- Email: [email protected]
hypoxic injury
Brain Research 1038 (2005) 41 – 49 www.elsevier.com/locate/brainres
Research report
Neural precursor cells division and migration in neonatal rat brain after ischemic/hypoxic injury Takeshi Hayashia,T, Masanori Iwaia, Tomoaki Ikedab, Guang Jina, Kentaro Deguchia, Shoko Nagotania, Hanzhe Zhanga, Yoshihide Seharaa, Isao Naganoa, Mikio Shojia, Tsuyomu Ikenoueb, Koji Abea a
Department of Neurology, Okayama University Graduate School of Medicine and Dentistry, 2-5-1 Shikata-cho, Okayama 700-8558, Japan b Department of Gynecology and Obstetrics, Miyazaki University School of Medicine, Miyazaki, Japan Accepted 17 December 2004
Abstract Ischemia/hypoxia (I/H) causes severe perinatal brain disorders such as cerebral palsy. The neonatal brain possesses much plasticity, and to enhance new cell production would be an innovative means of therapy for such disorders. In order to elucidate the dynamic changes of neural progenitor cells in the neonatal brain after ischemia, we investigated new cells production in the subventricular zone and subsequent migration of these cells to the injured area. Newly produced cells were confirmed by incorporation of bromodeoxyuridine (BrdU), and attempt for differentiation was investigated by immunohistochemistry for molecular markers of each cellular lineage. In the sham-control brain, there were many BrdU-labeled cells which gradually decreased as the animal becomes older. Many of these cells were oligodendroglial progenitor or microglial cells. Although there were only few neuronal cells labeled for BrdU in the sham-control, they dramatically increased after I/H. They were located at just beneath the subventricular zone where the progenitor cells reside and to the injured area, indicating that newly produced cells migrated to the infarct region and differentiated into neuronal precursor cells in order to compensate the lost neural cells. We found that BrdU-labeled astroglial, oligodendroglial progenitor, and microglial cells were also increased after I/H, suggesting that they also play active roles in recovery. Progenitor cells may have potential for treating perinatal brain disorders. D 2005 Elsevier B.V. All rights reserved. Theme: Disorders of the nervous system Topic: Developmental disorders Keywords: Hypoxia; Ischemia; Migration; Neonate; Progenitor cell; Rat
1. Introduction Hypoxic injury is the major cause of perinatal brain disorders such as cerebral palsy [18]. The incidence of cerebral palsy is 2.0 to 2.5 per 1000 live births in developed countries, and the total number of patients is still increasing Abbreviations: BrdU, bromodeoxyuridine; DCX, doublecortin; GFAP, glial fibrillary acidic protein; Iba1, ionized calcium-binding adapter molecule-1; I/H, ischemia/hypoxia; MCA, middle cerebral artery; PBS, phosphate buffered saline; SD, standard deviation; SVZ, subventricular zone T Corresponding author. Fax: +81 86 235 7368. E-mail address: [email protected] (T. Hayashi). 0006-8993/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.brainres.2004.12.048
over the world (reviewed by Reddihough and Collins [25]). In spite of marvelous progress in medicine, to treat cerebral palsy and other neurological deficits caused by hypoxic injury is quite difficult and still remains palliative [12]. To develop a novel means of therapy for these disorders is, therefore, an urgent big subject. Recent studies showed that the adult brain possesses ability to produce new neurons; there are stem or neural progenitor cells in the subventricular zone (SVZ) and in the subgranular zone of the dentate gyrus [3,9,22]. It was reported that division of these cells was enhanced by ischemic injuries, and newly produced cells became integrated in neural network and participated in recovery
42
T. Hayashi et al. / Brain Research 1038 (2005) 41–49
from neurological deficit [22]. Therapeutic use of these intrinsic cells for stroke attracts a great deal of attention, and is under intensive investigation now [10]. In the neonatal brain, on the other hand, less attention had been paid for therapeutic use of the intrinsic stem or neural progenitor cells. As hypoxia and ischemia underlay many perinatal brain diseases [13,18], neurogenesis might be exploited for treating these disorders as for stroke. Furthermore, the neonatal brain could possess larger ability to produce new cells than the adult brain does [29]. Just after the birth, in particular, neuronal loss caused by ischemia was completely recovered by new cell production [7]. If the newly produced cells migrate and be integrated in neural network, the role of neurogenesis should be bigger in the functional recovery from perinatal brain disease than from stroke. Not only production but also migration and differentiation of new cells are important for brain tissue recovery. In order to evaluate significance of progenitor cell activation, we need to confirm the fate of the newly produced cells. In the present study, therefore, we investigated cell division and migration in the neonatal brain after hypoxic injury. As glial cells which are vulnerable to hypoxic injury in the immature brain also play active roles in brain function [31], we investigated the attempt of the cells to differentiate into not only neuronal but also glial cells.
2. Materials and methods 2.1. Ischemic/hypoxic injury and bromodeoxyuridine labeling All procedures were approved by Animal Research Committee of Okayama University Graduate School of Medicine and Dentistry. Pregnant Wistar rats were purchased from Japan Charles River (Shizuoka, Japan), and day of birth was defined as P1 for pups. On P7, each pup was subjected to modified Levine procedure for producing ischemic/hypoxic brain injury (I/H) [23]. In brief, pups were anesthetized with dithylether and the left common artery was sectioned between double ligatures with 4-0 surgical silk. The pups were allowed to recover for 2 h, and then exposed to additional 2 h of hypoxia (8% oxygen and 92% nitrogen) at 33 8C. Sham control animals were treated identically except for both the carotid artery sectioning and hypoxia exposure. In order to identify the newly produced cells, we exploited bromodeoxyuridine (BrdU) labeling method. BrdU (Sigma, St. Louis, MO) dissolved in saline was intraperitoneally injected (50 mg/kg) at 50, 38, 26, 14, and 2 h before the decapitation [16]. With this method, we can label the cells which newly synthesized DNA in the last 2 days before decapitation. At 7, 14, and 21 days after the I/H (P14, P21, and P28, respectively) or sham-operation, the animals were sacrificed with the method described below (n = 4 or 5 for
each I/H group, and n = 3 for each sham-control group). The experimental paradigm is summarized in Fig. 1. 2.2. Tissue preparation and single immunohistochemical analysis Under deep anesthesia, the pups were transcardially perfused with ice-cold heparinized phosphate buffered saline (PBS) followed by 4% paraformaldehyde in phosphate buffer. The whole brain was subsequently removed and immersed in the same fixative for 12 h. After washing the paraformaldehyde out, the brains were cryoprotected and then rapidly frozen by 2-methylbutane chilled in liquid nitrogen. Coronal brain sections with 10 Am thickness at the caudate level was prepared using a cryostat, and mounted on a silane-coated slide glass. As described in our previous report [30], some brains showed severe histological damages although others showed only mild or no damages. Only the brains with more than 50% of cerebral cortical neurons degeneration were included in the present experiment. For single immunohistochemical analysis which uses peroxidase reaction, the slides were incubated in 0.3% hydrogen peroxide in methanol in order to quench endogenous peroxidase activity. For detection of BrdU, the slides were incubated in 1 mol/L HCl at 65 8C for 1 h to denature DNA and then rinsed in 0.1 mol/L boric acid (pH 8.5) at 25 8C for 10 min [16]. After blocking non-specific reaction by bovine serum albumin, the slides were incubated with the first antibody at 4 8C for 12 h. The first antibody used and each dilution were as follows: mouse monoclonal anti-BrdU antibody (Oncogene Research Products, Boston, MA) at 1:200, and goat polyclonal anti-doublecortin (DCX) antibody (Santa Cruz Biotechnology, Santa Cruz, CA) at 1:200. The slides were washed by PBS, and then incubated with biotinylated anti-mouse IgG antibody (Vector Laboratories, Burlingame, CA) or biotinylated anti-goat IgG antibody (Vector laboratories) at 1:200 for 1 h at room temperature. They were subsequently incubated with avidin–biotin–peroxidase complex (Vector Laboratories) for 30 min and then developed using diaminobenzidine as a peroxidase substrate. In each study, a set of sections were stained in a similar way without the primary antibody.
Fig. 1. Experimental paradigm of ischemic/hypoxic brain injury (I/H), BrdU injection, and decapitation. The pups underwent I/H or shamoperation on the postnatal day 7 (P7), and were sacrificed at 7, 14, or 21 days later (n = 4 or 5 for each I/H group, and n = 3 for each sham-control group). At 38, 26, 14, and 2 h before the decapitation, 50 mg/kg of BrdU was intraperitoneally injected.
