Joiirtid 01Neiirociicmi.s~r.r Raven Press, Ltd.. New York 0 1992 International Society for Neurochemistry

N-Methyl-D-Aspartate-Mediated Injury Enhances -Stimulated Phosphoinositide Turnover in Perinatal Rats

Chu-Kuang Chen, *-/-FayeS. Silverstein, and $Michael V. Johnston

Neuroscience Program, uncl Dqmimcws of'*Pediuirics uncl tNmrolog!,, Uriivtwitj-(?f Michigan. Ann .Arbor. Michigun, and $The Kmne& RtJseurch Instiiiiir and Johns Hopkins lJnivtwiij9Mdical Insiifirf ions, Baliirnore. Mur)jlund. U.S.A.

Abstract: Previous work in our laboratory demonstrated glutamate receptors because carbachol (CARB) failed to that ischemic-hypoxic brain injury in postnatal day 7 rats elicit preferential enhanced stimulation. To investigate the causes a substantial increase in phosphoinositide (PPI) turn- possibility that alterations in the release ofendogenous neu- over stimulated by the glutamate analogue quisqualic acid rotransmitters had a role in potentiating QUIS-stimulated (QUIS) in the hippocampus and striatum. To examine this PPI turnover after NMDA injection. we examined the effect phenomenon in more detail. we performed similar experi- of tetrodotoxin. Tetrodotoxin (0.5 pA4) did not alter QUIS- ments after producing injury by unilateral intracerebral in- or CARB-stimulated PPI hydrolysis in the lesioned or unle- jections of the glutamate analogue N-methyl-D-aspartate sioned tissue. The influence of extracellular calcium con- (NMDA). The 7-day-old rodent brain is hypersensitive to centration on QUIS-stimulated [3H]IP,formation was also NMDA neurotoxicity and NMDA injection causes histopa- examined after the NMDA lesion. Moderate reduction of thology that closely resembles that produced by ischemia- calcium in the buffer (1 pM) enhanced the lesion effect. hypoxia. NMDA, 17 nmol in 0.5 PI, was injected into the Low calcium buffer enhanced QUIS-stimulated PPI turn- right posterior striatum of 7-day-old rat pups and they were over in the lesioned hippocampal slices, but reduced QUIS killed 3 days later. Hippocampal or striatal tissue slices were stimulation in contralateral slices and controls. In contrast, prepared from ipsilateral and contralateral hemispheres CARB-stimulated PPI turnover was not enhanced in low from vehicle-injected control and from noninjected control Ca2' buffer. A similar pattern of Ca2+dependency was ob- rat pups. Slices were then incubated with mj~-[~H]inositol served in striatal slices. Calcium-free (

In the developing nervous system, activation of a sitol phospholipids (PPIs), especially phosphatidylino- subpopulation of excitatory amino acid (EAA) recep- sitol4,5-bisphosphate, and yields two intracellular sec- tors is coupled to activation of phospholipase C ond messenger molecules, inositol 1,4,5-trisphos- (PLC). PLC catalyzes the hydrolysis of membrane ino- phate (IP,) and diacylglycerol. IP, liberates Ca2+from

Received November 12, 199 I; revised manuscript received Feb- L-2-amino-4-phosphonobutyic acid: AP7. ~.~-2-amino-7-phos- ruary 6. 1992: accepted March 2, 1992. phonoheptanoic acid: CARB. carbachol: CPP. 3-[(&)-2-carboxypi- Address correspondence and reprint requests to Dr. M. V. John- perazin-4-$1-propyl- I-phosphoric acid: DNQX. 6.7-dinitroqui- ston at Kennedy Research Institutc. Johns Hopkins University. 707 noxaline-2.3-dione: EAA. excitatory amino acid: Glu. L-glutamic N. Broadway. Baltimore, MD 2 1205. U.S.A. acid: G protein. GTP-binding protein: IP,. inositol I,4.5-trisphos- The present addressofDr. C.-K. Chen is Kennedy Research Insti- phate: MK-801. (+)-5-niethyl-l0. I I-dihydro-51/-dibenzo[n.r/]- tute. 707 N. Broadway, Baltimore. MD 21205. U.S.A. cyclohepten-5.10-imine hydrogen maleate: NMDA. /~-methyl-D- ..Ihhl.~i'i~[i~ii~usrd AMPA. (RS)-~amino-3-hydroxy-5-methyl- aspartate: PLC. phospholipase C: PPI. inositol phospholipid: QUIS. 4-isoxazolepropionic acid: ANOVA. analysis of variance: AP4. D. quisqualic acid: TTX. tetrodotoxin.

963 964 C.-K. CHEN ET AL.

