Expression of neuronal nitric oxide synthase corresponds to regions of selective vulnerability to...

Neurobiology of Disease1995, 2, 145–155

Stephen M. Black 1,2

Melanie A. Bedolli 2

Salvador Martinez 5

James D. Bristow 2,4

Donna M. Ferriero 2,3

Scott J. Soifer 2,4

Expression of neuronal nitric oxide synthasecorresponds to regions of selective vulnerability tohypoxia–ischaemia in the developing rat brain

Departments of 1Pharmaceutical Chemistry, 2Pediatrics and 3Neurology and the 4Cardiovascular Research Institute,University of California, San Francisco,San Francisco, CA 94143-0106, USA 5Department of Morphological Sciences, Faculty of Medicine,University of Murcia, 30100, Murcia, Spain

Nitric oxide (NO) has been implicated in the pathogenesis of brain injury fromhypoxia-ischaemia. In the brain, the enzyme responsible for NO synthesis is neu-ronal nitric oxide synthase (nNOS). Using in situ hybridization, immunohistochem-istry and NADPH diaphorase histochemistry, we examined the spatial and temporalexpression of nNOS during development of the rat brain to determine whether theexpression of nNOS delineates the areas of the brain that are selectively vulnerable tohypoxic-ischaemia injury.The expression of nNOS was localized to discrete areas ofthe brain. nNOS could be detected in the developing forebrain in the 10-day-oldembryo (E10). From E14 to E18, the highest level of expression was in the corticalplate, where the majority of neurons were positive. However, this expression dimin-ished with time; in the adult there were only a few nNOS-positive neurones in thedeep layers of the cortex. Expression of nNOS was not detected prenatally in thebasal ganglia. There was transient high-level expression during the first postnatalweek. Thereafter, the basal ganglia exhibited the adult pattern of expression.Expression of nNOS could be detected in the hippocampus at E16.This expressionremained constant with regional localization in layers CA1 and CA3 in the adult.Similarly, nNOS expression in the developing cerebellum was observed only afterbirth. From the first day after birth (P1) to P6, expression was limited to the molecu-lar cell layer.As the cerebellum matured, nNOS expression could be detected in theinner granular layer. By P21, the adult distribution of nNOS expression wasobserved. All regions expressing nNOS mRNA also demonstrated nNOS proteinexpression and NADPH diaphorase catalytic activity. Our results demonstrate thatnNOS expression in the developing brain correlates with regions of selective vulner-ability to hypoxic-ischaemic injury, and, therefore, supports a role for NO inhypoxic-ischaemic injury in the developing brain.

hypoxia-ischaemia, in situ hybridization, immunohistochemistry, neuronal nitricoxide synthase, nitric oxide

Received 15 December 1994, accepted for publication 15 June 1995

Introduction

Nitric oxide synthase (NOS, EC 1.14.23) is the enzymeresponsible for the biosynthesis of nitric oxide (NO) fromL-arginine. NO is a free radical that has several importantbiological functions. It acts as a vasodilator, as a toxic agentin inflammatory responses and as a neurotransmitter. NOS

has at least three isoforms, the endothelial isoform, which ismembrane associated and found in endothelial cells, theinducible isoform found predominantly in macrophages,and the neuronal isoform found in neurones of the centraland peripheral nervous systems. The neuronal isoform(nNOS) has been purified from rat brain (Bredt & Snyder1990) and the corresponding cDNA has been cloned (Bredtet al. 1991). NOS activity co-localizes with neuronal nicoti-namide adenine dinucleotide phosphate (NADPH)

145Copyright © 1995 Academic Press, Inc.All rights of reproduction in any form reserved.

Correspondence: Scott J. Soifer M-680, Department of Pediatrics,University of California, San Francisco, San Francisco, CA 94143-0106,USA.

Summary

Keywords

diaphorase catalytic activity (Dawson et al. 1991a,b, Hope etal. 1991).

NO is thought to act as a retrograde secondary messenger(Garthwaite 1991).The solubility of NO allows free move-ment across cell membranes, negating the requirement forvesicular release. NO exerts its effect by activating solubleguanylate cyclase (Garthwaite 1991), leading to an increasein the intracellular concentration of cyclic 3',5'-guanosinemonophosphate (cGMP). In a limited population of neu-rones, glutamate activation of the N-methyl-D-aspartate(NMDA) receptor results in an increased concentration ofintracellular calcium, which stimulates NOS activity (East& Garthwaite 1991, Agullo & Garcia 1992, Okada 1992).Thus, excitotoxic activation of the NMDA receptor (pro-duced by hypoxia-ischaemia, for example) can lead toexcessive production and release of NO (Malinski et al.1993), initiating a cascade that may be responsible for thedelayed neurotoxicity seen after acute neurological insults(hypoxia-ischaemia, hypoglycemia, seizures or trauma)(Moncada et al. 1991).