T. Hayashi et al. / Brain Research 1038 (2005) 41–49
2.3. Double fluorescent immunohistochemistry In order to make it clear whether BrdU-labeled cells differentiate into neuronal, astroglial, or oligodendroglial cells, we performed double immunofluorescent studies for BrdU and DCX, glial fibrillary acidic protein (GFAP), and NG2; DCX, GFAP, and NG2 are the markers of differentiating neuronal, astroglial, and oligodendrocyte progenitor cells, respectively [8,11,21]. Microglial cells also play important roles in brain tissue repair, and therefore, we carried out double fluorescent study for BrdU and ionized calcium-binding adapter molecule-1 (Iba1) that is the wellknown marker of microglial cells [15,26]. The slides were first treated with HCl and boric acid as described previously in order to make BrdU easily detected [16]. The slides were then incubated with mouse monoclonal anti-BrdU antibody (Sigma) at 1:200 and antibody for each cellular marker, that is, goat polyclonal anti-DCX antibody (Santa Cruz Biotechnology) at 1:200, goat polyclonal anti-GFAP antibody (Santa Cruz Biotechnology) at 1:200, rabbit polyclonal anti-NG2 antibody (Chemicon, Temecula, CA) at 1:400, or rabbit polyclonal anti-Iba1 antibody (gift from Dr. Imai) at 1:250 [15]. After washing in PBS, the slides were incubated with FITC-labeled anti-mouse IgG antibody (Vector Laboratories) at 1:500 simultaneously with Alexa Fluor 546labeled anti-goat IgG antibody (Molecular Probe, Eugene, OR) at 1:500 or Rhodamine-labeled anti-rabbit IgG (Chemicon) at 1:500. The slides were then covered with VECTASHIELD mounting medium (Vector Laboratories). In order to confirm the specificity of the primary antibody, a set of sections were stained in a similar way without the anti-DCX, anti-GFAP, anti-NG2, or anti-Iba1 antibody, respectively. The sections were scanned with confocal microscope equipped with argon and HeNe1 laser (LSM-510, Zeiss, Jena, Germany). Sets of fluorescent images were acquired sequentially for the red and green channels to prevent crossover of signals from green to red or red to green channels. 2.4. Cell counting In order to evaluate the results of immunohistochemical analysis quantitatively, we counted the positively stained cells. For quantification of BrdU-labeled cells, we counted the positively stained cells in 25 pixels of 660X660 Am from 5 consecutive coronal sections. These sections were obtained from the striatum level, and were identical among the investigated animals. Cell counting for cerebral cortex and striatum was performed using same sections. The pixels included both the core and the boundary zone of the ischemic region, which were identical among all investigated animals. In the double fluorescent studies, we counted the double-positive cells in the same manner but the size of each pixel was 461 461 Am. Results were
43
expressed as mean F standard deviation (SD), and compared between the sham-control brains and those after I/H at each time point using Student’s t test, with P b 0.01 considered statistically significant.
3. Results 3.1. Histological outcome As our previous report [30], some animals showed only mild (defined as less than 50% of cerebral cortical degeneration) or no neuronal damages. We excluded these animals from the present study because the degree of brain damage would affect the response of SVZ progenitor cells. With these exclusion criteria, only one third of the animals which underwent I/H were included for this study. At each time point after I/H, the cerebral cortex, the caudate, and the hippocampus of the carotid arteryoccluded hemisphere showed severe tissue damage; neuronal as well as glial cells were degenerated, and the tissue became hypotrophic. SVZ was spared histologically, which was different from previous study by Levison et al [19]. Although the reason was uncertain, the temperature during hypoxia might have caused the difference of SVZ morphological change. No tissue changes were observed in the contralateral hemisphere as well as in the brains of the sham-control animals (data not shown). The results were in accordance with our previous report [23,30]. 3.2. Chronological change in BrdU-labeled cells Several cells were labeled for BrdU in the sham-control brain, both in the cerebral cortex (Figs. 2A, B, C) and in the striatum (Figs. 2J, K, L). The number of labeled cells gradually decreased from 7 days (P14) to 21 days (P28), indicating that stem cell proliferation becomes less active as the animal becomes older. In the animal of 7 days after I/H, a large number of cells were labeled for BrdU (Fig. 2D), both in the cerebral cortex (Fig. 2D) and in the striatum (Fig. 2M). At 14 days, the number of positively labeled cells became smaller both in the cerebral cortex (Fig. 2E) and in the striatum (Fig. 2N), but were still larger than those of the sham-control animals at the same time point. The number of labeled cells further decreased at 21 days (Figs. 2F, O), but were again larger than those of the sham-control animals at the same time point. In the contralateral hemisphere of which the carotid artery was not occluded, several cells were labeled for BrdU with gradual decrease from 7 to 21 days both in the cerebral cortex (Figs. 2G, H, I) and in the striatum (Figs. 2P, Q, R). The results of the contralateral hemisphere were comparable with those of the sham-control animals. With quantitative analysis (Fig. 3), it was revealed that 87.5 F 25.5 cells per 1 mm2 were labeled for BrdU in the
44
T. Hayashi et al. / Brain Research 1038 (2005) 41–49
Fig. 2. BrDU labeling in the brain with sham-operation or ischemic/hypoxic injury (I/H). Panels A to I show the cerebral cortex, and J to R show the striatum. Several cells were labeled in the sham-control groups, which was gradually decreased from 7 days to 21 days (A, B, C). In the ischemic side cortex after I/H, a large number of cells became positively labeled at 7 days (D). The numbers of labeled cells were decreased thereafter, but they still remained significantly larger than the sham-control groups (E, F). In the contralateral side, the numbers of labeled cells were comparable with those of the sham-group (G, H, I). In the striatum also, a number of cells were labeled at 7 days after sham-operation and were dramatically decreased thereafter (J, K, L). In the ischemic side striatum after I/H, many cells became labeled (M). Although they gradually decreased, they remained significantly larger in number throughout the investigated time course (N, O). In the contralateral side, the numbers of positively stained cells were comparable with those of the sham-groups (P, Q, R). Scale bar: 200 Am.
sham-control cerebral cortex at 7 days. They were decreased to 25.7 F 9.6 at 14 days and 7.3 F 5.3 at 21 days. In the ischemic cerebral cortex after I/H, the number of labeled cells were 456.4 F 71.6, 258.0 F 130.4, and 105.8 F 65.7
at 7, 14, and 21 days, respectively, that was significantly larger than that of the sham-control at each time point ( P b 0.01). In the contralateral cerebral cortex, 105.6 F 18.4, 37.2 F 23.6, and 11.9 F 4.8 cells per one mm2 were labeled
Fig. 3. Changes in number of BrdU-labeled cells. In the cortex and the striatum with sham-operation, the numbers of labeled cells were largest at 7 days and were decreased thereafter. In the ischemic side cortex with ischemia–hypoxia (I–H), about 5 times number of cells were labeled at 7 day. The labeled cells gradually decreased, but were far larger in number than those in the sham-group cortex at each time point. The numbers of labeled cells in the contralateral cortex were comparable with those in the sham-group. In the sham-control striatum, cells were labeled only at 7 days. About 6 times numbers of cells were labeled in the ischemic side striatum at 7 days after I/H. The labeled cells gradually decreased thereafter, but significantly larger in number throughout the investigated time course (TP b 0.01 compared to the sham-group of the same time point).
T. Hayashi et al. / Brain Research 1038 (2005) 41–49
at 7, 14, and 21 days, which was similar with the results of the sham-control. In the sham-control striatum, the numbers of positively stained cells were 127.9 F 28.2, 2.3 F 1.6, and 0.9 F 2.3 at 7, 14, and 21 days. It became 679.1 F 84.9, 183.7 F 102.6, and 118.0 F 4.8 in the ischemic striatum, all of which were significantly larger than those of the shamcontrol ( P b 0.01). It was 147.6 F 80.3, 5.1 F 3.7, and 3.0 F 4.1 in the contralateral striatum, which showed no statistical significance against the sham-control striatum of the same time points. 3.3. DCX expression of newly produced cells We confirmed that progenitor cell division was enhanced in the neonatal brain after I/H, but it remained uncertain whether these newly produced cells would participate in tissue repair. In order for the newly produced cells to be involved in tissue repair process, they need to differentiate into neuronal cells. With use of the brain sections same with those used for BrdU immunohistochemistry, we performed immunohistochemical analysis for DCX as well as double fluorescent study for DCX and BrdU.
45
Representative photomicrographs of the immunohistochemistry using brains 7 days after I/H are shown in Fig. 4. DCX-positive cells were extensively distributed to the cerebral cortex (Fig. 4B) and the striatum of the ischemic side (Figs. 4E, H, J). In particular, there were lots of DCXpositive cells around the peri-infarct area (Fig. 4B). Double fluorescent study revealed that many but not all BrdUlabeled cells at that region expressed DCX (Fig. 4C). Tomographic analysis with confocal microscopy further showed that signals for BrdU and DCX actually overlapped in some cells (Fig. 4D). In the striatum close to the lateral ventricle also, there were lots of DCX-positive cells (Fig. 4E). There were many cells which were stained for BrdU and DCX, both in the region just beneath the white matter (Fig. 4F) and a bit distant from the lateral ventricle wall (Fig. 4G). DCX-expressing cells were distributed also to the striatum close to the ventral pallidum (Fig. 4H), and again, there were many cells which were labeled for DCX and BrdU (Fig. 4I). To our surprise, there were DCX-expressing cells in the cerebral cortex of the ischemic core region (Fig. 4J). Although the number was substantially smaller, DCXand BrdU-positive cells were confirmed in this region as
Fig. 4. Immunohistochemical analysis for DCX (red) and BrdU (green) in the brain of 7 days after ischemia/hypoxia. Many cells in the cerebral cortex at the MCA boundary area expressed DCX (B), some of which were labeled also for BrdU (C). Tomographic analysis with higher magnification, in which smaller panels demonstrate the deep axis of the line-indicated section, further showed some cells stained for both DCX and BrdU (D). In the striatum close to the lateral ventricle (schematically indicated in panel A), many cells expressed DCX (E) and were labeled also for BrdU (F). The double-stained cells were distributed to the area distant from the lateral ventricle (G). DCX-expressing cells were also confirmed at the striatum close to the ventral pallidum (H), again some of which were labeled for BrdU (I). In the cerebral cortex of the ischemic core, smaller numbers of cells were stained for DCX (J), though some cells were labeled for BrdU (K). Scale bar: B, E, H, J (in J), 100 Am; C, I, K (in K), 100 Am; D, F (in F), 10 Am; G, 20 Am.
