume, the inflammatory response, neuronal cell death and apoptosis at 7 DPI. Finally, we evaluated the effect of cAMP restoration by Rm-PgP treatment on motor function recovery by beam walk and beam balance test at 1, 3, 5 and 7 DPI.

Materials & methods

Rm loading to PgP

PgP was synthesized using PLGA (4 kDa, 50:50, Durect Corporation, CA, USA) containing carboxylic end groups and branched PEI (bPEI, 25 kDa, Sigma-Aldrich, MO, USA) as previously described [27]. Rm was loaded to PgP micelles by a solvent evaporation method as described in our previous publication [29]. Briefly, Rm was dissolved in ethanol to get a final concentration of 10 mg/ml. Rm (100 μl, 1 mg Rm) stock solution was then slowly added into 1 ml of PgP solution (1 mg PgP/ml water), and incubated for 4 h to load Rm in the PgP core while stirring. After incubation, ethanol was removed by evaporation overnight. Samples were syringe filtered (0.2 μm) to remove unloaded Rm and then the amount of Rm loaded in PgP was measured by HPLC (Waters Corp., MA, USA) using a C18 column (Symmetry C18 3.5 μm, 4.6 × 75 mm, Waters) with a mobile phase of water:acetonitrile (60:40) and flow rate of 0.6 ml/min. Rm was detected by UV absorbance at 280 nm. The percent Rm loading efficiency was calculated as:

$$\%Loadingefficiency=\frac{AmountofrolipramloadedinPgP}{Amountofrolipramadded}\times 100$$

CCI TBI & Rm-PgP administration

All surgical procedures and postoperative care were conducted according to NIH guidelines for the care and use of laboratory animals under the supervision of the Clemson University Animal Research Committee (Approved animal protocol AUP #2015-082). Before surgery, male Sprague–Dawley rats (∼350 g, Charles Rivers) were anesthetized with 2–4% isoflurane gas. The head was shaved and secured in a stereotaxic frame (Kopf Instruments, CA, USA). A midline incision was made in the scalp, and the skin retracted using a microretractor to expose the skull. A circular craniotomy (5-mm diameter) was made over the right hemisphere centered at 3.5-mm lateral and 3.5-mm posterior of the bregma without disturbing the underlying dura. All injuries were performed with a TBI Impactor (PSI, VA, USA) using a 3-mm diameter flat impacting tip (impact speed = 3.5 m/s, depth = 2 mm, dwell time = 250 ms).

Rats were randomly divided into three groups: saline-injected TBI group (untreated TBI), Rm-PgP-treated group (TBI + Rm-PgP, 16 μg Rm/rat) and an uninjured sham surgery group (Sham). Immediately after injury a total of 20 μl Rm-PgP solution (16 μg Rm, 0.8 μg Rm/ml PgP) was administered by intraparenchymal injection at four predetermined positions (5 μl per injection: [lateral × caudal] 3 mm × 3 mm; 3 mm × 4 mm; 4 mm × 3 mm; and 4 mm × 4 mm) surrounding the lesion epicenter using a Hamilton syringe (30 gauge, beveled tip) at 2-mm depth (Figure 1). Injections were performed at a rate of 1 μl/min using a Legato 130 microinjector (KD Scientific, MA, USA) attached to the Kopf stereotaxic frame. Untreated TBI animals received 20-μl saline solution injected in the same manner. After injection, the skull cap was replaced, but not secured with bone cement to reduce intracranial pressure. The scalp was closed with 4–0 vicryl suture and animals were warmed with a heating pad for recovery.

Figure 1. . Location and parameters of injury and injection of rolipram-loaded poly(lactide-co-glycolide)-graft-polyethylenimine.

(A) Schematic drawing depicting the craniotomy location (red circle). (B) Images showing the location of injection (green dots) in the lesion site. (C) The injury parameters and injection locations.