Neuronal death and the neuropathological sequelae asso-ciated with hypoxia-ischaemia occur selectively in well-defined areas of the brain, which vary according to the stageof neuronal development (McDonald & Johnston 1990).For example, the striatum becomes selectively vulnerableonly after birth (Greenamyre et al. 1987). In the adult brain,there are distinct regions vulnerable to hypoxic-ischaemicinjury located in the cortex, cerebellum and hippocampus(Rothman & Olney 1986). These areas are known to berich in the expression of glutamate receptors (Rothman &Olney 1986, Greenamyre et al. 1987).

Recent studies have supported the hypothesis that NO isa mediator of neuronal injury after hypoxia-ischaemia(Ferriero et al. 1990, Dawson et al. 1991a,b, Burke &Baimbridge 1993). Since NO is a gas, it can diffuse freelyout of the neurones producing it, react with superoxide toform peroxynitrite and other free radicals, and induce celldeath in the surrounding neurones of the cerebral cortex,cerebellum and hippocampus (Dawson et al. 1991a,b,Garthwaite 1991, Beckman & Crow 1993, Maiese et al.1993). Inhibition of NOS reduces infarct volume afterhypoxic-ischaemic injury in the rat (Nowicki et al. 1991,Yamamoto et al. 1992). Similarly, hypoxic-ischaemicinjury is reduced in mice deficient in nNOS (Huang et al.1993, Huang et al. 1994). However, it is unclear why theNO-producing neurones are spared after hypoxic-ischaemic injury or NMDA receptor activation (Ferrante etal. 1985, Ferriero et al. 1988).

Since NO is likely to play an important role as a mediatorof excitotoxic neuronal injury after hypoxia-ischaemia, it isimportant to determine the spatial and temporal expressionof nNOS during development and establish whether thisexpression predicts the areas of the brain that are vulnerableto injury from hypoxia-ischaemia. Thus, we have used insitu hybridization, immunohistochemistry and NADPHdiaphorase histochemistry to localize nNOS expression

during the development of the rat brain.We report that theexpresion of nNOS is extensive within the rat brain andthat this expression appears to define the areas of the brainthat are most susceptible to injury from hypoxia-ischaemiaduring development.

Materials and methods

Generation of specific neuronal NOS probe

Sequence comparisons between the 3 isoforms of NOS(endothelial, inducible and neuronal) were performed usingdot-matrix homology plots to determine a region of mini-mal homology between them.This region corresponded toan area within the heme binding domain. Oligonucleotideswere synthesized to allow amplification of this regionwithin the rat neuronal nitric oxide synthase (nNOS)sequence. The sequences of the oligonucleotides were 5'-CTCATCTATGGCGCCAAGCATGCCTGG-3' foroligonucleotide number 1 and 5’-AGTTGT-CACAGTAGTCACGGACGCCGA-3’ for oligonu-cleotide number 2. The region amplified corresponds toamino acids 400–600 of the heme binding domain of thenNOS protein.

Total RNA was prepared from adult rat cerebellum usingthe acid-phenol method (Chomczynski & Sacchi 1987).This was used in RT/PCR reactions (kit from Perkin-Elmer). Random hexamers were used in the first cDNAstrand synthesis and then oligonucleotides numbers 1 and 2were used in the PCR reaction to amplify the nNOSsequence.The cDNA fragment generated (600bp) was thencloned directly into the pCR II vector (Invitrogen),sequenced (Sequenase kit from USB) and then subclonedinto pBluescript KS+ (Stratagene). The plasmids were lin-earized with the appropriate restriction enzyme (Gibco-BRL). Sense and antisense radiolabelled RNA probes weresynthesized using either T3 or T7 RNA polymerases(Boerhinger-Mannheim) with 35S-labelled UTP (NewEngland Nuclear). Southern blot analysis confirmed thespecificity of the nNOS probe (Fig. 1A).

Preparation of a neuronal NOS antiserum

The nNOS cDNA fragment was subcloned in frame intothe pET23a expression vector (Novagen) to overexpress thecorresponding protein. Bacterial extracts were separated on12% NaDodSO4 polyacrylamide gel and the bands visual-ized with KCl/DTT staining (Black et al. 1993).The bandcorresponding to the nNOS protein was excised, minced,lyophilized and injected into New Zealand white rabbits(Animal Pharm. Services) to produce a specific nNOSpolyclonal antiserum.The specificity of the antiserum wasassessed by western blot analysis on protein extracts pre-pared from a variety of tissues. nNOS protein was expressedin foetal and adult cerebellum, COS-7 cells transfected withthe full length rat brain nNOS cDNA (Fig. 1b) and cortex,

146 S. M. Black et al.

Copyright © 1995 Academic Press, Inc., Neurobiology of Disease, 2, 145–155All rights of reproduction in any form reserved.

but not in lung, liver, kidney (data not shown) or COS-7cells transfected with vector alone.

