A Rat Model of HIV-1 Encephalopathy Using Envelope Glycoprotein gp120 Expression Delivered by SV40 Vectors

Abstract

Human immunodeficiency virus (HIV)-1 encephalopathy (HIVE) is thought to result in part from the toxicity of HIV-1 envelope (Env) glycoprotein gp120 for neurons. Experimental systems for studying the effects of gp120 and other HIV proteins on the brain have been limited to the acute effects of recombinant proteins in vitro or in vivo or simian immunodeficiency virus-infected monkeys. We describe an experimental rodent model of ongoing gp120-induced neurotoxicity in which HIV-1 Env is expressed in the brain using a SV40-derived gene delivery vector, SV(gp120). When it is inoculated stereotaxically into the rat caudate putamen (CP), SV(gp120) caused a partly hemorrhagic lesion in which neuron and other cell apoptosis continue for at least 12 weeks. HIV gp120 is expressed throughout this time and some apoptotic cells are gp120-positive. Malonaldehyde and 4-hydroxynonenal assays indicated that there was lipid peroxidation in these lesions. Prior administration of recombinant SV40 vectors carrying antioxidant enzymes, Cu/Zn superoxide dismutase or glutathione peroxidase was protective from SV(gp120)-induced oxidative injury and apoptosis. Thus, in vivo inoculation of SV(gp120) into the rat CP causes ongoing oxidative stress and apoptosis in neurons and may therefore represent a useful animal model for studying the pathogenesis and treatment of HIV-1 Env-related brain damage.

INTRODUCTION

Human immunodeficiency virus (HIV) enters the brain very early after its initial entry into the body. The principal manifestations of central nervous system (HIV) infection result from neuronal injury and loss and from extensive damage to the dendritic and synaptic structures in the absence of neuronal loss. Neurons are rarely infected by HIV-1 and damage to neurons is felt mainly to be indirect (1, 2). HIV-1 infects resident microglia, periventricular macrophages, and some astrocytes (3) leading to increased production of cytokines and the release of cytotoxic HIV-1 proteins (4). Several HIV gene products have been identified as neurotoxins; these include the envelope (Env) proteins gp120 and gp41 and the nonstructural proteins Tat, Nef, Vpr and Rev. Gp120-induced apoptosis has been demonstrated in cortical cell cultures, in rat hippocampal slices and by intracerebral injections in vivo (5). Soluble gp120 triggers oxidative stress and leads to neuron cell death after mitochondrial permeabilization, cytochrome C release and activation of caspases and endonucleases (6).

Several animal models have been used to study the pathogenesis of HIV-1-induced neurological disease. Many of them are based on other lentiviruses. Feline immunodeficiency virus infection of cats (7, 8), Visna-Maedi virus infection of sheep (9), and simian immunodeficiency virus (SIV) infection of macaques (10, 11) have all greatly contributed to our present knowledge. In some of these models, however, only small percentages of animals develop neurological manifestations despite the development of more reproducibly neurovirulent strains (12); moreover, animal costs for these species may be high (13). Rodent models have also been developed. Transgenic expression of gp120 (14) has been used but gp120 in transgenic mice is expressed chiefly in astrocytes, whereas HIV-1 mainly infects microglial cells in humans. The introduction of HIV-infected macrophages into severe combined immunodeficient (SCID) mice brains induces gliosis (15), microglial activation and neuronal death (16), but this model suffers from the issue of the incompatibility of human macrophages delivered into a murine brain. Transplanted human cells into immunosuppressed mice and the use of hybrid viruses have also led to experimental central nervous system (CNS) infections that resemble neuroAIDS (17–20).