46
T. Hayashi et al. / Brain Research 1038 (2005) 41–49
well (Fig. 4K). From these results, we can suppose that some of the newly produced cells migrate to the injured area and attempted to differentiate into the neuronal cells. 3.4. Glial differentiation of newly produced cells With double staining study for DCX and BrdU, we found that substantial number of newly produced cells did not express DCX, indicating that some cells differentiated into glial cells or remained undifferentiated. In order to investigate glial differentiation of newly produced cells, we carried out double florescent study for BrdU and GFAP or NG2. In addition, we investigated the possible presence of BrdU-labeled microglial cells.
Representative photomicrographs of the immunohistochemistry using brains 7 days after I/H are shown in Fig. 5. In the cerebral cortex (Fig. 5A) and the striatum (Fig. 5B), there were many BrdU-labeled cells and GFAP expressing cells at the peri-infarct area, but a small number of cells were positive for both signals. Double staining for BdU and NG2 revealed that a few cells at the cerebral cortex (Fig. 5C) and the striatum (Fig. 5D) of the peri-infarct area expressed both signals. Double staining for BrdU and Iba1 showed that a substantial number of cells were stained for BrdU or Iba1. In the cerebral cortex of the peri-infarct area (Fig. 5E), there were large numbers of cells which expressed both signals. In the striatum of the peri-infarct area (Fig. 5F), however, the number of
Fig. 5. Immunohistochemical analysis for GFAP, NG2, Iba1 (red), and BrdU (green) in the MCA boundary area of 7 days after ischemia/hypoxia. A few BrdUlabeled cells expressed GFAP both in the cerebral cortex (A) and the striatum (B). There also were cells which was positive for BrdU and NG2 in the cerebral cortex (C) and the striatum (D), and larger numbers of cells were positive for BrdU and Iba1 (E, F). Bar: larger photomicrographs (in larger one of F) 100 Am; smaller windows (in smaller one of F) 10 Am.
T. Hayashi et al. / Brain Research 1038 (2005) 41–49
47
Table 1 Quantitative analysis of BrdU-labeled cells and their phenotypes of cortex after hypoxic/ischemic brain injury in neonates +
(a) Total BrDU cells (b) BrDU+ DCX+ cells % of (b)/(a) (c) BrDU+ GFAP+ cells % of (c)/(a) (d) BrDU+ NG2+ cells % of (d)/(a) (e) BrDu+ + Iba1 cells % of (e)/(a)
Sham 7DAH/I
7DAH/I
Ipsi-H/I 14DAH/I
21DAH/I
Contra-H/I 7DAH/I
87.5 0.5 0.6 2.4 2.7 29.1 33.3 19.6 22.4
456.4 108.4 23.8 10.5 2.3 38.5 8.4 149.3 32.7
258.0 49.2 19.1 5.8 2.2 7.9 3.1 51.8 20.1
105.8 27.7 26.2 3.1 2.9 3.7 3.5 40.3 38.1
105.6 0.6 0.6 2.4 2.3 19.6 18.6 17.3 16.4
F 25.5 F 0.5 F 1.1 F 3.7 F 4.1
F 71.6 F 12.7 F 3.1 F 7.4 F 21.6
double-positive cells were smaller than that of the cerebral cortex. 3.5. Quantification of phenotypes of BrdU-labeled cells In order to estimate the ratio of each cellular differentiation, we counted the number of double-positive cells in each brain. The results of the cerebral cortex are summarized in Table 1, and those of the striatum are in Table 2. In the sham-control cerebral cortex in which smaller numbers of BrdU-labeled cells were confirmed, there were only few cells which expressed BrdU and DCX. BrdU- and GFAP-positive cells were also small in number. On the contrary, many BrdU-labeled cells expressed NG2 or Iba1, indicating that oligodendroglial and microglial lineage cells were actively produced in normal condition at this age. After I/H in which total number of BrdU-labeled cells were increased, DCXpositive cells were prominently increased though NG2positive cells were relatively decreased. The number of BrdU- and Iba1-positive cells was increased after I/H, but was gradually decreased later. The results of the contralateral cerebral cortex were similar with those of the sham-control. In the striatum, only small numbers of BrdU-labeled cells expressed DCX or GFAP in normal condition, as seen in the cerebral cortex. After I/H, the number of DCX-positive cells were pronouncedly increased, though those of NG2-positive cells were relatively decreased. The results of the contralateral striatum were similar with those of the sham-control.
F 130.4 F 8.2 F 2.5 F 2.6 F 15.4
F 65.7 F 8.1 F 0.8 F 1.0 F 10.2
F 18.4 F 0.6 F 1.6 F 4.5 F 4.7
4. Discussion In the present study, we found that the cell proliferation in the SVZ was enhanced by I/H. Scheepens et al [27] showed that the SVZ cells showed no response against ischemia, but our and other group’s study indicated that SVZ are sensitive to ischemic injury in neonates [24]. As shown in Figs. 2 and 3, the new cell production gradually ceases after the birth in normal condition. This may reflect that the development of brain approaches completion as the animal becomes older [2]. Most of the BrdU-labeled cells in the sham-control brain were not stained for DCX (Tables 1 and 2). We could speculate that neurogenesis was not so active in this age; even at 50 h after the BrdU injection, many cells should have been stained for BrdU and DCX if neurogenesis had been as active as in the brains after I/H (Tables 1 and 2). On the other hand, many BrdU-labeled cells in the sham-control brains were stained for NG2 (Tables 1 and 2), which may indicate that myeline formation was still under an active stage in neonates [31]. In human being also, myelination does not complete in infancy, which is compatible with the findings obtained by the present study [5]. The next abundant cells labeled for BrdU in shamcontrol brains were microglial cells, which was stained for Iba1 (Tables 1 and 2). Because the microglial cells are mesodermal origin [14], the results presented here does not necessarily mean that the stem or progenitor cells at the SVZ differentiated into microglia; rather, it is more feasible to consider that BrdU was incorporated into microglial cells intra- or extra-medullarily. As BrdU-positive microglial
Table 2 Quantitative analysis of BrDU-labeled cells and their phenotypes of striatum after hypoxic/ischemic brain injury in neonates (a) Total BrDU+ cells (b) BrDU+ DCX+ cells % of (b)/(a) (c) BrDU+ GFAP+ cells % of (c)/(a) (d) BrDU+ NG2+ cells % of (d)/(a) (e) BrDu+ + Iba1 cells % of (e)/(a)
Sham 7DAH/I
7DAH/I
Ipsi-H/I 14DAH/I
21DAH/I
Contra-H/I 7DAH/I
129.7 1.0 0.6 0.8 0.6 18.1 14.0 8.6 6.4
679.1 1885.5 27.8 74.2 10.9 34.6 5.1 62.1 9.1
183.7 56.6 30.8 5.2 2.8 0.8 0.4 8.9 4.8
118.0 F 28.8 F 24.4 3.7 F 3.1 2.1 F 1.8 8.4 F 7.1
147.6 1.2 0.8 1.6 1.1 16.5 11.2 17.3 11.7
F 28.2 F 0.7 F 0.8 F 1.9 F 3.3
F 84.9 F 17.0 F 16.9 F 5.5 F 21.2
F 102.6 F 18.1 F 2.7 F 0.8 F 1.7
102.6 13.4 1.3 0.8 3.1
F 80.3 F 0.8 F 1.0 F 4.8 F 2.6
48
T. Hayashi et al. / Brain Research 1038 (2005) 41–49
cells were pronouncedly increased at 7 days (Fig. 5, Tables 1 and 2), these cells may have been implicated in tissue repair. Indeed, it is known that microglia enhances neural stem cell proliferation and subsequent migration to the injured area by excreting diffusible factors [1]. This cell should thus be a quite important supporting actor for neurological recovery. Compared to the number of BrdU-labeled NG2- or Iba1expressing cells, there were only small numbers of BrdUlabeled GFAP-expressing cells in the sham-control brain (Tables 1 and 2), indicating that supplying new astrocytes are not necessary in quiescent state. Once the I/H was imposed, however, marked astrocytosis occurred in the injured brain (Fig. 5). The origin of BrdU-labeled GFAP-expressing cells cannot exactly be made clear only by the present study; they might have arisen from stem cells at the SVZ, but de novo mitosis of astrocytes should also be considered [20]. Indeed, activation and division of astrocytes are the histological features of the subacute cerebral infarction [6]. As the increase and subsequent decrease of BrdU-labeled GFAPexpressing cells coincided with those of total BrdU-positive cells, however, we believe that at least some of the BrdUlabeled GFAP-expressing cells came from the SVZ. The geographical pattern which suggested migration of BrdUlabeled cells from the SVZ to the injured area (Fig. 4) also supports this speculation. We noticed that the degree of new astrocytes formation was different between the cerebral cortex and striatum (Fig. 5, Tables 1 and 2). This may be due to the intrinsic nature of each anatomical structure, but it is possible to consider that the striatum suffered more severe damage and massive astrocytosis occurred in the striatum [23]. Moreover, it should also be possible to speculate that the striatum was more close to the SVZ and many newly produced astroglial cells reached the injured area. Although there were only few cells which showed positive for BrdU and DCX in the sham-control brain, they dramatically increased in the cerebral cortex and the striatum after I/H (Fig. 4, Tables 1 and 2). As DCX-positive cells distributed from the SVZ to the injured cerebral cortex and the striatum, this increase should be an attempt to repair the damaged tissue by supplementing new neurons. To our surprise, we confirmed BrdU-labeled DCX-expressing cells also at the ischemic core region. Previous report using neonatal mice brain ischemia model also demonstrated DCX expressing cells in the ischemic striatum [24]. As tissue would become pannecrosis at this region [23], most of the DCX-expressing cells should undergo apoptosis in this ischemic core region. Previous reports also demonstrated that most of the newly produced cells would be eliminated before they would be integrated in the neural network [4,28]. In the peri-infarct area, on the other hand, ischemic damage would not result tissue pannecrosis and these newly produced cells might be implicated in repair process. In the adult rat brain after ischemic injury, it was estimated that at most 0.2% of damaged neurons were replaced by newly produced cells [4]. In the neonatal I/H model we used here, we cannot precisely calculate how many neurons were
supplemented by newly produced cells. However, the number of supplemented cells should not be a negligible one because there were so many DCX-positive neuronal precursor cells migrating to the injured area; we found that so many as 456 F 71.6 (cerebral cortex) and 679 F 84.9 (striatum) cells were labeled for BrdU per 1 mm2. Although simple comparison is difficult, neural supplementation should be more active than adult brains after ischemia; up to 400 cells were labeled for BrdU per one section in adult rat brain after focal ischemia [17]. In order to conclude whether these cells actually differentiate into mature neurons, further study would be required. Hypoxic damage causes severe perinatal brain diseases such as cerebral palsy [18]. To treat these disorders is a quite big challenge, but still remains palliative now [12]. As presented here, neonatal brain could possess plasticity and produces new cells. It might be possible to alleviate the neurological disability if we could exploit these characters of the neonatal brains.
