Experimental Neurology 269 (2015) 213–223

Contents lists available at ScienceDirect

Experimental Neurology journal homepage: www.elsevier.com/locate/yexnr

Regular Article

Chondroitin sulfate proteoglycans impede myelination by oligodendrocytes after perinatal white matter injury Ying-Ping Deng a,b, Yi Sun a, Lan Hu a, Zhi-Hua Li a, Quan-Mei Xu b, Yi-Ling Pei c, Zhi-Heng Huang a,b, Zhen-Gang Yang d, Chao Chen a,b,⁎ a

Department of Neonatology, Children's Hospital of Fudan University, Shanghai, China Key Laboratory of Neonatal Disease, Ministry of Health, Shanghai, China School of Public Health, Fudan University, Shanghai, China d Institute of Brain Science, Fudan University, Shanghai, China b c

a r t i c l e

i n f o

Article history: Received 19 May 2014 Revised 9 March 2015 Accepted 31 March 2015 Available online 8 April 2015 Keywords: Hypoxia–ischemia Chondroitin sulfate proteoglycan Hypomyelination Cognitive dysfunction Newborn Rat

a b s t r a c t Hypomyelination is the major cause of neurodevelopmental deficits that are associated with perinatal white matter injury. Chondroitin sulfate proteoglycans (CSPGs) are known to exert inhibitory effects on the migration and differentiation of oligodendrocytes (OLs). However, few studies describe the roles of CSPGs in myelination by OLs and the cognitive dysfunction that follows perinatal white matter injury. Here, we examined the alterations in the expression of CSPGs and their functional impact on the maturation of OLs and myelination in a neonatal rat model of hypoxic–ischemic (HI) brain injury. Three-day-old Sprague–Dawley rats underwent a right common carotid artery ligation and were exposed to hypoxia (6% oxygen for 2.5 h). Rats were given chondroitinase ABC (cABC) via an intracerebroventricular injection to digest CSPGs. Animals were sacrificed at 7, 14, 28 and 56 days after HI injury and the accompanying surgical procedure. We found that the expression of CSPGs was significantly up-regulated in the cortical regions surrounding the white matter after HI injury. cABC successfully degraded CSPGs in the rats that received cABC. Immunostaining showed decreased expression of the preoligodendrocyte marker O4 in the cingulum, external capsule and corpus callosum in HI + cABC rats compared to HI rats. However HI + cABC rats exhibited greater maturation of OLs than did HI rats, with increased expression of O1 and myelin basic protein in the white matter. Furthermore, using electron microscopy, we demonstrated that myelin formation was enhanced in HI + cABC rats, which had an increased number of myelinated axons and decreased G-ratios of myelin compared to HI rats. Finally, HI + cABC rats performed better in the Morris water maze task than HI rats, which indicates an improvement in cognitive ability. Our results suggest that CSPGs inhibit both the maturation of OLs and the process of myelination after neonatal HI brain injury. The data also raise the possibility that modifying CSPGs may repair this type of lesion associated with demyelination. © 2015 Elsevier Inc. All rights reserved.

Introduction Perinatal white matter injury (WMI) is the leading cause of adverse neurodevelopmental outcomes in premature infants. Clinical conditions related to critically ill premature infants such as perinatal asphyxia, sepsis and pulmonary dysfunctions are all predisposing factors for WMI (Logitharajah et al., 2009; Pogribna et al., 2013; Resch et al., 2012). In one study, approximately 90% of the premature infants who experienced perinatal asphyxia and hypoxia–ischemia (HI) suffered from WMI (Logitharajah et al., 2009). In perinatal WMI, a decreased expression of mature myelin proteins and decreased volume of the corpus callosum were found to be associated with hypomyelination and longterm cognitive dysfunction (Huang et al., 2009). This failure of ⁎ Corresponding author at: Department of Neonatology, Children's Hospital of Fudan University, Shanghai, China. Fax: +86 21 64931916 E-mail address: [email protected] (C. Chen).

http://dx.doi.org/10.1016/j.expneurol.2015.03.026 0014-4886/© 2015 Elsevier Inc. All rights reserved.

myelination was reported to be the result of defective development or loss of oligodendrocytes (OLs) (Back et al., 2002; Segovia et al., 2008). Oligodendrocyte precursors (OPCs) derived from neural stem cells undergo the complex and precise processes of proliferation, migration and differentiation to generate mature OLs that are capable of myelination. The selective vulnerability of pre-oligodendrocytes (pre-OLs) to HI insult has revealed a possible pathogenic mechanism that underlies the occurrence of hypomyelination (Back et al., 2002). Although reduced numbers of pre-OLs were identified in the acute stage of perinatal WMI because of degeneration or apoptosis (Back et al., 2005), the surviving pre-OLs regenerated robustly and attempted to repair myelin in lesions in the chronic stage of perinatal WMI (Segovia et al., 2008). However, these surviving pre-OLs that proliferated displayed arrested maturation and failed to differentiate and produce compact myelin (Billiards et al., 2008; Segovia et al., 2008). Thus, although a regenerative capacity is still retained in perinatal WMI, why this remyelination process is stunted remains poorly understood.

214

Y.-P. Deng et al. / Experimental Neurology 269 (2015) 213–223

Reactive gliosis typically occurs after a CNS injury. Glial cells with a hypertrophied morphology proliferate and infiltrate the lesion site and form glial scars in an attempt to demarcate the injured area, limit cell damage and restrain the spread of inflammation (Myer et al., 2006; Sofroniew and Vinters, 2010). The glial scar is rich in chondroitin sulfate proteoglycans (CSPGs), which are secreted by the reactive glial cells, especially astrocytes (McKeon et al., 1999; Tang et al., 2003). CSPGs constitute a complex family of macromolecules in the extracellular matrix of the CNS (Rauch, 2007). These proteoglycans are characterized by a core protein into which one or more glycosaminoglycan chains (GAGs) are covalently attached. The major CSPGs that are expressed in the CNS include three members of the lectican family, namely, neurocan (NC), brevican (BC) and phosphacan (PC) (Morgenstern et al., 2002; Yamaguchi, 2000). The enzyme chondroitinase ABC (cABC) can break down the GAGs of CSPGs into disaccharides. It has been widely reported that CSPGs exert an inhibitory effect on axon sprouting and reorganization because the enzymatic degradation of CSPGs enhances plasticity within the CNS (Barritt et al., 2006; Harris et al., 2013; Jefferson et al., 2011). Moreover, in vitro studies have implicated CSPGs in the inhibition of cell adhesion, proliferation and the outgrowth of processes of OPCs (Lau et al., 2012). However, whether CSPGs contribute to hypomyelination in perinatal WMI is unknown. Our previous work described that glial scars formed in the damaged brain regions using a neonatal rat model of HI brain injury (Huang et al., 2009). We also found that the HI rats developed hypomyelination and cognitive dysfunction. Hence, we hypothesized that hypomyelination in perinatal WMI was related to astrogliosis and accumulation of CSPGs. In addition, the degradation of CSPGs by cABC might attenuate hypomyelination and improve neurological functions. In the current study, we used a neonatal rat model of HI brain injury to investigate the functional roles of CSPGs in hypomyelination of the white matter after perinatal WMI.

