gp120, an HIV-1 protein, increases susceptibility to hypoglycemic and ischemic brain injury in perinatal rats

EXPERIMENTAL

NkJROLOGY

132, 123-133 (1995)

gpl20, an HIV-1 Protein, Increases Susceptibility to Hypoglycemic and lschemic Brain Injury in Perinatal Rats JOHN D. E. BARKS,* RONG SUN,* CHRISTA Departments of *Pediatrics and tNeurology,

MALINAK,*

AND FAYE S. SILVERSTEIN*+

University of Michigan, Ann Arbor, Michigan

INTRODUCTION

Recent data suggest that gp120, a glycoprotein secreted by HIV-l-infected macrophages, is neurotoxic, and that toxicity is mediated, at least in part, by overactivation of NMDA-type excitatory amino acid receptors. In experimental animals, considerable evidence indicates that hypoglycemic and ischemic neuro. nal injury are mediated by endogenous excitatory amino acids. We hypothesized that in the presence of gp120 the severity of brain injury resulting from hypoglycemia and cerebral ischemia would increase. To test this hypothesis in uiuo, we evaluated the influence of gp120 on the extent of brain injury resulting from these two clinically relevant pathophysiological insults in ?-day-old (P7) rats, the developmental stage of peak susceptibility to NMDA neurotoxicity. We compared the severity of hippocampal injury resulting from right intrahippocampal injections of gp120 (50 ng> in P7 rats rendered markedly hypoglycemic (n = 10) and in controls (n = 12). We also determined the influence of gp120 administration on the severity of hypoxic-ischemic injury, using a perinatal rat stroke model. P7 rats received intrahippocampal injections of gp120 (50 ng) (n = 23) or saline (n = 181 and then underwent right carotid ligation, followed by 2 h exposure to 8% oxygen. Brain injury was evaluated 5 days later, based on neuropathology evaluation and measurements of bilateral regional cross-sectional areas. The severity of hippocampal iqjury, based on cross-sectional area measurements, was considerably greater in animals from the hypoglycemic group than in litter-mate gpl20injected controls. Among the animals that underwent hypoxic-ischemic lesioning, the severity of ir+y, based on histopathology scoring and regional volume measurements, was considerably greater in animals that received gpl20 than in those that received saline. These results provide support for the hypothesis that locally secreted HIV peptides, such as gp120, may potentiate the neurotoxicity of endogenous excitatory amino acid neurotransmitters in HIV-infected brain. B uus Academic mess. IDC.

l To whom correspondence and reprint requests should be addressed at Room 8301, MSRB III Building, University of Michigan, Ann Arbor, MI 48109-0646. Fax: (313) 764-4279.

Progressive encephalopathy is a prominent and devastating complication of congenital HIV infection (4, 10). The severity and rate of progression of CNS disease in affected children vary widely; the factors that account for this variability are unknown. HIV-infected neonates are often at high risk to incur asphyxial insults, hypoglycemia, or exposure to potentially neurotoxic drugs in the perinatal period; the extent and mechanism(s) by which such adverse events influence progression of HIVrelated disease are unknown. Nor is it known if susceptibility to HIV neurotoxicity is influenced by specific developmental features expressed in the fetus and infant brain. The paradox that encephalopathy is a prominent symptom in patients with AIDS, despite little evidence of direct neuronal infection by HIV-l, initially focused attention on indirect mechanisms of HIV neurotoxicity (11,12,17,X3). Considerable experimental dataindicate that the HIV-l-derived envelope protein gp120, which is shed from HIV-infected macrophages during viral replication, may play a pivotal role in mediation of HIV neurotoxicity (6,17, 18). Brenneman et al. 6rst demonstrated that gp120 was toxic to cultured fetal hippocampal neurons (6). Subsequent studies suggested that gp120 neurotoxicity was mediated, at least in part, by overactivation of ZV-methyl-D-aspartate WMDAMype excitatory amino acid (EAA) receptors (18). Developmental stage is a critical determinant of EAA agonist-mediated neurotoxicity (231. In rat brain, susceptibility to the EAA agonist NMDA peaks in 7-day-old (P7) animals; in addition, recent data suggest that gp120 has intrinsic neurotoxic properties in the immature rodent brain (13). Thus, the immature animal may be particularly suitable for experimental studies of the role of excitotoxic mechanisms in gp120 neurotoxicity in uivo. Furthermore, animal models of perinatal brain injury may provide the optimal setting in which to test the hypothesis that gp120 could increase susceptibility to neuronal injury in uiuo by mechanisms involving NMDA receptor overactivation by endogenous EAA. Using in vivo microdialysis, we previously found that in P7 rats hypoglycemia and focal cerebral ischemia

123 All

Copyright 8 rights

of

1995

00144886/96 $6.00 by Academic Press, Inc.

reprcductionin any

form reserved.

124

BARKS ET AL.