Acknowledgments This work was partly supported by Grant-in Aid for Scientific Research (B) 15390273, (Hoga) 15659338, and National Project on Protein Structural and Functional Analyses from the Ministry of Education, Science, Culture and Sports of Japan, and grants (Itoyama Y, Kimura I and Kuzuhara S) from the Ministry of Health and Welfare of Japan.
References [1] J. Aarum, K. Sandberg, S.L. Haeberlein, M.A. Persson, Migration and differentiation of neural precursor cells can be directed by microglia, Proc. Natl. Acad. Sci. U.S.A. 100 (2003) 15983 – 15988. [2] J. Altman, G.D. Das, Autoradiographic and histological studies of postnatal neurogenesis: I. A longitudinal investigation of the kinetics, migration and transformation of cells incorporating tritiated thymidine in neonate rats, with special reference to postnatal neurogenesis in some brain regions, J. Comp. Neurol. 126 (1966) 337 – 389. [3] A. Alvarez-Buylla, J.M. Garcia-Verdugo, Neurogenesis in adult subventricular zone, J. Neurosci. 22 (2002) 629 – 634. [4] A. Arvidsson, T. Collin, D. Kirik, Z. Kokaia, O. Lindvall, Neuronal replacement from endogenous precursors in the adult brain after stroke, Nat. Med. 8 (2002) 963 – 970. [5] S.A. Back, N.L. Luo, N.S. Borenstein, J.M. Levine, J.J. Volpe, H.C. Kinney, Late oligodendrocyte progenitors coincide with the developmental window of vulnerability for human perinatal white matter injury, J. Neurosci. 21 (2001) 1302 – 1312. [6] A. da Cunha, J.J. Jefferson, W.R. Tyor, J.D. Glass, F.S. Jannotta, L. Vitkovic, Gliosis in human brain: relationship to size but not other properties of astrocytes, Brain Res. 600 (1993) 161 – 165. [7] J.L. Daval, G. Pourie, S. Grojean, V. Lievre, C. Strazielle, S. Blaise, P. Vert, Neonatal hypoxia triggers transient apoptosis followed by neurogenesis in the rat CA1 hippocampus, Pediatr. Res. 55 (2004) 561 – 567. [8] M.R. Dawson, A. Polito, J.M. Levine, R. Reynolds, NG2-expressing glial progenitor cells: an abundant and widespread population of
T. Hayashi et al. / Brain Research 1038 (2005) 41–49
[9]
[10] [11]
[12] [13]
[14]
[15]
[16]
[17]
[18]
[19]
cycling cells in the adult rat CNS, Mol. Cell. Neurosci. 24 (2003) 476 – 488. F. Doetsch, J.M. Garcia-Verdugo, A. Alvarez-Buylla, Cellular composition and three-dimensional organization of the subventricular germinal zone in the adult mammalian brain, J. Neurosci. 17 (1997) 5046 – 5061. M. Fisher, Stem cell transplantation for stroke: does it work, and if so, how? Stroke 34 (2003) 2083. F. Francis, A. Koulakoff, D. Boucher, P. Chafey, B. Schaar, M.C. Vinet, G. Friocourt, N. McDonell, O. Reiner, A. Kahn, S.K. McCornell, Y. Berwald-Netter, P. Denoulet, J. Chelly, Doublecortin is a developmentally regulated, microtubule-associated protein expressed in migrating and differentiating neurons, Neuron 23 (1999) 247 – 256. M. Goldstein, The treatment of cerebral palsy: what we know, what we don’t know, J. Pediatr. 145 (2 Suppl.) (2004) S42 – S46. J. Grow, J.D. Barks, Pathogenesis of hypoxic–ischemic cerebral injury in the term infant: current concepts, Clin. Perinatol. 29 (2002) 585 – 602. G.J. Guillemin, B.J. Brew, Microglia, macrophages, perivascular macrophages, and pericytes: a review of function and identification, J. Leukocyte Biol. 75 (2004) 388 – 397. Y. Imai, I. Ibata, D. Ito, K. Ohsawa, S. Kohsaka, A novel gene iba1 in the major histocompatibility complex class III region encoding an EF hand protein expressed in a monocytic lineage, Biochem. Biophys. Res. Commun. 224 (1996) 855 – 862. M. Iwai, K. Sato, N. Omori, I. Nagano, Y. Manabe, M. Shoji, K. Abe, Three steps of neural stem cells development in gerbil dentate gyrus after transient ischemia, J. Cereb. Blood Flow Metab. 22 (2002) 411 – 419. K. Jin, M. Minami, J.Q. Lan, X.O. Mao, S. Batteur, R.P. Simon, D.A. Greenberg, Neurogenesis in dentate subgranular zone and rostral subventricular zone after focal cerebral ischemia in the rat, Proc. Natl. Acad. Sci. U.S.A. 98 (2001) 4710 – 4715. M.V. Johnston, W.H. Trescher, A. Ishida, W. Nakajima, Neurobiology of hypoxic–ischemic injury in the developing brain, Pediatr. Res. 49 (2001) 735 – 741. S.W. Levison, R.P. Rothstein, M.J. Romanko, M.J. Snyder, R.L. Meyers, S.J. Vannucci, Hypoxia/ischemia depletes the rat perinatal
[20]
[21] [22]
[23]
[24]
[25] [26]
[27]
[28] [29] [30]
[31]
49
subventricular zone of oligodendrocyte progenitors and neural stem cells, Dev. Neurosci. 23 (2001) 234 – 247. B. Lin, M.D. Ginsberg, R. Busto, W.D. Dietrich, Sequential analysis of subacute and chronic neuronal, astrocytic and microglial alterations after transient global ischemia in rats, Acta Neuropathol. (Berlin) 95 (1998) 511 – 523. A. Messing, M. Brenner, GFAP: functional implications gleaned from studies of genetically engineered mice, Glia 43 (2003) 87 – 90. H. Nakatomi, T. Kuriu, S. Okabe, S. Yamamoto, O. Hatano, N. Kawahara, A. Tamura, T. Kirino, M. Nakafuku, Regeneration of hippocampal pyramidal neurons after ischemic brain injury by recruitment of endogenous neural progenitors, Cell 110 (2002) 429 – 441. A. Ota, T. Ikeda, X.Y. Xia, Y.X. Xia, T. Ikenoue, Hypoxic– ischemic tolerance induced by hyperthermic pretreatment in newborn rats, J. Soc. Gynecol. Invest. 7 (2000) 102 – 105. J.M. Plane, R. Liu, T.W. Wang, F.S. Silverstein, J.M. Parent, Neonatal hypoxic–ischemic injury increases forebrain subventricular zone neurogenesis in the mouse, Neurobiol. Dis. 16 (2004) 585 – 595. D.S. Reddihough, K.J. Collins, The epidemiology and causes of cerebral palsy, Aust. J. Physiother. 49 (2003) 7 – 12. Y. Sasaki, K. Ohsawa, H. Kanazawa, S. Kohsaka, Y. Imai, Iba1 is an actin-cross-linking protein in macrophages/microglia, Biochem. Biophys. Res. Commun. 286 (2001) 292 – 297. A. Scheepens, G. Wassink, M.J. Piersma, W.D. Van de Berg, C.E. Blanco, A delayed increase in hippocampal proliferation following global asphyxia in the neonatal rat, Brain Res., Dev. Brain Res. 142 (2003) 67 – 76. E.Y. Snyder, K.I. Park, Limitations in the brain repair, Nat. Med. 8 (2002) 928 – 930. F.M. Vaccarino, L.R. Ment, Injury and repair in developing brain, Arch. Dis. Child., Fetal Neonatal. Ed. 89 (2004) F190 – F192. Y.X. Xia, H. Sameshima, T. Ikeda, T. Higo, T. Ikenoue, Cerebral blood flow distribution and hypoxic–ischemic brain damage in newborn rats, J. Obstet. Gynaecol. Res. 28 (2002) 320 – 326. A.U. Zaidi, D.A. Bessert, J.E. Ong, H. Xu, J.D. Barks, F.S. Silverstein, R.P. Skoff, New oligodendrocytes are generated after neonatal hypoxic–ischemic brain injury in rodents, Glia 46 (2004) 380 – 390.