Results CSPGs accumulate in cortical regions surrounding the white matter after HI injury We previously reported that glial cells began to proliferate and form glial scars at 3 days post-injury in a neonatal rat model of HI brain injury (Huang et al., 2009). The constituents of glial scars were not investigated in the previous study. We therefore examined the effect of HI injury on the CSPG expression profiles at 7 days post-injury. Using a primary antibody to neurocan (NC), one of the most abundant types of CSPGs in the CNS, immunofluorescence analysis revealed that there were substantial punctate or plaque-like deposits of NC in areas surrounding the corpus callosum and external capsule of the ipsilateral white matter of the HI rats (Figs. 1A, B). NC was significantly increased at 7 days postinjury, and the fluorescence intensity was greater than that in the corresponding regions in the sham rats ((9.77 ± 0.52) × 103 a.u. vs. (6.15 ± 0.29) × 103 a.u.; p b 0.05; Fig. 1C). Glial fibrillary acidic protein (GFAP) and Iba1 immunostaining was performed to assess the presence of activated astrocytes and microglia (macrophages), which are considered to be the major cellular components of glial scars (Rolls et al., 2009). After HI injury, astrocytes became hypertrophic and accumulated primarily at the margins of the lesion areas in the ipsilateral white matter, especially in the corpus callosum and external capsule (Fig. 2A). Activated astrocytes co-expressed NC and exhibited similar distribution patterns to NC expression in the HI group, and the number of GFAP-positive cells was significantly increased compared with that in the corresponding regions of the sham rats (995.34 ± 27.00 cells/mm2 vs. 509.16 ± 17.17 cells/mm2; p b 0.01; Figs. 2B, C). The number of the microglia that expressed Iba1 was also significantly increased both within and at the margins of the ipsilateral white matter in the HI rats compared with the sham rats (288.00 ± 29.60 cells/mm2 vs. 93.47 ± 16.71 cells/mm2; p b 0.01;

Fig. 1. CSPGs accumulated in the cortical regions adjacent to the white matter after HI injury. (A) Diagram of the brain regions analyzed for the expression of neurocan (NC). (B) Immunofluorescence images for NC in HI and sham rat brains are shown. Regions of white matter are outlined by white dashed lines. NC was up-regulated in the cortical regions surrounding the corpus callosum (CC) and external capsule (EC) of the ipsilateral white matter (boxed areas in (A)) in HI rats. The scale bar represents 200 μm. (C) At 7 days post-injury, a quantification of fluorescence intensity in the white matter and in the adjacent regions revealed a significant increase in NC density in the brains of HI rats compared with that in the sham rats.

Y.-P. Deng et al. / Experimental Neurology 269 (2015) 213–223

215

Fig. 2. Glial scars were induced after HI injury. (A) GFAP- and Iba1-positive cells accumulated in the glial scars with similar distribution patterns to that of NC after HI injury. Boxed areas are magnified and shown in (B). The scale bar represents 200 μm. (B) GFAP- and Iba1-positive cells co-expressed NC after HI injury (arrowhead, magnified and displayed as insets). The scale bar represents 50 μm. (C) At 7 days post-injury, GFAP- and Iba1-positive cells were significantly increased in HI rats. Histograms are the mean ± SEM, n = 8 to 10 animals/group; **p b 0.01, compared with the sham group, Student's t-test.

216

Y.-P. Deng et al. / Experimental Neurology 269 (2015) 213–223

Figs. 2A, C). Iba1-positive activated microglia, which are characterized by hypertrophic cell bodies and thick spine-like processes, also coexpressed NC in the brains of HI rats, whereas resting-state microglia were identified in the brains of the sham group ( Fig. 2B). As shown in Fig. 1, we demonstrated that NC accumulated in the cortical regions surrounding the white matter. To further explore the expression profiles of CSPGs after HI injury, we used tissue that contained the right/ipsilateral cortex for protein assays. With the same antibody to NC that was used in the immunofluorescence analysis, western blots revealed a significant increase in the expression of intact NC (275 kDa) at 7 and 14 days post-injury (p b 0.05; Figs. 3A, B). Similar effects were observed with an anti-brevican monoclonal antibody, which detects brevican (BC, ~ 140 kDa). HI injury induced a significant increase in the levels of BC at 7 days post-injury, which was sustained until 14 days post-injury (p b 0.05; Figs. 3C, D). The third type of CSPG, phosphacan (PC, ~ 400 kDa), was also significantly increased in the ipsilateral cortex of rats in the HI group at 7 days post-injury (p b 0.05). Although no significant changes in the PC levels was found at 14 days post-injury, a tendency for increased protein expression was shown, and the p value is close to 0.05 (p = 0.054; Figs. 3E, F). cABC digests CSPGs effectively in HI-induced white matter injury We hypothesized that the digestion of CSPGs would facilitate remyelination after HI injury. cABC can digest the GAGs of the CSPGs, so we used intraventricular injections of cABC to disrupt the CSPGs. A

single in vivo intraventricular injection of 2 μl of 0.02 U cABC or vehicle was performed at 3 days post-injury. Tissues were harvested 4 days later, at 7 days post-injury, to evaluate the efficacy of cABC. Using an antibody to C4S, which recognizes the protein fragment that results from a successful cleavage of GAGs from CSPGs, we confirmed the enzymatic digestion of CSPGs by immunofluorescence (Figs. 4A, B). However, vehicle-injected sham or HI rats exhibited no apparent C4S immunoreactivity (Fig. 4B). The expression of C4S was significantly increased in HI + cABC rats ((4.29 ± 0.46) × 103 a.u.) when compared to sham + cABC ((2.13 ± 0.38) × 103 a.u.), HI ((0.48 ± 0.17) × 103 a.u.) and sham ((0.19 ± 0.05) × 103 a.u.) rats (p b 0.05, respectively; Fig. 4C). In sham + cABC rats, C4S expression was also upregulated when compared to HI and sham rats (p b 0.05, respectively; Fig. 4C). No significant difference in C4S expression was found between HI and sham rats. Degradation of CSPGs promotes the development of OLs To determine the fate of cells of the OL lineage after the degradation of CSPGs, we stained with an antibody to O4 that identifies pre-OLs (Segovia et al., 2008). O4-positive fragmented pre-OLs were observed in the HI group at 7 days post-injury, but the pre-OLs in HI + cABC rats showed densely stained soma and ramified processes (Fig. 5A) and the HI + cABC group had fewer O4-positive cells than did HI rats (358.89 ± 33.39 cells/mm2 vs. 436.04 ± 17.32 cells/mm2; p b 0.05; Fig. 4B). The number of pre-OLs was greater in HI + cABC rats compared with those in sham and sham + cABC rats (358.89 ± 33.39 cells/mm2

Fig. 3. Semiquantitative cortical protein expression analysis of neurocan, brevican, and phosphacan from the right/ipsilateral cortex lysates. Western blot illustrated a major band at 275 kDa for NC, ~140 kDa for BC, and ~400 kDa for PC (A, C, E). The corresponding blot of GAPDH was shown below as a loading control. Gels shown were representative of the time point of 7 days post-injury. The expression of CSPGs increased significantly in the ipsilateral cortex at 7 days post-injury, and these changes were sustained until 14 days post-injury (B, D, F). Histograms for the relative densities of the bands represent the mean ± SEM, n = 6 animals/group, for 3 independent experiments; *p b 0.05 compared to the sham group, Student's t-test.

