protection of striatal neurons by joint blockade of d1 and d2 receptor subtypes in an in vitro model...

Experimental Neurology 176, 229–236 (2002)doi:10.1006/exnr.2002.7926

Protection of Striatal Neurons by Joint Blockade of D1 and D2Receptor Subtypes in an in Vitro Model of Cerebral Hypoxia

Sue Davis,* Jonathan Brotchie,† and Ioan Davies‡*Wolfson Research Centre, Institute for Aging and Health, University of Newcastle upon Tyne, United Kingdom; †School of Biological

Sciences, University of Manchester; and ‡Schools of Medicine and Biological Sciences, University of Manchester, United Kingdom

Received April 2, 2001; accepted March 27, 2002

Massive increases in extracellular dopamine havebeen reported in the ischemic rodent striatum, impli-cating this neurotransmitter in toxic events. We haveexamined whether dopamine receptor antagonists areprotective against hypoxic insult, using brain slicescontaining the rostral striatum obtained from adultmale C57/BLIcrfat mice. Slices were subjected in vitroto 20 min nitrogen hypoxia, with or without additionof: (i) 50 �M haloperidol (D2 receptor antagonist andsigma ligand), (ii) 10 �M SCH23390 (selective D1 recep-tor antagonist), (iii) 10 �M eticlopride (selective D2receptor antagonist), (iv) 10 �M SCH23390 and 10 �Meticlopride in combination, and (v) 10 �M MK-801(noncompetitive NMDA receptor antagonist). Subse-quently, slices were reoxygenated, fixed 2 h postinsult,and processed for light microscopy. Damage was as-sessed by calculating pyknotic profiles as a percentageof total neuronal profiles present. No pyknotic profileswere detected in normoxic control tissue, but this phe-notype predominated in most slices subject to hypoxiaalone (60.1 � 30.6% pyknotic profiles). Marked protec-tion was produced by haloperidol (7.1 � 7.6%, P �0.002), MK-801 (8.6 � 6.9%, P � 0.007), and the com-bined application of SCH23390 and eticlopride (5.9 �9.4%, P � 0.001). No protection was demonstrated forSCH23390 or eticlopride when applied separately.These data suggest that hypoxic damage in the rostralmouse striatum is mediated via NMDA, D1, and D2receptors. Protection against hypoxic damage by do-pamine receptor antagonists requires the combinedblockade of both classes of dopamine receptor.© 2002 Elsevier Science (USA)

Key Words: dopamine; striatum; hypoxia; brain slice;neuroprotection.

INTRODUCTION

Significant increases in extracellular dopamine havebeen recorded in the striatal brain region in in vivomodels of global ischemia (16), focal ischemia (22),cerebral hypoxia (35), and cardiac arrest (47). Wheresuch increases have been prevented by the destruction

229

of the nigral dopaminergic neurons which inner-vate the striatum, striatal tissue has been protectedagainst ischemic insult (10, 15). Thus, increased levelsof dopamine have been implicated in mechanisms ofhypoxic–ischemic damage in rodent striatum. Althoughthe mechanisms that mediate dopamine toxicity in theischemic brain are not fully understood, evidence existsfor two major possibilities. First, the pro-oxidant proper-ties of catecholamine neurotransmitters are well known,and oxidative damage has been implicated in mecha-nisms of dopamine toxicity following its direct applica-tion to central nervous system (CNS) neurons in bothin vivo and in vitro models under normoxic conditions(20, 23, 32). Dopamine-induced oxidative damage hasalso been detected following intrastriatal injection ofthe mitochondrial inhibitor malonate, an insult whichsets up conditions analagous to those found in ischemicbrain (13). Second, transmission involving dopaminereceptors plays a role in neuronal death in the ischemicbrain, since dopamine receptor antagonists are neuro-protective against both ischemic and excitotoxic insults(6, 19, 48). Thus, a key question is what is the identityof the dopamine receptor subtypes mediating this tox-icity? In rat striatum D2 class receptors have beenimplicated since the D2 antagonists spiperone andsulpiride protected against a 20-min global ischemicinsult (19) and against excitotoxicity following intra-striatal injection of the NMDA agonist quinolinic acid(48), respectively. Some controversy exists concerningthe role of D1 class receptors. The D1 receptor antag-onist SCH23390 reduced infarct volume following per-manent focal cerebral ischemia in the mouse (6). Inaddition, striatal neurons expressing D1 receptorswere preferentially lost after a global ischemic insult inthe rat (17), suggesting that dopamine receptors in thisclass also mediate ischemic damage. However, a studyby Globus and coworkers failed to observe protection ofstriatal neurons by SCH23390 following a 20-minglobal ischemic insult in the rat (18). Further complex-ity is introduced by the study of Hashimoto et al. (19),which found not only that SCH23390 failed to protectstriatal neurons against ischemic insult, but that the

0014-4886/02 $35.00© 2002 Elsevier Science (USA)

All rights reserved.

D1 antagonist, when coadministered with the D2 an-tagonist spiperone, negated the neuroprotection pro-vided by spiperone when given separately. Hashimotoand coworkers concluded that the toxic effects of dopa-mine in the rodent striatum exposed to ischemia aremediated by receptors in the D2 class, since processesmediating neurotoxicity are positively inhibited by theactivation of receptors in the D1 class.

