Early induction of neuronal lipocalin-type prostaglandin D synthase after hypoxic-ischemic injury in developing brains

Neuroscience Letters 420 (2007) 39–44

Early induction of neuronal lipocalin-type prostaglandin D synthase after hypoxic-ischemic injury in developing brains Hidetoshi Taniguchi a , Ikuko Mohri a,b,c , Hitomi Okabe-Arahori a , Takahisa Kanekiyo a , Kuriko Kagitani-Shimono a , Kazuko Wada a , Yoshihiro Urade c , Masahiro Nakayama d , Keiichi Ozono a , Masako Taniike a,b,∗ a

Department of Pediatrics, Osaka University Graduate School of Medicine, 2-2 Yamadaoka, Suita-shi, Osaka 565-0871, Japan b Department of Mental Health and Environmental Effects Research, The Research Center for Child Mental Development, Osaka University Graduate School of Medicine, Suita, Osaka 565-0871, Japan c Department of Molecular Behavioral Biology, Osaka Bioscience Institute, 6-2-4 Furuedai, Suita, Osaka 565-0874, Japan d Division of Clinical Laboratory Medicine and Anatomic Pathology, Osaka Medical Center and Research Institute for Maternal and Child Health, 840 Murodou, Izumi, Osaka 594-1101, Japan Received 18 December 2006; received in revised form 27 February 2007; accepted 1 April 2007

Abstract Lipocalin-type prostaglandin (PG) D synthase (L-PGDS) is up-regulated in oligodendrocytes (OLs) in mouse models for genetic neurological disorders including globoid cell leukodystrophy (twitcher) and GM1 and GM2 gangliosidoses and in the brain of patients with multiple sclerosis. Since L-PGDS-deficient twitcher mice undergo extensive neuronal death, we concluded that L-PGDS functions protectively against neuronal degeneration. In this study, we investigated whether L-PGDS is also up-regulated in acute and massive brain injury resulting from neonatal hypoxic-ischemic encephalopathy (HIE). Analysis of brains from human neonates who had died from HIE disclosed that the surviving neurons in the infarcted lesions expressed L-PGDS. Mouse models for neonatal HIE were made on postnatal day (PND) 7. Global infarction in the ipsilateral hemisphere was evident at 24 h after reoxygenation in this model. Intense L-PGDS immunoreactivity was already observed at 10 min after reoxygenation in apparently normal neurons in the cortex, and thereafter, in neurons adjacent to the infarcted area. Quantitative RT-PCR revealed that the L-PGDS mRNA level of the infarcted hemisphere was 33-fold higher than that of the sham-operated mouse brains at 1 h after reoxygenation and that it decreased to the normal level by 24 h thereafter. Furthermore, in both human and mouse brains, many of L-PGDS-positive cells were also immunoreactive for p53; and some of these expressed cleaved caspase-3. The expression of L-PGDS in degenerating neurons implies that L-PGDS functions as an early stress protein to protect against neuronal death in the HIE brain. © 2007 Elsevier Ireland Ltd. All rights reserved. Keywords: Hypoxic-ischemic encephalopathy; Neonate; Prostaglandin D2 ; Lipocalin-type prostaglandin D synthase; Apoptosis

Lipocalin-type prostaglandin D synthase (L-PGDS) is expressed in oligodendrocytes (OLs) after the commencement of myelination [8,12] L-PGDS is a unique bifunctional protein, acting as both a prostaglandin (PG) D2 -producing enzyme and a lipophilic ligand-carrier protein of the lipocalin family [11]. We previously reported that L-PGDS was up-regulated in the perineuronal OLs in the twitcher mouse, a model of a genetic demyelinating disease, in which OLs are depleted by ∗