TBI: Traumatic brain injury.

cAMP measurement in brain tissue

Rats were euthanized by CO₂ overdose at 1 DPI (n = 3 rats/group), 3 DPI (n = 3 rats/group) and 7 DPI (n = 3 rats/group). Brains were retrieved and the ipsilateral cortical tissue surrounding the injury site collected and snap-frozen with liquid nitrogen. The tissue was homogenized in 1:5 (w/v) volumes of ice-cold 0.1 N HCI with a motorized pestle homogenizer, centrifuged at 10,000× g for 10 min and the supernatants collected and neutralized with 1 N NaOH. Samples were diluted with Calibrator Diluent RD5-55 and assayed for cAMP using a Mouse/Rat cAMP Parameter Assay (R&D Systems). Streptavidin-coated microplates were incubated with biotinylated mouse monoclonal antibodies for 1 h at room temperature, thoroughly washed and then mouse/rat cAMP conjugate (cAMP conjugated to horseradish peroxidase) were added. Standard, controls or samples were added to the appropriate wells and incubated for 2 h at room temperature under constant shaking on an orbital shaker. After washing, a chromogenic substrate solution was added, incubated for 30 min and the reaction halted by acidic stop solution. The optical density was measured at 450 nm using a BioTek Synergy microplate reader (Synergy HT, BioTek). To correct for background, the optical density at 570 nm was subtracted from the optical density at 450 nm.

Tissue preparation for immunohistochemistry

Rats (n = 5 rats/group) were euthanized by cardiac perfusion with 0.9% saline and 4% paraformaldehyde (PFA) under deep anesthesia with isoflurane gas at 7 DPI. The brain was retrieved and postfixed in 4% PFA at 4°C followed by sucrose solution gradient (10, 20 and 30% sucrose). Brains were rapidly frozen using Ice-IT (Thermo Fisher Scientific, MA, USA) spray and coronal sections were cut (30 μm) on a cryostat and stored in cryopreservation solution (30% sucrose, 1% polyvinylpyrrolidone, 30% ethylene glycol, 0.05 M phosphate-buffered saline [PBS]) at -20°C.

Nissl staining & lesion volume measurement

To evaluate the effect of Rm-PgP treatment on lesion cavity volume, evenly spaced (0.25 mm) coronal sections were selected from 2 to 4.5 mm posterior of the bregma (n = 10 sections/rat; n = 5 rats/group). Sections were stained for Nissl bodies with 0.1% cresyl violet acetate (Polyscientific R&D Corp, NY, USA) for 20 min, rinsed with tap water, dehydrated by ethanol gradient and cleared in xylene. The slides were coverslipped with a resinous mounting medium (Permount, Azer Scientific, PA, USA), and imaged using an inverted bright-field microscope (Leica Microsystems, IL, USA). Cavity areas were measured for each section using ImageJ software and the lesion volumes were calculated by using Cavalieri’s approximation $V=d\left({\sum }_{i=1}^{n}yi\right)-\left(t\right)y_{max}$ where d = distance between sections, y_i = cross-sectional area in the i-th section, n = total number of sections, t i = section thickness and y_max = maximum possible value of y (cross-sectional area of the impacting tip at depth of 2 mm) [31].

Histological analysis for secondary injury

The effects of Rm-PgP treatment on inflammatory cell infiltration, glial scar formation, apoptosis and neuron cell survival were evaluated at 7 DPI (n = 5 rats/group). Briefly, coronal sections were randomly selected between 3 and 4 mm posterior of bregma (n = 3 sections/rat) prior to staining. For immunohistochemistry (IHC), coronal sections were postfixed with 4% PFA, washed with PBS and incubated using blocking solution (5% bovine growth serum and 0.05% Triton X-100 in PBS) for 1 h at room temperature. Sections were rinsed with PBS and incubated overnight at 4°C with either mouse monoclonal anti-macrophage/monocyte CD68 (clone ED1, 1:200; Cat. no. MAB1435, Millipore, MA, USA) for activated microglia and macrophages, rabbit monoclonal anti-CD163 for M2-like microglia/macrophages (1:200, Cat. no. ab182422, Abcam, MA, USA), rabbit polyclonal anti-GFAP (1:200; Cat. no. ab7260, Abcam) for activated astrocytes or mouse monoclonal anti-NeuN antibodies (1:200; Cat. no. MAB377, Millipore) for neuronal nuclei. Sections were washed with PBS and then incubated with goat antimouse Cy3-conjugated secondary antibodies (1:200; Cat. no. 115-165-003, Jackson ImmunoResearch, PA, USA) or AlexaFluor^® 488-conjugated goat antirabbit secondary antibodies (1:500, Cat. no. A-11008, Thermo Fisher Scientific, NH, USA) for 1 h at room temperature. Sections were washed with PBS and then coverslipped with VectaShield mounting media containing DAPI (Vector Laboratories, CA, USA). Images of the stained tissue were captured using an Axiovert 40 CFL microscope (Carl Zeiss, Oberkochen, Germany).