We (21–23) and others (24, 25) have used model systems in which recombinant gp120 or Tat proteins are injected directly into the striatum. The neurotoxicity of such recombinant proteins is highly reproducible but is too acute to be useful either for studying chronic HIV-related tissue injury or for testing most therapeutic interventions. The lesions of HIV-associated dementia (HAD) reflect chronic injury caused by the ongoing production of gp120 and other substances by HIV-1-infected cells. Here, we report an experimental model of chronic HIV-1 Env-induced neurotoxicity based on recombinant SV40 (rSV40) vector-mediated expression of gp120 in the brain.

MATERIALS AND METHODS

Animals

Female Sprague-Dawley rats (200–250 g) were purchased from Charles River Laboratories (Wilmington, MA). Nude rats were kindly provided by the National Institutes of Health (NIH). Protocols for injecting and euthanizing animals were approved by the Thomas Jefferson University Institutional Animal Care and Use Committee and are consistent with Association for Assessment and Accreditation of Laboratory Animal Care standards. Because estrogens can regulate microglial activation in some conditions, experiments were done in female rats at similar points of their estrous cycle. The diet that the animals received specifically avoided components that might cause oxidative stress. Numbers of animals used in experiments are indicated in the “Experimental Design” section.

Antibodies

Different primary antibodies were used: mouse anti-rat CD68/ED1 (IgG1; 1:50), a marker of activated microglial cells in a phagocytic state (Serotec, Oxford, UK), rabbit anti-Iba1 (IgG; 1:100), a marker of quiescent and active microglia (Waco Chemicals, Osaka, Japan), mouse anti-GFAP (IgG2b; 1:100) (Becton Dickinson, Franklin Lakes, NJ), rabbit anti-N-acetyllysine-4-hydroxy-2-nonenal (HNE) (IgG; 1:50), a marker of lipid peroxidation (Calbiochem, La Jolla, CA), rabbit anti-caspase-3 (IgG; 1:100) (this antibody detects the active form of caspase-3) (Santa Cruz Biotechnology, Santa Cruz, CA), rat anti-dinitrophenol (DNP) (IgG1; 1:50) (Zymed Laboratories, San Francisco, CA), mouse anti-proliferating cell nuclear antigen (PCNA) (IgG2a; 1:100), goat anti-CD68 (IgG; 1:100) (Santa Cruz Biotechnology), rabbit anti-laminin (IgG; 1:100) (Sigma), mouse anti-NeuN (IgG1; 1:100) (Chemicon International, Temecula, CA). The monoclonal antibody against gp120 (monoclonal antibody to HIV-1 V3, 257-D IV; IgG1; 1:100) was obtained through the AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH and was a generous gift from Dr. S. Zolla-Pazner. This antibody reacts with HIV-1MN V3 epitope KRIHI (26). Secondary antibodies were used at 1:100 dilution: FITC and TRITC-conjugated goat anti-mouse IgG (γ-chain specific and against whole molecule respectively), FITC-conjugated mouse anti-human IgG (γ-chain specific), TRITC-conjugated goat anti-rabbit IgG (whole molecule), FITC-conjugated sheep anti-rabbit IgG (whole molecule), FITC-conjugated rabbit anti-goat IgG (whole molecule), Cγ3-conjugated rabbit anti-goat IgG (whole molecule) (Sigma), FITC and TRITC-conjugated donkey anti-mouse IgG (whole molecule), Cγ3-conjugated donkey anti-rabbit IgG (whole molecule) and anti-goat IgG (whole molecule) (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA), FITC-conjugated goat anti-rat IgG H&L ((Fab)2 fragment; Abcam, Cambridge, MA).