intracellular sites and modulates the intracellular cal- Intracerebral NMDA lesions and sham injections cium level. Compared with other glutamate ana- In three assay runs, we compared [3H]IP, accumulation logues, quisqualic acid (QUIS), a non-N-methy1-D- in vehicle-injected and untreated littermate control rat aspartate (NMDA) agonist, is one pups. In 7-day-old Sprague-Dawley rat pups. maintained of the most potent activators of PLC in neonatal rat under ether anesthesia, HEPES (0.5 M, pH 7.4) was stereo- 1 taxically injected into the posterior striatum by Hamilton brain. QUIS-stimulated PPI hydrolysis is 0-fold syringe over 1.5 min. The stereotaxic coordinates were AP greater in the neonatal hippocampus and striatum = +2, L = 2.5, D = +4.0 in relation to bregma. These than in the adult (Nicoletti et al., 1986~). vehicle-injected rats and untreated littermates were killed 3 Injury appears to enhance the ability of EAAs to days later. Samples of tissue from 10 rats were pooled for stimulate PLC activity in both mature and immature each group and were assayed in triplicate. For NMDA le- brains. We demonstrated that ischemic-hypoxic sion studies, NMDA ( I7 nmol, dissolved in 0.5 pl of HEPES brain injury in 7-day-old rats enhances QUIS-stimu- buffer, pH 7.4) was stereotaxically injected into the poste- lated PPI turnover in the injured hippocampus and rior striatum. In most experiments. NMDA-injected (18-24 striatum, whereas cholinergic receptor coupled PPI rats/assay pooled) rats and untreated littermate control rats turnover is not changed (Chen et al., 1988). In adult (12 rat pups/assay pooled) were killed 3 days later. Increased rats, a similar pattern of enhanced ibotenate-stimu- animal numbers in lesioned groups were used to provide enough tissue to assay the pooled sample in triplicate across lated PPI turnover is detected in striatum lesioned all parameters. For some experiments, additional groups of with (Nicoletti et al., 1987). The mecha- rats were killed 1, 6, and 33 days after NMDA injection. nism for this selective enhancement of EAA-stimu- To examine whether an NMDA lesion in the left hemi- lated PLC activity by injury is unknown. In this study, sphere affected receptor-coupled PPI hydrolysis to the same we examined this phenomenon by studying another degree as right-sided injury. in one assay run (16 rat pups type of injury in the developing brain mediated by the pooled). we injected NMDA (1 7 nmol) into the left striatum EAA analogue NMDA. Intrastriatal injection of of 7-day-old rat pups with the same coordinates as described NMDA into perinatal rats produces a brain lesion above but left of bregma. Three days later, hippocampal and that has a pathology similar to that observed after isch- striatal tissue was collected and assayed with untreated con- emic-hypoxic brain injury (McDonald et al., 1988; trols in the same way as described above. Histopathology was assessed in Nissl-stained coronal Ikonomidou et al., 1989) and NMDA neurotoxicity is brain sections (two animals each) at days 1,3,6. and 33 after enhanced on postnatal day 7 compared with the intrastriatal injection of 17 nmol of NMDA. Histology of adult. The lesion may reproduce the overstimulation vehicle-injected rats was also done. of EAA receptors that is thought to play a role in isch- emic-hypoxic injury. Tissue preparation and assay of PPI metabolism In these experiments, lesions were produced by The hippocampus and striatum were dissected out and NMDA microinjection in 7-day-old rat pups, and the cross-chopped brain slices (350 X 350 pm) were prepared pups were killed 1-33 days later. In slices prepared with a McIlwain chopper. For each anatomic region. this tissue was used to prepare three pooled samples from the from these animals, QUIS-stimulated formation of side of injection, the contralateral hemisphere, and control [3H]IP, was assayed in the presence of Li+. Experi- pups. Slices were suspended and washed three times with ments were aimed at quantifying the degree to which warm (37°C) buffer (142 mM NaC1, 5.6 mM KCI, 2.2 mM PPI turnover was enhanced after injury, the time CaCI,, 3.6 mM NaHCO,, 1 mM MgCI,, 5 mM D-glucose. course of the effect, and the biochemical receptor and 30 mMNa+-HEPES, pH 7.4). After the slices settled by mechanisms responsible for the enhancement. gravity, excess buffer was aspirated and a 204 aliquot of tissue (protein content, 0.1-0.4 mg) was transferred to test MATERIALS AND METHODS tubes containing 4.0 pCi of n~!,o-[2-~H]inositol,LiCl (10 mM), and specific agonists or antagonists in a final volume Materials of 500 p1 of buffer. This mixture was then incubated at 37°C rn~yo-[2-3H]Inositol( 15 Ci/mmol) was purchased from for 2 h. This continuous labeling paradigm was used be- American Radiolabeled Chemicals (St. Louis, MO, U.S.A.). cause of its simplicity and sensitivity (Berridge et al., 1982; Dowex- 1 X8 ion-exchange resin ( 100-200 mesh; formate Fisher and Bartus. 1985: Heacock et al., 1987). form) was obtained from Bio-Rad (Richmond, CA, U.S.A.). Two hours after incubation, reactions were terminated by QUIS was obtained from Cambridge Research Biochemi- adding 1.7 ml of /methanol ( 1 :2 by volume) and cals (Cambridge, U.K.). MK-801 [( +)-5-methyl- 10,ll- then left overnight. [,H]IP, accumulation was monitored as dihydro-5 H-dibenzo[a,rl]cyclohepten-5,lO-iminehydrogen described previously (Bemdge, 1982; Fisher and Bartus, maleate] was a gift of Dr. P. Anderson from Merck Sharp 1985). In brief, 1 ml of chloroform and 0.5 ml of water were and Dohme (West Point, PA, U.S.A.). (RS)-a-Amino-3-hy- added and the water-soluble and organic phases were sepa- droxy-5-methyl-4-isoxazolepropionicacid (AMPA), D,L-~- rated by centrifugation. The aqueous samples were re- amino-7-phosphonoheptanoic acid (AP7), 3-[(2)-2-carbox- moved, diluted with 1.5 ml of water. and then applied to ypiperazin-4-yll-propyl-I -phosphoric acid (CPP), and 6.7- 0.5-ml columns of AG IX-8 resin (formate form) equili- dinitroquinoxaline-2,3-dione (DNQX) were purchased brated with distilled water. The columns were first washed from Tocris Neuramin (Essex, U.K.). Kynurenic acid, car- with 42 ml of 5 mM sodium tetraborate/60 mM sodium bachol (CARB: carbamoylcholine), NMDA. ~.~-2-amino- formate to elute glycerophosphorylinositol. The inositol 4-phosphonobutyric acid (AP4). and other chemicals were monophosphate fraction was then eluted with 18 ml of 0.1 obtained from Sigma Chemical (St. Louis, MO, U.S.A.). M formic acid/0.2 M ammonium formate solution, and