Y.-P. Deng et al. / Experimental Neurology 269 (2015) 213–223

217

Fig. 4. In vivo administration of cABC degraded CSPGs. (A) Diagram of the brain regions analyzed for the expression of C4S. (B) Double immunofluorescence for digested CSPGs (C4S) and astroglial cells (GFAP). Robust expressions of C4S were observed in the EC and the adjacent regions (boxed area in (A)) of HI + cABC rats, as well as in the corresponding regions of sham + cABC rats. Significant co-expression of C4S and GFAP was shown in HI + cABC rats (arrowhead, magnified and displayed as inset). No apparent C4S immunoreactivity was detected in HI and sham rats. The scale bar represents 50 μm. (C) Quantification of fluorescence intensity of C4S. Histogram is the mean ± SEM, n = 8 to 10 animals/group; *p b 0.05, one-way analysis of variance, post hoc SNK test.

vs. 235.66 ± 8.91 cells/mm2 and 170.76 ± 10.43 cells/mm2; p b 0.05 respectively; Fig. 5B). At 14 days post-injury, O4-positive cells were still present and the number of these cells was greater in HI + cABC rats compared with those in sham and sham + cABC rats (272.94 ± 9.37 cells/mm 2 vs. 190.06 ± 6.07 cells/mm 2 and 159.86 ± 27.83 cells/mm2; p b 0.05 respectively; Fig. 5B). However, the expression of O4 was not significantly different between the HI and HI + cABC animals at this time point (285.97 ± 22.79 cells/mm2 vs. 272.94 ± 9.37 cells/mm2; p N 0.05; Fig. 5B). We next detected immature OLs with an antibody to O1 to evaluate the early myelination that occurred at 7 and 14 days post-injury. At 7 days post-injury, while the number of O1-stained cells was fewer in the HI rats than in the sham rats (175.01 ± 15.94 cells/mm 2 vs. 300.67 ± 22.14 cells/mm2 ; p b 0.05; Fig. 5C), we found more O1-positive cells with diffuse distribution in the white matter of HI + cABC rats compared to HI and sham + cABC rats (391.13 ± 51.46 cells/mm2 vs. 175.01 ± 15.94 cells/mm2 and 141.98 ± 26.76 cells/mm2; p b 0.05 respectively; Fig. 5C). There was no significant difference in the expression of O1 between HI and sham + cABC rats. At 14 days post-injury, the O1-positive cells exhibited a similar tendency to those in the earlier phase at 7 days post-injury; the O1positive cell density was significantly decreased in HI rats compared to sham rats (188.27 ± 11.49 cells/mm2 vs. 324.02 ± 20.26 cells/mm2; p b 0.05; Figs. 5A, C), and HI + cABC rats had a significantly increased

number of O1-positive cells compared with the HI and sham + cABC rats (394.69 ± 36.50 cells/mm2 vs. 188.27 ± 11.49 cells/mm2 and 132.06 ± 6.34 cells/mm2; p b 0.05 respectively; Fig. 5C). No significant difference in the expression of O1 was found between the HI + cABC rats and sham rats at either time point. However, at both time points, sham rats had more O1-positive cells than did sham + cABC rats. To determine the long-term effect of the degradation of CSPGs on the myelination process after HI injury, we evaluated the expression of myelin basic protein (MBP), a major protein component of the myelin sheath, at 28 days post-injury. We performed an immunohistochemical analysis using an antibody to MBP. The MBP-positive area was found to be significantly increased in the external capsule and corpus callosum in HI + cABC rats compared to that in the corresponding regions of HI rats (0.93 ± 0.06 mm2 vs. 0.66 ± 0.08 mm2; p b 0.05; Figs. 5A, D). An apparent loss of MBP as well as hypomyelinated projection fibers and decreased connections to the cortex were observed in the white matter of the HI rats (Fig. 5A). The MBP-positive area was significantly decreased in HI rats compared with sham rats (0.66 ± 0.08 mm2 vs. 0.94 ± 0.06 mm2; p b 0.05; Figs. 4A, D). No significant difference in MBP expression was found between the HI + cABC and sham rats, nor between the HI and sham + cABC rats (0.66 ± 0.08 mm2 vs. 0.59 ± 0.06 mm2 p b 0.05; Figs. 5A, D). Both the HI + cABC and sham rats had greater MBP-positive areas than did sham + cABC rats.

218

Y.-P. Deng et al. / Experimental Neurology 269 (2015) 213–223

Fig. 5. Degradation of CSPGs by cABC promotes myelination after HI brain injury. (A) Representative figures of O4, O1, and MBP immunostaining. Staining for O4 was markedly increased in HI rats and HI + cABC rats relative to the sham rats. Note that O4-labeled cells appeared to be fragmental within the lesions in HI rats but displayed densely stained soma and ramified processes in the corresponding regions of HI + cABC rats. The number of O1-labeled cells was significantly increased with diffuse distribution in the white matter in HI + cABC rats relative to HI rats. For O4 and O1 staining, the scale bars represent 50 μm. Staining for MBP showed increased myelin expression in the white matter in HI + cABC rats compared to HI rats, in which projection fibers were hypomyelinated with decreased connectivity to the cortex. The scale bar represents 100 μm. (B–D) Histograms for cell counts of O4, O1, and MBP markers. Data represent the mean ± SEM, n = 8 to 10 animals/group; *p b 0.05, one-way analysis of variance, post hoc SNK test.

Compact myelin formation is enhanced after degradation of CSPGs in HI rats We have demonstrated that the expression of MBP was significantly increased in HI + cABC rats compared to HI rats. To further explore whether the structure and density of myelin were influenced by HI injury and degradation of CSPGs with cABC, we analyzed sections of the corpus callosum at 8 weeks post-injury by electron micrography (EM). EM showed a pronounced lack of axonal myelination in HI rats (Fig. 6A, top). The myelinated fibers were sparsely distributed with fewer myelin wraps in HI rats, while myelin appeared more tightly wrapped in HI + cABC, sham + cABC and sham rats (Fig. 5A, bottom). The numbers of myelinated fibers in the corpus callosum in HI + cABC, HI and sham + cABC groups were all significantly decreased relative to that in the sham group ((7.81 ± 0.74) × 105/mm2, (4.27 ± 0.62) × 105/mm2 and (7.59 ± 0.39) × 105/mm2 respectively vs. (10.01 ± 0.58) × 105/mm2; p b 0.05; one-way ANOVA; Fig. 6B). Nevertheless, significantly more myelinated fibers were identified in HI + cABC and sham + cABC rats compared to HI rats (p b 0.05, respectively; one-way ANOVA; Fig. 6B). No differences in the number of unmyelinated fibers were noted between HI and HI + cABC rats ((6.53 ± 0.84) × 105/mm2 vs. (4.99 ± 1.56) × 105/mm2; p N 0.05; one-way ANOVA; Fig. 5C), but both groups had more unmyelinated fibers relative