stimulate striatal(28) and hippocampal(29) extracellular fluid accumulation of glutamate acutely. In this study, we examined the impact of gp120 on endogenous EAA-mediated neurotoxicity in uiuo in these two animal models of clinically relevant pathophysiological insults. We focused primarily on analysis of hippocampal neurotoxicity because of both the observation that gp120 was toxic to fetal hippocampal neurons (6) and our recent finding that intrahippocampal coinjection of gp120 with NMDA in P7 rats exacerbated excitotoxic injury (30). Hypoglycemia, defined as a blood glucose level below 40 mg %, is a common complication in human newborn infants (8); severe hypoglycemia is a well-recognized cause of neonatal seizures. Whether hypoglycemia, in the absence of seizures, has deleterious effects on the neonatal nervous system is an unresolved issue; there is no evidence that hypoglycemia, alone, causes neuronal injury in the human neonate, or in newborn experimental animals. We hypothesized, first, that profound hypoglycemia could increase susceptibility to gp120 neurotoxicity in P7 rats. To test this hypothesis, we compared the severity of hippocampal injury that evolved in two groups of P7 animals that received intrahippocampal injections of the same amounts of gp120, followed by intraperitoneal injections of either insulin or saline. We also hypothesized that in a perinatal rodent stroke model, in which endogenous excitatory amino acids contribute to the pathogenesis of neuronal injury (3,7), gp120 could increase the severity of brain injury. To test this hypothesis, we compared the severity of ischemic injury in P7 rats that received intrahippocampal injections of gp120 or saline and then underwent right carotid artery ligation and subsequent timed exposure to moderate hypoxia. Our results demonstrated that in P7 rats: (i) hypoglycemia increased the neurotoxic potency of gp120; and (ii) intrahippocampal injection of gp120 before hypoxicischemic lesioning markedly increased the severity of resulting ischemic forebrain injury. These data provide support for the hypothesis that locally secreted gp120 could influence outcome in clinical disorders that result in increased synaptic accumulation of endogenous EAA. METHODS Animal

Lesioning

All surgical procedures were performed in P7 unsexed Sprague-Dawley rats (Charles River Laboratories). The surgery protocols were approved by the University of Michigan Committee on Care and Use of Animals. To optimize consistency, a single individual (R.S.) performed all intrahippocampal injections. Intrahippocampal gp120 injection. In ether-anesthetized P7 rats, skull surface landmarks were exposed with a midline scalp incision and a al-gauge needle was

used to make a burr hole in the skull over the targeted injection site, following which a 25-gauge needle attached to a Hamilton syringe was inserted through the burr hole, as previously described (23, 24). Using injection coordinates (relative to bregma: AP, -2.0 mm; ML, 2.5 mm; D, 4.0 mm) targeted to right dorsolateral hippocampus, 50 ng/0.5 ~1 gp120 was injected over 2 min and the needle was left in place for a further 2 min. The gp120 dose selected for these experiments was based on several factors, including the limited volumes that can be reliably injected into the P7 hippocampus (maximum 1 ~11, the concentration of the commercially available gp120 preparation (100 ngl ~11,and our preliminary data from previous studies which indicated that intrahippocampsl injection of 50 ng gp120 in P7 rats did not elicit focal hippocampal neuronal injury, but, if coinjected with a dose of NMDA which was close to the threshold to elicit excitotoxic injury (5 nmol), increased the extent of resulting hippocampal atrophy (30). To determine if intrahippocampal gp120 injection raised body temperature, which could be a confounding factor in interpretation of outcome, in four additional animals that received intrahippocampal injections of 100 nggp120, surface temperature was measured Q5-10 min for 3 h postinjection with an electronic thermometer; the temperature range was the same as in a saline-injected control group. Insulin-induced hypoglycemia. In prior studies to evaluate the impact of insulin-induced hypoglycemia on EAA efllux using in uivo microdialysis in P7 rats, we found consistent stimulation of EAA efflux only when sustained blood glucose levels < 20 mg/dl were attained (28); we also found that hypoglycemia of this severity was invariably lethal within 24 h. To meet the objectives of this study, it was necessary to develop a protocol that elicited profound (blood glucose = 0 mg/dl) and relatively sustained hypoglycemia but still permitted at least 5-day survival, so that histopathology could be assessed. Several insulin regimens were evaluated in preliminary experiments; porcine crystalline insulin (diluted in saline) was routinely administered intraperitoneally tip); blood glucose was estimated in blood obtained from auricular puncture, using Chemstrips bG (Boehringer-Mannheim) interpreted visually. In five normal P7 rats, median blood glucose was 80 mg/dl (range 80-120) after a 2-h period in which they were separated from the dam (and thus fasted), and 80 mg/dl (range 40-120) after a 4-h interval. A single dose of 1 or 5 u/kg of insulin (n 2 S/dose) or 1.5 u/kg divided in three serial doses (n = 9) did not consistently induce a blood glucose of 0 mg/dl 2-3 h later. Regimens with higher insulin doses 115 u/kg single dose (n = 6) or 10 u/kg divided in two doses (n = 811 were lethal in the majority of animals. We determined empirically that in P7 rats fasted for 2 h, 7-8 u/kg of insulin, administered

gp120

AUGMENTS

as a single dose, consistently induced blood glucose values of 0 mg/dl 2 h later; recovery was ensured if animals were treated with supplemental glucose (10% glucose, 0.1-0.2 ml) by intraperitoneal injection 3.5 h after the insulin injection. This insulin dose is lower than that reported to elicit reproducible hypoglycemic neuronal injury in adult rats supported with assisted ventilation (l), but is consistent with doses used by other investigators to induce hypoglycemia in P7 rats (37). Using this regimen, two independent experiments were performed to evaluate the influence of hypoglycemia on gp120 neurotoxicity. Animals were fasted for 2 h and anesthetized by ether inhalation, and received injections of 50 ng gp120 into right dorsal hippocampus. Fifteen minutes later, rats were injected intraperitoneally with insulin (7-8 u/kg, in 0.1 ml, n = 12) or an equal volume of saline (n = 12). Animals were closely monitored and kept warm under a heating lamp over the next 4 h. Hypoglycemic animals were pale and lethargic during this period; after ip injections of 10% dextrose/ water (0.1-0.2 ml) 3.5 h later, their color and activity levels rapidly normalized. In 2/12 insulin-injected animals blood glucose did not decline to 0 mg %; these two were excluded from further analysis. Animals were then returned to the dams, and all survived. Animals were sacrificed 5 days later, on Postnatal Day 12 (P12) by decapitation; brains were removed and frozen under powdered dry ice. gp120 + carotid ligation

+ 8% oxygen exposure.