J. Neurochnn.. l’ol. 59. No. 3. 1992 NMDA AND PHOSPHOINOSITIDE TURNOVER IN RATS 965

radioactivity determined. To monitor the incorporation of lished that there was no difference between vehicle-injected labeled myo-inositol into the phospholipid fraction, a 0.2- controls, the contralateral side of NMDA-injected animals, ml aliquot from the organic phase was removed and radioac- and uninjected control pups. Paired and Student’s t test and tivity was measured (Berridge et al., 1982; Fisher and Bar- analysis of variance (ANOVA) were used for data analysis. tus, 1985). Estimation of EC,, for QUIS-stimulated PPI turnover was Three groups of experiments were performed for both performed using a pharmacologic calculation compu- injured and uninjured hippocampal and striatal tissue. ter program, PHARM/PCS (Microcomputer Specialist. First, we examined the effects ofglutamate agonists, antago- PA, U.S.A.). nists, and other neurotransmitter agonists on PPI turnover. Then, we examined the effect of tetrodotoxin (TTX) on QUIS-stimulated PPI turnover. Finally, we determined how RESULTS the calcium content of the incubation buffer modulated the extent of QUIS-stimulated PPI hydrolysis. Histologic changes after NMDA injection In rats killed 1 day after injection, there was both Pharmacologic studies of QUIS-stimulated PPI gross and microscopic evidence of tissue injury in the hydrolysis after NMDA lesions NMDA-injected hemisphere. Nissl-stained sections Rats were killed at four time points after NMDA injec- showed loss of staining in striatum, hippocampus, tion on postnatal day 7 (1 day after, two assay runs; 3 days, nine assays; 6 days, one assay; 33 days, two assays; 16-24 thalamus, and overlying cortex limited to the lesioned pups/assay). Hippocampal and striatal tissue slices were hemisphere. In hippocampus, all of the CAI, CA2, prepared from tissue of the injected side, noninjected side. CA3, and dentate were involved. In animals killed 3 and control pups, and incubated with glutamate agonist days after injection, the lesioned hemisphere was pale QUIS (lo-’ M)as described above. Dose-response studies and edematous, whereas the contralateral hemisphere were performed on hippocampal tissue prepared 3 days did not show any gross damage. Coronal brain sec- after NMDA injection. These tissue slices were incubated tions showed that the injury and tissue loss occurred with to MQUIS for 2 h and the [3H]IP, formation in the striatum, hippocampus, and cortex ipsilateral was monitored (four assays, 16-24 pups per assay pooled). to the side of NMDA injection (Fig. I). Six days after We tested five glutamate agonists [QUIS, L-glutamic acid (Glu), NMDA, AMPA, and AP4], three NMDA antagonists injection, the degree of edema decreased. Microscopi- (AP7. CPP. and MK-801), and two non-NMDA antago- cally, the loss of Nissl staining was greater than 3 days nists (kynurenic acid and DNQX) to pharmacologically after injection. Thirty-three days after injection, the characterize QUIS-stimulated PPI hydrolysis. CARB, a injected hemisphere appeared atrophied and there muscarinic agonist that stimulates PPI hydrolysis, was also was a great loss (-80%) of tissue in the hippocampus, examined concurrently and used as a comparison. striatum, and surrounding cortex. In vehicle-injected We investigated whether QUIS and CARB receptors are animals. there was no gross damage in the hemisphere located on the same cell population by examining [3H]IP, ipsilateral to injection. formation elicited by the combined addition of QUIS and CARB for both hippocampal and striatal tissue slices pre- Effect of vehicle injection on PPI turnover pared from the unlesioned control pups. In a pilot study, PPI turnover was measured in ve- In an experiment to block endogenous neurotransmitter hicle-injected and uninjected littermate rats (three as- release, TTX was dissolved in HEPES (50 mM. pH 7.4) says, 10 pups/assay for vehicle-injected or uninjected solution and added at the beginning of the QUIS incubation group) to determine if the vehicle itself enhanced PPI period (final concentration of TTX, 0.5 pM). turnover. QUIS- and CARB-stimulated and basal PPI Incubation of samples in Ca(+), Ca(-), and Ca(-) hydrolysis in hippocampal and striatal slices was mea- + EGTA buffers sured 3 days later. There was no significant difference In five assays, we examined the influence of changes in in [3H]IP, formation between tissue ipsilateral and the concentration ofCa2+in the incubation buffer on QUIS- contralateral to the vehicle injection and untreated stimulated PPI hydrolysis. Tissue aliquots were incubated animals in hippocampus or striatum (by ANOVA, with regular buffer containing CaCI,: Ca( +); Ca(-) buffer data not shown). in which CaCI, was replaced by water; or Ca2+-freebuffer Because there was no difference between brain tis- containing 0.5 mM EGTA: Ca(-) + EGTA. The calcium sue from animals injected with vehicle and uninjected concentrations in these three buffers were 2.2 mM, 1 pM, animals, tissue from uninjected littermate animals and < 10 nM, respectively, as measured by using Fluo-3 fluo- served as a control for NMDA-injected animals. Hip- rescence measurement (courtesy of Dr. S. K. Fisher). Tissue pocampal and striatal tissue slices were prepared from slices were washed three times with warm (37°C) Ca(-) buffer, and aliquots were then distributed into tubes con- tissue of the NMDA-injected side, the uninjected side, taining each of these buffers and incubated with agonist as and untreated littermate control pups. described above. QUIS-stimulated PPI hydrolysis is enhanced after Data analysis NMDA injection Data were normalized by protein content in each sample Rats were killed at four time points after NMDA tissue and expressed as disintegrations per minute per milli- lesioning on postnatal day 7 ( 1 day after, two assays; 3 gram protein. Protein concentration was determined ac- days, nine assays; 6 days, one assay; 33 days, two as- cording to Lowry et al. (195 I). Experiments (below) estab- says: 16-24 pups/assay run). At each time point,

