Delayed expression of cell cycle proteins contributes to astroglial scar formation and chronic inflammation after rat spinal cord contusion

Contusive SCI was performed in adult male Sprague-Dawley rats weighing 275–325 g as previously described [24, 37]. Briefly, rats were deeply anesthetized with sodium pentobarbital (65 mg/kg i.p.), and a moderate spinal cord contusion injury was induced at vertebral level T8 by dropping a 10-g weight from a height of 25 mm onto an impounder positioned on the exposed dura. Sham animals underwent the same procedure as injured rats, but received a laminectomy only, without weight drop. After injury, rats were placed into a heated cage to maintain normal core temperature until fully alert, and their bladders were manually expressed twice a day until a reliable bladder emptying reflex was established (10–14 days after SCI). The experimental protocols were approved by the University of Maryland School of Medicine Animal Care and Use Committee and met all NIH guidelines.

Following SCI, rats were randomly and blindly assigned to either drug or vehicle treatment group. Rats received intra-peritoneal (IP) injection of CR8 (second-generation roscovitine analog, 1 mg/kg, Tocris Bioscience) or an equal volume of vehicle once daily beginning 3 h post-injury and continuing for 7 days. CR8 was dissolved in sterile saline. This dose of CR8 was based on the results obtained from pilot studies in vitro and in vivo. More specifically, anti-apoptotic concentrations of CR8 in cultured cortical neurons were similar to those of flavopiridol, a potent pan-CDK inhibitor. Rats were sacrificed for histological/immunohistochemical study at 4 months post-injury.

At 4 weeks and 4 months post-injury, rat spinal cord tissue (5 mm) centered on the injury site was collected and frozen on dry ice for Western analysis [38, 39] with n = 4–5 rats per time point plus four laminectomy controls. Briefly, tissue was lysed in radioimmunoprecipitation assay (RIPA) buffer (Sigma) supplemented with 100 mM phenylmethylsulfonyl fluoride, 1× protease inhibitor cocktail, phosphatase inhibitor cocktail II and III (Sigma), then homogenized and sonicated. After centrifugation at 20,600 × g for 20 min, protein concentrations in supernatant were determined by the Pierce BCA method (Thermo Scientific). Normalized protein samples were denatured in LDS loading buffer. Each sample was from a different subject and run in an individual lane on 4 to 12% NuPAGE Novex Bis-Tris gradient gels (Invitrogen), and then transferred to nitrocellulose membranes (Invitrogen). After blocking in 5% nonfat milk for 1 h at room temperature, membranes were probed with antibodies against CDK4 (polyclonal, 1:1,000, Santa Cruz Biotechnology), cyclin D1 (polyclonal, 1:500, Neomarker), cyclin E (monoclonal, 1:500, Santa Cruz Biotechnology), E2F5 (polyclonal, 1:500, Santa Cruz Biotechnology), PCNA (polyclonal, 1:500, Santa Cruz Biotechnology), ionized calcium-binding adaptor molecule 1 (Iba-1,polyclonal, 1:1,000, Wako Chemicals) and p22^phox (polyclonal, 1:500, Santa Cruz Biotechnology) overnight at 4 °C followed by horseradish peroxidase-conjugated secondary antibodies (GE Healthcare) for 1 h at room temperature. The immunocomplexes were then visualized using SuperSignal West Dura Extended Duration Substrate (Thermo Scientific) and quantified by band densitometry of scanned films using the Gel-Pro Analyzer program (Media Cybernetics, Inc.) in the linear detection range. GAPDH was used as control for gel loading and protein transfer. Each sample was repeatedly run three times using the same blot and a pooled average was taken. The error bars in the Western blot quantification reflect variance across subjects and repeated runs of the same blots.

At 4 weeks and 4 months after injury, rats were anesthetized and intracardially perfused with 200 ml of saline followed by 300 ml of 10% buffered formalin. The dissected spinal cords were post-fixed for 2 h and cryoprotected through a sucrose gradient. A 1.5-cm segment of spinal cord centered at the injury area was sectioned at 20-μm thickness and thaw-mounted onto Superfrost Plus slides (Fisher Scientific) by placing them serially on sequential sets of ten slides, each set representing a 200-μm length of spinal cord. A representative slide from each set was stained with eriochrome cyanine (ECRC) for myelinated white matter (blue). The lesion epicenter was identified as the section with the least amount of spared white matter [40].

Lesion volume was assessed using the Stereologer 2000 software (Systems Planning and Analysis, Alexandria, VA). Sections spaced 1 mm apart from 5 mm caudal to 5 mm rostral from the injury epicenter were stained with GFAP and DAB as the chromogen for lesion volume assessment based on the Cavalieri stereology method with a grid spacing of 200 μm. Lesion volume where GFAP is absent is expressed as a percentage of total volume including both areas of GFAP present and absent [24].