Corresponding author at: Department of Mental Health and Environmental Effects Research, The Research Center for Child Mental Development, Osaka University Graduate School of Medicine, D-5, 2-2 Yamadaoka, Suita-shi, Osaka 565-0871, Japan. Tel.: +81 6 6879 3932; fax: +81 6 6879 3939. E-mail address: [email protected] (M. Taniike). 0304-3940/$ – see front matter © 2007 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.neulet.2007.04.016

apoptosis [9]. The fact that neuronal apoptosis, which is never observed in twitcher, is an additional conspicuous feature in LPGDS-deficient twitcher, suggests that L-PGDS expressed in perineuronal OLs plays a neuroprotective role [8]. Furthermore, we found out that L-PGDS was up-regulated in OLs and hypertrophied astrocytes in the demyelinated plaques of patients with multiple sclerosis [4]. These L-PGDS-positive OLs and astrocytes expressed ␣B-crystallin, which is a stress protein. Finally, we disclosed that L-PGDS expression was increased in OLs and neurons in a restricted area of the brain in patients with various lysosomal storage disorders (LSDs) presenting neurodegeneration [6]. These lines of evidence indicate that up-regulation of L-PGDS occurs as a stress response in these chronic neurological disorders.

40

H. Taniguchi et al. / Neuroscience Letters 420 (2007) 39–44

Presently we investigated whether L-PGDS is also upregulated in the case of an acute brain insult, specifically hypoxic-ischemic encephalopathy (HIE). HIE in neonates, some of which cases are due solely to a perinatal hypoxic-ischemic event (asphyxia), is a major cause of acute mortality and life-long neurological disability in the survivors. Vannucci and Vannucci established rodent models for neonatal HIE in which neuropathological events proceed even after the cession of hypoxia, i.e., after reoxygenation due to subsequent reperfusion [13]. The aim of this present study was three-fold: (1) to examine if L-PGDS is increased in the HIE brain; (2) to identify LPGDS-immunoreactive cells in HIE brains; and (3) to clarify the spatiotemporal L-PGDS expression in the mouse model for HIE. We considered that evaluation of this HIE model would allow us to suggest the pathophysiological role of L-PGDS, which might open the door to a novel therapy for alleviating the severe neurological sequelae in HIE patients. Human brain tissues were obtained from Osaka Medical Center and Research Institute for Maternal and Child Health. Paraffin sections of formalin-fixed brains were examined in the present

work. Autopsy was performed after written informed consent had been obtained from the bereaved families. Brains were obtained postmortem from eight HIE patients and eight agematched control patients who had died from non-neurological diseases. This study was approved by the institutional review boards of Osaka University Graduate School of Medicine and Osaka Medical Center and Research Institute for Maternal and Child Health. All animal experiments conformed to the Japanese Law for the Protection of Experimental Animals and followed the protocols approved by the Institutional Animal Care and Use Committee at the Osaka Bioscience Institute, where the experiments were carried out. A mouse model for neonatal HIE was prepared by following the method previously reported [7,13]. Briefly, the left common carotid arteries of 10 C57Bl/6J mice (Japan SLC, Inc., Shizuoka, Japan) were permanently ligated at postnatal day (PND) 7 under halothane anesthesia. After a recuperation period of 1–2 h at 37 ◦ C, the pups were placed in containers through which humidified 8% oxygen and balanced nitrogen flowed. The containers were placed on a 37 ◦ C

Fig. 1. L-PGDS immunocytochemistry on human neonatal brains. (A) Cingulate gyrus of a control neonate. (B) Cingulate gyrus and hippocampus of a neonate who had died from HIE (“a” and “b”, respectively) and negative control for B-b, in which the first antibody had been omitted (c). Arrows indicate L-PGDS-immunopositive cells (a and b). Scale bar = 50 ␮m. (C–N) Double immunofluorescence for L-PGDS and Olig2, an OL marker (C–E), NeuN, a neuronal marker (F–H), CD68, a microglial marker (I–K), and GFAP, an astroglial marker (L–N), in an HIE brain. Scale bar = 5 ␮m.