Tissue samples

Sprague-Dawley pregnant and nonpregnant adult rats andrat pups were killed by cervical dislocation or decapitation.Rat embryos (10, 12, 14, 16 and 18 days gestation[E10–18]) and rat brains (1, 6, and 21 days old [P1–21], andadult) were harvested and fixed overnight in 4%paraformaldehyde in phosphate buffered saline (PBS). Forin situ hybridization, rat embryos were dehydrated inethanol, cleared in toluene and embedded in paraffin; ratbrains were placed in 20% phosphate buffered sucrose forcryoprotection and then embedded in OCT (Tissue Tek).For immunohistochemistry and NADPH diaphorase histo-chemistry, rat embryos and brains were cryoprotected.Ten-micron sagittal or coronal sections were mounted onaminoalkylsilane-treated slides and stored at 4°C (for paraf-fin sections) or −70°C (for frozen sections) until used(Nelken et al. 1991, 1992, Soifer et al. 1994).

In situ hybridization

In situ hybridization was performed as described previously(Nelken et al. 1992, Soifer et al. 1994). Serial sections of ratembryos (E10–18) and rat brains (P1–21 and adult) werestudied using antisense probes or the corresponding non-hybridizing sense probe as a control. Rat embryo sectionswere deparaffinized and rat brain sections were thawed.Allsections were then fixed in 4% paraformaldehyde in PBS,treated with proteinase K and acetylated (0.25% aceticanhydride in 0.1 M triethalolamine-HCl [pH 7.5]). Afterwashing in 0.5 × SSC, the sections were covered withhybridization solution (50% deionized formamide, 0.3 M

NaCl, 20mM Tris [pH 8], 5mM EDTA, 1× Denhardt’s,10% dextran sulfate and 10 mM DDT) containing radiola-belled RNA probes (5000 cpm ml−1) and then incubated for15–18 h at 55°C. After hybridization, the sections werewashed for 20 min (2 × SSC, 10 mM DTT and 1 mM

EDTA), treated with RNAse A (20 mg ml−1) for 30 min atroom temperature, then washed at high stringency (50%deionized formamide, 2 × SSC and 0.1 M DTT) for 30 minat 65°C. The sections were washed, dehydrated, dipped in

147nNos in the developing brain


Fig. 1. Specificity of neuronal NOS (nNOS) probe and antiserum. Southern blot analysis (Panel A) showing the specificity of the nNOSprobe.Vectors containing the full length cDNA for rat nNOS, bovine endothelial NOS (bovine eNOS) and murine macrophage NOS(mouse iNOS) were digested to release the full length cDNAs.These were separated on a 0.8% agarose gel, transferred to Hybond andhybridized with a 32P-labelled probe made from the cloned rat 600 bp nNOS cDNA fragment.The full length rat nNOS cDNA wasspecifically detected with no cross-hybridization to either the full length bovine eNOS or murine iNOS.Western blot analysis (Panel B)showing the expression of an approximately 150 kD nNOS protein in fetal and adult cerebellum.Also shown (as a positive control) isnNOS expression by COS-7 cells transfected with the pECE vector containing the full length rat nNOS cDNA (Vector + cDNA) butnot by COS-7 cells transfected with the pECE vector alone (Vector).

photographic emulsion (Ilford) and stored at 4°C.The sec-tions were then developed after approximately 3 weeks ofexposure and counterstained with haematoxylin/eosin.Theresults shown are representative of those obtained from threesets of rat embryos (E10–18) and rat brains (P1-adult). Foreach experiment, four antisense and four sense slides (con-taining one to four sections) at each age were studied. Foreach experiment, new radiolabelled probes were synthe-sized.

Immunohistochemistry

Frozen sections of rat embryos (E10–18) and rat brains(P1–21 and adult) were reacted with the nNOS polyclonalantiserum (dilution 1:1000) overnight at 4°C.The primaryantibody was detected using biotin-conjugated secondary

antibody and avidin-horseradish peroxidase (VectorLaboratories, Burlingame, CA) as described previously(Nelken et al. 1992). Elimination of the primary antibodyserved as a control.

NADPH diophorase histochemistry

NADPH diaphorase histochemistry was performed onfrozen sections of rat embryos (E16) and rat brains (P1 andP6) to determine catalytic activity of the nNOS enzyme.The frozen sections were incubated at 37°C for 2h in asolution containing 0.1 M Tris HCl (pH 8), 0.2 mM NADP,15 mM Na malate, 0.2 mM nitro blue tetrazolium, 1 mMMnCl2 and 0.1% Triton X-100 as described previously(Ferriero et al. 1988).



Fig. 2. nNOS in the embryonic rat brain. Sagittal sections of rat embryo brains at 14 (Panels A and B) and 16 (Panels C and D) dayspost coital (E14 and E16) were hybridized with an antisense probe for rat nNOS. Panel A is a dark field micrograph showing nNOSexpression in the developing brain of an E14 rat.There is expression in the hippocampal formation (open small arrow) and hypothalamus(open large arrow) but no expression in the basal ganglia or cerebellum (arrowheads). Panel B shows the area represented by the whitebox in Panel A. nNOS expression is in the superficial or mantle layer of the developing cerebral cortex. Panel C shows bilaminarexpression of nNOS in the forebrain of an E16 rat. Panel D shows the area represented by the white box in Panel C. nNOS expressionin the cortical and intermediate layers is separated by a narrow plexiform layer (white arrow). Scale bar = 40 µm.