to sham + cABC or sham rats ((1.56 ± 0.10) × 105/mm2 or (2.19 ± 0.18) × 105/mm2 respectively; p b 0.05; one-way ANOVA; Fig. 6C). A G-ratio (the ratio of the inner axonal diameter to the outer diameter of the fiber) analysis revealed that compact myelin was thinner in HI rats compared to HI + cABC, sham + cABC and sham rats (G-ratio: 0.731 ± 0.005 vs. 0.711 ± 0.004, 0.694 ± 0.004 and 0.665 ± 0.004 respectively; p b 0.05; one-way ANOVA; Fig. 5D). HI + cABC rats showed a significantly higher G-ratio compared to sham + cABC or sham rats (p b 0.05 respectively; one-way ANOVA; Fig. 5D). Interestingly, G-ratio in sham + cABC rats was also increased relative to sham rats (p b 0.05; one-way ANOVA; Fig. 5D). All of these results confirmed that myelination was enhanced by the degradation of CSPGs after neonatal HI brain injury. Degradation of CSPGs by cABC improves the cognitive function after HI injury The Morris water maze (MWM) task is one of the most widely used tests for the assessment of spatial learning and memory, which are associated with cognitive function (Vorhees and Williams, 2006). A repeated measure analysis of variance (ANOVA) was performed for the navigation trials, with the “group” as the between-subjects factor and the navigation day as the within-subjects factor. All experimental

Y.-P. Deng et al. / Experimental Neurology 269 (2015) 213–223

219

Fig. 6. Alteration of myelin ultrastructure after degradation of CSPGs in HI rats. EM analysis was performed in different treatment groups at 8 weeks post-injury. (A) Electron micrographs of the corpus callosum showed that in HI + cABC rats, hypomyelination was attenuated, with more myelinated axons and more myelin wraps compared to that in the HI rats. The scale bar represents 1 μm. (B, C) Histograms for the densities of myelinated/unmyelinated axons in electron micrographs of the corpus callosum. (D) Histograms for G-ratio of the myelinated axons. Data represent the mean ± SEM, n = 8 animals/group; *p b 0.05, one-way analysis of variance, post hoc SNK test.

groups exhibited the ability to learn because the escape latency (EL) and swimming distance decreased over time. A significant effect was found in terms of the navigation day (p b 0.01). There was also a significant effect in terms of the different groups and the escape performance of the rats in these groups (p b 0.01). Post hoc analysis revealed that from the second day until the end of the navigation test, the HI + cABC rats had a significantly shorter EL (p b 0.05) and swimming distance (on the fourth day, p b 0.01, and on the second and third days, p b 0.05) relative to HI rats (Figs. 7A and B). The HI rats demonstrated long-term cognitive deficits, as the EL and swimming distance were significantly greater compared to sham rats (on the first day, p b 0.05, and on the second through fourth days, p b 0.01, Figs. 7A and B). Although there was a significant difference in the EL and swimming distance between HI + cABC and sham rats on the first or second day (p b 0.05), this difference was no longer significant on the third and fourth days (Figs. 7A and B). These results indicate that the learning ability was improved in HI + cABC rats. In the MWM probe test, HI rats spent less time in the target quadrant compared to HI + cABC, sham + cABC or sham rats (p b 0.05 respectively; Fig. 7C). Additionally, HI rats had significantly fewer platform crossings compared to

HI + cABC, sham + cABC or sham rats (p b 0.05 respectively; Figs. 7D and E). Thus, spatial memory and cognition were impaired after exposure to HI and were improved by the degradation of CSPGs. HI + cABC rats showed no significant differences in the number of platform crossings or the time spent in the target quadrant compared to sham + cABC or sham rats (p N 0.05 respectively; Figs. 7C and D). Discussion HI caused acute degeneration and the subsequent arrested maturation of the pre-OLs which are thought to be the key cellular mechanisms of the hypomyelination following neonatal HI brain injury (Back et al., 2002; Segovia et al., 2008). Pre-OLs were observed to accumulate within and surrounding the lesion area in an attempt to repair and remyelinate the axons after the CNS injury (Zaidi et al., 2004). However, the regenerated OPCs failed to mature and differentiate into functional OLs. The cause of this hindered process was rarely studied. It has been suggested that glial scars, enriched in CSPGs, are produced after spinal cord injury (SCI) and perinatal WMI in humans (Levine et al., 2001; Sherman and Back, 2008). Recently, CSPGs were found to inhibit both

220

Y.-P. Deng et al. / Experimental Neurology 269 (2015) 213–223

Fig. 7. Cognitive function was assessed by the Morris water maze test. (A) During the visible and hidden platform training, HI + cABC rats showed an enhanced ability to locate the platform relative to HI rats during the second to the fourth navigation day. HI rats had a greater EL compared to sham and sham + cABC rats. (B) HI + cABC rats swam shorter escape distances relative to HI rats during the second to the fourth navigation day. HI rats showed an impaired learning ability relative to sham and sham + cABC rats. (C, D) In the probe test, HI + cABC rats had improved spatial memory with more platform crossings and a higher percentage of time spent in the target quadrant compared to HI rats. (E) Swim path traces from the probe test. Error bars indicate the SEM. *p b 0.05 and **p b 0.01 compared with the sham group; #p b 0.05 and ##p b 0.01 compared with the HI group; repeated measures and multivariate analysis of variance (ANOVA), post hoc LSD test.

the adhesion of the pre-OLs and the outgrowth of their processes in culture (Lau et al., 2012; Siebert and Osterhout, 2011). However, we are not clear about the roles of CSPGs in hypomyelination after neonatal HI brain injury. Thus, the present study tested the hypothesis that CSPGs may be a potent inhibitor of myelination by OLs after HI brain injury. To address this issue, it is important to examine the temporal and spatial expression patterns of the CSPGs after HI injury. Our data suggested that CSPGs were up-regulated in the acute phase of HI brain injury and were sustained for at least 14 days. Notably, CSPGs were deposited primarily in the area adjacent to the white matter. Immunohistochemistry revealed high-intensity staining of neurocan in the glial scar regions where reactive astroglia and microglia accumulated. This is consistent with the western blot data that showed an increased expression of NC, BC and PC proteins at 7 and 14 days post-injury. The

CSPGs accumulated mainly in the ipsilateral cortex that overlies the white matter, which was in agreement with the results of other studies of CNS injury. Disparate regional changes in the expression of CSPGs were observed after spinal cord contusion injury. It was noted that CSPGs were not present in the lesion core, which likely resulted from progressive cell death, but up-regulated in the glial scars that surrounded the primary injury site (Harris et al., 2009). Both BC and PC were found to be significantly increased in the infarcted core and the adjacent brain regions in the 7-day-old rat model of HI brain injury (Leonardo et al., 2008; Matsui et al., 2005). In contradiction to these data, the western blot analysis using this same model demonstrated a reduction in the expression of both BC and NC in the injured rat brains (Aya-ay et al., 2005; Matsui et al., 2005). It is not surprising that the tissues analyzed in the above western blot experiments contained both the