H. Taniguchi et al. / Neuroscience Letters 420 (2007) 39–44

warm plate to maintain normothermia. The pups were put in this hypoxic environment for 30 min and then returned to their dam until they were killed. Mice with obviously decreased cerebral blood flow, as assessed by use of a Laser Doppler Blood Flow Meter (TBF-LC1, Unique Medical Co. Ltd., Tokyo, Japan), were subjected to subsequent analyses. Rabbit polyclonal and rat monoclonal antibodies against mouse or human L-PGDS were raised in Osaka Bioscience Institute [4,8]. The other primary antibodies used in this study were as follow: anti-human GFAP antibody (1:500 dilution; DakoCytomation, Glostrup, Denmark), anti-human NeuN antibody (1:100 dilution; Chemicon International, Inc., Temecula, CA), and anti-human Olig2 antibody (1:50 dilution; IBL, Gunma, Japan), for identification of astrocytes, neurons, and OLs, respectively. Anti-human/mouse HIF-1␣ antibody (Clone 241809; R&D Systems, Inc. Minneapolis, MN) and anti-human p53 antibody (Clone DO-7; 1:200 dilution, DakoCytomation) were used to identify pro-apoptotic cells [14], and anti-human

41

cleaved caspase-3 (Asp175; 1:200 dilution; Cell Signaling Technology, Inc., Danvers, MA) was used to detect apoptotic ones. After deparaffinized sections had been preincubated at 37 ◦ C for 15 min with 0.1% trypsin containing 0.1% CaCl2 , or autoclaved at 120 ◦ C for 10 min with 10 mM sodium citrate buffer (pH 6.0), according to the sample preparation or antibodies used, they were sequentially incubated with primary antibody, the appropriate biotinylated secondary antibody (2 ␮g/ml; Vector Laboratory, Burlingame, CA), and avidin–biotin-complex (ABC) by using an ABC elite kit (Vector Laboratory) according to the manufacturer’s protocol. For double immunofluorescence, deparaffinized sections were incubated at 4 ◦ C overnight with either anti-GFAP, anti-Olig2 or anti-NeuN antibody followed by rat antimouse or anti-human L-PGDS antiserum. FITC, Cy3 (Jackson ImmunoResearch Laboratories, Inc., West Groove, PA), Texas red, Alexa 488, and Alexa 546 (Invitrogen, Carlsbad, CA) were used as fluorescent labels.

Fig. 2. Temporospatial expressional change in L-PGDS immunostaining in the premotor area of HIE model mice (A–H), and quantitative RT-PCR for L-PGDS mRNA (I). (A–D) H&E staining. (E–H) L-PGDS immunostaining of sections adjacent to those in “A”–“D.” The time after the onset of reoxygenation is shown at the bottom of the figures. The area encircled by the dotted line in the ipsilateral cortex at 1 h represents an edematous lesion colored pale pink (C). Degenerating neurons, shown by the arrowheads, were seen after 24 h of reoxygenation (D). Scale bar = 100 ␮m. (I) Values are the copy number of L-PGDS mRNA molecules in the ipsilateral hemisphere. Bars represent the mean ± S.E. (n = 3). ** P < 0.01, as indicated by the bracket.