Immunohistochemistry was performed on spinal cord coronal sections at specified distances rostral and caudal to the injury epicenter. Standard fluorescent immunocytochemistry on serial, 20-um-thick sections was performed as described previously [38]. The following primary antibodies were used: rabbit anti-CDK4 (1:100, Santa Cruz Biotechnology), rabbit anti-cyclin D1 (1:50, Neomarker), mouse anti-cyclin E (1:100, Santa Cruz Biotechnology), rabbit anti-E2F5 (1:100, Santa Cruz Biotechnology), rabbit anti-PCNA (1:100, Santa Cruz Biotechnology), rabbit or mouse anti-GFAP (1:500, Chemicon), rabbit anti-Iba-1 (1:500, Wako Chemicals) and rabbit anti-p22^phox (1:100, Santa Cruz Biotechnology). Fluorescent-conjugated secondary antibodies (Alexa 488-conjugated goat anti-mouse or rabbit, 1:400, Molecular Probes) were incubated with tissue sections for 1 h at room temperature. Cell nuclei were labeled with bis-benzimide solution (Hoechst 33258 dye, 5 ug/ml in PBS, Sigma). Finally, slides were washed and mounted with an anti-fading medium (Invitrogen). Immunofluorescence was visualized by tile scan using a Leica TCS SP5 II Tunable Spectral Confocal microscope (Leica Microsystems Inc., Bannockburn, IL). The images were processed using Adobe Photoshop 7.0 software (Adobe Systems, San Jose, CA). All immunohistological staining experiments were carried out with appropriate positive control tissue as well as primary/secondary-only negative controls.

Rat hind limb locomotor recovery was assessed at 1 day post injury and weekly thereafter for up to 16 weeks using the Basso, Beattie and Bresnahan (BBB) open field expanded locomotor score [41]. In the BBB test, normal animals are given a score of 21, while animals with no hind-limb function are given a 0, with any combination of indicators of paralysis or regained function yielding scores in between. Rats were also scored on a battery of tests to determine recovery of hind limb motor and sensory function including: open field locomotion (motor score); withdrawal reflex to hind limb extension, pain and pressure; foot placing, toe spread and righting reflexes; maintenance of position on an inclined plane and swimming tests. Results of these tests are reported as a Combined Behavioral Score (CBS) [42]. Rats with normal function receive a score of 0, while rats with abnormal scores on all tests receive a score of 100. All rats were tested without knowledge of treatment group.

All data are plotted as mean ± SEM where “n” is the number of individual animals. Western blot, lesion volume and stereological analyses were performed by an investigator blinded to treatment group. The expressions of various proteins (% of sham) were analyzed by using Kruskal-Wallis one-way ANOVA based on ranks, followed by Dunnett’s or Tukey’s post-hoc test (Sigma Stat Program, Version 3.5, Systat Software). BBB and CBS scores were analyzed with two-way ANOVA and repeated measures. All other statistical tests were performed using the GraphPad Prism Program, version 3.02 for Windows (GraphPad Software). A p < 0.05 was considered statistically significant.

Quantitative analysis of Western blots showed that CDK4 expression was significantly increased at both 4 weeks (3.5 fold of sham) and 4 months (4.5 fold of sham) post SCI (p < 0.05, respectively Figure 1). The expressions of cyclin D1 and E were significantly increased at 4 weeks and 4 months post injury (Figure 1C and D). E2F5 expression levels were three fold that of sham at 4 weeks post injury and remained elevated (1.8 fold of sham) at 4 months. Levels of PCNA protein in spinal cord tissue were approximately two fold that of sham from 28 days through 4 months post-injury (Figure 1 A and B).

To determine the distribution and the cellular localization of increased cell cycle proteins in the injured rat spinal cord, we performed immunofluorescent double labeling of key cell cycle molecules and several cell-specific markers. In the intact spinal cord, immunoreactivity of CDK4, E2F5, cyclin D1 and E was weakly detected in neurons across the gray matter (Figure 2A, E, I and M). Four weeks after injury, expression was increased for each of these proteins in the spared tissues surrounding the central lesion at 1 mm rostral to the epicenter, especially in the border between spared tissue and the lesion (Figure 2 B, F, J and N). At 4 months post-injury, there was an increase in immunolabeling of cyclin D1 and E, E2F5 and CDK4 in contrast to sham tissue (Figure 3 Ab and Bd-f). Double-immunolabeling demonstrated that cyclin D1⁺/GFAP⁺ and cyclin E⁺/GFAP⁺ cells were readily apparent in the spared tissue (Figure 2H and L and Figure 3 Bb, e, h), whereas CDK4 and E2F5 were not only expressed by GFAP⁺ hypertrophic astrocytes in the lesion scar border, but also in the central lesion (Figure 2 C-D, G-H, K-L and O-P).