Six

independent experiments were done in which littermates (5-6/group) received intrahippocampal injections of 50 ng gp120 or equal volumes of saline, and then underwent right carotid ligation and timed exposure to 8% Oz. To elicit unilateral forebrain ischemic injury, P7 ether-anesthetized rats underwent right carotid ligation, followed by exposure to 8% oxygen (02) for 1.5-2 h, using previously published methods (3, 22, 29). The intrahippocampal injection was done 1 h before ligation. In this stroke model, features of ischemic injury range from selective neuronal loss to hemispheric infarction; the duration of exposure to hypoxia is a major determinant of severity of injury, and there is also some unexplained inter- and intraexperiment variability in the severity of injury (25,34). The threshold duration of 8% Oz exposure reported to elicit any evidence of neuronal injury is about 1.5 h. Initially, we tested relatively short durations of 8% 02 exposure, since we were concerned that the addition of a second surgical procedure to the experimental protocol would increase susceptibility to the deleterious effects of carotid ligation and hypoxia. Severity of ipsilateral forebrain injury was assessed by evaluation of histopathology on P12. In the first experiment, animals were exposed to 8% 02 for 1.5

NEUROTOXICITY

125

h and none had evidence of injury on P12; in the second experiment, animals were exposed to 1.75 h of 8% 0s and again there was no injury on P12 (results not shown). In next four experiments, animals were exposed to 2 h 8% 0s and all animals had evidence of ischemic injury when evaluated 5 days later, on P12. Histopathology

Twenty-micrometer coronal frozen brain sections, postfixed over paraformaldehyde vapors, were stained with cresyl violet. After intrahippocampal injection of saline, the needle track may be identifiable, but there is no adjacent cell loss or distortion of regional anatomy; similarly, intrahippocampal injection of 50 ng gp120 results in minimal hippocampal injury, with focal pyramidal cell loss immediately adjacent to the injection track, and no apparent hippocampal atrophy. Hippocampal atrophy and pyramidal cell loss are reproducible findings in models of E&A-mediated perinatal brain injury (e.g., unilateral intracerebral injection of NMDA, 23, 24), and gp120 potentiation of NMDA toxicity is manifest primarily by a reduction in volume of the injected hippocampus, reflecting injury to the dendritic arbor. Initial review of histopathology in the gp120/ hypoglycemia experiments indicated that there was only subtle evidence of injury, which could not be reliably scored. It was, however, feasible to evaluate histopathology systematically in the animals that had undergone hypoxic-ischemic lesioning; severity of ischemic injury was scored by an observer blinded to animal identity (F.S.S.) using an arbitrary O-3 point scale (summarized in Table 2). Image Analysis

Since injury is incurred during a phase of rapid brain growth, cross-sectional areas and derived volume estimates on P12 (5 days after injury) provide reliable objective measures of the severity of tissue damage; neurochemical measurement, e.g., choline acetyl transferase activity, confirm that loss of mass closely parallels loss of specific neuronal markers (24). A computerized image analysis system (MCID, Imaging Research, St. Catharines, Ontario) was used to estimate volumes of structures of interest; regional cross-sectional areas of every sixth coronal 20-cl,rn brain section were measured, and values/region were summated and multiplied by the distance between sections (120 p,m) (16, 32). Measurements were obtained only from regions with intact Nissl-staining. Dorsal hippocampal volumes were estimated, based on measurements beginning at the rostral tip and continuing posteriorly to the level of the caudal limit of the corpus callosum. To quantify severity of injury after hypoxic-ischemic lesioning, volumes of cerebral hemisphere, caudate, dorsal hippocampus (to the

126

BARKS ET AL.

FIG. 1. This montage compares the features of hipp~~p~ his~patho~o~ elicited by right ~ntrah~ppo~p~ injection of 50 ng gp120 in two groups of P7 animals, lesioned in a single experiment, and sacrificed 5 days later. In 6/12 animals, 7-8 u/kg of insulin was injected intraperitoneally (ip) 15 min after the gp120 injection, and these animals received an ip injection of 10% glucose (0.1 ml) 3 h later; 6/12 litter-mates received ip injections of 0.1 mI saline (PBS) 15 min after the intracerebral gp120 injections. The coronal section presented from each brain demonstrated the most severe hippocampal injury observed. fn sections from all 6 animals in the gpl20 + hypoglycemia group (A-F) right hippocampal atrophy is visually apparent; in contrast, focal atrophy in detectable in only l/6 animals (G) from the gp120 + saline group (G-L). Even in animals with prominent hippocampal atrophy, only minimal pyramidal cell loss (arrowheads), limited to regions immediately adjacent to the injection track, is evident. The arrows identify the mechanical defects through cortex and external capsule resulting from the injection procedure. Scale bar: 1 mm.

g-p120

AUGMENTS

NEUROTOXICITY

127

FIG. 2. This montage demonstrates the antero-posterior extent of hippocampal atrophy elicited by right intrahippocampal injection of 50 ng gp120 in a P7 rat that was rendered hypoglycemic by ip injection of 8 u/kg insulin 15 min after gp120 administration; the animal received an ip injection of 10% glucose (0.1 ml) 3 h later, and was subsequently sacrificed on P12. Unilateral hippocampal atrophy is visually apparent throughout the dorsal hippocampal; arrowheads identify focal neuronal loss adjacent to the injection track (D, E, and G). Scale bar: 2 mm.

level of the caudal limit of the corpus callosum), and cortex (at the level of dorsal hippocampus) were measured bilaterally.

changeably; gp120 was thawed and aliquoted (100 ng/ l.~l saline) upon receipt, stored at -70°C and rethawed immediately before use. Other chemicals were purchased from Sigma (St. Louis, MO).