Vector Production

The general principles for making recombinant Tag-deleted, replication-defective SV40 viral vectors have been previously reported (27). SV(gp120) is a recombinant Tag-deleted SV40-derived vector that expresses HIV-1NL4-3 gp120 under the control of the cytomegalovirus immediate early promoter (28). Copper/zinc superoxide dimutase-1 (SOD1) and glutathione peroxidase-1 (GPx1) transgenes were subcloned into pT7[RSVLTR], in which transgene expression is controlled by the Rous Sarcoma Virus long terminal repeat (RSV-LTR) as a promoter. The cloned rSV40 genome was excised from its carrier plasmid, gel-purified, recircularized and then transfected into COS-7 cells. These cells supply large T-antigen (Tag) and SV40 capsid proteins in trans, which are needed to produce recombinant replication-defective SV40 viral vectors (29). Crude virus stocks were prepared as cell lysates, band-purified by discontinuous sucrose density gradient ultracentrifugation and then titered by quantitative PCR (30). SV(human bilirubin-uridine 5′-diphosphate-glucuronosyl-transferase) (BUGT) was used as negative control vector with a non-toxic byproduct (31).

Experimental Design

Injection of SV(gp120) into the Caudate Putamen

SV(gp120) was injected into the caudate putamen (CP) of Sprague-Dawley rats and the brains were harvested 1, 2, 4 and 12 weeks after injection, with 6 rats at each time point (n = 24). Controls received saline or SV(BUGT) instead of SV(gp120) in the CP (3 controls for each time point, n = 12 for SV(BUGT), n = 12 for saline). SV(gp120) or either SV(BUGT) or saline were also injected into the CP of nude rats. Brains of nude rats were harvested 1, 2, 4 and 12 weeks after the injection into the CP (6 rats for each time point for the animals injected with SV(gp120) (n = 24), 3 rats at each time point for the animals injected with SV(BUGT) (n = 12) and saline (n = 12)).

Challenge with SV(gp120) after Administration of SV(GPx1)/SV(SOD1)

To study possible neuroprotection by rSV40-delivered expression of SOD1 and GPx1 from SV(gp120)-related apoptosis, we first injected the CP of rats with SV(SOD1) (n = 6) and SV(GPx1) (n = 6). One month later, the CP in which SV(SOD1) or SV(GPx1) had been administered was injected with 5 μl SV(gp120). Brains were harvested 1 week after injection and studied for apoptosis by terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay and lipid peroxidation. Controls received SV(BUGT) in the CP instead of SV(SOD1) and SV(GPx1) (n = 6).

In Vivo Transduction

Rats were anesthetized with isofluorane UPS (BaxterHealthcare Corp., Deerfield, IL) (1.0 unit isofluorane/1.5 liters O₂ per minute) and placed in a stereotaxic apparatus (Stoelting Corp., Wood Dale, IL) for cranial surgery. Body temperature was maintained at 37°C using a feedback-controlled heater (Harvard Apparatus, Boston, MA). Glass micropipettes (1.2 mm outer diameter; World Precisions Instruments, Inc., Sarasota, FL) with tip diameters of 15 μm were backfilled with either 5 μl of SV(gp120), SV(SOD1) or SV(GPx1) viral vector, which contains approximately 10⁷ infectious units. The vector-filled micropipettes were placed in the CP using coordinates obtained from the rat brain atlas of Paxinos and Watson (32). For injection into the CP, a burr hole was placed +0.48 mm anterior to bregma and −3.0 mm lateral to the sagittal suture. Once centered, the micropipette was placed 6.0 mm ventral from the top of the brain. The vector was given by a Picospritzer II (General Valve Corp., Fairfield, NJ) pulse of compressed N2 duration 10 ms at 20 psi until the fluid was completely ejected from the pipette. As positive control for immunocytochemistry for gp120, 500 ng/μl recombinant gp120 were injected into the CP as described and the brains were harvested 4 days after injection and analyzed for gp120 immunoreactivity. Following surgery, animals were housed individually with free access to water and food.

After variable survival periods, rats were anesthetized via intraperitoneal injection of sodium pentobarbital (Abbott Laboratories, North Chicago, IL) at 60 mg/kg and perfused transcardially though the ascending aorta with 10 ml heparinized saline followed by ice-cold 4% paraformaldehyde (Electron Microscopy Sciences, Fort Washington, PA) in 0.1M phosphate buffer (pH 7.4). Immediately following perfusion-fixation, the rat brains were dissected out, placed in 4% paraformaldehyde for 24 hours, then in a 30%-sucrose solution for 24 hours, then frozen in methyl butane cooled in liquid nitrogen. The samples were cut on a cryostat (10-μm-thick sections).