Y.-P. Deng et al. / Experimental Neurology 269 (2015) 213–223

injury core and the surrounding glial scars, which were different from ours. Therefore, the observed reduction of CSPGs suggests the net effect of the HI injury. Intervention with cABC was the most widely adopted strategy for modification of CSPGs. cABC is an enzyme that cleaves the CS-GAG chains from the CSPG core protein and results in the liberation of disulfate disaccharides (Suzuki et al., 1968; Yamagata et al., 1968). There is abundant evidence that the cABC-mediated digestion of CSPGs can increase brain plasticity (Barritt et al., 2006; Harris et al., 2013; Jefferson et al., 2011). In addition, cABC retains its activity for the digestion of chondroitin sulfate chains for 7–10 days after injection into the injured brain (Lin et al., 2008). Therefore, in the present study, we injected a single dose of cABC into the ipsilateral ventricle of the rat brain which differed from other protocols. In adult models of brain injury, cABC was continuously or interventionally infused into the lesioned brain (Chen et al., 2014; Jefferson et al., 2011). The brain's plasticity is greater during development than it is during adulthood, as is the influence of the environment (Hensch, 2004). Thus, we reduced the number of cABC injections to investigate the resultant effects. CSPGs have been extensively studied for their inhibitory effects on the regeneration of axons after SCI. It is the first study to explore the role of CSPGs in hypomyelination after perinatal WMI. Our data showed that in vivo treatment with cABC had successfully digested the GAGs of CSPGs. The number of pre-OLs was significantly reduced in HI + cABC rat brains at 7 days post-injury, and the number of immature OLs was significantly increased at both 7 and 14 days post-injury in this group. The enhanced myelination observed in HI + cABC rats was most likely caused by the improved differentiation of the regenerated pre-OLs. Recent work has demonstrated that the degradation of CSPGs can increase the migration of endogenous OPCs to the site of the demyelinating lesion following SCI (Siebert et al., 2011). Similarly, increased numbers of OLs and enhanced myelination were also found in a model of multiple sclerosis after the administration of cABC to suppress the synthesis of CSPGs (Lau et al., 2012). These results showed the different roles of CSPGs in the functions of OLs, which can be attributed to the different mechanisms of demyelination in these models of CNS injury. CSPGs can interact with a large number of binding partners, such as cell adhesion molecules, growth factors, chemokines and other types of receptors, to mediate an inhibitory effect on axons and OLs (Bartus et al., 2012). CSPGs may block neurite outgrowth through the activation of the mitogen-activated protein kinase (MAPK) pathway, which is dependent on the epidermal growth factor receptor (EGFR) (Kaneko et al., 2007). The epidermal growth factor receptor ErbB1 was found to be an important mediator of the outgrowth-inhibitory cues (Leinster et al., 2013). In addition to MAPK, several groups have also shown that CSPGs may signal through the Rho/ROCK pathway to exert inhibitory roles in the growth and regeneration of the axons (Lingor et al., 2007; Monnier et al., 2003). Recently, a receptor for CSPGs, PTPσ, was identified; it was determined that this receptor participates in the Rho/ROCK signaling pathway and allows CSPGs to suppress the outgrowth of OL processes and their capacity to remyelinate axons after SCI (Pendleton et al., 2013). However, the molecular basis for the interaction of CSPGs with the pre-OLs in the developing brain requires further exploration. In the present study, the MWM experiments showed decreased EL of the HI + cABC group compared to the HI group in reference and working memory protocols. This implies that CSPG degradation improved the functions of learning, memory and cognition in HI rats. Many brain structures are involved in learning and memory processes and cognition (D'Hooge and De Deyn, 2001). It has been well established that the hippocampus has an important role in MWM performance (Redish and Touretzky, 1998). However, rats with damage to the white matter, such as the fimbria and fornix, showed deficits in spatial memory and learning as well (Nilsson et al., 1987). The functions of the hippocampus are interrupted by the damaged fiber connections and pathways that run through the fimbria and fornix, and

221

disconnecting rather than destroying brain regions that contribute to the spatial learning may affect the MWM acquisition as well. WMI in the corpus callosum was reported to be associated with impaired learning and memory (Cengiz et al., 2011; Huang et al., 2009). The degradation of CSPGs with cABC was observed to restore sensory function and promote behavior recovery after SCI (Cafferty et al., 2008; Starkey et al., 2012). We are the first to reveal that the degradation of CSPGs can improve myelination and cognitive functions in neonatal HIinduced brain injury. Our data suggest that the modification of CSPGs may have possible effects on brain plasticity and functional recovery in perinatal WMI. In conclusion, we have demonstrated that CSPGs play a key role in inhibiting myelination after neonatal HI brain injury because treatment with cABC promotes myelin formation and cognitive development. Therefore, strategies that prevent the synthesis of CSPGs, promote their degradation, or block their signaling might enhance brain plasticity and functional recovery after neonatal HI brain injury. Further investigations are needed to explore the regulatory mechanisms of CSPG synthesis and degradation, the receptors that the OLs use to respond to CSPGs and the associated signal transduction pathways. The development of therapeutic agents targeting CSPGs could provide a novel class of neuroprotective agents that may alleviate hypomyelination after neonatal HI brain injury. Materials and methods Animal protocols All procedures were approved by the Animal Ethics Committee of Fudan University. The right common carotid artery was double-ligated and cut between the ligatures in 3-day-old (P3) Sprague–Dawley rats (10–12 pups/litter) (Back et al., 2002; Vannucci et al., 1999). The total surgery time was limited to 5 min. After a recovery period of 1–1.5 h, the pups were placed in the containers that were submerged in a water bath to maintain normothermia (37 °C). To induce moderate HI cerebral injury, the container was perfused with humidified oxygen (6%) in a nitrogen gas mixture that flowed at 3 l/min for 2.5 h. Thereafter, the pups were returned to their respective dams until they were sacrificed. Sham operation was performed with neither ligating the common carotid artery nor exposure to hypoxia. HI rat pups were randomly allocated to the HI group or to the HI + cABC group. Intracerebroventricular injection Three days after surgery and HI (P6), a single dose of cABC (HI + cABC and sham + cABC groups) or vehicle (HI and sham groups) was injected into the ventricles of the HI and sham operated rats. The animals were fixed in a stereotaxic frame (51600; Stoelting, Wood Dale, IL, USA), and the heads were retracted. The injection site was identified as 0.8 mm posterior from the lambda, 1.0 mm lateral from the sagittal suture, and 2.5 mm deep from the dura. A 2 μl injection of 10 mU/μl cABC (Sigma-Aldrich; St. Louis, MO, USA) solution containing 50 mM Tris (Sangon Biotech; SH, SHN), 60 mM sodium acetate (Sangon Biotech; SH, CHN), and 0.02% bovine serum albumin (Usen Biotech; SH, CHN) at a pH of 8.0 or vehicle alone (the same buffer as described in the cABC solution) was delivered via a glass 34-gauge needle at a rate of 0.5 μl/min. There was a 40 s interval before the needle was retracted from the head. Histology At 7, 14, and 28 days after surgery and HI (P10, P17, and P31), the rats were euthanized and perfused with ice-cold 4% paraformaldehyde in 0.1 M phosphate buffer of pH 7.4 before the brains were extracted. The brains were then sequentially immersed in 4% paraformaldehyde as described above for 24 h and then were immersed in a 20% sucrose