The work presented in the current paper reexaminedthe question of which dopamine receptor subtypes me-diate toxicity in the hypoxic striatum, using an in vitropreparation, the corticostriatal brain slice. The roles ofthe dopamine receptor subtypes were examined usingSCH23390, the potent D2 receptor antagonist andsigma ligand haloperidol, and the selective D2 receptorantagonist eticlopride. In addition, the involvement ofthe NMDA receptor in the development of hypoxicdamage in the rostral striatum was investigated usingthe noncompetitive NMDA receptor antagonist MK-801. We show that hypoxic damage in this brain regionis mediated via NMDA receptors and via both classesof dopamine receptor.

MATERIALS AND METHODS

Animals. Three-month-old C57BL/Icrfat male micewere obtained from the Biological Services Unit at theUniversity of Manchester. These animals are barrier-reared and maintained to FELASA guidelines for mi-crobiological status. All experiments were carried outin strict compliance with Home Office guidelines.

Buffer. A standard buffer was used for perfusion,preparation of slices, and incubation, comprising NaCl,118 mM; MgSO4, 1.2 mM; KCl, 4.8 mM; glucose, 11mM; CaCl2, 2mM; NaHCO3, 24.8 mM; and KH2PO4, 0.5mM. The solution was filtered prior to use through a0.22-�m filtration disc (Millipore Corporation, Bedford,UK) and preequilibrated with a gas mixture compris-ing 95%O2:5%CO2. The buffer was prepared usingAnalaR-grade chemicals (BDH Laboratory Supplies,Poole, UK).

Preparation and incubation of brain slices. Micewere deeply anesthetized in 3% halothane in a mixtureof O2:NO2, (1:1, v/v) and then perfused transcardiallywith the ice-cold oxygenated buffer (�1°C) describedabove. Perfusion was carried out with the animalplaced on a metal tray set on ice and was ceased whenthe temperature of perfusate leaving the heart hadfallen to 15°C. The animal was then decapitated andthe brain removed into buffer chilled to 15°C. Thebrain was prepared for sectioning by removing thecerebellum and caudal aspect of the forebrain andglued on its caudal surface to the chuck of a Vibroslicetissue slicer (Camden Instruments, Sileby, Leics, UK).The block was divided sagittally so that each pass ofthe blade would produce two paired coronal slices, onefrom each hemisphere. Tissue was removed from the

rostral end of the forebrain until the rostral end of thestriatum was just uncovered, after which a further 200�m tissue was removed, and one pair of 400-�m slicescontaining rostral striatum and associated cortex wascollected for experimental use. The buffer in the slicingbath was maintained at 15°C during slice preparation.Brain slices were incubated in a closed system (Fig. 1)in a shaking water bath (model SB-16, Techne, Cam-bridge, UK).

Experimental studies. Five sets of experiments ex-amined the neuroprotective effects of 50 �M haloperi-dol (n � 6), 10 �M MK-801 (n � 5), 10 �M SCH23390(n � 8), 10 �M eticlopride (n � 8), and 10 �MSCH23390 � 10 �M eticlopride (n � 6), against a20-min hypoxic insult. MK-801 (10 �M) and haloperi-dol (50 �M) have been shown to offer marked neuro-protection in in vitro models of hypoxia–ischemia (4,38, 30), as 10 �M SCH23390 and 10 �M eticlopridehave been shown to block dopamine-mediated effects inrat striatal brain slices (3, 46). After preparation, sliceswere allowed a recovery period of 60 min at 25°C, afterwhich bath temperature was raised over 15 min to afinal incubation temperature of 34°C. Hypoxia was im-posed immediately after the period for slice recovery asdescribed below. Drugs were added to the buffer 10 minbefore the imposition of hypoxia and left on the tissuefor the remainder of the experimental period. In allcases, brain slices were reoxygenated after the hypoxic

FIG. 1. System for brain slice incubation. Chambers consist of100-mL straight-sided “Duran” glass beakers (Fisher Scientific,Loughborough, UK) cut down to a height of 5 cm. Lids cast from RTVRubberise (R. S. Components, Stockport, UK), bearing holes for gasentry and exit, seal the chamber. Slices are incubated between twolayers of soft nylon netting on a stainless steel platform. The stain-less steel base and top plates indicated each have a center holepunched. Buffer level is set at �1 mm above the platform. Gas isbubbled directly into incubation buffer via Monoject U-100 insulinneedles (bore, 0.4 mm) (Sherwood Medical, St. Louis, MO) andvented via the trimmed barrel of a 1-mL rigid plastic syringe (Com-ing Costar, High Wycombe, Bucks, UK).

230 DAVIS, BROTCHIE, AND DAVIES

period and incubated for a further 2 h. The capacity of50 �M haloperidol to protect against a 30-min hypoxicinsult was also assessed (n � 7). Finally, slices wereincubated under normoxic conditions for an experi-mental period equivalent to that in which 20-min hyp-oxia was imposed, in order to assess neuronal preser-vation in optimal conditions (n � 5). Incubation beakerswere shaken gently throughout the experiments, todiscourage collection of buffer on the underside of thebeaker lid. Gas flow rates of 120 mL/min were used.

Imposition of hypoxia. In the set of experimentsdescribed above, hypoxia was imposed by switching thegas supply from the standard mixture of 95%O2:5%CO2

(BOC Gases, Worsley, Manchester, UK), to 95%N2:5%CO2 (BOC Special Gases, Guildford, Surrey, UK). Inan additional set of experiments (n � 5), the degree ofhypoxia produced by this method was checked by mea-suring percentage oxygen saturation in buffer with aJenway 9200 dissolved O2 meter (Spectronic AnalyticalInstruments, Leeds, UK). Before obtaining measure-ments, buffer was equilibrated with 95%O2:5%CO2 for45 min at 34°C, after which hypoxia was imposed for20 min.