42

H. Taniguchi et al. / Neuroscience Letters 420 (2007) 39–44

By using quantitative RT-PCR analysis, we measured the mRNA level of L-PGDS. Glyceraldehyde-3-phosphate dehydrogenase (G3PDH) was used as an internal standard. Total RNA was prepared from sagitally dissected ipsilateral and contralateral cerebral hemispheres of HIE mice by using Isogene according to the manufacturer’s instructions (Nippon Gene, Tokyo, Japan). First-strand cDNAs were synthesized from 1 ␮g of total RNA by using avian myeloblastosis virus reverse transcriptase and oligo dT-adaptor primer (Takara, Kyoto, Japan) at 50 ◦ C for 40 min after denaturation at 72 ◦ C for 3 min. cDNA was amplified by use of a real-time PCR LightCycler system, a LightCycler-FastStart DNA Master SYBR Green I (Roche Diagnostics, Indianapolis, IN), and gene-specific primers under the following conditions: an initial denaturation at 95 ◦ C for 10 min, followed by 40 cycles of denaturation at 95 ◦ C for 11 s, annealing at 59 ◦ C for 5 s, and extension at 72 ◦ C for 10 s. The products for L-PGDS and G3PDH were detected at 89 ◦ C and 87 ◦ C, respectively. Gene-specific primers used were as follow: L-PGDS forward, 5 -CAGGAAAAACCAGTGTGAGAGACC3 ; L-PGDS reverse, 5 -AGAGGGTGGCCATGCGGAAG-3 ; G3PDH forward, 5 -TGAACGGGAAGCTCACTGG-3 ; and G3PDH reverse, 5 -TCCACCACCCTGTTGCT-3 . The PCR products were evaluated by melting-curve analysis according to the manufacturer’s instructions, and checked after agarose gel electrophoresis. All values were corrected with reference to the value for G3PDH. Statistical comparisons were made by using Student’s t-test. Values of P < 0.05 were considered to be significant. We performed L-PGDS immunostaining of the cingulate gyrus, temporal cortex, and hippocampus of eight human neonatal brains and eight control brains. Subjects with HIE were 0–9 days old (2.9 days old on average), and the age of control patients was 0–7 days old (2 days old on average).

In the control brains, L-PGDS immunoreactivity was rarely observed (Fig. 1A) in all the regions of the brain we examined, and the few immunoreactive cells that were found were small and reminiscent of OLs. In the HIE brains, however, the L-PGDS immunoreactivity was markedly increased in regions known to be susceptible to a hypoxic-ischemic insult, such as the cingulate gyrus and hippocampus (Fig. 1B-a and b). In the cingulate gyrus, many of L-PGDS-positive cells had shrunken cytoplasm (Fig. 1B-a); whereas in the hippocampus, the majority of the L-PGDS-immunoreactive cells were relatively large and reminiscent of neurons (arrows, Fig. 1B-b). Double immunostaining revealed that L-PGDS-positive cells in the cingulate gyrus were positive for either Olig2, a marker for OLs (Fig. 1C–E), or NeuN, a marker for neurons (Fig. 1F–H). L-PGDS-positive cells were rarely positive for CD68 or GFAP, markers for microglia or astrocytes, respectively (Fig. 1I–N). To validate the human data, we generated an animal HIE model by using mouse neonates at PND7. Hematoxylin and eosin (H&E) staining did not reveal any apparent change on the ipsilateral side at 10 min after the start of reoxygenation (Fig. 2B) as compared with the histology of the contralateral brain (Fig. 2A). Immunocytochemistry showed that ipsilateral L-PGDS expression was already intense as early as 10 min in layers IV-VI of the premotor area (Fig. 2F), whereas LPGDS expression was not observed in the brain parenchyma and was restricted to the meninges on the contralateral side, in the HIE model mice (Fig. 2E). One hour after the start of reoxygenation, H&E staining revealed that an edematous change including the deep cortical layer was evident (Fig. 2C); and LPGDS-immunoreactive cells were recognized at the edge of this edematous area (Fig. 2G). After 24 h of reoxygenation, all of the six cortical layers in the premotor area had undergone infarction (Fig. 2D); and almost all of the neurons and neuronal layers

Fig. 3. Double immunostaining for L-PGDS and Olig2 or NeuN in the cortex of HIE mice. (A–C) Double immunostaining for L-PGDS and Olig2. The arrowhead points to an L-PGDS+ Olig2+ OL. Scale bar = 10 ␮m. (D–I) Double immunostaining for L-PGDS and NeuN. The arrows in “D”–“F” indicate L-PGDS+ NeuN+ neurons. The asterisks in “G” and “I” indicate the infarct area, which shows extensive neuronal loss. Scale bars = 10 ␮m (D–F) and 100 ␮m (G–I).