J. Neirrochein.. l’ul. SY. Nu. 3. IYY2 966 C.-K. CHEN ET AL.

Three days after intrastriatal NMDA was adminis- tered by injection, QUIS ( M) stimulation of [3H]IP, release was greater in the lesioned ipsilateral hippocampal (by 123%)and striatal tissue than in the tissue from the contralateral side (by 156%)(Fig. 3). In one assay, we found that there was no difference in the effect when NMDA was injected into the opposite (left) side of the brain and the remainder of the experi- ments were performed with right-sided lesions. CARB-stimulated PPI hydrolysis was unchanged after NMDA injection In contrast to QUIS stimulation, CARB (lop2 M)-stimulated [3H]IP, release was the same in hippo- campal (seven assays) and striatal (eight assays) tissue from both hemispheres. To address the issue of whether QUIS and CARB receptors were present on the same cell population, we determined the stimulation of [3H]IP,release elic- ited by addition of either QUIS ( lop5M) or CARB (lo-* M) alone, or when combined in unlesioned tis- sue. The combination of both QUIS and CARB re- sulted in less than additive [3H]IP,.Thus, in 10-day- old rat hippocampus (n = 4 assays), the values for [3H]IP, release (dpm/mg of protein) elicited by the addition of QUIS or CARB alone were 52,080 * 1,680 and 30,900 f 1,500, respectively. In the pres- ence of both agonists, [3H]IP, formation was 57,250 * 3,400. In striatum (n = 2 assays), the comparable values were 35,260 * 1,150, 27,750 * 3,300, and 48,030 f 3,550. FIG. 1. Photographs of Nissl-stainedbrain sections from 1O-day- Dose responses for QUIS-stimulated 13H]IP, release old rats at striatal (A) and hippocampal (B) level. The animals re- ceived a unilateral intrastriatal NMDA (17 nmol) injection 3 days Dose-response curves for QUIS-stimulated PPI earlier (postnatal day 7). Extensive loss of substance is apparent turnover were constructed using tissue obtained ipsi- in ipsilateral striatum, hippocampus, and overlying cortex. S, laterally or contralaterally to the site of lesion (Fig. 4). striatum; C, cortex; H, hippocampus; Th, thalamus; L, left side; R, right side. Bar = 1 mm.

QUIS-stimulated PPI hydrolysis in the lesioned hip- 400000 pocampal tissue was considerably higher than that in the unlesioned tissue (Fig. 2). This enhanced QUIS- stimulated PPI breakdown was maximal 1 day after injection (2,000%of baseline), declining subsequently thereafter. Basal (unstimulated) [3H]IP, formation was also compared in these tissue slices. On postinjec- tion day 1, basal [3H]IP, release was higher in the le- sioned than in the contralateral group (p < 0.05). There was no difference in basal [3H]IP, formation between lesioned and unlesioned hippocampal slices at later time points. Also, there were no differences in 1 3 6 33 QUIS-stimulated and basal PPI hydrolysis between unlesioned and control pups at all time points (data DAYS, AFTER INJECTION not shown). In striatal slices, the time courses of FIG. 2. Stimulation of PPI hydrolysis by QUIS in hippocampal QUIS-stimulated and basal PPI hydrolysis were simi- slices at four times after intrastriatal NMDA (17 nmol) injection on lar to that observed for hippocampal slices (data not postnatal day 7. Slices were prepared from hippocampus ipsilat- shown). To facilitate the analysis of agonist-stimu- era1 (Ipsil) and contralateral (Contra) to NMDA injection and incu- bated with QUIS M, open and closed circles) or buffer (BA- lated [3H]IP, release, we used animals killed 3 days SAL; open and closed squares). Values (dprn/mg of protein) are after NMDA injection for subsequent experiments. expressed as means k SEM.

J. Niwroclzern , I'd 5Y. No. 3. lYY2 NMDA AND PHOSPHOINOSITIDE TURNOVER IN RATS 96 7

250000 A. * FIG. 3. A: QUlS and CARB-stimulated [3H]IP1forma- 200000 tion in hippocampal tissue 3 days after unilateral in- .-C trastriatal injection of 17 nmol of NMDA on postnatal -al day 7. Tissue slices were prepared from hippocam- 2 a pus ipsilateral(IPSIL) and contralateral to the injection 150000 (CONTRA), and from untreated controls (CON- F TROL). Tissue slices were incubated with rny0-[2-~H]- inositol in the presence of QUlS M) or CARB U5 (lo-' M) and buffers (BASAL) (see Materials and - 100000 Methods for details of assay). Values are t3H]IP, for- +-- mation (dpm/mg of protein) and expressed as the I means f SEM. These data are triplicate results from m 7 (QUIS) and 8 (CARB) assays, respectively. CARB 50000 stimulation in CONTROL slices was only assayed once. 6: Corresponding values for QUIS- and CARB- stimulated [3H]IP1release in striatum assayed concur- rently. 'p < 0.05, comparison of values from ipsilat- 0 era1 and contralateral hemispheres (post-hoc test, BASAL QUlS CARB BASAL QUlS CAR0 Fisher's PLSD method, after ANOVA).