Histology/morphological analysis. Slices were fixedovernight at 4°C in 4% formaldehyde in 0.1 M phos-phate buffer and then embedded in JB-4 plastic em-bedding medium (Warrington, PA). Sections (3 �m)were prepared on a LKB Historange microtome andstained with 0.1% toluidine blue in 0.1% borax. Onesection from the central 100 �m of each slice wasselected for analysis, which was carried out on a LeitzOrtholux light microscope at a magnification of �70(oil immersion), using an eyepiece grid of area 130 �m2.Nine fields in each section were selected for analysis bya systematic procedure designed to sample the fullexpanse of the striatum. Neuronal preservation wasassessed by classification of neuronal profiles accordingto morphological criteria, profiles being designated as(i) normal, (ii) pyknotic, and (iii) unclassified. Normalprofiles exhibited a palely staining nucleus, prominent,intensely staining nucleoli, differential staining of thenucleus and cytoplasm, with distinct and regular nu-clear and cellular profiles. Pyknotic profiles displayedan isolated pyknotic nucleus containing clumps of chro-matin, set in a vacuole separating it from surroundingneuropil. These profiles were considered to be neuronalbecause of their exclusive location within gray matterand their predominance in this region in slices subjectto hypoxia alone, just as the normal neuronal pheno-type predominated in oxygenated tissue. The nuclearalterations shown by this phenotype have been identi-fied by light and electron microscopy as a morphologi-cal characteristic of neurons which had undergone hyp-oxia–ischemia-induced edematous degeneration in therat (24). Unclassified profiles comprised those profilesidentifiable as neurons by criteria of size and the pres-

ence of nucleoli, but showing severe morphological ab-normality and/or increased density of stain, and thosestructures large enough to represent dead neurons, butlacking the morphological detail necessary for identifi-cation as such. Profiles in both the pyknotic and un-classified classes were considered to represent dead orirreversibly damaged cells. Profiles exhibiting the mor-phological and staining properties of glia were ex-cluded from the count. A minimum of 100 profiles werecounted in each section. Analysis was carried out un-blinded.

Drugs. MK-801 and SCH23390 hydrochloride 1/4H2O were obtained from Tocris Cookson (Langford,Bristol, UK), eticlopride hydrochloride from ResearchBiochemicals International (Natick, MA), and haloper-idol from Sigma, (St. Louis, MO), MK-801, SCH23390HCl, and eticlopride HCl were made up in distilledwater and stored at �20°C as 10 mM stock. Haloperi-dol was made up as 25 mM stock in 100% ethanolimmediately prior to use. The final concentration ofethanol in incubation buffer in experiments involvinghaloperidol was 0.2%.

RESULTS

Measurements of dissolved oxygen during hypoxia.When the buffer was perfused with 95%N2:5%CO2, thedissolved O2 concentration fell rapidly, reaching valuesclose to zero within 10 min (Fig. 2).

Morphological characteristics of incubated tissue.Normal profiles were observed in normoxic slices (Fig.3a), and in slices protected against hypoxia by MK-801

FIG. 2. Dissolved oxygen in incubation buffer, measured in fiveseparate experiments. Buffer was equilibrated with 95%O2:5%CO2

for 45 min at 34°C, after which 20 min 95%N2:5%CO2 hypoxia wasimposed. Buffer was then reoxygenated. Gas flows were maintainedat 120 � 5 mL/min. As the upper limit of the instrument (199%dissolved O2) was rapidly reached when gassing with 95%O2:5%CO2,actual dissolved O2 was only recorded during hypoxia. Data pointsrepresent mean � SD. *, Dissolved O2 � 199%.

231PROTECTION OF THE HYPOXIC STRIATUM BY D1/D2 RECEPTOR BLOCKADE

and dopamine receptor antagonists (Figs. 3c–3e). How-ever, even in optimally preserved tissue (Fig. 3a) a cer-tain degree of vacuolization of the neuropil was evi-dent. Perivascular vacuoles were invariably present,and perineuronal vacuoles were associated with a mi-nority of normal neuronal profiles. Such swelling wasminor in slices oxygenated throughout incubation (Fig.3a), and ranged from zero to pronounced in slices ren-dered hypoxic. Most commonly, where cell death waspresent in slices subject to hypoxia alone, striatal neu-ronal profiles were predominantly pyknotic (Fig. 3b),while preserved neuronal profiles if present were con-fined to the perimeter of the striatum. The neuronalpyknotic phenotype shown in Fig. 3b was distinct fromthe dark phenotype observed within fibre bundles inoxygenated slices, which was considered to be glial(Fig. 3a).

Profile counts. Data obtained from slices exposed to20 min hypoxia in the presence or absence of dopaminereceptor antagonists or MK-801 and from slices incu-bated for an equivalent experimental period under nor-moxic conditions are shown in Table 1. Data obtainedfrom slices exposed to 30 min hypoxia in the presenceor absence of haloperidol are shown in Table 2. Thedistribution of all data sets was normal (KolgomorovSmirnov test). Control data in the three morphologicalcategories from the five sets of experiments in which 20

min hypoxia was imposed were pooled when one-wayANOVAs confirmed that there was no difference be-tween these data sets (pyknotic data, F � 1, df � 4, 28,P � 0.4; normal data, F � 0.9, df � 4, 28, P � 0.4;unclassified data, F � 0.3, df � 4, 28, P � 0.9). Sub-

FIG. 3. Brain slices from the rostral striatum (a) incubated under normoxic conditions; (b) exposed to 20 min hypoxia alone; (c) exposedto 20 min hypoxia in the presence of 50 �M haloperidol; (d) 10 �M MK-801; and (e) 10 �M SCH23390 � 10 �M eticlopride. ➝ , Normal profiles;3, vacuolization of the neuropil around neurons and blood vessels; ➛ , pyknotic profiles within the neuropil, considered to be neuronal; ��,dark profiles within fiber bundles, considered to be glial; —}, unclassified profiles. Striatal neuronal profiles in (a) are largely preserved, andprofiles in (b) are largely pyknotic. The majority of profiles in tissue treated with haloperidol (c), MK-801 (d), and SCH23390 � eticlopride(e) are preserved and would have been classified as normal. Unclassified profiles, perineuronal and perivascular vacuoles are seen in alltreatments. Photomicrographs are of 3-�m plastic sections stained with 0.1% toluidine blue in 0.1% borax. Scale bar, 50 �m.