Materials gp120 (strain III B, recombinant, glycosylated), purchased from American Bio-Technologies, Inc. (Cambridge, MA) and gpl20 (strain SF2, recombinant, glycosylated) obtained from the AIDS Research and Reference Reagent Program (NIH, Bethesda, MD) were used inter-

Data Analysis Microcomputer-based statistical programs [Statview II (ABACUS, Berkeley, CA) and Systat (Systat, Inc., Evanston, IL)] were used. Interhemispheric differences were calculated as the means of differences in bilateral

128

BARKS

volumes/animal in each group. Mean volume differences and histopathology score differences were compared by Mann-Whitney ranking tests or ANOVA.

RESULTS

gp120 + Hypoglycemia Overall, in animals that received intrahippocampal injections of 50 ng gp120 there was little evidence of tissue injury on P12. Figure 1 demonstrates the range of histopathologic abnormalities encountered in all 12 animals from one of the two experiments in which the impact of hypoglycemia on outcome was assessed; the sections presented demonstrate the most severe abnormalities that were detected in each brain. In animals that received intrahippocampal injections of gp120, followed by an ip saline injection (Figs. lG-lL), the only neuropathologic abnormality detected on P12 was focal hippocampal pyramidal cell loss adjacent to the injection track (as in Fig. 1G). In sections from animals that had undergone insulin-induced hypoglycemia following gp120 injection (Figs. lA-1F and Fig. 21, the most consistent abnormality was mild ipsilateral hippocampal atrophy; pyramidal neuron loss was only seen immediately adjacent to the injection track, even in cases where hippocampal atrophy was readily apparent (e.g., Fig. 1A and Fig. 2). The range of histopathological findings was the same in the second experiment. Comparison of bilateral hippocampal volumes demonstrated substantial reductions in right hippocampal volumes in the insulin-treated animals. Table 1 summarizes the hippdcampal volume measurements and calculated interhemispheric volume differences in 12 brains, corresponding with the samples presented in Fig. 1. In the hypoglycemic group, interhemispheric differences ranged from -8.8 to -25%; in the control group, values ranged from + 3.3 to - 11%. Analysis of pooled results from the two experiments performed demonstrated that in animals from the gpl20/saline group, mean right hippocampal volumes on P12 were minimally reduced (mean 2 SE: -3.0 + 1.2%) compared to left hippocampal volumes; in animals from the gpl20/insulin group, there was a much greater reduction in right hippocampal volume (- 18.4 ? 2.0%, P < 0.001, comparing interhemispheric differences by unpaired t test). Similarly, comparison of mean right hippocampal volumes [(mm31 f SE] in the two groups revealed a significant reduction in the hypoglycemic animals (gpl20/saline, 8.9 2 0.3; gpl20/insulin, 8.0 & 0.3; P < 0.05, unpaired t test). There was no reduction in left hippocampal volume in the hypoglycemic group (gpl20/saline, 9.2 + 0.9; gplaO/insulin, 9.9 r 1.0; P = NS).

ET AL.

gp120 + Hypoxia-Ischemia Intrahippocampal injection of 50 ng gp120 1 h before animals underwent right carotid artery ligation and 2-h exposure to 890 oxygen greatly increased the severity of resulting ipsilateral forebrain injury, in comparison with litter-mates that received saline injections, based both on histopathology assessment and on comparisons of bilateral regional volume measurements. Figure 3 compares the distribution of injury in two animals that underwent intrahippocampal injection of gp120 or of PBS, followed by right carotid artery ligation + 2 h 8% O2 exposure. In Figs. 3A and 3B, prepared from an animal that received an intrahippocampal saline injection 1 h preligation, there is subtle substance loss in the right cortex and hippocampus and diffuse pyramidal cell loss; major anatomic landmarks are readily discernible. Sections in Figs. 3C and 3D demonstrate marked striatal, hippocampal, and cortical injury; it is important to note that lesions of comparable severity are commonly seen in non-gpl20-injected animals that are exposed to longer durations of hypoxia (e.g., 3 h). In two of the four experiments that were done, the ischemic injury in both gp120- and saline-injected animals was less severe than in the other two experiments. TABLE

1

Comparison of Bilateral Hippocampal Volumes in Animals That Received Right Intrahippocampal Injections of 50 ng gp120” Hippocampal Identification A B C D E F G H I J K L

gp120 gp120 gp120 g-p120 gp120 gp120 gp120 gp120 gp120 gp120 gp120 gp120

group + + + + + + + + + + + +

insulin insulin insulin insulin insulin insulin saline saline saline saline saline saline

volume*

(mm3)

Left

Right

B Difference’

10.06 9.62 9.07 10.96 8.46 9.46 8.14 9.78 a.47 7.83 9.63 9.14

7.54 7.28 6.86 9.26 7.72 8.14 8.41 9.15 7.53 7.68 8.91 9.28

-25.08 -24.29 -24.30 - 15.56 -8.81 - 13.90 +3.33 -6.44 -11.09 - 1.90 -7.46 1.59

0 On P7 animals received right intrahippocampai injection of 50 ng gp120 (see text for details of methods); 15 min later animals received intraperitoneal injections of insulin (7-8 u/kg) or an equal volume of saline. Insulin-injected animals received a “rescue” injection of 10% glucose 3 h later. AI1 animals were sacrificed on P12. b Hippocampal volumes were estimated by summing dorsal hippocampal areas measured by image analysis in coronal sections 120 pm apart and multiplying by the distance between sections (see text for details of methods). c % difference = [(RH vol - LH vol)/LH voll x 100, where RH vol = right hippocampal volume, LH vol = left hippocampal volume.

gp120

AUGMENTS

129

NEUROTOXICI’IY

FIG. 3. Comparison of the severity of injury in two animals that underwent right carotid ligation and 2 h exposure to 87~ 0s on P7, preceded 1 h preligation by an intrahippocampal injection of saline (PBS) (A and Bj or of 50 nggpl20 (C and D). In panel A there is subtle substance loss in the right cortex; in panel B the black arrow demonstrates the site of the injection in the cortex, the arrowhead points to the hippocampus which demonstrates mild atrophy and diffuse subtle pyramidal cell loss, and the open arrow identifies a region of mild cortical cell loss and atrophy. In panel C, marked atrophy of the right hemisphere is apparent; the asterisks identify prominent striatal and cortical injury; in panel D, the large arrow identifies severe cortical infarction, and the arrowheads indicate marked hippocampal atrophy. C, cortex; S, striatum; H, hippocampus. Scale bar: 1 mm.