The effect of Rm-PgP on apoptosis was evaluated by terminal deoxynucleotidyl transferase-mediated dUTP nicked end labeling (TUNEL) assay in coronal brain sections using an ApopTag^® Plus Fluorescein in situ Apoptosis Detection Kit (S7111, EMD Millipore, CA, USA).

The number of ED1⁺, CD163⁺ and TUNEL⁺ cells were counted in the entire field of each 20 × (610 × 458 μm) image using ImageJ (n = 3 sections/rat, n = 5 rat/group). The number of DAPI + cell nuclei was used to calculate the percentage of ED1⁺ and TUNEL⁺ cells to total cells for each evaluated image. The NeuN⁺ cells were counted in a 610-μm width region of each 10 × image using ImageJ (n = 3 sections/rat, n = 5 rat/group). For activated astrocytes, the fluorescence intensity of GFAP-stained sections was measured and then normalized to intensity of sections in sham group (n = 3 sections/rat, n = 5 rat/group).

Effect of Rm-PgP treatment on motor function recovery

The effect of Rm-PgP treatment on motor function recovery was assessed by beam walk and beam balance tests at four postinjury time points (n = 5 rats/group; 1, 3, 5 and 7 DPI). Animals were trained for the baseline starting 3 days before injury. All tests were performed at the same time each day during the active period for the rats. A preinjury baseline was recorded 1 day before injury. Beams were cleaned in between rats using 70% ethanol and rats received at least 10 s rest between trials.

The beam walk test was performed using a beam (2.5 cm wide and 100 cm long) placed 60 cm above the floor. A bright light source was placed opposite a goal platform to act as an adverse stimulus and turned off when the rat reached the goal. All rats received three trials per day and the rats were placed on the goal platform for 10 min before the first trial. The time required to traverse the entire beam uninterrupted was recorded in addition to the number of foot slips. A foot slip was defined as passing of the foot below the midline of the beam. Rats were given 10 s rest in between trials. If the rat fell or could not cross the beam in 60 s it was placed at the goal platform and the time was recorded as 60 s.

For the beam balance test, a rod (1.5 cm diameter, 60 cm long) was placed at a height of 60 cm from the floor. For the baseline test and all proceeding trials (1, 3, 5 and 7 DPI), the rats received three trials per day and the time that they could remain on the beam without falling (latency) was recorded with a maximum of 60 s. If a rat remained for 60 s, then it was removed from the beam and placed back in its home box.

Statistical analysis

All data are presented as mean ± standard deviation (STD). The statistical analysis of data from lesion volume, inflammatory response and apoptosis was conducted by Student's t-test. The statistical analysis of neurite length, cAMP level (in vitro and in vivo), astrogliosis (GFAP) and motor recovery was conducted by one-way analysis of variance (ANOVA) with the multicomparisons between groups being evaluated by Tukey’s post hoc method. All statistical analyses were performed using GraphPad Prism 9.0. Significance was defined as a p-value of less than 0.05.

Results

Rm loading in PgP

The loading efficiency of Rm in the hydrophobic core of PgP nanocarriers was determined by HPLC. The loading efficiency was approximately 80% (0.8 mg Rm/ml PgP) and represented a fourfold increase above water solubility (∼0.2 mg/ml).