GROUP

Maximal [3H]IP, release ( 1,379 f 1 14%)of unstimu- QUIS-stimulated PPI turnover in control samples lated littermate control pups was obtained at M were similar (hippocampus, 4.6 X M, striatum, QUIS. Higher concentrations of QUIS ( 1 0-4 and 10-3 6.7 X M). M)showed a decreased stimulation. The EC50 values were calculated from the stimulation of PPI turnover Effects of EAA agonists and antagonists at concentrations between and M. These Several glutamate agonists and antagonists were values were the same in ipsilateral and contralateral tested for their abilities to modify the stimulation of hippocampus (1.2 X lo-' vs. 6.0 X M)and stria- [3H]IP, release elicited by the addition of QUIS (see tum ( 1.1 x vs. 5.7 x M).EC5o estimates for Table 1). There was no difference in the extent of QUIS-stimulated PPI hydrolysis between the tissue contralateral to NMDA injection and noninjected control pups: only values from the NMDA-lesioned and unlesioned tissue were therefore compared. NMDA antagonists, AP7 ( lop3M, two assays) CPP ( M, two assays), and non-NMDA antagonists, DNQX ( 1 0-4 M, two assays) and kynurenic acid ( 10-3 M, one assay) did not stimulate ['HIIP, release (data not shown). QUIS-stimulated PPI turnover was not affected by competitive NMDA antagonists AP7 and CPP and non-NMDA antagonists kynurenate and DNQX (data not shown). Only high concentrations of noncompetitive NMDA antagonist MK-80 1 ( 10-3 M)inhibited QUIS-stimulated PPI hydrolysis nonse- lectively in tissue from ipsilateral (by 80%)and contra- m 0;. , . , , , , , , , , , . , I lateral (by 90%) hemispheres (see Table 1). Lower -10 -9 -8 -7 -5 -4 -3 -6 concentrations of MK-801 (

TABLE I. Effects of QUIS and rdated agonists and antagonists oj.r3H]IP, accirmulation Hippocampus Striaturn

Contralateral lpsilateral Contralateral lpsilateral

~~ Basal 9 6,980 t 440 9,170 t 1,060" 8,670 * 2.240 9.890 f 2,960 MK-801 ( M) 3 4,410 t 1,560 8,690 f 520 6,080 * 2.940 18,300 _t 4,590 QUIS ( IO-~M) 4 28,5 10 _t 17,680 78.980 f 29,270".' 25,460 f 10.460' 92,340 t 23.070",' QUIS ( M) 9 36,960 t 9.3 10' 85,820 t- 18,450hd 30,990 t 6,480' 79,180 t 1 1.310Zd + MK-801 3 4,140 + 2,836 16,120 * 5,386 3,320 f 496 30.140 f 4,450' AMPA (10-3 M) 5 13.550 f 2.540 2 1.970 t 2,500',' 15,090 k 1,980' 32,330 k 3,760" t MK-801 I 5,350 17,720 2,140 17,000

Tissue was prepared from the rats killed 3 days after NMDA (1 7 nmol) injection on postnatal day 7. Data are ['HIIP, release (dpm/mg of protein) and presented as the means f SEM. n = number of assays. Data of the untreated control rat pups are not shown here. "p< 0.05; 'p < 0.001; 'p < 0.005, ipsilateral vs. contralateral; two-tailed. paired f test. 'p < 0.0 I; 'p < 0.05, vs. basal. .fp< 0.05. vs. QUIS M).

sue. AP4 itself weakly stimulated PPI hydrolysis, but corporation of my~-[~H]inositolinto lipid was similar markedly inhibited QUIS stimulation (-60%) in both to that in the unstimulated group. However, in the lesioned and unlesioned hippocampal and striatal tis- unlesioned hippocampal tissue, QUIS-stimulated sue (n = 1 assay, data not shown). my~-[~H]inositolwas lower (22%) than basal values. In the lesioned striatal slices, radiolabeled rny~-[~H]- Incorporation of my~-[~H]inositolinto phospholipid inositol incorporation was decreased significantly after NMDA injection (34%) in QUIS-stimulated group when compared The extent of incorporation of rny~-[~H]inositol with that in the basal group. In the unlesioned striatal into membrane phospholipid was examined in three tissue, there was no difference in incorporation of assay runs. Incorporation of my~-[~H]inositolinto lip- my~-[~H]inositolinto lipid between QUIS, CARB, ids increased significantly in the lesioned hippocam- and basal groups. pal (by 170%) and striatal (by 90%) tissue compared with the unlesioned tissue (see Fig. 5; lesioned vs. un- Influence of TTX and low calcium on QUIS- lesioned, ANOVA). In samples from the lesioned hip- stimulated PPI hydrolysis pocampal tissue incubated with QUIS and CARB, in- To explore the mechanism for the enhanced QUIS- stimulated PPI hydrolysis after excitotoxic injury, we examined the hypothesis that it is mediated by en- Hlppocampus Striaturn hanced release of an endogenous neurotransmitter in the damaged tissue. However, TTX (0.5 pM), which 4 EFl I inhibits neurotransmitter release by blocking Na+ channels at presynaptic terminals (Catterall, I980), failed to alter QUIS- or CARB-stimulated [3H]IP, re- lease in damaged or undamaged tissue. Because neuro- transmitter release is also sensitive to the extracellular calcium concentration, we assayed [3H]IP, release in the presence of three different concentrations of ex- lpsilateral Contralateral lpsilateral Contralaterd tracellular calcium (Fig. 6). In hippocampal tissue GROUP from animals killed 3 days after NMDA injection, the magnitude of QUIS-stimulated PPI turnover was fur- FIG. 5. Incorporation of rny0-[2-~H]inositolinto phospholipid in hippocampal and striatal tissue from animals killed 3 days after ther enhanced in the lesioned tissue in low calcium NMDA (17 nmol) injection on postnatal day 7. Tissue slices were Ca(-) buffer compared with the extent of [3H]IP, re- prepared from hemisphere ipsilateral and contralateral to the in- lease in Ca(+) buffer (by 59%, n = 5, p < 0.05). In the jection and incubated with rny0-[2-~H]inositoland QUIS (1 0-5 M), unlesioned hippocampal tissue slices, QUIS-stimu- CARB (lo-* M),or buffer (BASAL) for 2 h (see Materials and Methods for details). After extraction of water-soluble inositol lated [)H]IP, formation in Ca(-) buffer was sup- phosphates, 0.2-ml aliquots of the chloroform phase were taken pressed by 45%. [3H]IP, (dpm/mg of protein) stimu- and counted for radioactivity. Values (dpmlmg of protein) are the lated by QUIS was also enhanced in lesioned stnatal means k SEM of triplicate replicates from three separate experi- slices [Ca(-) buffer compared with that in Ca(+) ments. In ipsilateral tissue, the incorporation of rny~-[~H]inositol buffer: 259,2 10 vs. 170,280, n = 5, p < 0.051. In the into phospholipid in all groups is significantly higher than that ob- tained from contralateral tissue in both hippocampusand striatum unlesioned striatal slices, QUIS-stimulated PPI hydro- (p < 0.05), compared with corresponding values of BASAL (post- lysis was suppressed by 31% (p < 0.05, data not hoc test, Fisher's PLSD method) after ANOVA. shown). There was no difference between values from