Figure 4 presents representative hippocampal histopathology from the two experiments in which the overall severity of injury was less pronounced; hippocampal injury is compared in sections prepared from eight different animals; in gpl20-injected animals (Figs. 4A, 4C, 4E, and 4G1 CA3 pyramidal cell loss is more extensive and hippocampal atrophy is more pronounced than in saline-injected animals. The histopathology scores were analyzed separately for the two groups of experiments. Table 2 summarizes the scoring criteria and compares the scores in gpl20- and saline-injected animals; the severity of injury was greater in both groups of gp120treated animals (Group A: milder injury, P = 0.02; Group B: moderate-severe injury, P = 0.0001, 2-way ANOVA, factoring treatment and region). To quantify the severity of injury, volumes in four brain regions of the lesioned and contralateral hemispheres were measured in all animals, and data from all four experiments were included in the statistical analysis. Figure 5 compares the severity of injury (based on loss of tissue mass) in all animals that received right intrahippocampal injections of saline or of 50 ng gp120 before carotid ligation and exposure to 8% oxygen for 2

h; in the four brain regions compared, the extent of injury was considerably greater in the gpl20-treated group (P < 0.0005, 2-way ANOVA, factoring treatment and region). DISCUSSION

In perinatal rodents gp120 influences the impact of hypoglycemic and hypoxic-ischemic insults on neuropathological outcome. Induction of transient profound hypoglycemia (blood glucose = 0 mg/dl) after intrahippocampal injection of gp120 markedly worsened outcome. Hypoglycemia-mediated potentiation of gp120 neurotoxicity resulted in hippocampal atrophy with little pyramidal neuron loss, presumably reflecting damage to the dendritic arbor. Hippocampal cross-sectional area measures provided sensitive indicators of dendritic injury at this developmental stage. Our focus on analysis of hippocampal injury was based on the in vitro evidence that gp120 was toxic to hippocampal neurons (61, and our preceding data demonstrating potentiation of hippocampal NMDA neurotoxicity by gp120 (301. The neuropathology of combined gpl20lhypoglycemic injury closely parallels the features of hippocampal histopathol-

130

FIG. 4. Comparison of the severity of hippocampal injury in eight P12 animals that underwent right carotid ligation and 2 h 8% oxygen exposure on P7, preceded 1 h earlier by an intrahippocampal injection of 50 ng gp120 (A, C, E, G) or saline (PBS) (B, D, F, H). In the two experiments which included these eight animals the lesions in both the saline and gpl20-injected animals were more circumscribed than in the two preceding experiments (from which representative lesions were presented in Fig. 2). In each coronal brain section, the arrowhead identifies the needle track in overlying cortex and/or white matter. In animals that received gp120 hippocampal atrophy is more prominent than pyramidal cell loss; in panels C, E, and G, asterisks identify the CA3 subfield where pyramidal cell loss is most extensive. More subtle CA3 cell loss is also detectable in 2/4 of the PBS-injected animals (B and H). Scale bar: 1 mm.

ogy observed after intrahippocampal coinjection of the same dose of gp120 and a dose of NMDA close to the threshold to elicit injury in P7 rats (5 nmol) (30). This pattern of injury is also consistent with evidence that loss of dendritic markers is a feature of neurotoxicity resulting from overexpression of gp120 in transgenic mice (331, neurotoxicity elicited by repeated systemic administration of gp120 (131, and HIV-1 encephalopathy (20,21,36). The impact of hypoglycemia on the perinatal brain may depend on the rapidity, duration, and means of induction of the insult; for example, fasting-induced hypoglycemia reduced focal hypoxic-ischemic injury in P7 rats while a similar degree of insulin-induced hypoglycemia had no effect (37). Such differences in outcome relate to other metabolic effects of fasting, e.g., ketonemia, which may provide the immature brain with alter-

native energy substrates. The hypoglycemia protocol we employed combined brief fasting and insulin administration, and we did not attempt to characterize changes in cerebral glucose metabolism elicited by this complex insult. Our primary rationale for selection of this method was to replicate conditions that had resulted in stimulation of glutamate efflux (281, and yet enable survival of a majority of animals. With the exception of a single abstract in which the age of rats, insulin dose, and the timing and extent of “acute neuronal swelling” and “hyperchromatic” changes are not stated (141, there are no published reports suggesting that insulin-induced hypoglycemia induces neuronal injury in immature animals. The unilateral hippocampal atrophy that we observed clearly cannot be attributed solely to hypoglycemia, since the volume of the contralateral noninjected hippocampus was not reduced.

gp120 AUGMENTS

TABLE 2

ip-

Comparison of Histopathology Scores in Animals That Underwent Hypoxic-Ischemic Lesioning Preceded by Intr&ppmpal Injection of 50 ng gp120 or Saline Region