Research report
Neural precursor cells division and migration in neonatal rat brain after ischemic/hypoxic injury Takeshi Hayashia,T, Masanori Iwaia, Tomoaki Ikedab, Guang Jina, Kentaro Deguchia, Shoko Nagotania, Hanzhe Zhanga, Yoshihide Seharaa, Isao Naganoa, Mikio Shojia, Tsuyomu Ikenoueb, Koji Abea a
Department of Neurology, Okayama University Graduate School of Medicine and Dentistry, 2-5-1 Shikata-cho, Okayama 700-8558, Japan b Department of Gynecology and Obstetrics, Miyazaki University School of Medicine, Miyazaki, Japan Accepted 17 December 2004
Abstract Ischemia/hypoxia (I/H) causes severe perinatal brain disorders such as cerebral palsy. The neonatal brain possesses much plasticity, and to enhance new cell production would be an innovative means of therapy for such disorders. In order to elucidate the dynamic changes of neural progenitor cells in the neonatal brain after ischemia, we investigated new cells production in the subventricular zone and subsequent migration of these cells to the injured area. Newly produced cells were confirmed by incorporation of bromodeoxyuridine (BrdU), and attempt for differentiation was investigated by immunohistochemistry for molecular markers of each cellular lineage. In the sham-control brain, there were many BrdU-labeled cells which gradually decreased as the animal becomes older. Many of these cells were oligodendroglial progenitor or microglial cells. Although there were only few neuronal cells labeled for BrdU in the sham-control, they dramatically increased after I/H. They were located at just beneath the subventricular zone where the progenitor cells reside and to the injured area, indicating that newly produced cells migrated to the infarct region and differentiated into neuronal precursor cells in order to compensate the lost neural cells. We found that BrdU-labeled astroglial, oligodendroglial progenitor, and microglial cells were also increased after I/H, suggesting that they also play active roles in recovery. Progenitor cells may have potential for treating perinatal brain disorders. D 2005 Elsevier B.V. All rights reserved. Theme: Disorders of the nervous system Topic: Developmental disorders Keywords: Hypoxia; Ischemia; Migration; Neonate; Progenitor cell; Rat
1. Introduction Hypoxic injury is the major cause of perinatal brain disorders such as cerebral palsy [18]. The incidence of cerebral palsy is 2.0 to 2.5 per 1000 live births in developed countries, and the total number of patients is still increasing Abbreviations: BrdU, bromodeoxyuridine; DCX, doublecortin; GFAP, glial fibrillary acidic protein; Iba1, ionized calcium-binding adapter molecule-1; I/H, ischemia/hypoxia; MCA, middle cerebral artery; PBS, phosphate buffered saline; SD, standard deviation; SVZ, subventricular zone T Corresponding author. Fax: +81 86 235 7368. E-mail address: [email protected] (T. Hayashi). 0006-8993/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.brainres.2004.12.048
over the world (reviewed by Reddihough and Collins [25]). In spite of marvelous progress in medicine, to treat cerebral palsy and other neurological deficits caused by hypoxic injury is quite difficult and still remains palliative [12]. To develop a novel means of therapy for these disorders is, therefore, an urgent big subject. Recent studies showed that the adult brain possesses ability to produce new neurons; there are stem or neural progenitor cells in the subventricular zone (SVZ) and in the subgranular zone of the dentate gyrus [3,9,22]. It was reported that division of these cells was enhanced by ischemic injuries, and newly produced cells became integrated in neural network and participated in recovery
42
T. Hayashi et al. / Brain Research 1038 (2005) 41–49
from neurological deficit [22]. Therapeutic use of these intrinsic cells for stroke attracts a great deal of attention, and is under intensive investigation now [10]. In the neonatal brain, on the other hand, less attention had been paid for therapeutic use of the intrinsic stem or neural progenitor cells. As hypoxia and ischemia underlay many perinatal brain diseases [13,18], neurogenesis might be exploited for treating these disorders as for stroke. Furthermore, the neonatal brain could possess larger ability to produce new cells than the adult brain does [29]. Just after the birth, in particular, neuronal loss caused by ischemia was completely recovered by new cell production [7]. If the newly produced cells migrate and be integrated in neural network, the role of neurogenesis should be bigger in the functional recovery from perinatal brain disease than from stroke. Not only production but also migration and differentiation of new cells are important for brain tissue recovery. In order to evaluate significance of progenitor cell activation, we need to confirm the fate of the newly produced cells. In the present study, therefore, we investigated cell division and migration in the neonatal brain after hypoxic injury. As glial cells which are vulnerable to hypoxic injury in the immature brain also play active roles in brain function [31], we investigated the attempt of the cells to differentiate into not only neuronal but also glial cells.
2. Materials and methods 2.1. Ischemic/hypoxic injury and bromodeoxyuridine labeling All procedures were approved by Animal Research Committee of Okayama University Graduate School of Medicine and Dentistry. Pregnant Wistar rats were purchased from Japan Charles River (Shizuoka, Japan), and day of birth was defined as P1 for pups. On P7, each pup was subjected to modified Levine procedure for producing ischemic/hypoxic brain injury (I/H) [23]. In brief, pups were anesthetized with dithylether and the left common artery was sectioned between double ligatures with 4-0 surgical silk. The pups were allowed to recover for 2 h, and then exposed to additional 2 h of hypoxia (8% oxygen and 92% nitrogen) at 33 8C. Sham control animals were treated identically except for both the carotid artery sectioning and hypoxia exposure. In order to identify the newly produced cells, we exploited bromodeoxyuridine (BrdU) labeling method. BrdU (Sigma, St. Louis, MO) dissolved in saline was intraperitoneally injected (50 mg/kg) at 50, 38, 26, 14, and 2 h before the decapitation [16]. With this method, we can label the cells which newly synthesized DNA in the last 2 days before decapitation. At 7, 14, and 21 days after the I/H (P14, P21, and P28, respectively) or sham-operation, the animals were sacrificed with the method described below (n = 4 or 5 for
each I/H group, and n = 3 for each sham-control group). The experimental paradigm is summarized in Fig. 1. 2.2. Tissue preparation and single immunohistochemical analysis Under deep anesthesia, the pups were transcardially perfused with ice-cold heparinized phosphate buffered saline (PBS) followed by 4% paraformaldehyde in phosphate buffer. The whole brain was subsequently removed and immersed in the same fixative for 12 h. After washing the paraformaldehyde out, the brains were cryoprotected and then rapidly frozen by 2-methylbutane chilled in liquid nitrogen. Coronal brain sections with 10 Am thickness at the caudate level was prepared using a cryostat, and mounted on a silane-coated slide glass. As described in our previous report [30], some brains showed severe histological damages although others showed only mild or no damages. Only the brains with more than 50% of cerebral cortical neurons degeneration were included in the present experiment. For single immunohistochemical analysis which uses peroxidase reaction, the slides were incubated in 0.3% hydrogen peroxide in methanol in order to quench endogenous peroxidase activity. For detection of BrdU, the slides were incubated in 1 mol/L HCl at 65 8C for 1 h to denature DNA and then rinsed in 0.1 mol/L boric acid (pH 8.5) at 25 8C for 10 min [16]. After blocking non-specific reaction by bovine serum albumin, the slides were incubated with the first antibody at 4 8C for 12 h. The first antibody used and each dilution were as follows: mouse monoclonal anti-BrdU antibody (Oncogene Research Products, Boston, MA) at 1:200, and goat polyclonal anti-doublecortin (DCX) antibody (Santa Cruz Biotechnology, Santa Cruz, CA) at 1:200. The slides were washed by PBS, and then incubated with biotinylated anti-mouse IgG antibody (Vector Laboratories, Burlingame, CA) or biotinylated anti-goat IgG antibody (Vector laboratories) at 1:200 for 1 h at room temperature. They were subsequently incubated with avidin–biotin–peroxidase complex (Vector Laboratories) for 30 min and then developed using diaminobenzidine as a peroxidase substrate. In each study, a set of sections were stained in a similar way without the primary antibody.
Fig. 1. Experimental paradigm of ischemic/hypoxic brain injury (I/H), BrdU injection, and decapitation. The pups underwent I/H or shamoperation on the postnatal day 7 (P7), and were sacrificed at 7, 14, or 21 days later (n = 4 or 5 for each I/H group, and n = 3 for each sham-control group). At 38, 26, 14, and 2 h before the decapitation, 50 mg/kg of BrdU was intraperitoneally injected.