J. Niwochem.. Vol. 59. No. 3, 1992 NMDA AND PHOSPHOINOSITIDE TURNOVER IN RATS 969

fiooooa buffer, in the lesioned and unlesioned hippocampal Ipsilateral Contralateral tissue, QUIS-stimulated and basal incorporation of 500000 ._c my~-[~H]inositolinto phospholipid was significantly -a, * higher than that of corresponding groups in Ca( +) g 400000 T W-) buffer (p< 0.05). .F W Ca(-)+EGTA In striatum, the trend of increased incorporation of E 300000 a myo-[3H]inositolinto lipid in Ca(-) buffer, as well as U a further increase in Ca(-) + EGTA, were similar to &- 200000 those in hippocampus (p< 0.05, data not shown). I m T 100000 DISCUSSION

BASAL OUIS CARE BASAL OUIS CARE In this study, brain injury caused by unilateral in- GROUP jection of NMDA in 7-day-old rats enhanced QUIS- stimulated PPI hydrolysis in the striaturn and hippo- FIG. 6. Comparison of QUIS-stimulated release of [3H]IP, in hip- campus, but did not alter muscarinic-stimulated PPI pocampal slices incubated in buffers containing 2.2 mM CaCI,, no hydrolysis. The preferential enhancement of QUIS- added CaCI,, or no added CaCI, plus EGTA. Tissue slices were prepared from the hippocampus of ipsilateral and contralateral stimulated PPI hydrolysis appeared on the first day hemispheres of NMDA (17 nmol)-injected rats killed on postnatal after NMDA injection and persisted for I1 month. A day 10 and from untreated control rat pups. Slices were washed similar effect has been reported in neonatal animals three times with warm (37°C) buffer in which the CaCI, was re- after ischemic-hypoxic brain injury (Chen et al., placed with water, then incubated with QUIS M) or CARB 1988) and in adult rats after kainate-induced hippo- (lo-' M) or buffer (BASAL) in regular buffer containing 2.2 mM CaCI, [Ca(+)], CaCI,-free buffer [Ca(-)], or CaCI,-free + 0.5 mM campal injury (Nicoletti, 1987). EGTA [Ca(-) + EGTA] for 2 h. Values (dpm/mg of protein) are the Pharmacologic studies indicated that the enhance- means -+ SEM of triplicate results from five different experiments. ment of PPI turnover after injury was largely re- The stimulation of untreated control rat pups was not different by stricted to the QUIS receptor. The glutamate agonists one-way ANOVA from that of contralateral hippocampal slices and was therefore not shown. CARB-stimulatedr3H]IP, formation Glu and QUIS and the non-NMDA agonist AMPA in Ca(-) + EGTA buffer was only assayed once in both hemi- increased PPI hydrolysis in the lesioned tissue from spheres. "p < 0.05, comparing r3H]IP, release in hippocampal both hippocampus and striaturn, but NMDA was in- tissue incubated in Ca(-) or Ca(-) + EGTA buffer with that in corresponding Ca(+) buffer (post-hoc test, Fisher's PLSD method, after ANOVA).