Results

In situ hybridization, immunohistochemistry and NADPHdiaphorase histochemistry detected expression of nNOS inmany areas of the developing rat brain. The first expression,detected at E10 by in situ hybridization, was limited to thedeveloping forebrain (data not shown). At E14, there wasexpression of nNOS in the developing cerebral cortex andhypothalamus and throughout the developing hippocampus(Fig. 2A,B). At E16, nNOS was distributed between thecortical primordial plexiform and intermediate layers (Figs2C,D and 4D). However, no expression could be detectedwithin either the basal ganglia or cerebellum prior to birth(Figs 2A, 4D, 5A and 7). After birth, nNOS expression waslimited to certain populations of neurones in the cerebralcortex (Figs 3 and 4B,C). There was also a transient highlevel of nNOS expression in the basal ganglia during thefirst week postpartum which became more discretely local-ized as the brain matured (Figs 4E,F and 5B). During thisperiod, nNOS expression in the hippocampus remainedconstant (Figs 4H,I and 6). nNOS expression in the cere-bellum increased after birth to P21 (Fig. 7).The followingsections detail the expression of nNOS within the areas thatare found to be selectively vulnerable to injury fromhypoxia-ischaemia during the development of the rat brain.

nNOS expression during the development of the cerebralcortex

Definitive expression of nNOS was detected within thedeveloping cerebral cortex at E14 (Fig. 2A). nNOS expres-sion was homogeneously distributed within the mantlelayer of the cortical anlage (Fig. 2B). By E16, the expressionof nNOS within the neocortex was apparent in two layers,separated by a narrow plexiform layer of cells that lackedexpression (Fig. 2C,D). At E18, nNOS expression wasmore diffuse within the cortical plate, whereas there wasintense expression in the intermediate zone.

At P1, a more homogeneous pattern of expression wasseen throughout the cortex (Fig. 3A). By P6, discrete areasof nNOS expression were evident (Fig. 3B), and betweenP21 and the adult, the expression of nNOS changed, suchthat in the adult only a few neurones expressing NOS werefound and these were scattered throughout the deep layersof the cortex (Fig. 3C). Catalytic activity of nNOS in thecerebral cortex, represented by NADPH diaphorase histo-chemical staining, was seen at E16 in the small neuroblastsof the superficial cortical layer in a pattern that corre-sponded to that of nNOS mRNA localization by in situhybridization (Fig. 4A). With cortical development (Fig.4B), the morphology of the neuroblasts became distinct.



Fig. 3. nNOS expression in the cerebral cortex of the rat after birth.Sagittal sections of rat brains at 1 (PanelA) and 6 days (Panel B) afterbirth (P1 and P6) and from an adult rat (Panel C) were hybridized with an antisense probe for rat nNOS. Panel A, a high powermagnification dark field micrograph, shows the expression of nNOS at P1 is throughout the cortical layers (WM, white matter). Panel Bshows discrete areas of the cerebral cortex expressing nNOS at P6. Panel C shows large discrete neurones in the adult brain expressionnNOS. Scale bar = 40 µm.

Indeed, by P6 there was golgi-like NADPH diaphorasestaining of the large neurones (Fig. 4C).

nNOS expression during the development of the basal ganglia

No expression of nNOS prior to birth in the caudate-puta-men was detected by in situ hybridization (Fig. 5A),immunohistochemistry (data not shown) or NADPHdiaphorase histochemistry (Fig. 4D) At P1, nNOS mRNAexpression was seen in clumps of neurones throughout thenuclei corresponding to the NADPH diaphorase positiveneurones in the nuclei (Figs 4E and 5B). Golgi-like stainingof large interneurones was seen at P6 (Fig 4F). By P21,nNOS mRNA expression was more discrete (Fig. 5C).There was no difference in the morphological pattern ofexpression of nNOS between P21 and the adult. In the

adult, nNOS protein expression was localized to the largeinterneurones (Fig. 5D) which also contain somatostatin,neuropeptide Y and NADPH diaphorase catalytic activity(Ferrante et al. 1985).

nNOS expression during the development of the hippocampus

At E16, the archicortex (the anlage of the hippocampus)stained homogeneously in a manner similar to that of thedeveloping cerebral cortex (Figs 2A, C, 4G and 6A) At P1,nNOS expression was homogeneous throughout the hip-pocampal zones (Fig 6B,C). NADPH diaphorase histo-chemical staining of neurones was seen in a pattern identicalto nNOS mRNA expression by in situ hybridization (Fig.4H). At P6, nNOS expression was seen in the CA1, CA2and CA3 layers (Fig. 6E). Immunohistochemical analysis