Effect of Rm-PgP on cAMP level after TBI

The effect of Rm-PgP treatment on cAMP level in vivo was evaluated at 1, 3 and 7 DPI. Following injury, the cAMP level in the untreated TBI group was significantly lower than that in the sham animal group at 1, 3 and 7 DPI. We observed that the cAMP level in the Rm-PgP treatment group was restored to levels not significantly different from the sham animal group at 1, 3 and 7 DPI (Figure 2).

Figure 2. . Restoration of cyclic adenosine monophosphate level in the ipsilateral cortex by rolipram-loaded poly(lactide-co-glycolide)-graft-polyethylenimine after traumatic brain injury.

Fresh cortical tissue was harvested at 1, 3 and 7 days post-injury, homogenized and cAMP concentration was evaluated by ELISA assay (n = 3 rats/group for each time point).

***p < 0.001 compared with sham.

^#p < 0.05 compared with TBI.

Rm-PgP: Rolipram-loaded poly(lactide-co-glycolide)-graft-polyethylenimine; TBI: Traumatic brain injury.

Effect of Rm-PgP on lesion volume after TBI

The effect of Rm-PgP on lesion cavity formation at 7 DPI was determined by calculating the lesion volume for each brain using evenly spaced area measurements from 2 to 4.5 mm posterior of the bregma. Figure 3A shows representative images of Nissl-stained brain sections. We observed that Rm-PgP treatment significantly reduced cavity volume compared with the untreated TBI group at all positions except 4.25 mm posterior to the bregma (Figure 3B). The average lesion volume of the Rm-PgP-treated group was significantly lower than the lesion volume of the untreated TBI group (Figure 3C).

Figure 3. . Lesion area and volume following cyclic adenosine monophosphate restoration by rolipram-loaded poly(lactide-co-glycolide)-graft-polyethylenimine at 7 days postinjury.

(A) Representative images of immunohistochemical staining for Nissl bodies in sections selected from 2 to 4.5 mm posterior of the bregma. Original magnification: 25 ×. Black scale bars = 2 mm. (B) Lesion areas were measured from ten evenly spaced (0.25 μm spacing) Nissl-stained coronal brain sections using ImageJ (n = 5 rats/group). (C) Lesion volumes were calculated by Cavalieri approximation using all lesion area measurements and the spacing between sections (0.25 μm; n = 5 rats/group).

^#p < 0.05 compared with TBI.

Rm-PgP: Rolipram-loaded poly(lactide-co-glycolide)-graft-polyethylenimine; TBI: Traumatic brain injury.

Effect of Rm-PgP on inflammatory responses after TBI

The effect of Rm-PgP on the inflammatory response following TBI was evaluated by immunostaining for ED1⁺ (M1 phenotype) and CD163⁺ (M2 phenotype) activated microglia/macrophage cells. Figure 4A shows representative images of ED1⁺ and CD163⁺ cells in the ipsilateral cortex, corpus callosum and hippocampus (original magnification: 10×). The percent of ED1⁺ cells in the Rm-PgP treatment group was significantly lower in the cortex and the corpus callosum compared with the untreated TBI group. The percent of ED1⁺ cells in the Rm-PgP treatment group was not significantly different in the hippocampus compared with untreated TBI group (Figure 4B).

Figure 4. . Inflammatory response following cyclic adenosine monophosphate restoration by rolipram-loaded poly(lactide-co-glycolide)-graft-polyethylenimine at 7 days post-injury.

(A) Representative images of immunofluorescent staining for activated microglia (ED1, red) and M2-like microglia (CD163, green) in the (1) ipsilateral cortex, (2) corpus callosum and (3) hippocampus in the perilesional area denoted by red boxes in Nissl-stained images. (B) The average percentage of ED1⁺ activated microglia/macrophages to total DAPI + cell nuclei in each image area (n = 3 sections/rat, n = 5 rats/group). (C) The average percentage of CD163⁺ M2-like cells normalized to the total ED1⁺ cells in each image area (n = 5 rats/group, n = 3 sections/rat). Black scale bar = 1 mm. White scale bars = 100 μm.