I lpsilateral I Contralateral I contralateral and control groups (by one-way AN- OVA). In the same experiments, CARB-stimulated PPI hydrolysis in Ca(-) buffer was not increased in any tissue. In the presence of Ca2+-freebuffer, Ca( -) + EGTA, [3H]IP, formation was markedly suppressed (60- 90%)in basal, QUIS, and CARB groups in hippocam- pus (Fig. 6). A similar degree of repression in PPI hy- drolysis in all three groups was also observed in stria- tal slices (data not shown). BASAL auis CARE BASAL QUIS CARB GROUP Influence of Caz+ on the incorporation of myo- [3H]inositol into phospholipid FIG. 7. Incorporation of my0-[2-~H]inositolinto phospholipid in hippocampal tissue in Ca(+), Ca(-), and Ca(-) + EGTA (see Fig. The effect of the three different Ca2+ concentra- 6 legend). Hippocampal slices were prepared 3 days after NMDA tions on the incorporation of my~-[~HJinositol into (17 nmol) injection on postnatal day 7, from ipsilateral and contra- phospholipid was assayed (Fig. 7). After incubation in lateral hemispheres and incubated with my0-[2-~H]inositoland Ca(-) buffer, QUIS-stimulated incorporation of myo- QUIS M), CARB (lO-'M), or buffer (BASAL) in one of three [3H]inositolinto phospholipid in the lesioned (Fig. 7. incubation buffers Ca(+), Ca(-), and Ca(-) + EGTA for 2 h. After extraction of water-soluble inositol phosphates, 0.2-ml aliquots of Ipsilateral) hippocampal tissue was further increased the organic phase were taken and counted for radioactivity (see compared with that in Ca( +) buffer. In the unlesioned Materials and Methods for details). Values (dpm/mg of protein) (Fig. 7, Contralateral) hippocampal tissue. QUIS- and are shown as the means * SEM of triplicates from five separate CARB-stimulated and basal incorporation of myo- experiments. CARB-stimulated labeling into lipid in Ca(-) + EGTA buffer was only assayed once in both hemispheres. *p < 0.05, [3H]inositolinto lipid was significantly increased in comparing [3H]-labelinginto lipid in hippocampal tissue incubated Ca( -) buffer compared with corresponding values in in Ca(-) or Ca(-) + EGTA buffer with that in corresponding Ca(+) Ca(+) buffer. After incubation in Ca(-) + EGTA buffer (post-hoctest, Fisher's PLSD method, after ANOVA). 9 70 C.-K. CHEN ET AL.

effective. QUIS-stimulated PPI hydrolysis was not af- lipid fraction) may not accurately reflect the much fected by competitive NMDA antagonists AP7 and smaller pool of phospholipids from which [3H]IP, is CPP or the noncompetitive NMDA antagonist MK- generated. 80 1 at pharmacologically relevant concentrations. Several potential mechanisms may be responsible However, a very high concentration of MK-801 ( for the specific enhancement of QUIS-stimulated PPI M) could reduce the response, probably through a turnover after NMDA-induced injury. An increased nonspecific effect. The finding that AP4 inhibited number of QUIS receptors coupled to PPI hydrolysis QUIS stimulation of PPI hydrolysis was consistent caused by NMDA lesion is one possibility. However, with previous results (Schoepp and Johnson, 1988), quantitative autoradiography of glutamate binding though its mechanism is still unclear. showed that both QUIS- and NMDA-binding sites These observations are consistent with previous decrease in number in NMDA-lesioned tissue (F. S. findings that among glutamate receptors, only the Silverstein, unpublished observation). This suggests QUIS receptor is coupled to PPI hydrolysis in the hip- that the number of QUIS- or NMDA-binding sites in pocampus and striatum (Nicoletti et al., 1986a). The these experiments was decreased after the NMDA le- finding that the AMPA-stimulated PPI turnover was sion, and therefore the enhanced QUIS-stimulated rather low in the unlesioned tissue has also been PPI turnover observed for the NMDA-lesioned brain shown in previous studies (Palmer et al., 1988; Weiss tissue is unlikely to be the result of an increased num- et al., 1988). After brain injury, AMPA-stimulated ber of QUIS receptors. [3H]IP, formation could be further enhanced by A second possibility is that this enhancement might about onefold in the lesioned tissue when compared be caused by an increased ability of the QUIS receptor with that in the unlesioned tissue. The mechanism of to couple to PLC activity. Recent data suggest that a this enhanced AMPA-stimulated PPI hydrolysis is GTP-binding protein (G protein) links the glutamate still unclear. It has been suggested that the QUIS re- receptors to PLC activation (Nicoletti et al., 1988; ceptor can be further subdivided into AMPA-sensi- Ambrosini and Meldolesi, 1989). QUIS and musca- tive and non-AMPA-sensitive receptors (Fletcher et rinic receptors appear coupled to PLC by two differ- al., 1988; Neuman et al., 1988; Palmer et al., 1988). ent G proteins (Ambrosini and Meldolesi, 1989). Be- The AMPA-sensitive receptor is coupled to mem- cause we observed that QUIS-stimulated, but not brane depolarization, whereas the non-AMPA-sensi- CARB-stimulated, PPI turnover was enhanced after tive receptor may be preferentially coupled to PPI the NMDA lesion, this enhanced coupling may occur turnover. We also found that the specific AMPA an- specifically at the QUIS receptor-coupled G protein. tagonist DNQX failed to block QUIS-stimulated PPI In addition to neurons, large numbers of glia in the hydrolysis in any tissue. These data may suggest that tissue slices may possess QUIS receptors that couple there is some interaction between the AMPA-sensi- to PPI hydrolysis (Pearce et al., 1986). However, after tive and non-AMPA-sensitive QUIS receptors. lesioning rats on postnatal day 7, there was relatively Excitotoxic injury also increased the amount of little reactive gliosis in NMDA-lesioned tissue on rny~-[~H]inositolincorporated into the lipid pool in postlesion day 5 (F. S. Silverstein, unpublished obser- both hippocampus and striatum, possibly reflecting vation). Muscarinic receptors on glia are also coupled enhanced lipid turnover of injured membranes. We to PPI hydrolysis (Pearce et al., 1986). However, in considered the possibility that this increase in lipid the lesioned tissue, CARB addition did not elicit an labeling may have contributed to the QUIS-stimu- enhanced stimulation of [3H]IP, release, suggesting lated enhancement of [3H]IP, release. However, sev- that glia may not contribute to the enhanced QUIS- eral observations make this unlikely. If the increase in stimulated PPI hydrolysis in the lesioned tissue dur- [3H]IP, was related simply to an increase in the spe- ing the first week after NMDA injection. cific activity of radiolabeled phospholipids, an in- We considered the possibility that QUIS-stimu- crease in [3H]IP, release would also be expected in lated potentiation of PPI turnover in the injured tis- unstimulated injured tissue slices. However, there was sue was related to increased presynaptic release of en- no difference in [3H]IP, release between unstimulated dogenous excitatory neurotransmitters. It was hy- injured and control tissue. Furthermore, QUIS, but pothesized that decreased release of an inhibitory not CARB, stimulated [3H]IP, release in injured tis- neurotransmitter might alter the extent of QUIS- sue with identical elevations in my~-[~H]inositolin- stimulated PPI turnover. However, we found that the corporation. Experiments in which QUIS and CARB addition ofTTX, which inhibits voltage-sensitive neu- were added together to the incubation medium indi- rotransmitter release, failed to influence the enhance- cated that the response was not additive, suggesting ment of QUIS-stimulated PPI hydrolysis. These that both agonists stimulate the same lipid pool. The results suggest that the release of endogenous data from experiments conducted in low calcium neurotransmitters does not explain the enhanced buffer (see below) also suggest that the preferential QUIS-stimulated PPI turnover observed in NMDA- stimulation of [3H]IP, release by QUIS is not related lesioned tissue. to increased lipid labeling. Measurement of myo- Experiments in which the calcium concentration of [3H]inositol-labeled lipids (dpm/mg of protein of the the medium was varied were also performed to ex-