TABLE 1

Profile Counts in Oxygenated Slices and in Slices Exposedto 20 Min Hypoxia in the Presence or Absence of DopamineReceptor Antagonists or MK-801

Treatment % Normal % Pyknotic % Unclassified

Normoxic [n � 5] 85.8 � 4.4 0 14.2 � 4.4Hypoxia alone

[n � 33]18.8 � 29.3 60.1 � 30.6 21 � 8.8

50 �M haloperidol[n � 6]

74.9 � 9.7b 7.1 � 7.6a 18 � 4.6

10 �M MK-801[n � 5]

55.8 � 22.7 8.6 � 6.9a 35.6 � 16.9

10 �M SCH23390 �10 �M eticlopride[n � 6]

80.2 � 16.5b 5.9 � 9.4a 14 � 7.5

10 �M SCH23390[n � 8]

32.8 � 31 33.5 � 35.7 33.7 � 21.2

10 �M eticlopride[n � 8]

34.4 � 32.7 40.8 � 39.7 24.8 � 11

a P � 0.01 as compared to control.b P � 0.001 as compared to control.


stantial cell death was observed in control slices ex-posed to 20 min hypoxia (60.1 � 30.6% pyknotic pro-files). Protection against this insult was conferred by50 �M haloperidol (P � 0.002), 10 �M MK-801 (P �0.007), and by the combined application of 10 �MSCH23390 � 10 �M eticlopride (P � 0.001), all ofwhich reduced the proportion of pyknotic profiles tounder 10%. No protection was demonstrated for treat-ment with 10 �M SCH23390 alone (P � 0.5) or 10 �Meticlopride alone (P � 0.5). Haloperidol (50 �M) alsoproduced marked protection where 30 min hypoxia wasimposed (P � 0.001). Normal profiles formed the ma-jority in slices protected against 20 min hypoxia by 50�M haloperidol, 10 �M SCH23390 � 10 �M eticlo-pride, and 10 �M MK-801. However, only the first twoof these data sets were significantly different fromcontrol (haloperidol, P � 0.001; 10 �M SCH23390 � 10�M eticlopride, P � 0.001). No difference was demon-strated between the control value and and that seen inslices treated with 10 �M MK-801 (P � 0.07), 10 �MSCH23390 alone (P � 0.9), or 10 �M eticlopride alone(P � 0.7). Finally, no significant difference was foundbetween the control value for unclassified profiles andvalues for any of the pharmacological interventionswhere 20 min hypoxia was imposed. P values citedabove were obtained by post hoc analysis using theTukey HSD test, after a one-way ANOVA had beenrun. Slices oxygenated throughout incubation con-tained high levels of normal neuronal profiles (85.8 �4.4%), with no pyknotic profiles being detected in thistissue.

DISCUSSION

Data presented in this paper have shown the in-volvement of NMDA receptors, and of both classes ofdopamine receptor, in the mediation of hypoxic damagein the rostral mouse striatum. Prior to discussing thesefindings, some comment should be passed on moregeneral aspects of this study. First, quantitative histo-logical techniques are used relatively infrequently inthe brain slice preparation, but in this case were jus-tified by the reproducibly high levels of neuronal pres-ervation obtained in normoxic slices. It should be noted

that the same consistency was not observed in controlslices subject to hypoxia, pyknotic profiles forming�50% total profiles in 24/33 slices, but �10% totalprofiles in 5/33 slices, after 20 min hypoxia. This “all-or-nothing” response to insult, though undesirable, hasbeen noted previously in the slice preparation (1, 49).The causes of this phenomenon are unclear. The capac-ity of a minority of neurons in control tissue to surviveextended periods of oxygen deprivation may be attrib-uted to the continued availability of glucose duringinsult, and to the subphysiological bath temperature.Both glucose and hypothermia have been shown toprotect functional parameters in the hypoxic slice (27,34, 43). Hypothermic temperatures are routinely usedin the slice preparation to enhance viability (28) and inour case were found necessary to achieve high levels ofpreservation of striatal neurons in normoxic conditions(unpublished observations). Second, the presence ofperineuronal and perivascular vacuoles in normoxictissue may suggest a less-than-satisfactory standard ofmorphological preservation under optimal conditions:this has been considered to be a common feature of thispreparation (39, 51). Vacuolization of the slice is likelyto have reflected swelling as a result of water gain (5,33). Third, the model of cerebral hypoxia used in thisstudy has the limitation inherent to this preparation ofbeing short-term. It is unable, therefore, to support therange of tests possible in in vivo models; for example itcannot be used to test whether application of drug attime points significantly postinsult is protective orwhether application of the pharmacological agents un-der investigation definitively prevent or merely delayneuronal death. Nevertheless, we consider that ourmodel provides a useful means of establishing whetherdrugs are able to afford short-term neuroprotectionand at the very least provides a relatively inexpensiveand ethical means of screening drugs prior to moreextensive testing in in vivo models. Finally, the use ofthe pyknotic profile (Fig. 3b) as the criterion to assessdamage and neuroprotection in slices subject to hyp-oxia was justified by its absence from normoxic controlslices.