222

Y.-P. Deng et al. / Experimental Neurology 269 (2015) 213–223

solution (in 4% paraformaldehyde) and a 30% sucrose solution (in 0.1 M phosphate buffer, pH 7.4) at 4 °C until the brains descended to the bottom of the tubes. Coronal brain sections (30 μm) were cut sequentially using a cryostat (CM 1950; Leica, Nussloch, Germany) and stored in a cryoprotectant solution (30% ethylene glycol, 30% sucrose in 0.1 M phosphate buffer, pH 7.4) at −20 °C. A total of 8–10 brains were analyzed for each group at each time point. Immunofluorescence Free-floating tissue sections were washed in PBS three times and then saturated with blocking buffer (10% goat serum, 0.3% Triton-X100 in PBS) for 30 min at 37 °C, followed by incubation with the primary antibody overnight at 4 °C. The following primary antibodies were used: mouse monoclonal anti-neurocan antibody (1:1000, MAB5212; Chemicon, CA, USA), mouse monoclonal anti-chondroitin-4-sulfate (also called C4S, 1:1000, MAB2030; Chemicon, CA, USA), rabbit polyclonal anti-GFAP (1:1000, AB5804; Chemicon, CA, USA), and rabbit polyclonal anti-IBA1 (1:1000, 019-19741; Wako, Tokyo, Japan). After the sections were rinsed, they were incubated for 1 h at 37 °C with the appropriate Alexa Fluor-conjugated secondary antibodies (1:200, Invitrogen, CA, USA). Sections were mounted with Vectashield (H1200; Vector, CA, USA), which contains DAPI to visualize the nuclei. Immunofluorescence was visualized and photographed using a Leica TCSSP8 fluorescence microscope. The sections were analyzed in the regions at the level of the midseptal nuclei, which encompassed the supracallosal radiation, the corpus callosum, and the adjacent external capsule (Back et al., 2002). For each section, a minimum of three fields were analyzed, and the average values of staining intensity and positive cell density per section were calculated in a blinded manner using the Image-Pro Plus 6.0 software (Media Cybernetics; Rockville, MD, USA).

(5%–6%) and transferred to a nitrocellulose membrane. To detect CSPG proteins, the following primary antibodies were used: mouse monoclonal anti-neurocan (1:1000, MAB5212; Chemicon, CA, USA), mouse monoclonal anti-phosphacan (1:1000, MAB5210; Chemicon, CA, USA), mouse monoclonal anti-brevican (1:1000, MABN491; Chemicon, CA, USA) and mouse monoclonal anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (1:1000, KC-5G4; Kangchen, SH, CHN). The membranes were saturated with blocking buffer (5% skim milk in Tris-buffered saline), incubated with horseradish peroxidase-conjugated goat antimouse IgG (1:5000, PK-6102; Vector, CA, USA), and then developed using a chemiluminescent detection system (ChemiQ 4600; Bioshine, SH, CHN). The intensities of the bands were analyzed using the Image J 1.46 software (National Institutes of Health; Bethesda, MD, USA). Six animals from each group at each time point were used in this analysis. Electron microscopy (EM) Eight weeks post-injury (P59), rats were anesthetized and perfused with 4% paraformaldehyde and 1% glutaraldehyde in 0.1 M phosphate buffer (pH 7.2). After perfusion, the brains were dissected out, and the corpus callosum was post-fixed in 4% paraformaldehyde and 2.5% glutaraldehyde in 0.1 M phosphate buffer (pH 7.2). Tissue samples were transferred to 0.1 M PBS and immersed in 1% OsO4 in 0.1 M PBS, dehydrated in ethanol, and then embedded with epoxy resin. Ultrathin sections (70 nm) were cut with an ultramicrotome (4802A; LKBProdukter AB, Bromma, Sweden) and placed onto copper grids. After staining with uranyl acetate, followed by lead citrate, the sections were examined using a transmission electron microscope (JEM-1400; JEOL, Tokyo, Japan). Randomly selected images were analyzed using the NIH Image J 1.46 software (National Institutes of Health; Bethesda, MD, USA). Eight animals of each group were used in this analysis.

Immunohistochemistry

Morris water maze

Tissue sections were washed three times in PBS, incubated at room temperature (RT) in 1% hydrogen peroxide in PBS for 30 min, and were rinsed three times again in PBS. The sections were then saturated with blocking buffer (10% goat serum, 0.3% Triton-X-100 in PBS) for 30 min at 37 °C, followed by incubation with the primary antibody overnight at 4 °C. The following primary antibodies were used: mouse monoclonal anti-O4 (1:200, MAB345; Chemicon, CA, USA), mouse monoclonal anti-O1 (1:200, MAB344; Chemicon, CA, USA), and mouse anti-MBP (1:800, SMI-94R; Covance, CA, USA). After the sections were rinsed, they were incubated for 1 h at 37 °C with the following biotinylated secondary antibodies: goat anti-mouse IgG (1:200, PK-6102; Vector, CA, USA), goat anti-rabbit IgG (1:200, PK-6101; Vector, CA, USA), and goat anti-mouse IgM (for O4 and O1, 1:200, M31515; Invitrogen, CA, USA). Immunostaining was performed using an ABC kit (1:200, PK-6101 and PK-6102; Vector, CA, USA); sections were incubated with ABC solution for 1 h at RT and developed with 0.05% diaminobenzidine as the chromogen (0430; Amresco, OH, USA). Sections were then transferred to gelatinized slides, dehydrated, and mounted in neutral balsam. Immunohistochemical staining was visualized and photographed using a Leica DMRA2 microscope. The positive cell density and staining area were analyzed as described above.

Learning and memory tests were performed on P52–56 (n = 15 for the HI and HI + cABC groups respectively, and n = 10 for the sham and sham + cABC groups respectively). For this experiment, the Morris water maze (MWM) consisted of a circular black pool (180 cm in diameter, 60 cm high, and 45 cm deep) filled with water (24 ± 0.5 °C), and a circular platform (15 cm in diameter and 27 cm high). The platform was located in the center of the southwest quadrant of the pool and was submerged 1 cm below the surface of the water. For the navigation test, the animals performed four trials daily for four consecutive days. During each trial, the rats were placed into the water and released facing the wall of the pool at each of the four quadrants (labeled North (N), East (E), South (S), and West (W)), which were randomly predetermined. In each trial, the rat was allowed to swim for 60 s until it found the platform and remained there for 15 s. If the rat did not locate the platform within 60 s, it was guided onto the platform and allowed to remain there for 30 s to enhance its spatial memory. The movement of the rats within the maze was recorded by a computerized tracking system and software (ANY-maze 4.98; Stoelting, Wood Dale, IL, USA). The EL (time to reach the platform) and the swimming distance were analyzed to evaluate the learning capabilities of the animals. On the fifth day (at P56), the platform was removed, and a probe test was performed. The number of platform area crossings and the amount of time spent in the platform quadrant were analyzed to evaluate the memory capacity of the animals. The navigation and probe tests were performed by individuals who were blinded to the experimental group.