Fig. 4. NADPH diaphorase histochemical staining in rat brain before and after birth. Coronal sections of rat brains at E16 (Panel A, Dand G). P1 (Panel B, E and H) and P6 (Panel C, F and I) were stained for NADPH diaphorase catalytic activity. In the cerebral cortex(Panel A–C), there is NADPH diaphorase staining of neuroblasts (arrow) at E16 (Panel A) small neurones (arrow) at P1 (Panel B) andlarge interneurones at P7 (Panel C) in a pattern similar to nNOS mRNA expression demonstrated by in situ hybridization at similar agesScale bar = 100 µm. In the basal ganglia (Panels D–F), there is no NADPH diaphorase staining prior to birth (Panel D), but there areclumps of NADPH diaphorase positive neurones (arrow) at PI (Panel E) and single discrete NADPH diaphorase positive interneuronesat P6 (Panel F) Scale bars = 10 µm. In the hippocampus (Panels G–I), there is dense nonspecific NADPH diaphorase staining (openarrows) prior to birth in the hippocampal unlarge (Panel G) and specific NADPH diaphorase staining throughout CA1, CA3 and DG(superior blade of the dentate granule cell layer) at P1 (Panel H) and P6 (Panel I) in a pattern similar to nNOS mRNA expressiondemonstrated by in situ hybridization at similar ages Scale bars = 200 µm.

revealed nNOS protein expression in similar layers (Fig.6D). NADPH diaphorase catalytic activity was seen in neu-rones expressing nNOS mRNA (Fig. 4I). In the adult hip-pocampus, nNOS expression was highest in CA1 followedby CA3, with no expression detected in CA2 (Fig. 6F).

nNOS expression during the development of the cerebellum

The developing cerebellar cortex did not express nNOSuntil P1 (Fig. 7A). At P6, nNOS was expressed in the pos-

terior folia (Fig. 6B).This zone of expression continued toexpand during postnatal cerebellar development until theadult expression pattern was achieved by P21 (Fig. 6C).Expression of nNOS during the first postnatal week waslocated primarily in the molecular/Purkinje cell layer andin the incipient granular cell layer (Fig. 6B). The externalgranular layer and the developing white matter exhibitedno nNOS expression at this time (Fig. 6B).Immunohistochemical analysis of the latter developmentalstages (P21 to adult) revealed nNOS expression in basket,



Fig. 5. nNOS expression in the basal ganglia of the rat before and after birth. Sagittal sections of E16 rat embryo brains (Panel A) and ofP1 (Panel B), P21 (Panel C) and adult (Panel D) rat brains were hybridized with an antisense probe for rat nNOS mRNA (Panel A–C) orstained with an nNOS specific antibody (Panel D). Panel A shows no nNOS expression prior to birth. Scale bar = 60 µm. Panel Bdemonstrates nNOS expression throughout the nucleus. Panel C shows that the expression of nNOS is more discrete by P21. Panel Dshows large interneurones in the adult brain expressing nNOS protein. Scale bar = 30 µm for Panels B–D.

stellate and granule cells (Fig 6E) No nNOS expressioncould be detected in the other cell types within the cerebel-lum.

Discussion

Cerebral hypoxia–ischaemia remains a major cause of acuteperinatal brain injury initiated when either placental or pul-monary gas exchange is compromised. This leads to neu-ronal death and, ultimately, to neurologic dysfunction(cerebral palsy, mental retardation or epilepsy), which variesaccording to the stage of neuronal development duringwhich the injury occurs.Although the processes that lead to

neuronal death after hypoxic–ischaemic injury are likely tobe multifactorial, evidence suggests that a key component isan increased production and release of NO. Thus, anunderstanding of the developmental expression of nNOSmay lead to insights into how hypoxia–ischaemia producesselective neuronal loss in different brain regions at differentdevelopmental stages.

Our results show that the expression of nNOS within thedeveloping brain is both extensive and variable. nNOSexpression accurately predicts the regions of the brain thatcontain the selectively vulnerable neuronal populations thatdie after the excitotoxic injury of hypoxia-ischaemia.Thus,the expression of nNOS throughout the developing cere-



Fig. 6. nNOS expression in the hippocampus of the rat before and after birth. Sagittal sections of E16 rat embryo brains (Panel A) andP1 (Panels B and C), P6 (Panels D and E) and adult (Panel F) rat brains were hybridized with an antisense probe for rat nNOS mRNA(Panel A–C, E, F) or with an nNOS specific antibody (Panel D). Panel A shows that the anlage of the hippocampus (box) has diffusenNOS expression. Scale bar = 50 µm. Panel B shows homogenous nNOS expression throughout the hippocampal formation. Scale bar= 100 µm Panel C, a higher magnification of Panel B, shows nNOS expression in the pyramidal cell layer of CA1, CA2, CA3 andCA3c. Scale bar = 50 µm. Panel D shows nNOS immunoreactive cells in CA3. Scale bar = 25 µm. Panel E shows nNOS expressionthroughout the hippocampal formation at P6 Scale bar = 100 µm Panel F shows nNOS expression in the adult brain in CA1 (openarrow) and CA3 (solid arrow) but not in CA2. Scale bar = 100 µm.