Involvement of dopamine receptors in hypoxia-in-duced neuronal death. Our finding that haloperidolprotected striatal neurons against hypoxic insult isconsistent with those of several studies. Haloperidolhas been observed to extend the survival time of miceexposed in vivo to a low-oxygen atmosphere (36), and toprotect cortical neurons in cell cultures against expo-sures to glutamate, NMDA, and hypoxia–hypoglyce-mia (11, 30). Our data provide further confirmation ofthe neuroprotective efficacy of this drug, in a model inwhich this property has not previously been demon-strated. However the broad pharmacological profile ofthis drug, which in addition to its potent affinity for D2and sigma receptors, exhibits antagonistic actions at

TABLE 2

Profile Counts in Slices Exposed to 30 Min Hypoxiain the Presence or Absence of Haloperidol

Treatment % Normal % Pyknotic % Unclassified

Hypoxia alone[n � 7]

7.6 � 15.2 69.8 � 27.8 22.5 � 12.9

50 �M haloperidol[n � 7]

46.1 � 22.2a 12.3 � 24.5b 41.7 � 9.5a

a P � 0.01 as compared to control.b P � 0.001 as compared to control.


NMDA receptors (31, 50) and D1 receptors (44) andblocks flow through voltage-gated Ca2� channels (14),precludes the attribution of its neuroprotective efficacysolely to D2 receptor blockade. Data obtained using theselective dopamine receptor antagonists SCH23390and eticlopride demonstrate dopamine receptor-medi-ated toxicity to striatal neurons in the hypoxic murinecorticostriatal brain slice preparation and show that,in the rostral striatum, this toxicity is mediated byboth D1- and D2-class receptors. The requirement forblockade of both D1 and D2 receptors to achieve pro-tection has not previously been reported and requirescomment. As both classes of dopamine receptor havebeen implicated in mechanisms of hypoxia–ischemia-induced toxicity in the striatum by this and by previousstudies, it is not unreasonable to hold that blockade ofboth D1 and D2 receptor subtypes should be moreeffective than blockading than individually. Reports ofprotection by individual treatment with D1 and D2antagonists (6, 19) may simply indicate that the patho-physiological processes triggered at the unantagonizedreceptor subtype by the level of insult applied were notsufficient to produce observable damage. It should alsobe restated that protection of striatal neurons has notalways been conferred by blockade of a single dopa-mine receptor subtype (18). No study other than ourshas confined its experimentation to the rostral stria-tum, and it is possible to speculate that the require-ment to block both D1 and D2 receptor subtypes in thisregion may reflect a heightened sensitivity to dopa-mine receptor-mediated toxicity here, conferred bygreater concentrations rostrally of dopamine (25) andof D1 and D2 receptors (41, 2). At the current state ofknowledge, it is impossible to explain why one study(19), in opposition to our findings and those of others(6, 18), found that activation of D1 receptors was pos-itively beneficial in hypoxia–ischemia. Our data indi-cate clearly that, in our model, the separate activationof both D1 and D2 receptors during hypoxia can pro-duce significant damage in the rostral murine stria-tum. It may be suggested that our failure to observeprotection where SCH23390 and eticlopride were ap-plied individually resulted from the use of insufficientconcentrations of drug. No similar in vitro study existsfor comparison, and testing this proposal would requireadditional experiments. Nevertheless, it is significantthat the same concentrations of antagonist which wereineffective when administered separately were effec-tive when administered together. The pathways bywhich dopamine receptors mediate toxicity in hypoxia–ischemia are currently unknown. Extrapolating fromthe literature it is possible to speculate that D1 recep-tor-mediated toxicity in rodent striatum may involvethe enhancement of NMDA receptor-mediated toxicity,as the D1 agonist SKF38393 has been observed toincrease NMDA-induced swelling of neostriatal neu-rons in rat striatal slices (7) and the modulation of

L-type Ca2� conductances, which are enhanced by D1agonists at depolarized membrane potentials in ratstriatal slices (8, 21). However, all of these studies wereconducted in continuously oxygenated slices. The elu-cidation of the mechanisms of D1 and D2 receptor-mediated toxicity under hypoxic conditions is clearly asubject for future research.