Treatment

Cortex, anterior

gp120 Saline gp120 Saline gp120 Saline gp120 Saline gp120 Saline gp120 Saline

Cortex, posterior Striatum Hippocampus, anterior Hippocampus, posterior Thalamus

Histopathologic

Group Aa

Group Bb

0.5 0.0 1.4 0.8 0.8 0.7 1.8 1.2 2.0 1.7 0.8 0.8

2.6 1.0 2.8 1.7 2.5 1.3 2.8 2.3 2.5 1.8 2.5 1.0

2 0.3 + 0 + 0.3 2 0.2 2 0.3 + 0.3 2 0.2 ?z 0.4 + 0.1 2 0.3 f 0.3 f 0.2

1

2

Cortex

Normal

Striatum

Normal

Slight pallor, columnar cell loss Slight pallor; minimal atrophy

Hippocampus

Normal

Thalamus

Normal

Thinning of cortical mantle Atrophy ( ventricular diameter Cell loss; overt atrophy; landmarks identifiable Cell loss + atrophy

Subtle pallor

r 0.3 If: 0.4 -t 0.2 2 0.2 2 0.2 -r 0.3 2 0.2 d 0.3 f 0.2 f 0.2 f 0.2 f 0.4

scoring criteria

0

Cell loss; no atrophy

Lo IB” w zm $120 ,_. __. _.. : : ,.:, ,..’ ./...::..:: g&NJ ;$$I ,.: _,.. ‘,..l ,..::..::.. ._.‘...~‘..~_ _.: ,... . ,.. ;. . ,: ...: .._. _::_:,..~:_ EB .::. ‘_.‘... in d 131

NEUBOTOXICITY

3 Infarction Atrophy (>1/2sectionsl; ventricu1ar > striatal diameter Infarction; loss of major anatomic landmarks Loss of identifiable anatomic landmarks

Note. Cortex, anterior: cortex overlying striatum; cortex, posterior: cortex overlying hippocampus; hippocampus, anterior: anterior to habenula; hippocampus, posterior: at level of habenula. Scoring was done by an individual unaware of animal identity, based on assessment of 2 60 sections/brain. Values are expressed as mean 2 SEM. o Group A: pooled results from two experiments in which mild hypoxic-ischemic lesions evolved; there were significant outcome differences between gp120 and saline-treated animals (2-way ANOVA, P = 0.02). b Group B: pooled results from two experiments in which moderatesevere hypoxic-ischemic lesions evolved; there were significant outcome differences between gp120 and saline-treated animals (2-way ANOVA, P = 0.0001).

Since hypoglycemia increases extracellular fluid glutamate accumulation, it is attractive to speculate that exposure to gp120 potentiated the deleterious effects of endogenous EAA. If we could provide evidence that treatment with an excitatory amino acid receptor antagonist (e.g., MK-801) blocked hypoglycemic potentiation of

CORTRX

STRIATUM

FIG. 5. Quantitative measurements of severity of ipsilateral forebrain injury in all animals that underwent right carotid ligation and 2 h 8% oxygen exposure on P7, preceded 1 h earlier by a right intrahippocampal injection of 1 ul saline (PBS group, n = 18) or 50 ng gp120 (GPl20 group, n = 23). Animals were sacrificed on P12, and severity of injury was assessed by comparison of interhemispheric differences in volumes of cerebral hemispheres, cortex, hippocampus, and striatum. Bilateral cross-sectional areas (measured by computerized image analysis, see Methods) were used to estimate volumes of the four structures indicated; cortex was measured at the level of hippocampus. Values are means + SEM. *P < 0.001, S-way ANOVA, comparing overall severity of injury in the gp120 and PBS-injected groups. Post hoc Fisher LSD tests confirmed signiticant differences between the two groups (P I 0.0051 in severity of injury in cortex, hippocampus, and striatum.

gp120 neurotoxicity (22), this would provide considerable support for the latter hypothesis. However, the combined cardio-respiratory depressant effects of profound insulin-induced hypoglycemia and MK-801 treatment limit the technical feasibility of such experiments in P7 rats. In adult rat brain, intracerebroventricular injection of gp120 (l-10 pmol) resulted in global reductions in cerebral metabolic rate for glucose acutely (15), and considerable recent data indicate that neuronal energy deprivation increases susceptibility to excitotoxic injury (26). Thus, both disruption of energy metabolism and increased accumulation of and/or susceptibility to endogenous EAA may contribute to gp120 neurotoxicity in this experimental paradigm. Pretreatment with gp120 increased the extent and severity of hypoxic-ischemic injury relative to littermate controls; it did not reduce the temporal threshold for onset of ischemic injury, which suggests that its pathogenetic role involves augmentation rather than initiation of injury. The deleterious effects of gp120 administration prior to hypoxic-ischemic lesioning extended beyond hippocampus to include adjacent brain regions. gp120 may stimulate production of diffusible factors [such as cytokines and/or arachidonic acid (31, 391 that contribute to the progression of ischemic

132

BARKS ET AL

injury (27). Although the relatively widespread exacerbation of ischemic injury was unexpected, it is reminiscent of the histopathologic features of excitotoxic injury observed in P7 rats, in that focal injections of NMDA (at high doses, e.g., 25 nmol) also elicit widespread ipsilateral forebrain injury (23, 24). Although questions could be raised as to whether the deleterious effects of gp120 were a nonspecific effect of injecting a foreign protein into the brain, the results of the preceding hypoglycemia experiments, in which all animals received the same gp120 dose, would argue strongly against this interpretation. Mechanisms underlying potentiation of ischemiainduced neurotoxicity by gp120 could include direct agonist effects or enhanced functional activity of EAA. In vitro data indicate that gp120 has no direct agonist effects at EAA recognition sites or at allosteric regulatory sites of the NMDA receptor (17). If gp120 suppressed presynaptic and/or glial EAA reuptake activity, this would augment synaptic concentrations of the endogenous agonist. In cultured astrocytes, gp120 activates the amiloride-sensitive Na+-H+ transporter, potassium conductance, and glutamate efflux (5); if gp120 elicits these effects in Go, they could contribute to neuronal depolarization and sustained elevations in synaptic concentrations of glutamate. In vitro and in uiuo studies suggest several possible indirect mediators of gp120 neurotoxicity. HIV-infected macrophages secrete unidentified heat-stable factor(s) whose neurotoxic actions are blocked by NMDA antagonists; in vitro, gp120 neurotoxicity is dependent on the presence of macrophages (11, 12, 19). Nitric oxide (generated by NMDA receptor activation and/or by nonneuronal cells) may also contribute directly to gp120/ EAA neurotoxicity (9). In adult rodents, intraventricular injection of gp120 (100 ng4 kg) stimulated induction of interleukin-19 activity (31), and this cytokine may be an important mediator of ischemic brain injury (27). Our results illustrate some of the critical factors that may influence outcome and interpretation of results of studies of gp120 neurotoxicity in uiuo. A recent report concluded that gp120 was not neurotoxic in hippocampus of adult rats; several factors could account for this discrepant finding (2). In that study, gp120 was administered by continuous infusion through an implanted pump, yet the stability of gp120 at body temperature over many hours has not been documented. Injections were made into the lateral ventricle, and rapid diffusion may have limited the amount of gp120 to which the hippocampus was exposed. Furthermore, the outcome measure that was used, CA1 neuronal cell counts, may not be the most sensitive indicator of gpl20-mediated neurotoxicity; our data suggest that dendritic injury is more readily discerned than neuronal loss. In addition,