Western blot analysis Animals were sacrificed at 7 and 14 days post-injury (P10 and P17) by decapitation. The freshly dissected brain cortices were washed and stored at −80 °C. Frozen tissue samples were pulverized into powdered tissues with a homogenizer (Tiss-24; Jingxing, SH, CHN) and dissolved with RIPA lysis buffer (P0013B; Beyotime, SH, CHN) containing 1 mM PMSF (phenylmethanesulfonyl fluoride, M145; Amresco, OH, USA) for 2 min. The protein samples were then denatured at 100 °C for 5 min. The denatured protein samples were separated using SDS-PAGE

Data analysis Immunofluorescence and immunohistochemistry images were analyzed at 400× magnification, and positive cells were counted by an assessor who was blinded with regard to the animals and their respective groups. Data were presented as the mean ± standard error.

Y.-P. Deng et al. / Experimental Neurology 269 (2015) 213–223

Unpaired t-tests were performed to compare the measurements of NC, GFAP, and Iba1 staining intensities and the western blot data between the HI and sham animals. A comparison of the immunostaining data across the different groups was conducted using one-way analysis of variance (ANOVA) followed by a post hoc SNK test. So did the EM data. Repeated measurements and a multivariate ANOVA, the general linear model in the SPSS software, were used to analyze MWM data among the different groups and time points. Acknowledgments This work was supported by the National Natural Science Foundation of China (grant number 81070520 to Chao Chen). References Aya-ay, J., Mayer, J., Eakin, A.K., Muffly, B.G., Anello, M., Sandy, J.D., Gottschall, P.E., 2005. The effect of hypoxic–ischemic brain injury in perinatal rats on the abundance and proteolysis of brevican and NG2. Exp. Neurol. 193, 149–162. Back, S.A., Han, B.H., Luo, N.L., Chricton, C.A., Xanthoudakis, S., Tam, J., Arvin, K.L., Holtzman, D.M., 2002. Selective vulnerability of late oligodendrocyte progenitors to hypoxia–ischemia. J. Neurosci. 22, 455–463. Back, S.A., Luo, N.L., Mallinson, R.A., O'Malley, J.P., Wallen, L.D., Frei, B., Morrow, J.D., Petito, C.K., Roberts Jr., C.T., Murdoch, G.H., Montine, T.J., 2005. Selective vulnerability of preterm white matter to oxidative damage defined by F2-isoprostanes. Ann. Neurol. 58, 108–120. Barritt, A.W., Davies, M., Marchand, F., Hartley, R., Grist, J., Yip, P., McMahon, S.B., Bradbury, E.J., 2006. Chondroitinase ABC promotes sprouting of intact and injured spinal systems after spinal cord injury. J. Neurosci. 26, 10856–10867. Bartus, K., James, N.D., Bosch, K.D., Bradbury, E.J., 2012. Chondroitin sulphate proteoglycans: key modulators of spinal cord and brain plasticity. Exp. Neurol. 235, 5–17. Billiards, S.S., Haynes, R.L., Folkerth, R.D., Borenstein, N.S., Trachtenberg, F.L., Rowitch, D.H., Ligon, K.L., Volpe, J.J., Kinney, H.C., 2008. Myelin abnormalities without oligodendrocyte loss in periventricular leukomalacia. Brain Pathol. 18, 153–163. Cafferty, W.B., Bradbury, E.J., Lidierth, M., Jones, M., Duffy, P.J., Pezet, S., McMahon, S.B., 2008. Chondroitinase ABC-mediated plasticity of spinal sensory function. J. Neurosci. 28, 11998–12009. Cengiz, P., Uluc, K., Kendigelen, P., Akture, E., Hutchinson, E., Song, C., Zhang, L., Lee, J., Budoff, G.E., Meyerand, E., Sun, D., Ferrazzano, P., 2011. Chronic neurological deficits in mice after perinatal hypoxia and ischemia correlate with hemispheric tissue loss and white matter injury detected by MRI. Dev. Neurosci. 33, 270–279. Chen, X.R., Liao, S.J., Ye, L.X., Gong, Q., Ding, Q., Zeng, J.S., Yu, J., 2014. Neuroprotective effect of chondroitinase ABC on primary and secondary brain injury after stroke in hypertensive rats. Brain Res. 1543, 324–333. D'Hooge, R., De Deyn, P.P., 2001. Applications of the Morris water maze in the study of learning and memory. Brain Res. Rev. 36, 60–90. Harris, N.G., Carmichael, S.T., Hovda, D.A., Sutton, R.L., 2009. Traumatic brain injury results in disparate regions of chondroitin sulfate proteoglycan expression that are temporally limited. J. Neurosci. Res. 87, 2937–2950. Harris, N.G., Nogueira, M.S., Verley, D.R., Sutton, R.L., 2013. Chondroitinase enhances cortical map plasticity and increases functionally active sprouting axons after brain injury. J. Neurotrauma. 30, 1257–1269. Hensch, T.K., 2004. Critical period regulation. Annu. Rev. Neurosci. 27, 549–579. Huang, Z., Liu, J., Cheung, P.Y., Chen, C., 2009. Long-term cognitive impairment and myelination deficiency in a rat model of perinatal hypoxic–ischemic brain injury. Brain Res. 1301, 100–109. Jefferson, S.C., Tester, N.J., Howland, D.R., 2011. Chondroitinase ABC promotes recovery of adaptive limb movements and enhances axonal growth caudal to a spinal hemisection. J. Neurosci. 31, 5710–5720. Kaneko, M., Kubo, T., Hata, K., Yamaguchi, A., Yamashita, T., 2007. Repulsion of cerebellar granule neurons by chondroitin sulfate proteoglycans is mediated by MAPK pathway. Neurosci. Lett. 423, 62–67. Lau, L.W., Keough, M.B., Haylock-Jacobs, S., Cua, R., Doring, A., Sloka, S., Stirling, D.P., Rivest, S., Yong, V.W., 2012. Chondroitin sulfate proteoglycans in demyelinated lesions impair remyelination. Ann. Neurol. 72, 419–432. Leinster, V.H., Joy, M.T., Vuononvirta, R.E., Bolsover, S.R., Anderson, P.N., 2013. ErbB1 epidermal growth factor receptor is a valid target for reducing the effects of multiple inhibitors of axonal regeneration. Exp. Neurol. 239, 82–90. Leonardo, C.C., Eakin, A.K., Ajmo, J.M., Gottschall, P.E., 2008. Versican and brevican are expressed with distinct pathology in neonatal hypoxic–ischemic injury. J. Neurosci. Res. 86, 1106–1114. Levine, J.M., Reynolds, R., Fawcett, J.W., 2001. The oligodendrocyte precursor cell in health and disease. Trends Neurosci. 24, 39–47.