Involvement of NMDA receptors in hypoxia-inducedneuronal death. Although all three major subclassesof ionotropic glutamate receptor mediate hypoxic–isch-emic damage to CNS neurons, the relative contributionof individual receptor subtypes to such damage varieswith the model employed. Cell death in primary neu-ronal cultures exposed to short periods of hypoxia isreported to be mediated primarily by NMDA receptors(9), whereas injury to hippocampal neurons following aglobal ischemic insult in vivo is mediated mainly byAMPA receptors (12, 45). Several studies, using a va-riety of techniques, have reported that NMDA recep-tors are involved in mechanisms of neuronal death inbrain slice models of hypoxia–ischemia. For example, astudy using the calcium indicator dye calcium greenhas shown that increases in cytosolic Ca2� occurring inhippocampal neurons in the first 5 min of in vitroischemia are mediated primarily by the NMDA recep-tor (52), whereas both MK-801 and a low-calcium en-vironment were observed to prevent a neuronal depo-larization that otherwise inevitably followed with-drawal of oxygen and glucose and persisted afterreoxygenation (38). Again, several reports have indi-cated that NMDA receptor antagonists protect postin-sult recovery of synaptic function in brain slice modelsof hypoxia–ischemia (4, 26, 29, 37, 42). MK-801 hasappeared as a particularly potent neuroprotectant, en-abling, for instance, 100% recovery of synaptic functionin CA1 neurons exposed to 20 min hypoxia as com-pared to 20% recovery in control slices (42). Thus,acutely occurring, NMDA receptor-mediated increasesin (Ca2�)i, are implicated in the impairment of neuro-nal function in slice models of hypoxia–ischemia. Ourdata, obtained by morphological analysis, have shownMK-801, equally with haloperidol and SCH23390 �eticlopride, to be effective in preventing that form ofhypoxia-induced cell death of which the morphologicalmanifestation is a pyknotic phenotype. It is thereforelikely that the NMDA receptor is involved in the me-diation of this type of neuronal death. However, MK-801 data differ from those of the other two protectivetreatments in that the percentage of normal profiles inMK-801-treated slices was not different from the per-centage of normal profiles in the hypoxic control, al-though this value approached significance (P � 0.07).Nevertheless, it may be argued that in this study MK-801 has proved a less effective neuroprotectant thaneither haloperidol or SCH23390 � eticlopride. Multiplemechanisms have been implicated in hypoxia–isch-


emia-induced death of neurons (40), and it is possiblethat MK-801 was less effective than the other twosuccessful strategies because of the greater specificityof this treatment. Lower levels of normal profiles inMK-801-treated slices were accompanied by a percent-age of unclassified profiles that was higher than in thehypoxic control and in slices treated with haloperidolor SCH23390 � eticlopride, although this was signifi-cant only for MK-801 vs SCH23390 � eticlopride (P �0.03). It may be suggested that some unclassified pro-files counted in slices treated with MK-801 may haverepresented neurons in which pyknotic cell death hadbeen blocked by the drug, but which were neverthelesssufficiently damaged to progress to another necroticend point. This pattern was particularly evident in oneexperiment in which the MK-801-treated slice showedsignificant numbers of dead neurons, most of whichwere unclassified profiles (65.2%) rather than pyknoticprofiles (19.4%).

In summary, this study has shown that both D1 andD2 classes of dopamine receptor and the NMDA recep-tor mediate damage in the rostral mouse striatum ex-posed to hypoxia. Though these findings may suggestan interaction between the striatal glutamatergic anddopaminergic systems in mechanisms of hypoxic dam-age, such an effect has not been examined in this study,and further work would be necessary to test for such aninteraction. Nevertheless, to our knowledge this studyconstitutes the first demonstration that the toxicity ofdopamine to striatal neurons, previously shown in invivo models of cerebral ischemia and excitotoxicity, isreplicated in the corticostriatal brain slice exposed invitro to hypoxia. Therefore the model used in this studyshould prove to be a useful and valid model with whichto study dopamine toxicity in the hypoxic–ischemicstriatum.

ACKNOWLEDGMENTS

We acknowledge the expert technical assistance of Andy Fother-ingham in the design and construction of the system for brain sliceincubation used in this study.

REFERENCES

1. Aitken, P. G., M. Balestrino, and G. G. Somjen. 1988. NMDAantagonists: Lack of protective effect against hypoxic damage inCA1 region of hippocampal slices. Neurosci. Lett. 89: 187–192.

2. Altar C. A., H. Kim, J. F. Marshall. 1985. Computer imagingand analysis of dopamine (D2) and serotonin (S2) binding sitesin rat basal ganglia or neocortex labeled by [3H] spiroperidol.J. Pharmacol. Exp. Ther. 233(2): 527–538.

3. Aosaki T., K. Kiuchi, and Y. Kawaguchi. 1998. Dopamine D1-like receptor activation excites rat striatal large aspiny neuronsin vitro. J. Neurosci., 18(14): 5180–5190.

4. Arias, R. L., J. R. Tasse, and M. R. Bowlby. 1999. Neuroprotectiveinteraction effects of NMDA and AMPA receptor antagonists in anin vitro model of cerebral ischemia. Brain Res. 816: 299–308.

5. Banay-Schwartz, M., A. Gergely, and A. Lajtha. 1974. Indepen-dence of amino acid uptake from tissue swelling in incubatedslices of brain. Brain Res. 65(2): 265–276.

6. Bielenberg, G. W., and M. Burkhardt. 1991. Dopamine D1 an-tagonist SCH23390 reduces infarct size in a mouse model offocal cerebral ischemia. J. Cereb. Blood Flow Metab. 11(S2):S726.

7. Cepeda C., C. S. Colwell, J. N. Itri, E. Gruen, M. S. Levine.1998. Dopaminergic modulation of early signs of excitotoxicityin visualized rat neostriatal neurons. Eur. J. Neurosci. 10:3491–3497.

8. Cepeda C., C. S. Colwell, J. Itri, S. H. Chandler, M. S. Levine.1998. Dopaminergic modulation of NMDA-induced whole cellcurrents in neostriatal neurons in slices: Contribution of cal-cium conductances. J. Neurophysiol. 79: 82–94.

9. Choi, D. 1998. Antagonizing excitotoxicity: A therapeutic strat-egy for stroke? Mt. Sinai J. Med., 65(2): 133–138.

10. Clemens, J. A. and L. A. Phebus. 1988. Dopamine depletionprotects striatal neurons from ischemia-induced cell death. LifeSci. 42: 707–713.

11. DeCoster, M. A., K. L. Klette,. E. S. Knight, and F. C. Tortella.1995. � receptor-mediated neuroprotection against glutamatetoxicity in primary rat neuronal cultures. Brain Res. 671(1):45–53.