J. Ncirrochcm.. Vol. 59. No. 3. 1992 NMDA AND PHOSPHOINOSITIDE TURNOVER IN RATS 9 71

plore possible alterations in neurotransmitter release Chen C.-K., Silverstein F. S., Fisher S. K.. Statman D.. and John- in the injured slices. Unexpectedly, we found that ston M. V. (1988) Perinatal ischemic-hypoxic brain injury en- hances quisqualic acid-stimulated phosphoinositide turnover. moderately low calcium [Ca(-), 1 pM] markedly en- J. Nenrocheni. 51, 353-359. hanced the effect of injury to increase QUIS-stimu- Fisher S. K. and Bartus R. T. (1 985) Regional differences in the lated PPI hydrolysis. In contrast, the Ca(-) buffer at- coupling of muscarinic receptors to inositol phospholipid hy- tenuated the QUIS- and CARB-stimulated [3H]IP,re- drolysis in guinea pig brain. J. Neirrochein. 45, 1085-1095. lease in contralateral and unlesioned striatum and Fletcher E. J., Martin D., Aram J. A.. Lodge D.. and Honore T. ( 1988) Quinoxalinediones selectively block quisqualate and hippocampus. Like the effect of injury itself, augmen- kainate receptors and synaptic events in rat neocortex and hip- tation of PPI turnover by low calcium in lesioned tis- pocampus and frog spinal cord in viiro. Br. J. Phurmacol. 95, sue was selective for QUIS, but not CARB, stimula- 585-597. tion. Gonzales R. A. and Crews F. T. (1 984) Characterization of the cholinergic stimulation of phosphoinositide hydrolysis in rat Ca(-) buffer also increased incorporation of myo- brain slices. J. Nezrrosci. 4, 3 120-3 127. [3H]inositolinto total lipids in the lesioned slices and Heacock A. J., Fisher S. K., and Agranoff B. W. (1987) Enhanced this increase in labeling could contribute to the in- coupling of neonatal muscarinic receptors in rat brain to phos- creased QUIS-stimulated [3H]IP, formation in the phoinositide turnover. J. Neurochem. 48, 1904-19 I 1. Ikonomidou C., Mosinger J. L., Shahid Salles K., Labruyere J., and Ca(-) buffer. However, the modest rise in lipid label- Olney J. W. (1989) Sensitivity of the developing rat brain to ing is too small to explain the large increase in PPI hypobaric/ischemic damage parallels sensitivity to N-methyl- turnover. Furthermore, my~-[~H]inositolincorpora- Baspartate neurotoxicity. J. Neiirosci. 9, 2809-28 18. tion was greater in Ca(-) than in Ca(+) in the unle- Kendall D. A. and Nahorski S. R. (1984) Inositol phospholipid sioned hippocampal tissue, but QUIS-stimulated hydrolysis in rat cerebral cortical slices: 11. Calcium require- [3H]IP, release was actually decreased in the same tis- ment. J. Neiirochem. 42, 1388- 1394. Lowry 0. H., Rosebrough N. J., Farr A. L., and Randall R. J. ( I95 I ) sue in Ca(-) buffer. This result suggests that increased Protein measurements with the Folin phenol reagent. J. Bid. QUIS-stimulated PPI hydrolysis in Ca( -) buffer was Chem. 193,265-275. due to enhanced QUIS receptor-coupled events McDonald J. W.. Silverstein F. S.. and Johnston M. V. (1988) Neu- rather than an increased specific activity of the pre- rotoxicity of N-methyl-D-aspartate is markedly enhanced in cursor lipid pool. The mechanism for this interesting developing rat central nervous system. Bruin Res. 459, 200- 203. low calcium effect is unclear. The pronounced reduc- Neuman R. S.. Ben-Ari Y..and Cherubini E. ( 1988) Antagonism of tions observed in [3H]IP, release in the very low cal- spontaneous and evoked bursts by 6-cyano-7-nitroquinoxa- cium buffer (

J. Neitrocllein.. 1'01. SY. No. 3, I YY2