^#p < 0.05 and ^##p < 0.01 compared with TBI.

Rm-PgP: Rolipram-loaded poly(lactide-co-glycolide)-graft-polyethylenimine; TBI: Traumatic brain injury.

The effect of Rm-PgP on the percentage of CD163⁺/ED1⁺ cells was also assessed. The percentage of CD163⁺/ED1⁺ cells in the cortex was significantly higher in the Rm-PgP-treated group compared with the untreated TBI group, while the percentage of CD163⁺/ED1⁺ cells in the corpus collosum or hippocampus was not significantly different in the Rm-PgP-treated group compared with the untreated TBI group (Figure 4C).

Effect of Rm-PgP on astrogliosis after TBI

The effect of Rm-PgP on astrocyte activation and glial scar formation was evaluated by immunostaining for GFAP. We observed that the fluorescence intensity of GFAP staining in the ipsilateral cortex of the Rm-PgP-treated group was not significantly different with that of sham group, while the fluorescence intensity of GFAP staining was significantly higher in the ipsilateral cortex of the untreated TBI group compared with that of sham group (Figure 5).

Figure 5. . Reduced astrogliosis in the ipsilateral cortex following cAMP restoration by rolipram-loaded poly(lactide-co-glycolide)-graft-polyethylenimine at 7 days post-injury.

(A) Representative images of GFAP + reactive astrocytes (green) in the (1) ipsilateral cortex, (2) corpus callosum and (3) hippocampus in the perilesional area denoted by red boxes in Nissl-stained images. Black scale bar = 1 mm. White scale bars = 100 μm. (B) Average fluorescence intensity of GFAP-stained sections normalized to that of sham group at 7 DPI (n = 3 sections/rat, n = 5 rats/group).

*p < 0.05 compared with sham.

Rm-PgP: Rolipram-loaded poly(lactide-co-glycolide)-graft-polyethylenimine; TBI: Traumatic brain injury.

Effect of Rm-PgP on apoptosis after TBI

The effect of Rm-PgP treatment on apoptotic response was evaluated by TUNEL assay. Figure 6A shows representative images of TUNEL positive (TUNEL⁺) staining in the ipsilateral cortex, corpus callosum and hippocampus. We observed that the percentage of TUNEL⁺ cells was significantly lower in the ipsilateral cortex of the Rm-PgP-treated group compared with the untreated TBI group. However, the percentage of TUNEL⁺ cells in the corpus callosum and hippocampus of the Rm-PgP-treated group were not significantly different compared with the untreated TBI group (Figure 6B).

Figure 6. . Reduced apoptosis following cAMP restoration by rolipram-loaded poly(lactide-co-glycolide)-graft-polyethylenimine at 7 days post-injury by TUNEL assay.

(A) Representative images of TUNEL⁺ cells (green) captured in the (1) ipsilateral cortex, (2) corpus callosum and (3) hippocampus in the perilesional areas denoted by red boxes in Nissl-stained images. Black scale bar = 1 mm. White scale bars = 100 μm. (B) The average percentage of TUNEL⁺ cell nuclei was determined in each image (n = 5 rats/group, n = 3 sections/rat).
^#p < 0.05 compared with TBI.

Rm-PgP: Rolipram-loaded poly(lactide-co-glycolide)-graft-polyethylenimine; TBI: Traumatic brain injury.

Effect of Rm-PgP on neuronal survival after TBI

Neuron cell survival was evaluated by immunostaining for the neuronal nuclear marker, NeuN. Figure 7A shows the representative images of NeuN⁺ cells in the ipsilateral cortex. We observed that the average number of NeuN⁺ cells in the Rm-PgP-treated group was higher than in the untreated TBI group, but it was not significantly different (Figure 7B).