12. Diemer, N. H., M. B. Jorgensen, F. F. Johansen, M. Sheardown,and T. Honore. 1992. Protection against ischemic hippocampalCA1 damage in the rat with a new non-NMDA antagonist,NBQX. Acta Neurol. Scand. 86(1): 45–49.

13. Ferger, B., O. Eberhardt, P. Teismann, C. de Groote and J.Schulz. 1999. Malonate-induced generation of reactive oxygenspecies in rat striatum depends on dopamine release but not onNMDA receptor activation. J. Neurochem. 73(3): 1329–1332.

14. Fletcher, E. J., J. Church, and J. F. MacDonald. 1994. Haloper-idol blocks voltage-activated Ca2� channels in hippocampalNeurons. Eu. J. Pharmacol. 267(2): 249–252.

15. Globus, M. Y.-T., M. D. Ginsberg, W. D. Dietrich, R. Busto, andP. Scheinberg. 1987. Substantia nigra lesion protects againstischemic damage in the striatum. Neurosci. Lett. 80: 251–256.

16. Globus, M. Y.-T., R. Busto, W. D. Dietrich, E. Martinez, I.Valdes, and M. D. Ginsberg. 1988. Effect of ischemia on the invivo release of striatal dopamine, glutamate and �-aminobu-tyric acid studied by intracerebral microdialysis. J. Neurochem.51(5): 1455–1464.

17. Globus, M. Y.-T., D. C. Mash, W. D. Dietrich, R. Busto, I Valdes,and M. D. Ginsberg. 1988. Dopamine D1 but not D2 receptorsare selectively vulnerable to transient ischaemia. Soc. Neurosci.Abs. 14: 205.5.

18. Globus, M. Y.-T., W. D. Dietrich, R. Busto, I. Valdes, and M. D.Ginsberg. 1989. The combined treatment with a dopamine D1antagonist [SCH-23390] and NMDA receptor blocker [MK-801]dramatically protects against ischemia-induced hippocampaldamage. J. Cereb. Blood Flow Metab. 9(S1): 5.

19. Hashimoto, N., T. Matsumoto, H. Mabe, T. Hashitani, and H.Nishino. 1994. Dopamine has inhibitory and accelerating ef-fects on ischemia-induced neuronal cell damage in the rat stri-atum. Brain Res. Bull. 33: 281–288.

20. Hastings, T. G., D. A. Lewis, and M. I. Zigmond. 1996. Role ofoxidation in the neurotoxic effects of intrastriatal dopamineinjections. Proc. Natl. Acad. Sci. USA 93(5): 1956–1961.

21. Hernandez-Lopez, S., J. Bargas, D. J. Surmeier, A. Reyes, andE. Galarraga. 1997. D1 receptor activation enhances evokeddischarge in neostriatal medium spiny neurons by modulatingan L-type Ca2� conductance. J. Neurosci. 17(9): 3334–3342.

22. Hillered, L., A Hallstrom, S. Segersvard, L. Persson, and U.Ungerstedt. 1989. Dynamics of extracellular metabolites in the


striatum after middle cerebral artery occlusion in the rat mon-itored by intracerebral microdialysis. J. Cereb. Blood FlowMetab. 9: 607–616.

23. Hoyt, K. R., I. J. Reynolds, and T. G. Hastings. 1997. Mecha-nisms of dopamine-induced cell death in cultured rat forebrainneurons: interactions with and differences from glutamate-in-duced cell death. Exp. Neurol. 143(2): 269–281.

24. Ikonomidou, C., M. T. Price, J. L. Mosinger, G. Frierdich, J. La-bruyere, K. Shahid Salles, and J. W. Olney. 1989. Hypobaric–ischemic conditions produce glutamate-like cytopathology ininfant rat brain. J. Neurosci. 9(5): 1693–1700.

25. Joyce, J. N., S. K. Loeschen, and J. F. Marshall. 1985. Dopa-mine D2 receptors in the rat caudate-putamen: The lateral tomedial gradient does not correspond to dopaminergic innerva-tion. Brain Res. 338: 209–218.

26. Kass, I. S., G. Chambers, and J. E. Cottrell. 1989. The N-methyl-D-aspartate antagonists aminophosphonovaleric acidand MK-801 reduce anoxic damage to dentate granule and CA1pyramidal cells in the rat hippocampal slice. Exp. Neurol. 103:116–122.

27. Kurihara, E., A. Ishikawa, N. Tamaki, and Y. Okada. 1996. Theprotective effects of hypothermia on the recovery of neuralactivity after deprivation of oxygen and glucose: Study of slicesfrom the hippocampus and superior colliculus. Neurosci. Lett.204: 197–200.

28. Lipton, P., P. G. Aitken, F. E. Dudek, K. Eskessen, M. T.Espanol, P. A. Ferchmin, J. B. Kelly, N. R. Kreisman, P. W.Landfield, P. M. Larkman, L. Leybaert, G. C. Newman, K. L.Panizzon, R. S. Payne, P. Philips, K. M. Raley-Susman, M. E.Rice, R. Santamaria, J. M. Sarvey, A. Schurr, M. Segal, C. P.Taylor, T. J. Teyler, V. Y. Vasilenko, S. Veregge, S. H. Wu, andR. Wallis. 1995. Making the best of brain slices: Comparingpreparative methods. J. Neurosci. Methods 59(1): 151–156.

29. Lobner, D., and P. Lipton. 1990. �-Ligands and non-competitiveNMDA antagonists inhibit glutamate release during cerebralischemia. Neurosci. Lett. 117: 169–174.