H. Taniguchi et al. / Neuroscience Letters 420 (2007) 39–44

had disappeared (Fig. 2D). At this time, L-PGDS expression was almost undetectable in the devastating lesion in the brain parenchyma (Fig. 2H). Quantification of the L-PGDS mRNA revealed that its level at 1 h after the start of reoxygenation was 33 times higher than the pre-hypoxia value in the infarcted hemisphere and that it decreased to the normal level by 24 h and remained so by 48 h after reoxygenation had begun (Fig. 2I). Double immunofluorescence disclosed that only a few of the L-PGDS-immunoreactive cells were Olig2-positive OLs (Fig. 3A–C) and that the vast majority of L-PGDS-positive cells were NeuN-positive neurons (Fig. 3D–F) in the infarcted mouse brain. L-PGDS-positive neurons were restricted to the boundary of the infarct area (asterisk in Fig. 3G and I). There were no L-PGDS-positive astrocytes or microglia as judged from the results of double immunocytochemistry for L-PGDS and GFAP or RCA-1, a marker for astrocytes or microglia, respectively (data not shown). Since we hypothesized that L-PGDS was expressed in proapoptotic cells as an early stress protein from our data described above, we conducted double immunostaining for L-PGDS and p53, hypoxia-inducible factor 1␣ (HIF-1␣) or cleaved caspase3, p53 and HIF-1␣ identifying pro-apoptotic cells in the brain

43

and cleaved caspase-3, apoptotic ones [2]. In mouse and human brains, many of the L-PGDS-positive neurons expressed p53 (Fig. 4C and F, respectively; arrowheads), although there were some L-PGDS-positive p53-negative cells (Fig. 4C and F, arrows). In the HIE mouse brain, many of the L-PGDS-positive cells were immmunopositive for HIF-1␣ (Fig. 4I), whereas only a small portion of them were positive for cleaved caspase-3. These results clearly indicate that L-PGDS was expressed in pro-apoptotic neurons. In the present study, we showed that L-PGDS-positive cells were more abundant in human neonatal HIE brains than in nonHIE neonatal human brains. These cells were predominantly, if not exclusively, OLs and neurons. We reported earlier that L-PGDS expression in OLs is upregulated in brains affected by demyelinating diseases, such as the mouse model of globoid cell leukodystrophy (twitcher) [8], human multiple sclerosis [4], and mouse models of LSDs [6]. From the facts that neuronal apoptosis was significantly increased in L-PGDS-deficient twitcher mice and L-PGDSpositive OLs expressed ␣B-crystallin, an established stress marker, in patients with chronic multiple sclerosis [4], we hypothesized that L-PGDS is induced as an early stress pro-

Fig. 4. Double immunostaining for L-PGDS and p53 in the HIE mouse cerebral cortex after 24 h of reoxygenation (A–C) and in the cerebral cortex of a human HIE neonate who was born at 37 weeks of gestation and died on the day of birth (D–F). The arrowheads and the thin arrows point to L-PGDS+ p53+ and L-PGDS+ p53cells, respectively. (G–I) Double immunostaining for L-PGDS and HIF-1␣ in the HIE mouse cerebral cortex. The arrowheads, thin arrow, and bold arrow point to L-PGDS+ HIF-1␣+ , L-PGDS+ HIF-1␣− and L-PGDS- HIF-1␣+ cells, respectively. (J–L) Double immunostaining for L-PGDS and cleaved caspase-3 in the HIE mouse cerebral cortex. The arrowheads and thin arrows point to L-PGDS+ cleaved caspase-3+ and L-PGDS+ cleaved caspase-3− cells, respectively. Scale bars = 20 ␮m.