223

Lin, R., Kwok, J.C., Crespo, D., Fawcett, J.W., 2008. Chondroitinase ABC has a long-lasting effect on chondroitin sulphate glycosaminoglycan content in the injured rat brain. J. Neurochem. 104, 400–408. Lingor, P., Teusch, N., Schwarz, K., Mueller, R., Mack, H., Bahr, M., Mueller, B.K., 2007. Inhibition of Rho kinase (ROCK) increases neurite outgrowth on chondroitin sulphate proteoglycan in vitro and axonal regeneration in the adult optic nerve in vivo. J. Neurochem. 103, 181–189. Logitharajah, P., Rutherford, M.A., Cowan, F.M., 2009. Hypoxic–ischemic encephalopathy in preterm infants: antecedent factors, brain imaging, and outcome. Pediatr. Res. 66, 222–229. Matsui, F., Kakizawa, H., Nishizuka, M., Hirano, K., Shuo, T., Ida, M., Tokita, Y., Aono, S., Keino, H., Oohira, A., 2005. Changes in the amounts of chondroitin sulfate proteoglycans in rat brain after neonatal hypoxia–ischemia. J. Neurosci. Res. 81, 837–845. McKeon, R.J., Jurynec, M.J., Buck, C.R., 1999. The chondroitin sulfate proteoglycans neurocan and phosphacan are expressed by reactive astrocytes in the chronic CNS glial scar. J. Neurosci. 19, 10778–10788. Monnier, P.P., Sierra, A., Schwab, J.M., Henke-Fahle, S., Mueller, B.K., 2003. The Rho/ROCK pathway mediates neurite growth-inhibitory activity associated with the chondroitin sulfate proteoglycans of the CNS glial scar. Mol. Cell. Neurosci. 22, 319–330. Morgenstern, D.A., Asher, R.A., Fawcett, J.W., 2002. Chondroitin sulphate proteoglycans in the CNS injury response. Prog. Brain Res. 137, 313–332. Myer, D.J., Gurkoff, G.G., Lee, S.M., Hovda, D.A., Sofroniew, M.V., 2006. Essential protective roles of reactive astrocytes in traumatic brain injury. Brain 129, 2761–2772. Nilsson, O.G., Shapiro, M.L., Gage, F.H., Olton, D.S., Björklund, A., 1987. Spatial learning and memory following fimbria–fornix transection and grafting of fetal septal neurons to the hippocampus. Exp. Brain Res. 67, 195–215. Pendleton, J.C., Shamblott, M.J., Gary, D.S., Belegu, V., Hurtado, A., Malone, M.L., McDonald, J.W., 2013. Chondroitin sulfate proteoglycans inhibit oligodendrocyte myelination through PTPsigma. Exp. Neurol. 247, 113–121. Pogribna, U., Yu, X., Burson, K., Zhou, Y., Lasky, R.E., Narayana, P.A., Parikh, N.A., 2013. Perinatal clinical antecedents of white matter microstructural abnormalities on diffusion tensor imaging in extremely preterm infants. PLoS ONE. 8 e72974. Rauch, U., 2007. Brain matrix: structure, turnover and necessity. Biochem. Soc. Trans. 35, 656–660. Redish, A.D., Touretzky, D.S., 1998. The role of the hippocampus in solving the Morris water maze. Neural Comput. 10, 73–111. Resch, B., Neubauer, K., Hofer, N., Resch, E., Maurer, U., Haas, J., Muller, W., 2012. Episodes of hypocarbia and early-onset sepsis are risk factors for cystic periventricular leukomalacia in the preterm infant. Early Hum. Dev. 88, 27–31. Rolls, A., Shechter, R., Schwartz, M., 2009. The bright side of the glial scar in CNS repair. Nat. Rev. Neurosci. 10, 235–241. Segovia, K.N., McClure, M., Moravec, M., Luo, N.L., Wan, Y., Gong, X., Riddle, A., Craig, A., Struve, J., Sherman, L.S., Back, S.A., 2008. Arrested oligodendrocyte lineage maturation in chronic perinatal white matter injury. Ann. Neurol. 63, 520–530. Sherman, L.S., Back, S.A., 2008. A ‘GAG’ reflex prevents repair of the damaged CNS. Trends Neurosci. 31, 44–52. Siebert, J.R., Osterhout, D.J., 2011. The inhibitory effects of chondroitin sulfate proteoglycans on oligodendrocytes. J. Neurochem. 119, 176–188. Siebert, J.R., Stelzner, D.J., Osterhout, D.J., 2011. Chondroitinase treatment following spinal contusion injury increases migration of oligodendrocyte progenitor cells. Exp. Neurol. 231, 19–29. Sofroniew, M.V., Vinters, H.V., 2010. Astrocytes: biology and pathology. Acta Neuropathol. 119, 7–35. Starkey, M.L., Bartus, K., Barritt, A.W., Bradbury, E.J., 2012. Chondroitinase ABC promotes compensatory sprouting of the intact corticospinal tract and recovery of forelimb function following unilateral pyramidotomy in adult mice. Eur. J. Neurosci. 36, 3665–3678. Suzuki, S., Saito, H., Yamagata, T., Anno, K., Seno, N., Kawai, Y., Furuhashi, T., 1968. Formation of three types of disulfated disaccharides from chondroitin sulfates by chondroitinase digestion. J. Biol. Chem. 243, 1543–1550. Tang, X., Davies, J.E., Davies, S.J., 2003. Changes in distribution, cell associations, and protein expression levels of NG2, neurocan, phosphacan, brevican, versican V2, and tenascin-C during acute to chronic maturation of spinal cord scar tissue. J. Neurosci. Res. 71, 427–444. Vannucci, R.C., Connor, J.R., Mauger, D.T., Palmer, C., Smith, M.B., Towfighi, J., Vannucci, S.J., 1999. Rat model of perinatal hypoxic–ischemic brain damage. J. Neurosci. Res. 55, 158–163. Vorhees, C.V., Williams, M.T., 2006. Morris water maze: procedures for assessing spatial and related forms of learning and memory. Nat. Protoc. 1, 848–858. Yamagata, T., Saito, H., Habuchi, O., Suzuki, S., 1968. Purification and properties of bacterial chondroitinases and chondrosulfatases. J. Biol. Chem. 243, 1523–1535. Yamaguchi, Y., 2000. Lecticans: organizers of the brain extracellular matrix. Cell. Mol. Life Sci. 57, 276–289. Zaidi, A.U., Bessert, D.A., Ong, J.E., Xu, H., Barks, J.D., Silverstein, F.S., Skoff, R.P., 2004. New oligodendrocytes are generated after neonatal hypoxic–ischemic brain injury in rodents. Glia. 46, 380–390.