our data demonstrate that gp120 neurotoxicity is most readily documented in experimental models in which its pathogenetic role involves potentiation rather than initiation of injury. Finally, developmental stage may also be a determinant of susceptibility to gp120. Our experiments were restricted to a single developmental stage, selected based on the ontogenetic peak of susceptibility to NMDA-mediated brain injury. Maturational factors may be important determinants of susceptibility to gp120 neurotoxicity and the extent to which our findings our relevant to the adult brain remains to be determined. Identification of the mediators of HIV neurotoxicity in uiuo is an essential prerequisite for development of effective neuroprotective therapies. These results provide support for the hypotheses that clinically relevant pathophysiologic insults can increase susceptibility to gp120 neurotoxicity in uiuo and that locally secreted gp120 may worsen outcome in HIV-infected infants who are subject to hypoglycemia or cerebral hypoxiaischemia. ACKNOWLEDGMENTS This work was supported by Grants 500226 from the Pediatric AIDS Foundation and USPHS Grant NS 31054. Preliminary reports of these data were presented at the annual meetings of the Child Neurology Society, October, 1993 (Orlando, FL) and the Society for Neuroscience, November, 1993 (Washington, DC) and November, i994 (Miami. FL). REFERENCES 1. ALJER,R. N., T. WIELOCH, Y. OLSSON,AND B. K. SIESJO. 1984. The distribution of hypoglycemic brain damage. Acta Neuropathol. (Bet-l.1 64: 177-191. 2. BAGETTA, G., FINAZZI-AGRO,A., PALMA, E., ANDNISTICO, G. 1994. Intra-cerebral injection of human immunodeficiency virus type 1 coat glycoprotein gp120 does not produce neurodegeneration in rats. Neurosci. Lett. 176: 97-100. 3. BARKS, J. D. E., AND F. S. SILVERSTEIN. 1992. Excitatory amino acids contribute to the pathogenesis of perinatai hypoxicischemic brain injury. Brain Pathol. 2:235-243. 4. BELMAN, A. L. 1992. Acquired immunodeficiency syndrome and the child’s central nervous system. Pediatr. C&n. N. Am. 39: 691-714. 5. BENOS, D. J., B. H. HAHN, J. K. BIJBIEN, et al. 1994. Envelope glycoprotein gp120 of human immunodeficiency virus type 1 alters ion transport in astrocytes: Implications for AIDS dementia complex. Proc. Natl. Acod. Sci. USA 91: 494-498. 6. BRENNEMAN, D. E., G. L. WESTBROOK, S. P. FITZGERALD,D. L. ENNIST, K. L. ELINS, M. R. RUFF, AND C. B. PERT. 1988. Neuronai cell killing by the envelope protein of HIV and its prevention by vasoactive intestinal peptide. Nature 335: 639-642. 7. CHOI, D. W., AND S. M. ROTHMAN. 1990. The role of glutamate neurotoxicity in hypoxic-ischemic neuronal death. Annu. Rev. Neurosci. 13: 171-182. 8. COHEN, R. S., AND D. K. STEVENSON. 1989. Hypoglycemia and brain injury. In Fetal and Neonatal Brain Injury: Mechanisms,

~120 AUGMENTS and the Risks of Practice In (D. K. Stevenson and P. Sunshine, Eds.), pp. 141-146. Decker, Philadelphia. DAWSON,V. L., T. M. DAWSON,G. R. UHL, ANDS. H. SYNDER.1993. Human immunodeficiency virus type-l coat protein neurotoxicity mediated by nitric oxide in primary cortical cultures. Proc.

Management, 9.

Natl.

Acad.

Sci. USA

90: 3256-3259.

10. DECARLI, C., L. A. CIVITELLO, P. BROU~ERS, AND P. A. PIZZO. 1993. The prevalence of computed tomographic abnormalities of the cerebrum in 100 consecutive children symptomatic with the human immune deficiency virus. Ann. Neural. 34: 198-205. 11. GIULIAN, D., K. VACA, AND C. A. NOONAN. 1990. Secretion of neurotoxins by mononuclear phagocytes infected with HIV-l. Science 250: 1593-1596. 12. GIULIAN, D., E. WENDT, K. VACA, AND C. A. NOONAN. 1993. The envelope glycoprotein of HIV type 1 stimulates release of neurotoxins from monocytes. Proc. Natl. Acad. Sci. USA 90: 27692773. 13. HILL, J. M., R. F. MERVIS, R. AVIDOR, T. W. MOODY, AND D. E. BRENNEMAN. 1993. HIV envelope protein-induced neuronal damage and retardation of behavioral development in rat neonates. Brain

Res. 603:

222-233.