T. Hayashi et al. / Brain Research 1038 (2005) 41–49
2.3. Double fluorescent immunohistochemistry In order to make it clear whether BrdU-labeled cells differentiate into neuronal, astroglial, or oligodendroglial cells, we performed double immunofluorescent studies for BrdU and DCX, glial fibrillary acidic protein (GFAP), and NG2; DCX, GFAP, and NG2 are the markers of differentiating neuronal, astroglial, and oligodendrocyte progenitor cells, respectively [8,11,21]. Microglial cells also play important roles in brain tissue repair, and therefore, we carried out double fluorescent study for BrdU and ionized calcium-binding adapter molecule-1 (Iba1) that is the wellknown marker of microglial cells [15,26]. The slides were first treated with HCl and boric acid as described previously in order to make BrdU easily detected [16]. The slides were then incubated with mouse monoclonal anti-BrdU antibody (Sigma) at 1:200 and antibody for each cellular marker, that is, goat polyclonal anti-DCX antibody (Santa Cruz Biotechnology) at 1:200, goat polyclonal anti-GFAP antibody (Santa Cruz Biotechnology) at 1:200, rabbit polyclonal anti-NG2 antibody (Chemicon, Temecula, CA) at 1:400, or rabbit polyclonal anti-Iba1 antibody (gift from Dr. Imai) at 1:250 [15]. After washing in PBS, the slides were incubated with FITC-labeled anti-mouse IgG antibody (Vector Laboratories) at 1:500 simultaneously with Alexa Fluor 546labeled anti-goat IgG antibody (Molecular Probe, Eugene, OR) at 1:500 or Rhodamine-labeled anti-rabbit IgG (Chemicon) at 1:500. The slides were then covered with VECTASHIELD mounting medium (Vector Laboratories). In order to confirm the specificity of the primary antibody, a set of sections were stained in a similar way without the anti-DCX, anti-GFAP, anti-NG2, or anti-Iba1 antibody, respectively. The sections were scanned with confocal microscope equipped with argon and HeNe1 laser (LSM-510, Zeiss, Jena, Germany). Sets of fluorescent images were acquired sequentially for the red and green channels to prevent crossover of signals from green to red or red to green channels. 2.4. Cell counting In order to evaluate the results of immunohistochemical analysis quantitatively, we counted the positively stained cells. For quantification of BrdU-labeled cells, we counted the positively stained cells in 25 pixels of 660X660 Am from 5 consecutive coronal sections. These sections were obtained from the striatum level, and were identical among the investigated animals. Cell counting for cerebral cortex and striatum was performed using same sections. The pixels included both the core and the boundary zone of the ischemic region, which were identical among all investigated animals. In the double fluorescent studies, we counted the double-positive cells in the same manner but the size of each pixel was 461 461 Am. Results were
43
expressed as mean F standard deviation (SD), and compared between the sham-control brains and those after I/H at each time point using Student’s t test, with P b 0.01 considered statistically significant.
3. Results 3.1. Histological outcome As our previous report [30], some animals showed only mild (defined as less than 50% of cerebral cortical degeneration) or no neuronal damages. We excluded these animals from the present study because the degree of brain damage would affect the response of SVZ progenitor cells. With these exclusion criteria, only one third of the animals which underwent I/H were included for this study. At each time point after I/H, the cerebral cortex, the caudate, and the hippocampus of the carotid arteryoccluded hemisphere showed severe tissue damage; neuronal as well as glial cells were degenerated, and the tissue became hypotrophic. SVZ was spared histologically, which was different from previous study by Levison et al [19]. Although the reason was uncertain, the temperature during hypoxia might have caused the difference of SVZ morphological change. No tissue changes were observed in the contralateral hemisphere as well as in the brains of the sham-control animals (data not shown). The results were in accordance with our previous report [23,30]. 3.2. Chronological change in BrdU-labeled cells Several cells were labeled for BrdU in the sham-control brain, both in the cerebral cortex (Figs. 2A, B, C) and in the striatum (Figs. 2J, K, L). The number of labeled cells gradually decreased from 7 days (P14) to 21 days (P28), indicating that stem cell proliferation becomes less active as the animal becomes older. In the animal of 7 days after I/H, a large number of cells were labeled for BrdU (Fig. 2D), both in the cerebral cortex (Fig. 2D) and in the striatum (Fig. 2M). At 14 days, the number of positively labeled cells became smaller both in the cerebral cortex (Fig. 2E) and in the striatum (Fig. 2N), but were still larger than those of the sham-control animals at the same time point. The number of labeled cells further decreased at 21 days (Figs. 2F, O), but were again larger than those of the sham-control animals at the same time point. In the contralateral hemisphere of which the carotid artery was not occluded, several cells were labeled for BrdU with gradual decrease from 7 to 21 days both in the cerebral cortex (Figs. 2G, H, I) and in the striatum (Figs. 2P, Q, R). The results of the contralateral hemisphere were comparable with those of the sham-control animals. With quantitative analysis (Fig. 3), it was revealed that 87.5 F 25.5 cells per 1 mm2 were labeled for BrdU in the
44
T. Hayashi et al. / Brain Research 1038 (2005) 41–49
Fig. 2. BrDU labeling in the brain with sham-operation or ischemic/hypoxic injury (I/H). Panels A to I show the cerebral cortex, and J to R show the striatum. Several cells were labeled in the sham-control groups, which was gradually decreased from 7 days to 21 days (A, B, C). In the ischemic side cortex after I/H, a large number of cells became positively labeled at 7 days (D). The numbers of labeled cells were decreased thereafter, but they still remained significantly larger than the sham-control groups (E, F). In the contralateral side, the numbers of labeled cells were comparable with those of the sham-group (G, H, I). In the striatum also, a number of cells were labeled at 7 days after sham-operation and were dramatically decreased thereafter (J, K, L). In the ischemic side striatum after I/H, many cells became labeled (M). Although they gradually decreased, they remained significantly larger in number throughout the investigated time course (N, O). In the contralateral side, the numbers of positively stained cells were comparable with those of the sham-groups (P, Q, R). Scale bar: 200 Am.
sham-control cerebral cortex at 7 days. They were decreased to 25.7 F 9.6 at 14 days and 7.3 F 5.3 at 21 days. In the ischemic cerebral cortex after I/H, the number of labeled cells were 456.4 F 71.6, 258.0 F 130.4, and 105.8 F 65.7
at 7, 14, and 21 days, respectively, that was significantly larger than that of the sham-control at each time point ( P b 0.01). In the contralateral cerebral cortex, 105.6 F 18.4, 37.2 F 23.6, and 11.9 F 4.8 cells per one mm2 were labeled
Fig. 3. Changes in number of BrdU-labeled cells. In the cortex and the striatum with sham-operation, the numbers of labeled cells were largest at 7 days and were decreased thereafter. In the ischemic side cortex with ischemia–hypoxia (I–H), about 5 times number of cells were labeled at 7 day. The labeled cells gradually decreased, but were far larger in number than those in the sham-group cortex at each time point. The numbers of labeled cells in the contralateral cortex were comparable with those in the sham-group. In the sham-control striatum, cells were labeled only at 7 days. About 6 times numbers of cells were labeled in the ischemic side striatum at 7 days after I/H. The labeled cells gradually decreased thereafter, but significantly larger in number throughout the investigated time course (TP b 0.01 compared to the sham-group of the same time point).
T. Hayashi et al. / Brain Research 1038 (2005) 41–49
at 7, 14, and 21 days, which was similar with the results of the sham-control. In the sham-control striatum, the numbers of positively stained cells were 127.9 F 28.2, 2.3 F 1.6, and 0.9 F 2.3 at 7, 14, and 21 days. It became 679.1 F 84.9, 183.7 F 102.6, and 118.0 F 4.8 in the ischemic striatum, all of which were significantly larger than those of the shamcontrol ( P b 0.01). It was 147.6 F 80.3, 5.1 F 3.7, and 3.0 F 4.1 in the contralateral striatum, which showed no statistical significance against the sham-control striatum of the same time points. 3.3. DCX expression of newly produced cells We confirmed that progenitor cell division was enhanced in the neonatal brain after I/H, but it remained uncertain whether these newly produced cells would participate in tissue repair. In order for the newly produced cells to be involved in tissue repair process, they need to differentiate into neuronal cells. With use of the brain sections same with those used for BrdU immunohistochemistry, we performed immunohistochemical analysis for DCX as well as double fluorescent study for DCX and BrdU.
45
Representative photomicrographs of the immunohistochemistry using brains 7 days after I/H are shown in Fig. 4. DCX-positive cells were extensively distributed to the cerebral cortex (Fig. 4B) and the striatum of the ischemic side (Figs. 4E, H, J). In particular, there were lots of DCXpositive cells around the peri-infarct area (Fig. 4B). Double fluorescent study revealed that many but not all BrdUlabeled cells at that region expressed DCX (Fig. 4C). Tomographic analysis with confocal microscopy further showed that signals for BrdU and DCX actually overlapped in some cells (Fig. 4D). In the striatum close to the lateral ventricle also, there were lots of DCX-positive cells (Fig. 4E). There were many cells which were stained for BrdU and DCX, both in the region just beneath the white matter (Fig. 4F) and a bit distant from the lateral ventricle wall (Fig. 4G). DCX-expressing cells were distributed also to the striatum close to the ventral pallidum (Fig. 4H), and again, there were many cells which were labeled for DCX and BrdU (Fig. 4I). To our surprise, there were DCX-expressing cells in the cerebral cortex of the ischemic core region (Fig. 4J). Although the number was substantially smaller, DCXand BrdU-positive cells were confirmed in this region as
Fig. 4. Immunohistochemical analysis for DCX (red) and BrdU (green) in the brain of 7 days after ischemia/hypoxia. Many cells in the cerebral cortex at the MCA boundary area expressed DCX (B), some of which were labeled also for BrdU (C). Tomographic analysis with higher magnification, in which smaller panels demonstrate the deep axis of the line-indicated section, further showed some cells stained for both DCX and BrdU (D). In the striatum close to the lateral ventricle (schematically indicated in panel A), many cells expressed DCX (E) and were labeled also for BrdU (F). The double-stained cells were distributed to the area distant from the lateral ventricle (G). DCX-expressing cells were also confirmed at the striatum close to the ventral pallidum (H), again some of which were labeled for BrdU (I). In the cerebral cortex of the ischemic core, smaller numbers of cells were stained for DCX (J), though some cells were labeled for BrdU (K). Scale bar: B, E, H, J (in J), 100 Am; C, I, K (in K), 100 Am; D, F (in F), 10 Am; G, 20 Am.