Figure 7. . Perilesional neuronal cell survival following cAMP restoration by rolipram-loaded poly(lactide-co-glycolide)-graft-polyethylenimine at 7 days post-injury.

(A) Representative images of NeuN⁺ cells (red) in the ipsilateral cortex were obtained in the perilesional areas denoted by red boxes in Nissl-stained images. (B) The average number of NeuN⁺ cells in the ipsilateral in each image area (n = 5 rats/group, n = 3 sections/rat). Black scale bar = 1 mm. White scale bars = 100 μm.

*p < 0.05 and **p < 0.01 compared with sham.

Rm-PgP: Rolipram-loaded poly(lactide-co-glycolide)-graft-polyethylenimine; TBI: Traumatic brain injury.

Effect of Rm-PgP on motor function recovery after TBI

The effect of Rm-PgP treatment on motor function recovery was evaluated by the beam walk and beam balance test. We observed significant motor function deficits in the untreated TBI group compared with the sham group (Figure 8). For the beam walk test, the beam crossing time in the Rm-PgP-treated group was not significantly different from the sham group at all time points. We also observed that the beam crossing time in the Rm-PgP-treated group was significantly shorter than the untreated TBI group at 1 DPI, while the beam crossing time in the untreated TBI group was significantly longer than in the sham group at 1 DPI (Figure 8A). The number of foot slips in the Rm-PgP-treated group was not significantly different than the sham group, while the number of foot slips in the untreated TBI group was significantly increased relative to the sham group at 1 DPI (Figure 8B). Additionally, the number of foot slips in the Rm-PgP-treated group was significantly lower relative to the untreated TBI group at 1 DPI. For the beam balance test, the latency on the beam in the Rm-PgP-treated group was not significantly different from the sham group at 3, 5 and 7 DPI and the latency on the beam in the Rm-PgP treatment group was significantly increased relative to the untreated TBI at 7 DPI (Figure 8C). The untreated TBI group showed significantly shorter latency on the beam compared with the sham group at all time points (Figure 8C).

Figure 8. . Motor function recovery following cyclic adenosine monophosphate restoration by rolipram-loaded poly(lactide-co-glycolide)-graft-polyethylenimine after traumatic brain injury.

The effect of Rm-PgP on motor performance was evaluated by the beam walk: (A) crossing time and (B) number of foot slips and by beam balance: (C) latency to fall from the beam (n = 5 rats/group).

*p < 0.05, **p < 0.01 and ***p < 0.001 compared with sham.

^#p < 0.05 and ^##p < 0.01 compared with TBI.

Rm-PgP: Rolipram-loaded poly(lactide-co-glycolide)-graft-polyethylenimine; TBI: Traumatic brain injury.