14. JONES, E. L., AND W. T. SMITH. 1970. The effects of neonatal hypoglycemia on rat brain. J. Pathol. 101: Px. 15. KIMES, A. S., E. D. LONDON, G. SZABO, L. RAYMON, AND B. TABAKOFF. 1991. Reduction of cerebral glucose utilization by the HIV envelope glycoprotein gp-120. Exp. Neural. 112: 224-228. 16. KONIGSMARK, B. W. 1970. Methods for the counting of neurons. In Contemporary Research Methods in Neuroanatomy 0%‘. J. H. Nauta and S. 0. E. Ebbeson, Eds.), pp. 315-340. Springer, New York. 17. LIPTON, S. A., N. J. SUCHER, P. K. KAISER, AND E. B. DREYER. 1991. Synergistic effects of HIV coat protein and NMDA receptor mediated neurotoxicity. Neuron ?: 11 l-l 18. 18. LIPTON, S. A. 1992. Models of neuronal injury in AIDS: Another role for the NMDA receptor? Trends Neurosci 15: 76-80. 19. LIPTON, S. A. 1992. Requirement for macrophages in neuronal injury induced by HIV envelope protein gp120. NeuroReport 3: 913-915. 20. MASLIAH, E., N. GE, M. MOREY, R. DE TERESA, R. D. TERRY, AND C. A. WILEY. 1992. Cortical dendritic pathology in human immunodeficiency virus encephalitis. Lab. [west. 66: 285-291. 21. WLWI, E., M. GE, C. L. ACHIM, et al. 1992. Selective neuronal vulnerability in HIV encephalitis. J. Neuropathol. Exp. Neural. 51: 585593. 22. MCDONALD, J. W., F. S. SILVERSTEIN,AND M. V. JOHNSTON.1987. MK-801 protects the neonatal brain from hypoxic-ischemic damage. Eur. J. Pharmacol. 140: 359-361. 23. MCDONALD, J. W., F. S. SILVERSTEIN,AND M. V. JOHNSTON.1988. Neurotoxicity of NMDA is markedly enhanced in developing rat central nervous system. Brain Res. 469: 200-203.

NEUROTOXICITY

133

24. MCDONALD, J. W., N. F. ROESER, F. S. SILVERSTEIN, AND M. V. JOHNSTON. 1989. Quantitative assessment of neuroprotection against NMDA-induced brain injury. Exp. Neurol. 108: 289-296. 25. NELSON, C., AND F. S. SILVERSTEIN. 1994. Acute disruption of cytochrome oxidase activity in brain in a perinatal rat stroke model. Pediatr. Res. 36: 12-19. 26. NOVELLI, A., J. A. REILLY, P. G. LYSKO, AND R. C. HENNEBERRY. 1988. Glutamate becomes neurotoxic via the N-methyl-Daspartate receptor when intracellular energy levels are reduced. Brain Res. 451: 205-212. 27. RELTON, J. K., AND N. J. ROTHWELL. 1992. Interleukin-1 receptor antagonist inhibits ischemic and excitotoxic neuronal damage in the rat. Brain Res. Bull. 29: 243-246. 28. SILVERSTEIN,F. S., J. SIMPSON,AND K. E. GORDON.1990. Hypoglycemia alters striatal excitatory amino acid efflux in perinatal rats: An in viva microdialysis study. Ann. Neural. 28: 516-521. 29. SILVERSTEIN, F. S., B. NAIK, AND J. SIMPSON. 1991. Hy-poxiaischemia stimulates hippocampal glutamate efflux in perinatal rat brain: An in vivo microdialysis study. Pediatr. Res. 30: 587-590.

30.

31.

32. 33.

34. 35.

36. 37.

SILVERSTEIN,F. S., J. D. E. BARKS, AND R. SUN. 1993. GP120, an HIV-derived peptide, selectively augments NMDA-mediated hippocampal injury in perinatal rodents. Sot. Neurosci. Abstr. 19: 1772. SUNDAR, S. K., M. A. CIERPIAL, L. S. KAMARATU, et al. 1991. Human immunodeficiency virus glycoprotein (gp120) infused into rat brain induces interleukin 1 to elevate pituitary-adrenal activity and decrease peripheral cellular immune responses. Proc. Natl. Acad. Sci. USA 188: 11246-11250. SWANSON,R. A., M. T. MORTON, G. Ts~o-Wu, et al. 1990. A semiautomated method for measuring brain infarct volume. J. Cereb. Blood Flow Metab. 10: 290-294. Tocc~s, S. M., E. M,ULIAH, E. M. ROCKENSTEIN,et al. 1994. Central nervous system damage produced by expression of the HIV-l coat protein gp120 in transgenic mice. Nature 367: 188-193. TO~FIGHI, J., J. Y. YAGER, C. HOUSMAN, AND R. C. VANNUCCI. 1991. Neuropathology of remote hy-poxic-ischemic damage in the immature rat. Acta Neuropathol. 81: 578-587. WAHL, L. M., M. L. CORCORAN,S. E. PyLE, et al. 1989. HIV glycoprotein (gpl20) induction of monocyte arachidonic acid metabolites and interleukin 1. Proc. Natl. Acad. Sci. USA 86: 621-625. WILEY, C. A., E. MASLIAH, M. MOREY, C. LEMERE, R. DETERESA,M. GRAFE, L. HANSEN, AND R. TERRY. 1991. Neocortical damage during HIV infection. Ann. Neural. 29: 651-657. YAGER, J. Y., D. F. HEITJAN, J. TOWFIGHI, AND R. C. VANNUCCI. 1992. Effect of insulin-induced and fasting hypoglycemia on perinatal hypoxic-ischemic brain damage. Pediatr. Res. 31: 138142.