46
T. Hayashi et al. / Brain Research 1038 (2005) 41–49
well (Fig. 4K). From these results, we can suppose that some of the newly produced cells migrate to the injured area and attempted to differentiate into the neuronal cells. 3.4. Glial differentiation of newly produced cells With double staining study for DCX and BrdU, we found that substantial number of newly produced cells did not express DCX, indicating that some cells differentiated into glial cells or remained undifferentiated. In order to investigate glial differentiation of newly produced cells, we carried out double florescent study for BrdU and GFAP or NG2. In addition, we investigated the possible presence of BrdU-labeled microglial cells.
Representative photomicrographs of the immunohistochemistry using brains 7 days after I/H are shown in Fig. 5. In the cerebral cortex (Fig. 5A) and the striatum (Fig. 5B), there were many BrdU-labeled cells and GFAP expressing cells at the peri-infarct area, but a small number of cells were positive for both signals. Double staining for BdU and NG2 revealed that a few cells at the cerebral cortex (Fig. 5C) and the striatum (Fig. 5D) of the peri-infarct area expressed both signals. Double staining for BrdU and Iba1 showed that a substantial number of cells were stained for BrdU or Iba1. In the cerebral cortex of the peri-infarct area (Fig. 5E), there were large numbers of cells which expressed both signals. In the striatum of the peri-infarct area (Fig. 5F), however, the number of
Fig. 5. Immunohistochemical analysis for GFAP, NG2, Iba1 (red), and BrdU (green) in the MCA boundary area of 7 days after ischemia/hypoxia. A few BrdUlabeled cells expressed GFAP both in the cerebral cortex (A) and the striatum (B). There also were cells which was positive for BrdU and NG2 in the cerebral cortex (C) and the striatum (D), and larger numbers of cells were positive for BrdU and Iba1 (E, F). Bar: larger photomicrographs (in larger one of F) 100 Am; smaller windows (in smaller one of F) 10 Am.
T. Hayashi et al. / Brain Research 1038 (2005) 41–49
47
Table 1 Quantitative analysis of BrdU-labeled cells and their phenotypes of cortex after hypoxic/ischemic brain injury in neonates +
(a) Total BrDU cells (b) BrDU+ DCX+ cells % of (b)/(a) (c) BrDU+ GFAP+ cells % of (c)/(a) (d) BrDU+ NG2+ cells % of (d)/(a) (e) BrDu+ + Iba1 cells % of (e)/(a)
Sham 7DAH/I
7DAH/I
Ipsi-H/I 14DAH/I
21DAH/I
Contra-H/I 7DAH/I
87.5 0.5 0.6 2.4 2.7 29.1 33.3 19.6 22.4
456.4 108.4 23.8 10.5 2.3 38.5 8.4 149.3 32.7
258.0 49.2 19.1 5.8 2.2 7.9 3.1 51.8 20.1
105.8 27.7 26.2 3.1 2.9 3.7 3.5 40.3 38.1
105.6 0.6 0.6 2.4 2.3 19.6 18.6 17.3 16.4
F 25.5 F 0.5 F 1.1 F 3.7 F 4.1
F 71.6 F 12.7 F 3.1 F 7.4 F 21.6
double-positive cells were smaller than that of the cerebral cortex. 3.5. Quantification of phenotypes of BrdU-labeled cells In order to estimate the ratio of each cellular differentiation, we counted the number of double-positive cells in each brain. The results of the cerebral cortex are summarized in Table 1, and those of the striatum are in Table 2. In the sham-control cerebral cortex in which smaller numbers of BrdU-labeled cells were confirmed, there were only few cells which expressed BrdU and DCX. BrdU- and GFAP-positive cells were also small in number. On the contrary, many BrdU-labeled cells expressed NG2 or Iba1, indicating that oligodendroglial and microglial lineage cells were actively produced in normal condition at this age. After I/H in which total number of BrdU-labeled cells were increased, DCXpositive cells were prominently increased though NG2positive cells were relatively decreased. The number of BrdU- and Iba1-positive cells was increased after I/H, but was gradually decreased later. The results of the contralateral cerebral cortex were similar with those of the sham-control. In the striatum, only small numbers of BrdU-labeled cells expressed DCX or GFAP in normal condition, as seen in the cerebral cortex. After I/H, the number of DCX-positive cells were pronouncedly increased, though those of NG2-positive cells were relatively decreased. The results of the contralateral striatum were similar with those of the sham-control.
F 130.4 F 8.2 F 2.5 F 2.6 F 15.4
F 65.7 F 8.1 F 0.8 F 1.0 F 10.2
F 18.4 F 0.6 F 1.6 F 4.5 F 4.7
4. Discussion In the present study, we found that the cell proliferation in the SVZ was enhanced by I/H. Scheepens et al [27] showed that the SVZ cells showed no response against ischemia, but our and other group’s study indicated that SVZ are sensitive to ischemic injury in neonates [24]. As shown in Figs. 2 and 3, the new cell production gradually ceases after the birth in normal condition. This may reflect that the development of brain approaches completion as the animal becomes older [2]. Most of the BrdU-labeled cells in the sham-control brain were not stained for DCX (Tables 1 and 2). We could speculate that neurogenesis was not so active in this age; even at 50 h after the BrdU injection, many cells should have been stained for BrdU and DCX if neurogenesis had been as active as in the brains after I/H (Tables 1 and 2). On the other hand, many BrdU-labeled cells in the sham-control brains were stained for NG2 (Tables 1 and 2), which may indicate that myeline formation was still under an active stage in neonates [31]. In human being also, myelination does not complete in infancy, which is compatible with the findings obtained by the present study [5]. The next abundant cells labeled for BrdU in shamcontrol brains were microglial cells, which was stained for Iba1 (Tables 1 and 2). Because the microglial cells are mesodermal origin [14], the results presented here does not necessarily mean that the stem or progenitor cells at the SVZ differentiated into microglia; rather, it is more feasible to consider that BrdU was incorporated into microglial cells intra- or extra-medullarily. As BrdU-positive microglial
Table 2 Quantitative analysis of BrDU-labeled cells and their phenotypes of striatum after hypoxic/ischemic brain injury in neonates (a) Total BrDU+ cells (b) BrDU+ DCX+ cells % of (b)/(a) (c) BrDU+ GFAP+ cells % of (c)/(a) (d) BrDU+ NG2+ cells % of (d)/(a) (e) BrDu+ + Iba1 cells % of (e)/(a)
Sham 7DAH/I
7DAH/I
Ipsi-H/I 14DAH/I
21DAH/I
Contra-H/I 7DAH/I
129.7 1.0 0.6 0.8 0.6 18.1 14.0 8.6 6.4
679.1 1885.5 27.8 74.2 10.9 34.6 5.1 62.1 9.1
183.7 56.6 30.8 5.2 2.8 0.8 0.4 8.9 4.8
118.0 F 28.8 F 24.4 3.7 F 3.1 2.1 F 1.8 8.4 F 7.1
147.6 1.2 0.8 1.6 1.1 16.5 11.2 17.3 11.7
F 28.2 F 0.7 F 0.8 F 1.9 F 3.3
F 84.9 F 17.0 F 16.9 F 5.5 F 21.2
F 102.6 F 18.1 F 2.7 F 0.8 F 1.7
102.6 13.4 1.3 0.8 3.1
F 80.3 F 0.8 F 1.0 F 4.8 F 2.6
48
T. Hayashi et al. / Brain Research 1038 (2005) 41–49
cells were pronouncedly increased at 7 days (Fig. 5, Tables 1 and 2), these cells may have been implicated in tissue repair. Indeed, it is known that microglia enhances neural stem cell proliferation and subsequent migration to the injured area by excreting diffusible factors [1]. This cell should thus be a quite important supporting actor for neurological recovery. Compared to the number of BrdU-labeled NG2- or Iba1expressing cells, there were only small numbers of BrdUlabeled GFAP-expressing cells in the sham-control brain (Tables 1 and 2), indicating that supplying new astrocytes are not necessary in quiescent state. Once the I/H was imposed, however, marked astrocytosis occurred in the injured brain (Fig. 5). The origin of BrdU-labeled GFAP-expressing cells cannot exactly be made clear only by the present study; they might have arisen from stem cells at the SVZ, but de novo mitosis of astrocytes should also be considered [20]. Indeed, activation and division of astrocytes are the histological features of the subacute cerebral infarction [6]. As the increase and subsequent decrease of BrdU-labeled GFAPexpressing cells coincided with those of total BrdU-positive cells, however, we believe that at least some of the BrdUlabeled GFAP-expressing cells came from the SVZ. The geographical pattern which suggested migration of BrdUlabeled cells from the SVZ to the injured area (Fig. 4) also supports this speculation. We noticed that the degree of new astrocytes formation was different between the cerebral cortex and striatum (Fig. 5, Tables 1 and 2). This may be due to the intrinsic nature of each anatomical structure, but it is possible to consider that the striatum suffered more severe damage and massive astrocytosis occurred in the striatum [23]. Moreover, it should also be possible to speculate that the striatum was more close to the SVZ and many newly produced astroglial cells reached the injured area. Although there were only few cells which showed positive for BrdU and DCX in the sham-control brain, they dramatically increased in the cerebral cortex and the striatum after I/H (Fig. 4, Tables 1 and 2). As DCX-positive cells distributed from the SVZ to the injured cerebral cortex and the striatum, this increase should be an attempt to repair the damaged tissue by supplementing new neurons. To our surprise, we confirmed BrdU-labeled DCX-expressing cells also at the ischemic core region. Previous report using neonatal mice brain ischemia model also demonstrated DCX expressing cells in the ischemic striatum [24]. As tissue would become pannecrosis at this region [23], most of the DCX-expressing cells should undergo apoptosis in this ischemic core region. Previous reports also demonstrated that most of the newly produced cells would be eliminated before they would be integrated in the neural network [4,28]. In the peri-infarct area, on the other hand, ischemic damage would not result tissue pannecrosis and these newly produced cells might be implicated in repair process. In the adult rat brain after ischemic injury, it was estimated that at most 0.2% of damaged neurons were replaced by newly produced cells [4]. In the neonatal I/H model we used here, we cannot precisely calculate how many neurons were
supplemented by newly produced cells. However, the number of supplemented cells should not be a negligible one because there were so many DCX-positive neuronal precursor cells migrating to the injured area; we found that so many as 456 F 71.6 (cerebral cortex) and 679 F 84.9 (striatum) cells were labeled for BrdU per 1 mm2. Although simple comparison is difficult, neural supplementation should be more active than adult brains after ischemia; up to 400 cells were labeled for BrdU per one section in adult rat brain after focal ischemia [17]. In order to conclude whether these cells actually differentiate into mature neurons, further study would be required. Hypoxic damage causes severe perinatal brain diseases such as cerebral palsy [18]. To treat these disorders is a quite big challenge, but still remains palliative now [12]. As presented here, neonatal brain could possess plasticity and produces new cells. It might be possible to alleviate the neurological disability if we could exploit these characters of the neonatal brains.