Immunocytochemistry

For immunofluorescence, coronal cryostat sections were processed for indirect immunofluorescence. Blocking was performed by incubating 60 minutes with 10% goat serum or 10% donkey serum in phosphate buffer saline (PBS; pH 7.4). Sections then were incubated with antibodies diluted according to manufacturer’s recommendations: 1 hour with primary antibody, then 1 hour with secondary antibody, all at room temperature. Double immunofluorescence was performed as previously described (33). Detection of blood-brain barrier (BBB) dysfunction was assessed by employing a one-step immunohistochemical detection of IgG in which sections were incubated for 1 hour with the antibody (1:100). All incubations were followed by extensive washing with PBS. Medium containing DAPI (Vector Laboratories, Burlingame, CA) was used to stain nuclei. Specimens were examined under a Leica DMRBE microscope (Leica Microsystems, Wetzlar, Germany). Negative controls consisted of pre-incubation with PBS, substitution of non-immune isotype-matched control antibodies for the primary antibody, and/or omission of the primary antibody.

TUNEL Assay

TUNEL assay was performed according to the manufacturer’s recommendations (Roche Diagnostics, Indianapolis, IN) and following a previously described protocol (21, 23).

Staining of Neurons Using NeuroTrace

NeuroTrace (NT) (Molecular Probes, Inc., Eugene, OR) staining has been used as a neuronal marker in numerous studies focusing on neuron characterization (18, 25, 34–36) and was performed as previously reported (21, 23, 37, 38). After rehydration in 0.1 M PBS, pH 7.4, sections were treated with PBS plus 0.1 % Triton X-100 10 minutes, washed twice for 5 minutes in PBS then stained by NT (1:100) for 20 minutes at room temperature. Sections were washed in PBS plus 0.1 % Triton X-100 then x 2 with PBS, then let stand for 2 hours at room temperature in PBS before being counterstained with DAPI. Combination NT and antibody staining was performed using primary and secondary antibodies staining first (see above), followed by staining with the NT fluorescent Nissl stain. For antibody, TUNEL, and NT staining, immunohistochemistry was the first step, followed by TUNEL assay, then by NT staining. All experiments were repeated 3 times and test and control slides were stained the same day.

General Morphology

Microscopic morphology of the brains was assessed by hematoxylin and eosin (H&E) of cryostat sections (10 μm thick). Briefly, cryostat sections were stained with hematoxylin during 3 minutes, rinsed in deionized and tap water, dipped in acid ethanol, rinsed in tap and deionized water before being stained by eosin during 30 seconds. Sections were dehydrated in 95% then 100% ethanol before being cleared in xylene and mounted in Permount (Fischer Scientific, Pittsburgh, PA). Morphometric study of the brain lesion was assessed by neutral red (NR) staining of the sections. Briefly, the sections were stained by NR for 5 minutes, then washed, dehydrated in alcohols and finally cleared in xylene.