bral cortex delineates the sensitivity of this region tohypoxic-ischaemic injury. Our results show that the great-est expression of nNOS during development is in the cere-bral cortex at E14–18 when the majority of neuronesexpress nNOS. At this stage of development, the cerebralcortex is at the greatest risk from hypoxic-ischaemic injury(Williams et al. 1990). However, in the newborn (P1–6),the region of the cerebral cortex that is vulnerable tohypoxic-ischaemic injury is limited to the deeper layers in acolumnar pattern (Ferriero et al. 1988), corresponding tothe regions that express nNOS postnatally in our study. Inthe developing rat brain, the regions that express nNOS

correspond to the regions that also express NMDA recep-tors (McBain & Mayer 1994). Therefore, the presence ofnNOS appears to correlate with regions of excitatory aminoacid neurotoxicity both in vivo and in vitro (Ferriero et al.1990, Dawson et al. 1991a,b)

There is no prenatal nNOS mRNA or protein expres-sion or specific neuronal NADPH diaphorase catalyticactivity in the basal ganglia.This region is relatively invul-nerable to hypoxic-ischaemic injury prior to birth and onlybecomes vulnerable after birth. In the developing basal gan-glia, synaptosomal uptake of glutamate after hypoxic-ischaemic injury increases dramatically, especially during



Fig. 7. nNOS expression in the cerebellum of the rat after birth. Sagittal sections of P1 (Panel A), P6 (Panel B), P21 (Panel C) and adultrat brains (Panels D and E) were dybridized with an antisense probe for rat nNOS mRNA (Panel A-D) or with an nNOS specificantibody (Panel E) Panel A shows that there is nNOS expression in the cerebellar anlage at P1 Scale bar = 25 µm. In Panel B, nNOSexpression is in the superficial layers of the posterior folia (ML/PL, molecular layer/Purkinie cell layer;WM, white matter) Scale bar =100 µm Panel C shows nNOS expression in the molecular layer, but not in the developing Purkinje cells and white matter by P21. Scalebar = 200 µm. Panel D shows that the expression of nNOS in the adult cerebellum is similar to the expression at P21, with expression inthe molecular and granular cell layers. Panel E shows that these cells in the cerebellum also express nNOS protein Scale bar = 50 µm forPanels D and E.

the second week of life (Campochiaro & Coyle 1978,Ferriero et al. 1988). The increase in glutamate uptake isparalleled by an increase in nNOS expression and mayexplain the selective vulnerability of the basal ganglia tohypoxia-ischaemia in the early postnatal rat brain. Previousstudies in the embryonic mouse have shown NADPHdiaphorase activity in the incipient corpus striatum (Derer& Derer 1993), a difference that may be due to species dif-ferences.

nNOS expression in the hippocampus is of great interestbecause this is a region that demonstrates enhanced sensitiv-ity to postnatal excitotoxic neuronal injury (Tombaugh &Sapolsky 1990, Patel et al. 1993). Our results show that themost vulnerable regions in the hippocampus are againdelineated by their expression of nNOS. There is transientexpression of nNOS postnatally that parallels the increase inthe susceptibility of the hippocampus to NMDA toxicity.During the second postnatal week, the pyramidal cell layerin area CA1 becomes more sensitive to excitatory aminoacid stimulation (Hamon & Heinemann 1988). Our resultsdemonstrate a transient increase in nNOS expression,which may account for the narrow window of susceptibilityof this area to hypoxic-ischaemic injury. In the adult hip-pocampus, the regions corresponding to CA3 and CA1 arevulnerable, while CA2 is relatively spared afterhypoxia-ischaemia. Again, this pattern corresponds wellwith the distribution of nNOS expression that we havedetected, with the highest expression detected in layersCA1 and CA3, and no expression in CA2.

There is also no expression of nNOS in the cerebellaranlage. In this region, nNOS expression is present only afterthe first week of life, explaining the relative invulnerabilityof the cerebellum to hypoxic-ischaemic injury prior to thisstage of development. After the second week of life, thecerebellum becomes increasingly vulnerable to neuronaldeath caused by NMDA receptor activation (Watanabe etal. 1992). During this period of postnatal cerebellar devel-opment, marked changes occur in the distribution ofmRNAs encoding for the various NMDA-receptor sub-units. For example, before synapse formation, migratinggranule cells show distinct NMDA receptor-possessingproperties similar to many central neurones, whereasmature granule cells express a different type of NMDAreceptor (Farrant et al. 1994).The pattern of appearance ofnNOS expression in the postnatal cerebellum parallels thedevelopment of the NMDA receptors in these cells. In ratcerebellar slices, different cell types exhibit changing sensi-tivity to excitatory amino acid receptor agonists as postnataldevelopment proceeds (Garthwaite 1991). This changingsusceptibility correlates with the increased regional and cel-lular expression for nNOS that we have demonstrated.