To examine whether increased expression of cell cycle-related proteins was associated with chronic inflammation, Western blotting and immunohistochemistry were performed for inflammatory markers Iba-1, p22^PHOX and CD11b (OX42) at 4 weeks and 4 months after SCI. Western blot analysis of Iba-1 protein expression indicated a 2.5-3.0-fold increase in injured spinal cord extracts compared to sham tissue (Figure 4 C). We also found a significant increase in p22^PHOX protein expression at 4 weeks (3.5 fold of sham) followed by a prolonged upregulation for up to 4 months post injury (Figure 4 A and B), consistent with our prior report [17]. Immunohistochemistry at 28 days and 4 months post-injury demonstrated an increase in immunolabeling of Iba-1, p22^PHOX and OX42 in contrast to sham tissue (Figure 5 C, G, K, O and Figure 6Ae and B c-d). Moreover, double-labeling immunohistochemistry revealed that large numbers of PCNA⁺, cyclin E⁺ and D1⁺ cells in the injured coronal sections were co-labeled with OX42, p22^PHOX or Iba-1 at 1.5 mm rostral to the epicenter (Figure 5 and Figure 6Ab, f, h). Overall, these data suggest that SCI-induced upregulation of cell cycle-related proteins occurs persistently for up to 4 months, and is associated with reactive astrocytes and activated microglia/macrophages.

To further investigate the role of cell cycle pathway in SCI-induced astrogliosis and inflammation, injured rats were given the selective CDK inhibitor CR8 or vehicle by ip injection 3 h post-injury and daily thereafter for 7 days. Our pilot data showed that CR8 exhibits a 50-fold higher potency than roscovitine in in vitro neuroprotection. In the present study, rats were treated with 1 mg/kg CR8 or the equivalent volume of saline, and spinal cord sections were collected at 4 months after SCI for immunohistochemical study. GFAP is an indicator of astrocyte reactivity associated with glial scar formation [5]. Immunohistochemical analysis revealed increased expression of cyclin D1 (Figure 3Ab) and GFAP (Figure 3Ae), and their co-localization (Figure 3Ah) in the spared tissue surrounding the lesion site. In contrast, CR8-treated rats showed less immunoreactivity for both cyclin D1 and GFAP (Figure 3Ac, f, i). In addition, SCI resulted in increased expression of cyclin E, CDK4 and E2F5 at the site of the injury (Figure 3Bd-f), which were attenuated by CR8 treatment (Figure 3Bg-i).

Chronic inflammation was also evaluated by immunohistochemical analysis of spinal cord sections using Iba-1, p22^PHOX and OX42. Double-labeling immunohistochemistry revealed increased expression of PCNA (Figure 6Ab) and OX42 (Figure 6Ae), and their co-localization (Figure 6Ah) in the injured tissue at 4 months post-injury. CR8 treatment reduced immunoreactivity for both PCNA and OX42 (Figure 6Ac, f, i). SCI increased expression of Iba-1 and p22^PHOX as compared to sham (Figure 6Bc-d); these changes were attenuated by CR8 treatment (Figure 6Be-f). Taken together, these data demonstrated that delayed systemic treatment with a novel, selective and potent CDK inhibitor reduced astrogliosis and microglial proliferation at 4 months post-SCI.

SCI-induced lesion volume/cavity formation was measured with GFAP/DAB staining at 1 and 4 months after SCI and analyzed by unbiased stereological techniques. Histological assessment showed that tissue collected 4 months post-SCI possessed a larger lesion cavity (52.4 ± 5.0% of total volume) than that at 1 month after injury (28.4 ± 3.9% of total volume), indicating progressive damage in the injured spinal cord (Figure 7). Notably, a significant reduction in lesion volume was found in CR8 treated rats at 4 months post-injury (40.8 ± 3.6% of total volume, Figure 7A and B). These reductions occurred in both white and gray matter, with an overall decrease in cavity formation and tissue loss.

To further address whether inhibition of cell cycle-related proteins improves neurological outcome, we next examined whether this same treatment strategy could influence long-term functional recovery after SCI. We assessed the behavior of CR8 and vehicle-treated rats over the 16 weeks after injury. Testing was performed 1 day after injury and weekly thereafter using the BBB test of hind limb locomotor function [41] and the Combined Behavioral Score (CBS), an evaluation of overall hind limb sensory-motor deficits [42]. Both BBB and CBS tests showed significantly improved functional recovery after delayed systemic CR8 treatment (Figure 8). The effect of treatment was statistically significant beginning at week 2 as measured by BBB score (Figure 8 A) and beginning at week 3 as measured by CBS score (Figure 8 B). Both measures showed significantly improved recovery chronically from 4–16 weeks after SCI.