30. Lockhart, B. P., P. Soulard, C. Benicourt, A. Privat, and J.-L.Junien. 1995. Distinct neuroprotective profiles for � ligandsagainst N-methyl-D-aspartate (NMDA) and hypoxia-mediatedneurotoxicity in neuronal culture: toxicity studies. Brain Res.675(1–2): 110–120.

31. Lynch, D. R., and M. J. Gallagher. 1996. Inhibition of N-methyl-D-aspartate receptors by haloperidol: Developmental and phar-macological characterization in native and recombinant recep-tors. J. Pharmacol. Exp. Ther. 279(1): 154–161.

32. McLaughlin, B. A., D. Nelson, M. Erecinska, and M. F. Ches-selet. 1998. Toxicity of dopamine to striatal neurons in vitro andpotentiation of cell death by a mitochondrial inhibitor. J. Neu-rochem. 70(6): 2406–2415.

33. Newman, G. C., F. E. Hospod, H. Qi, and H. Patel. 1995. Effectsof dextran on hippocampal brain slice water, extracellularspace, calcium kinetics and histology. J. Neurosci. Methods61(1–2): 33–46.

34. Paschen, W., and B. Djuricic. 1995. Comparison of in vitroischemia-induced disturbances in energy metabolism and pro-tein synthesis in the hippocampus of rats and gerbils. J. Neu-rochem. 65(4): 1692–1697.

35. Pastuszko, A. 1994. Metabolic responses of the dopaminergicsystem during hypoxia in newborn brain. Biochem. Med. Metab.Biol. 51(1):1–15.

36. Pontecorvo, M. J., E. W. Karbon, S. Goode, D. B. Clissold, S. A.Borosky, R. J. Patch, and J. W. Ferkany. 1991. Possible cere-broprotective and in vivo NMDA antagonist activities of sigmaagents. Brain Res. Bull. 26: 461–465.

37. Radek, R. J., and W. J. Giardina. 1992. The neuroprotectiveeffects of dextromethorphan on guineapig-derived hippocampalslices during hypoxia. Neurosci. Lett. 139: 191–193.

38. Rader, R. K., and T. H. Lanthorn. 1989. Experimental ischemiainduces a persistent depolarization blocked by decreased cal-cium and NMDA antagonists. Neurosci. Lett. 99: 125–130.

39. Rice, M. E. 1999. Use of ascorbate in the preparation andmaintenance of brain slices. Methods: Companion Methods En-zymol. 18: 144–149.

40. Rosenblum, W. I. 1997. Histopathologic clues to the pathwaysof neuronal death following ischemia/hypoxia. J. Neurotrauma14(5): 313–326.

41. Savasta, M., A. Dubois, and B. Scatton. 1986. Autoradiographiclocalisation of D1 dopamine receptors in the rat brain with [3H]SCH23390. Brain Res. 375:291–301.

42. Schurr, A., R. S. Payne, and B. M. Rigor. 1995. Protection byMK-801 against hypoxia-, excitotoxin-, and depolarization-in-duced neuronal damage in vitro. Neurochem. Int. 26(5): 519–525.

43. Schurr, A., R. S. Payne, J. J. Miller, and B. M. Rigor. 1999.Study of cerebral energy metabolism using the rat hippocampalslice preparation. Methods: Companion Methods Enzymol. 18:117–126.

44. Seeman, P., and H. H. M. Van Tol. 1994. Dopamine receptorpharmacology. Trends Pharmacol. Sci. 15(7): 264–269.

45. Sheardown, M. J., P. D. Suzdak, and L. Nordholm. 1993.AMPA, but not NMDA, receptor antagonism is neuroprotectivein gerbil global ischaemia, even when delayed 24h. Eu. J. Phar-macol. 236(3): 347–353.

46. Thonsaard, W., D. A. Kendall, G. W. Bennett, and C. A. Mars-den. 1997. A Simple Method for Measuring Dopamine Releasefrom Rat Brain Slices. J. Pharmacol. Toxicol. Methods 37: 143–148.

47. Toner, C. C., and J. A. Stamford. 1996. “Real time” measure-ment of dopamine release in an in vitro model of neostriatalischaemia. J. Neurosci. Methods 67(2): 133–140.

48. Wahl, F., M. Plotkine, R. Boulu, T. Obrenovitch, and L. Symon.1993. Involvement of dopaminergic receptors in quinolinate-induced striatal lesions. Neuroreport 5(2):151–153.

49. Weber, M. L., and C. P. Taylor. 1994. Damage from oxygen andglucose deprivation in hippocampal slices is prevented by tetro-dotoxin, lidocaine and phenytoin without blockade of actionpotentials. Brain Res. 664(1–2): 167–177.

50. Whittemore, E. R., V. I. Ilyin, and R. M. Woodward. 1997.Antagonism of N-methyl-D-aspartate receptors by � site li-gands: Potency, subtype-selectivity and mechanisms of inhibi-tion. J. Pharmacol. Exp. Ther. 282: 326–338.

51. Whittingham, T. S., W. D. Lust, D. A. Christakis, and J. V.Passonneau. 1984. Metabolic stability of hippocampal slicepreparations during prolonged incubation. J. Neurochem. 43:689–696.

52. Zhang, Y., and P. Lipton. 1999. Cytosolic Ca2� changes duringin vitro ischemia in rat hippocampal slices: Major roles forglutamate and Na�-dependent Ca2� release from mitochondria.J. Neurosci. 19(9): 3307–3315.


protection of striatal neurons by joint blockade of d1 and d2 receptor subtypes in an in vitro model...

Documents