44

H. Taniguchi et al. / Neuroscience Letters 420 (2007) 39–44

tein and functions to protect neurons from apoptosis through some unknown mechanism. In this study, we demonstrated that L-PGDS expression was also induced in human HIE brains. Neuronal L-PGDS expression has been observed in the rat brain during the first 2 weeks of life [10] and in adult brains of mouse models for LSDs including GM1 gangliosidosis and Niemann-Pick type C [6]. These data indicate that neurons are potentially capable of producing LPGDS in both human and mouse brains. In this study, we showed that L-PGDS was expressed in neurons in neonatal HIE mouse brains. In this model, L-PGDS was at first present in neurons located in the deep layer of the cortex, where the cell injury seems to occur at an early phase (Fig. 2F and G); however, at 24 h after reoxygenation, when many of the cells in the deep cortex had been deleted by apoptosis or some other form of cell death, localization of L-PGDS-positive cells was restricted to the penumbra (Fig. 3I). Thus, L-PGDS seems to be induced where the cell injury takes place and it would disappear when those cells die. This expression change implies that L-PGDS is induced as a protective reaction of the cells; however, the insult is so devastating that the L-PGDS-mediated cell protection often may not succeed in this disease process. Furthermore, L-PGDS expression was co-localized with p53, HIF-1␣, and cleaved caspase-3 in proapoptotic/apoptotic neurons. p53 is a pro-apoptotic molecule, and its expression increases after hypoxia-ischemia in neonatal mice [1]. HIF-1␣ is a stress protein that acts in the initiation and commitment phase in apoptosis and is expressed upstream of p53; whereas cleaved caspase-3 is a well-known executioner of apoptosis [3]. In the HIE mouse brain, HIF-1␣ and p53 were often co-localized with L-PGDS. Furthermore, all p53-positive cells were immunoreactive for L-PGDS, whereas some of the LPGDS-positive cells were negative for p53. On the other hand, only a small number of the L-PGDS-positive cells expressed cleaved caspase-3 [2,15]. Our data imply that induction of LPGDS occurred at the same time window of that of HIF-1␣ and preceded that of p53. From these lines of evidence, we conclude that neuronal L-PGDS was induced at a relatively early stage of the degeneration process. L-PGDS is a unique bifunctional protein, which not only catalyzes the synthesis of PGD2 but also functions as a lipocalin as well. Liang et al. in 2005 reported the neuroprotective effect of PGD2 in vitro [5]. Therefore, it is possible that L-PGDS may protect neurons by local secretion of PGD2 in HIE brains. As a lipocalin, on the other hand, L-PGDS binds to and transports small hydrophobic molecules such as biliverdin, bilirubin, and retinoic acid [11]. We also reported that L-PGDS binds to harmful lipophilic storage substrates in mouse models for GM1 and GM2 gangliosidoses, in which L-PGDS expression is up-regulated [6]. This binding may explain the neuroprotective role of L-PGDS we had previously reported [8]. These lines of evidence indicate that up-regulation of L-PGDS expression in HIE brains may be a protective response. In conclusion, we showed that L-PGDS is closely related to the apoptotic process in neonatal HIE. We believe that accumulation of data about the physiological roles of L-PGDS in HIE brains, including our present data, will surely offer a new insight to develop novel neuroprotection strategies.