Discussion

Rm, a PDE4 inhibitor, can be used to effectively reduce neuronal damage for the treatment of traumatic CNS injury, but one challenge to the clinical use of Rm is its limited water solubility [19]. In our lab, we have developed a polymeric nanocarrier, PgP and reported that PgP can increase Rm water solubility approximately 6.8-times and the size of Rm-PgP was 121.8 ± 3.5 nm with polydispersity 0.28 ± 0.32 and Rm-PgP was stable at 37°C for over 5 weeks [29]. In our previous study, we evaluated the cytotoxicity of PgP/pDNA (N/P ratio of 30/1) and bPEI/pDNA (N/P ratio of 5/1) polyplexes as a function of polymer concentration in vitro. In both glioma (C6) cells and neuroblastoma (B35) cells, PgP/pDNA showed higher viability than bPEI/pDNA at all polymer concentrations [27]. In our other study, we evaluated the cytotoxicity of PgP on cerebellar granular neurons (CGNs) cultured in hypoxia condition and we found that PgP did not result in any cytotoxicity, but increased cAMP levels as well as neurite length of CGNs cultured in hypoxia condition, compared with the untreated CGNs in hypoxia group [29]. We also reported that a single injection of Rm-PgP in the rat spinal cord lesion site restored cAMP levels and reduced apoptosis and the inflammatory response in a rat SCI compression injury model [29]. In this study, we investigated the therapeutic effects of intraparenchymal injection of Rm-PgP on functional recovery and aspects of secondary injury such as the inflammatory response, neuronal survival and gliosis in a rat acute mild TBI model. A reduction in cAMP level following TBI is implicated in various aspects of CNS secondary injury pathology such as neuroinflammation and apoptosis [12,13]. A single injection of Rm-PgP (16 μg Rm/rat) in the lesion site restored the cAMP to a level that was not significantly different from the sham group up to 7 DPI (Figure 2). In addition, the cAMP level in the Rm-PgP group was significantly higher than the untreated TBI group. This was accompanied by reduction in lesion cavity volume in the Rm-PgP-treated group relative to the untreated TBI group at 7 DPI (Figure 3). This result is similar to our previous observations in a spinal cord injury model [29]. In contrast, Atkins et al. observed that treatment with Rm (1 mg/kg) by intraperitoneal injection at 30 min postinjury and then repeated once per day for 3 days in a mouse CCI TBI model led to a significant increase in cortical contusion volume [32]. In a separate study using a fluid percussion brain injury in rats, Atkins et al. observed that a single intraperitoneal injection of Rm (1 mg/kg) at 30 min postinjury followed by continuous administration (3 mg/kg per day) for 10 days by osmotic pump led to a significant increase in contusion volume in the ipsilateral cortex [33]. They speculated that this result may be due to differences in model-specific pathological and hemodynamic changes. Additionally, Atkins et al. suggested that PDE4 inhibition in endothelial cells could lead to vasodilatory effects and thereby increase hemorrhagic contusion volume [32,33]. In our study with a rat CCI model, we used a significantly lower Rm dose than that in the Atkins studies. Additionally, we administered a single intraparenchymal injection rather than using repeated large-dose intraperitoneal injections. These differences in Rm administration route and dose may explain the beneficial outcomes we observed compared with the detrimental effects observed by Atkins et al.

There is an established correlation between the cAMP signaling and the severity of inflammation [34,35]. Treatment with Rm-PgP significantly reduced infiltration of the ipsilateral cortex and corpus callosum by activated, ED1⁺ microglia/macrophages at 7 DPI compared with the untreated TBI group (Figure 4). This is consistent with previous reports that restoration of cAMP can reduce neuroinflammation following CNS injury [12,15,36]. We also probed for the presence of proregenerative (M2-like) microglia and macrophages using the marker CD163, which are associated with tissue regeneration and specifically the phagocytotic clearance of cell-free hemoglobin [37–39]. Treatment with Rm-PgP treatment led to a significant increase in the percentage of CD163 expressing ED1⁺ microglia/macrophage cells in the ipsilateral cortex. Together, these changes in inflammatory cell phenotype suggest a beneficial shift in the neuroinflammatory environment of the lesion.

The activation and proliferation of reactive astrocytes is key to the formation of an astrocytic scar. The astrocytic scar surrounds the lesion environment to protect adjacent neurons from further damage, however, it also creates a barrier to neuronal plasticity [40–42]. By staining tissue sections for GFAP, a marker for reactive astrocytes, we observed that Rm-PgP treatment significantly reduced GFAP expression compared with the untreated TBI group indicating a significant reduction in astrocytic scar formation (Figure 5). The relationship between the severity of neuroinflammation and astrocyte activation is well documented [40–42] and several studies have also demonstrated that cAMP can modulate astrocyte reactivity [43–45].