In conclusion, we have observed nNOS expression inmany areas of the developing rat brain through the use of insitu hybridization, immunohistochemistry and NADPHdiaphorase histochemistry. A similar pattern of nNOS pro-tein expression has recently been described (Bredt & Snyder

1994). The maturation of the expression corresponds wellwith the regions of selectivity vulnerable neuronal popula-tions that die after excitotoxic stimuli such ashypoxia-ischaemia. This developmental profile lends sup-port to the following hypothesis: that regional changes inneuronal phenotypes predict patterns of vulnerability, andexplain why, neuronal and glial cell damage to the develop-ing nervous system differs from damage to the mature ner-vous system after hypoxia-ischaemia. The nNOS expres-sion in these regions provides provocative preliminary datato support the hypothesis that NO production is a criticalstep in the pathophysiology of cell death afterhypoxia-ischaemia in the developing central nervous sys-tem.

Acknowledgements

This work was supported in part by a NATO Grant (S.M.),by the University of California San Francisco REAC grant(D.M.F.) and by grant HL 35518 from the National HeartLung and Blood Institute (S.J.S.).

References

Agullo L. & Garcia A. (1992) Different receptors mediate stimula-tion of nitric oxide-dependent cyclic GMP formation in neu-rons and astrocytes in culture. Biochem. Biophys. Res.Commun. 182, 1362–1368.

Beckman J.S. & Crow J.P. (1993) Pathological implications ofnitric oxide, superoxide and peroxynitrite formation. Biochem.Soc.Trans. 21, 330–334.

Black S.M., Szklarz G.D., Harikrishna J.A., Wolf C.R., Lin D. &Miller W.L. (1993) Regulation of proteins in the cholesteroside-chain cleavage system in JEG-3 and Y1 cells. Endocrinology132, 539–545.

Bredt D.S., Glatt C.E., Huang P.M., Fotuhi M., Dawson T.M. &Snyder S.H. (1991) Nitric oxide synthase protein and mRNAare discretely localized in neuronal populations of the mam-malian CNS together with NADPH diaphorase. Neuron 7,615–624.

Bredt D.S. & Snyder S.H. (1990) Isolation of nitric oxide syn-thetase, a calmodulin-requiring enzyme. Proc. Natl Acad. Sci.USA 87, 682–685.

Bredt D.S. & Snyder S.H. (1994) Transient nitric oxide synthaseneurons in embryonic cerebral cortical plate, sensory ganglia,and olfactory neurons. Neuron 13, 301–313.

Burke R.E. & Baimbridge K.G. (1993) Relative loss of the striatalstriosome compartment, defined by calbindin-D28kimmunostaining, following developmental hypoxic-ischaemicinjury. Neuroscience 56, 305–315.

Campochiaro P. & Coyle J.T. (1978) Ontogenetic developmentof kainate neurotoxicity: correlates with glutamatergic inner-vation. Proc. Natl Acad. Sci. USA 75, 2025–2029.

Chomcyznski P. & Sacchi N. (1987) Single-step method of RNAisolation by acid guanidinium thiocyanate-phenol-chloroformextraction. Anal. Biochem. 162, 156–159.

Dawson T.M., Bredt D.S., Fotuhi M., Huang P.M. & SnyderS.H. (1991a) Nitric oxide synthase and neuronal NADPH



diaphorase are identical in brain and peripheral tissues. Proc.Natl Acad. Sci. USA 88, 7779–7781.

Dawson V.L., Dawson T.M., London E.D., Bredt D.S. & SnyderS.H. (1991b) Nitric oxide mediates glutamate neurotoxicity inprimary cortical cultures. Proc. Natl Acad. Sci. USA 88,6368–6371.

Derer P. & Derer M. (1993) Ontogenesis of NADPH-diaphoraseneurons in the mouse forebrain. Neurosci. Lett. 152, 21–24.

East S.J. & Garthwaite J. (1991) NMDA receptor activation in rathippocampus induces cyclic GMP formation through the L-arginine-nitric oxide pathway. Neurosci. Lett. 123, 17–19.

Farrant M., Feldmeyer D., Takahashi T. & Cull-Candy S.G.(1994) NMDA-receptor channel diversity in the developingcerebellum. Nature 368, 335–339.

Ferrante R.J., Kowall N.W., Beal M.F., Richardson E.P. Jr. BirdE.D & Martin J.B. (1985) Selective sparing of a class of striatalneurons in Huntington’s disease. Science 230, 561–563.

Ferriero D.M., Arcavi L.J., Sagar S.M., McIntosh T.K. & SimonR.P. (1988) Selective sparing of NADPH-diaphorase neuronsin neonatal hypoxia-ischemia. Ann. Neurol. 24, 670–676.