Acknowledgements We thank Dr. Keiko Matsuoka for preparing autopsy samples and Dr. Nanae Nagata for technical assistance. This study was supported by a grant (No. 17591085, to M.T. and 18591218 to K.W.) from the program Grants-in-aid for Scientific Research (C) of the Ministry of Education, Culture, Sports, Science, and Technology of Japan, and by grants from Mitsubishi Foundation (to Y.U.), Japan Foundation for Applied Enzymology (to I.M and to Y.U.), Japan Aerospace Exploration Agency (to Y.U.), Program for Promotion of Fundamental Studies in Health Sciences of the National Institute of Biomedical Innovation (NIBIO; to Y.U.) and Osaka City (to Y.U.). References [1] J.W. Calvert, J. Cahill, M. Yamaguchi-Okada, J.H. Zhang, Oxygen treatment after experimental hypoxia-ischemia in neonatal rats alters the expression of HIF-1alpha and its downstream target genes, J. Appl. Physiol. 101 (2006) 853–865. [2] C. Culmsee, M.P. Mattson, p53 in neuronal apoptosis, Biochem. Biophys. Res. Commun. 331 (2005) 761–777. [3] M.W. Halterman, H.J. Federoff, HIF-1alpha and p53 promote hypoxiainduced delayed neuronal death in models of CNS ischemia, Exp. Neurol. 159 (1999) 65–72. [4] K. Kagitani-Shimono, I. Mohri, H. Oda, K. Ozono, K. Suzuki, Y. Urade, M. Taniike, Lipocalin-type prostaglandin D synthase (beta-trace) is upregulated in the alphaB-crystallin-positive oligodendrocytes and astrocytes in the chronic multiple sclerosis, Neuropathol. Appl. Neurobiol. 32 (2006) 64–73. [5] X. Liang, L. Wu, T. Hand, K. Andreasson, Prostaglandin D2 mediates neuronal protection via the DP1 receptor, J. Neurochem. 92 (2005) 477–486. [6] I. Mohri, M. Taniike, I. Okazaki, K. Kagitani-Shimono, K. Aritake, T. Kanekiyo, T. Yagi, S. Takikita, H.S. Kim, Y. Urade, K. Suzuki, Lipocalintype prostaglandin D synthase is up-regulated in oligodendrocytes in lysosomal storage diseases and binds gangliosides, J. Neurochem. 97 (2006) 641–651. [7] R.A. Sheldon, C. Sedik, D.M. Ferriero, Strain-related brain injury in neonatal mice subjected to hypoxia-ischemia, Brain Res. 810 (1998) 114–122. [8] M. Taniike, I. Mohri, N. Eguchi, C.T. Beuckmann, K. Suzuki, Y. Urade, Perineuronal oligodendrocytes protect against neuronal apoptosis through the production of lipocalin-type prostaglandin D synthase in a genetic demyelinating model, J. Neurosci. 22 (2002) 4885–4896. [9] M. Taniike, I. Mohri, N. Eguchi, D. Irikura, Y. Urade, S. Okada, K. Suzuki, An apoptotic depletion of oligodendrocytes in the twitcher, a murine model of globoid cell leukodystrophy, J. Neuropathol. Exp. Neurol. 58 (1999) 644–653. [10] Y. Urade, N. Fujimoto, T. Kaneko, A. Konishi, N. Mizuno, O. Hayaishi, Postnatal changes in the localization of prostaglandin D synthetase from neurons to oligodendrocytes in the rat brain, J. Biol. Chem. 262 (1987) 15132–15136. [11] Y. Urade, O. Hayaishi, Prostaglandin, D synthase: structure and function, Vitam. Horm. 58 (2000) 89–120. [12] Y. Urade, K. Kitahama, H. Ohishi, T. Kaneko, N. Mizuno, O. Hayaishi, Dominant expression of mRNA for prostaglandin D synthase in leptomeninges, choroid plexus, and oligodendrocytes of the adult rat brain, Proc. Natl. Acad Sci. U.S.A. 90 (1993) 9070–9074. [13] R.C. Vannucci, S.J. Vannucci, Perinatal hypoxic-ischemic brain damage: evolution of an animal model, Dev. Neurosci. 27 (2005) 81–86. [14] X. Wang, J.O. Karlsson, C. Zhu, B.A. Bahr, H. Hagberg, K. Blomgren, Caspase-3 activation after neonatal rat cerebral hypoxia-ischemia, Biol. Neonate 79 (2001) 172–179. [15] C. Zhu, L. Qiu, X. Wang, U. Hallin, C. Cande, G. Kroemer, H. Hagberg, K. Blomgren, Involvement of apoptosis-inducing factor in neuronal death after hypoxia-ischemia in the neonatal rat brain, J. Neurochem. 86 (2003) 306–317.