The progressive secondary cell death following TBI is primarily due to apoptosis and excessive apoptotic activity is associated with larger lesion volumes after CNS injury [6,24,25,46,47]. Active signaling through the cAMP pathway can promote cell survival and reduce apoptosis through activation of cAMP-responsive binding element and Akt [48–50]. Our analysis showed significantly reduced apoptotic activity in the ipsilateral cortex of the Rm-PgP treatment group compared with the untreated TBI group (Figure 6). A general promotion of cell survival and reduction in cellular apoptosis is important; however, promotion of neuroprotection is a more specific determination of therapeutic success following TBI [51–55]. Restoration of cAMP following CNS injury has been shown to promote neuroprotection [14,16,35,53,55]. Our analysis showed that TBI caused a significant decrease in neuron cell survival in the ipsilateral cortex and although the Rm-PgP treated group showed an increase in neuron survival at the border of the lesion compared with the untreated TBI group, this increase was not significant. It is important to note that we did show significantly reduced cortical lesion volume, which indicates tissue sparring and an increase in the overall number of surviving neurons due to preservation of cortical tissue. The significant tissue sparring and reduced neuroinflammation observed in the histological analysis also translated to improvement in motor function. Treatment with Rm-PgP significantly improved performance compared with the untreated TBI group at 1 DPI. Additionally, the beam balance latency of Rm-PgP-treated animals was not significantly different from the sham group at 3, 5 and 7 DPI, whereas the latency of animals in the untreated TBI group was significantly different from the sham group at all days tested (Figure 8). Though the degree of improvement differed between the two motor tests, the results indicate that Rm-PgP administration improved motor function, likely through reduced secondary injury resulting from cAMP restoration. One of the limitations of this study is the small sample size (n = 5 rats per group) for motor function study. In the future, we will investigate the effect of Rm-PgP on cognitive functional recovery using large-sample size by the Morris water maze test at 2 weeks postinjury in an acute moderate TBI model. We will repeat the effect of Rm-PgP on motor function study with large-sample size at 1, 3, 5 and 7 DPI before performing the Morris water maze test.

Conclusion

In this study, we demonstrate that the PgP nanocarrier is a promising platform for developing a successful therapy for TBI. The efficient loading of Rm in PgP nanocarriers facilitated low volume intraparenchymal injection of a therapeutically relevant Rm dosage without the need for organic solvents. Administration of Rm-PgP was able to restore cAMP level and motor function and significantly reduced lesion cavity volume, neuroinflammatory response and apoptotic response at 7 DPI in a rat CCI TBI model in vivo. The application of PgP nanocarriers provides the potential for combinatorial delivery of hydrophobic drugs such as Rm with therapeutic nucleic acids. This improves the potential for developing an efficacious treatment for TBI.

Future perspective

The complex pathology of TBI continues to pose challenges for successful clinical applications of single therapeutics. Combinatorial therapeutic strategies can address multiple pathways involved in TBI pathology and have the potential for improving therapeutic efficacy and successful clinical translation. In this study, we demonstrated that restoration of cAMP by Rm-PgP can reduce secondary injury and improve functional recovery after TBI. In our previous work, we have reported that RhoA knockdown by PgP/RhoA siRNA showed reduced neuroinflammation and apoptosis, and enhanced neuroprotection after TBI. Future work will aim to evaluate combinatorial delivery of Rm and RhoA siRNA using PgP as a nanocarrier. We will determine the therapeutic effects of Rm-PgP/siRhoA NP on progression of neuroinflammation, astrocytic scarring, neuroprotection and functional recovery after intrathecal injection in a chronic rat TBI model.

Summary points.

• The secondary messenger cAMP plays important roles in neural growth cone development, regenerative capacity of adult neurons and expression levels of proinflammatory cytokines.

• cAMP level decreases after traumatic brain injury (TBI) due to degradation by activated PDE4 and restoration of the cAMP level can be a promising target for the treatment of TBI.

• The poly(lactide-co-glycolide)-graft-polyethylenimine (PgP) nanocarrier can increase rolipram (Rm) water solubility by loading Rm within hydrophobic core.

• A single intraparenchymal injection of Rm-PgP restored the cAMP level to sham level after mild TBI in rats for up to 7 days postinjury.

• Rm-PgP treatment was able to reduce lesion volume, neuroinflammation and apoptosis at 7 days postinjury as assessed by immunohistochemistry.

• Administration of Rm-PgP improved recovery of motor function up to 7 days postinjury as assessed by the beam walk and beam balance, likely through mitigation of secondary injury resulting from cAMP restoration.