Ferriero D.M., Arcavi L.J. & Simon R.P. (1990) Ontogeny ofexcitotoxic injury to nicotinamide adenine dinucleotide phos-phate diaphorase reactive neurons in the neonatal rat striatum.Neuroscience 36, 417–424.

Garthwaite J. (1991) Glutamate, nitric oxide and cell-cell sig-nalling in the nervous system. Trends Neurosci. 14, 60–67.

Greenamyre T., Penney J.B., Young A.B., Hudson C., SilversteinF.S. & Johnston M.V. (1987) Evidence for transient perinatalglutamatergic innervation of globus pallidus. J. Neurosci. 7,1022–1030.

Hamon B. & Heinemann U. (1988) Developmental changes inneuronal sensitivity to excitatory amino acids in area CA1 ofthe rat hippocampus. Brain Res. 466, 286–290.

Hope B.T., Michael G.J., Knigge K.M. & Vincent S.R. (1991)Neuronal NADPH diaphorase is a nitric oxide synthase. Proc.Natl Acad. Sci. USA 88, 2811–2814.

Huang P.L., Dawson T.M., Bredt D.S., Snyder S.H. & FishmanM.C. (1993) Targeted disruption of the neuronal nitric oxidesynthase gene expression Cell 75, 1273–1286.

Huang Z., Huang P.L., Panahian N., Dalkara T., Fishman M.C.& Moskowitz M.A. (1994) Effects of cerebral ischemia inmice deficient in neuronal nitric oxide synthase. Science 265,1883–1885.

Maiese K., Boniece I.R., Skurat K. & Wagner J.A. (1993) Proteinkinases modulate the sensitivity of hippocampal neurons tonitric oxide toxicity and anoxia. J. Neurosci. Res. 36, 77–87.

Malinski T. Bailey F., Zhang Z.G. & Chopp M. (1993) Nitic

oxide measured by a porphyrinic microsensor in rat brain aftertransient middle cerebral artery occlusion. J. Cereb. Blood FlowMetab. 13, 355–358.

McBain C.J. & Mayer M.L. (1994) N-methyl-D-aspartic acidreceptor structure and function. Physiol. Rev. 74, 723–760.

McDonald J.W. & Johnston M.V. (1990) Physiological andpathophysiological roles of excitatory amino acids during cen-tral nervous system development. Brain Res. Rev. 15, 41–70.

Moncada S., Palmer R.M.J. & Higgs E.A. (1991) Nitric oxide:physiology, pathophysiology and pharmacology. Pharmacol.Rev. 93, 109–142.

Nelken N.A., Coughlin S.R., Gordon D. & Wilcox J.N. (1991)Monocyte chemoattractant protein-1 human atheromatousplaques. J. Clin. Invest. 88, 1121–1127.

Nelken N.A., Soifer S.J., O’Keefe J., Vu T.K., Charo I.F. &Coughlin S.R. (1992) Thrombin receptor expression in nor-mal and atherosclerotic human arteries. J. Clin. Invest. 90,1614–1621.

Nowicki J.P., Duval D., Poignet H. & Scatton B. (1991) Nitricoxide mediates neuronal death after focal cerebral ischemia inthe mouse. Euro. J. Pharmacol. 204, 339–340.

Okada D. (1992) Two pathways of cyclic GMP productionthrough glutamate receptor-mediated nitric oxide synthesis. J.Neurochem 59, 1203–1210.

Patel M.N., Yim G.K. & Isom G.E. (1993) N-mehtyl-D-asparatate receptors mediate cyanide-induced cytotoxicity inhippocampal cultures. Neurotoxicology 14, 5–40.

Rothman S. & Olney J. (1986) Glutamate and the pathophysiol-ogy of hypoxic-ischemic brain damage. Ann. Neurol. 19,105–111.

Soifer S.J., Peters K.G., O’Keefe J. & Coughlin S.R. (1994)Disparate temporal expression of the prothrombin and throm-bin receptor genes during mouse development. Am. J.Pathol. 144, 60–69.

Tombaugh G.C. & Sapolsky R.M. (1990) Mechanistic distinc-tions between excitotoxic and acidotic hippocampal damagein an in vitro model of ischemia. J. Cereb. Blood FlowMetab. 10, 527–535.

Watanabe M., Inoue Y., Sakimura K. & Mishina M. (1992)Developmental changes in distribution of NMDA receptorchannel subunit mRNAs. Neuroreport 3, 1138–1140.

Williams C.E., Gunn A.J., Synek B. & Gluckman P.D. (1990)Delayed seizures occurring with hypoxic-ischemicencephalopathy in the fetal sheep. Pediatr. Res. 27, 561–565.

Yamamoto S., Golanov E.V., Berger S.B. & Reis D.J. (1992)Inhibition of nitric oxide synthesis increases focal ischemicinfarction in rat. J. Cereb Blood Flow Metab. 12, 717–726.



Expression of neuronal nitric oxide synthase corresponds to regions of selective vulnerability to...

Documents

Transcript of Expression of neuronal nitric oxide synthase corresponds to regions of selective vulnerability to...