Acknowledgments This work was partly supported by Grant-in Aid for Scientific Research (B) 15390273, (Hoga) 15659338, and National Project on Protein Structural and Functional Analyses from the Ministry of Education, Science, Culture and Sports of Japan, and grants (Itoyama Y, Kimura I and Kuzuhara S) from the Ministry of Health and Welfare of Japan.
References [1] J. Aarum, K. Sandberg, S.L. Haeberlein, M.A. Persson, Migration and differentiation of neural precursor cells can be directed by microglia, Proc. Natl. Acad. Sci. U.S.A. 100 (2003) 15983 – 15988. [2] J. Altman, G.D. Das, Autoradiographic and histological studies of postnatal neurogenesis: I. A longitudinal investigation of the kinetics, migration and transformation of cells incorporating tritiated thymidine in neonate rats, with special reference to postnatal neurogenesis in some brain regions, J. Comp. Neurol. 126 (1966) 337 – 389. [3] A. Alvarez-Buylla, J.M. Garcia-Verdugo, Neurogenesis in adult subventricular zone, J. Neurosci. 22 (2002) 629 – 634. [4] A. Arvidsson, T. Collin, D. Kirik, Z. Kokaia, O. Lindvall, Neuronal replacement from endogenous precursors in the adult brain after stroke, Nat. Med. 8 (2002) 963 – 970. [5] S.A. Back, N.L. Luo, N.S. Borenstein, J.M. Levine, J.J. Volpe, H.C. Kinney, Late oligodendrocyte progenitors coincide with the developmental window of vulnerability for human perinatal white matter injury, J. Neurosci. 21 (2001) 1302 – 1312. [6] A. da Cunha, J.J. Jefferson, W.R. Tyor, J.D. Glass, F.S. Jannotta, L. Vitkovic, Gliosis in human brain: relationship to size but not other properties of astrocytes, Brain Res. 600 (1993) 161 – 165. [7] J.L. Daval, G. Pourie, S. Grojean, V. Lievre, C. Strazielle, S. Blaise, P. Vert, Neonatal hypoxia triggers transient apoptosis followed by neurogenesis in the rat CA1 hippocampus, Pediatr. Res. 55 (2004) 561 – 567. [8] M.R. Dawson, A. Polito, J.M. Levine, R. Reynolds, NG2-expressing glial progenitor cells: an abundant and widespread population of
T. Hayashi et al. / Brain Research 1038 (2005) 41–49
[9]
[10] [11]
[12] [13]
[14]
[15]
[16]
[17]
[18]
[19]
cycling cells in the adult rat CNS, Mol. Cell. Neurosci. 24 (2003) 476 – 488. F. Doetsch, J.M. Garcia-Verdugo, A. Alvarez-Buylla, Cellular composition and three-dimensional organization of the subventricular germinal zone in the adult mammalian brain, J. Neurosci. 17 (1997) 5046 – 5061. M. Fisher, Stem cell transplantation for stroke: does it work, and if so, how? Stroke 34 (2003) 2083. F. Francis, A. Koulakoff, D. Boucher, P. Chafey, B. Schaar, M.C. Vinet, G. Friocourt, N. McDonell, O. Reiner, A. Kahn, S.K. McCornell, Y. Berwald-Netter, P. Denoulet, J. Chelly, Doublecortin is a developmentally regulated, microtubule-associated protein expressed in migrating and differentiating neurons, Neuron 23 (1999) 247 – 256. M. Goldstein, The treatment of cerebral palsy: what we know, what we don’t know, J. Pediatr. 145 (2 Suppl.) (2004) S42 – S46. J. Grow, J.D. Barks, Pathogenesis of hypoxic–ischemic cerebral injury in the term infant: current concepts, Clin. Perinatol. 29 (2002) 585 – 602. G.J. Guillemin, B.J. Brew, Microglia, macrophages, perivascular macrophages, and pericytes: a review of function and identification, J. Leukocyte Biol. 75 (2004) 388 – 397. Y. Imai, I. Ibata, D. Ito, K. Ohsawa, S. Kohsaka, A novel gene iba1 in the major histocompatibility complex class III region encoding an EF hand protein expressed in a monocytic lineage, Biochem. Biophys. Res. Commun. 224 (1996) 855 – 862. M. Iwai, K. Sato, N. Omori, I. Nagano, Y. Manabe, M. Shoji, K. Abe, Three steps of neural stem cells development in gerbil dentate gyrus after transient ischemia, J. Cereb. Blood Flow Metab. 22 (2002) 411 – 419. K. Jin, M. Minami, J.Q. Lan, X.O. Mao, S. Batteur, R.P. Simon, D.A. Greenberg, Neurogenesis in dentate subgranular zone and rostral subventricular zone after focal cerebral ischemia in the rat, Proc. Natl. Acad. Sci. U.S.A. 98 (2001) 4710 – 4715. M.V. Johnston, W.H. Trescher, A. Ishida, W. Nakajima, Neurobiology of hypoxic–ischemic injury in the developing brain, Pediatr. Res. 49 (2001) 735 – 741. S.W. Levison, R.P. Rothstein, M.J. Romanko, M.J. Snyder, R.L. Meyers, S.J. Vannucci, Hypoxia/ischemia depletes the rat perinatal
[20]
[21] [22]
[23]
[24]
[25] [26]
[27]
[28] [29] [30]
[31]
49
subventricular zone of oligodendrocyte progenitors and neural stem cells, Dev. Neurosci. 23 (2001) 234 – 247. B. Lin, M.D. Ginsberg, R. Busto, W.D. Dietrich, Sequential analysis of subacute and chronic neuronal, astrocytic and microglial alterations after transient global ischemia in rats, Acta Neuropathol. (Berlin) 95 (1998) 511 – 523. A. Messing, M. Brenner, GFAP: functional implications gleaned from studies of genetically engineered mice, Glia 43 (2003) 87 – 90. H. Nakatomi, T. Kuriu, S. Okabe, S. Yamamoto, O. Hatano, N. Kawahara, A. Tamura, T. Kirino, M. Nakafuku, Regeneration of hippocampal pyramidal neurons after ischemic brain injury by recruitment of endogenous neural progenitors, Cell 110 (2002) 429 – 441. A. Ota, T. Ikeda, X.Y. Xia, Y.X. Xia, T. Ikenoue, Hypoxic– ischemic tolerance induced by hyperthermic pretreatment in newborn rats, J. Soc. Gynecol. Invest. 7 (2000) 102 – 105. J.M. Plane, R. Liu, T.W. Wang, F.S. Silverstein, J.M. Parent, Neonatal hypoxic–ischemic injury increases forebrain subventricular zone neurogenesis in the mouse, Neurobiol. Dis. 16 (2004) 585 – 595. D.S. Reddihough, K.J. Collins, The epidemiology and causes of cerebral palsy, Aust. J. Physiother. 49 (2003) 7 – 12. Y. Sasaki, K. Ohsawa, H. Kanazawa, S. Kohsaka, Y. Imai, Iba1 is an actin-cross-linking protein in macrophages/microglia, Biochem. Biophys. Res. Commun. 286 (2001) 292 – 297. A. Scheepens, G. Wassink, M.J. Piersma, W.D. Van de Berg, C.E. Blanco, A delayed increase in hippocampal proliferation following global asphyxia in the neonatal rat, Brain Res., Dev. Brain Res. 142 (2003) 67 – 76. E.Y. Snyder, K.I. Park, Limitations in the brain repair, Nat. Med. 8 (2002) 928 – 930. F.M. Vaccarino, L.R. Ment, Injury and repair in developing brain, Arch. Dis. Child., Fetal Neonatal. Ed. 89 (2004) F190 – F192. Y.X. Xia, H. Sameshima, T. Ikeda, T. Higo, T. Ikenoue, Cerebral blood flow distribution and hypoxic–ischemic brain damage in newborn rats, J. Obstet. Gynaecol. Res. 28 (2002) 320 – 326. A.U. Zaidi, D.A. Bessert, J.E. Ong, H. Xu, J.D. Barks, F.S. Silverstein, R.P. Skoff, New oligodendrocytes are generated after neonatal hypoxic–ischemic brain injury in rodents, Glia 46 (2004) 380 – 390.