Morphometry

Transduction was assessed for each injected brain (CP injections) by serial cryo-sectioning of the whole brain (10-μm-thick coronal sections), each slide being numbered, then by immunostaining of every tenth section for gp120 (i.e. at 100-μm intervals, spanning at least 3 mm of tissue rostral and 3 mm caudal to the injection site). This step determined the most anterior and posterior extent of the transduced area. Between 50 and 60 sections were immunostained for each rat. The total number (and not the number in random areas) of gp120-positive cells in one hemisphere for every tenth section was counted and summed using a computerized imaging system (Image-Pro Plus, MediaCybernetics, Bethesda, MD). The final number presented was an average of the results measured in the different sections examined. The total number of positive cells in a brain could be estimated by “multiplying” the cell counts by the length of the transduced area, assuming that the number of positive cells in the sections spanned between the first and tenth one, for example, will be close to the one measured on these first and tenth sections. This procedure, already described for assessment of numbers of transgene-positive cells in the brain (38, 39) allows quantitative and relative comparisons among different time points, although it does not reflect the total number of transduced cells in vivo. If these calculations of cell counts do not conform to unbiased stereological methods, however, that had been realized on 10-μm-thick sections and the total number of positive cells was counted in each section; stereological methods are usually applied to 40- to 50-μm-thick sections and random fields of the sections analyzed. Moreover, we used this method for enumerating total neurons in the rat normal striatum by immunocytochemistry (using anti-NeuN antibodies) and staining of neurons by NT. Preliminary results showed a total number of neurons close to the one measured using the optical fractionator method reported in the literature.

The longitudinal distribution of the transgene away from the injection site was assessed as the distance comprised between the most distant sections anterior and posterior to the injection site demonstrating transgene expression.

TUNEL-, Iba1- and CD68-positive cells were counted using Image-ProPlus in the whole CP (and not randomly chosen fields) after TUNEL assay or immunostaining with anti-CD68 and anti-Iba1 antibodies, respectively, in at least 5 consecutive sections. The final number was an average of results measured in the different sections. NT-positive cells in the CP were counted as described above and neuronal loss was estimated by calculating the ratio of the number of NT-positive cells on the injected side compared to the number of NT-positive side on the uninjected side.

The lesion areas were determined in NR-stained sections using the imaging system. Computer-assisted tracing of the perimeter of the lesion area surrounding the injection site as well as the whole CP was conducted to determine area measures. A ratio of the lesion area compared to the whole CP area was determined. A total of 20 sections (1 section every 100 μm; 10 sections rostral and 10 sections caudal to the injection site) per animal were used. The length of the extent of tissue damage was determined on serial sections of the whole brain stained by NR. This procedure was similar to one previously described (25).

Measurement of Malondialdehyde

Measurement of malondialdehyde (MDA) was used as an indicator of lipid peroxidation. MDA assay was performed using lipid peroxidation kit (Oxford Biomedical Research, Oxford, MI). The assay is based on reaction of chromogenic reagent, N-methyl-2-phenylindole with MDA and 4-hydroxyalkenals at 45°C to form a stable chromophore at 586 nm. MDA standards were prepared with target concentration in reaction mixture ranging from 0 to 4 μM and the assay was performed according to manufacturer’s instructions. Cryostat sections (50 μm thick) of the striatum areas injected with SV(gp120) or the control vector SV(BUGT) were harvested then homogenized in buffer and lysates were prepared as previously reported (21). Briefly, for 200 μl of sample, 650 μl of reagent R1 and 150 μl of 12N HCl were added and incubated for 45°C for 60 minutes. The absorbance was measured at 586 nm and the MDA levels of unknown samples were deduced from the standard curve.

Western Analysis

For in vivo protein analysis, brain sections were homogenized in buffer containing protease inhibitors. Lysates were prepared as previously reported (21). Total protein was estimated using BCA protein assay kit (Pierce, Rockford, IL). Total cell proteins (100 μg) were loaded in each well, then transferred to Immobilon-P Poly Vinylidine difluoride (PVDF) membrane (Millipore, Inc., Billerica, MA) and analyzed using BM chemiluminescence Western blotting kit (mouse/rabbit) (Roche, Indianapolis, IN). Equal lane loading was assessed using anti-glyceraldehyde-3-phosphate-dehydrogenase (GAPDH) antibody (1:5000). Anti-gp120 antibody was used at a dilution of 1:500, and anti-mouse/rabbit cocktail horseradish peroxidase was used at a dilution of 1:5,000 before addition of chemiluminescence substrate (Roche). The blot was visualized according to manufacturer’s instructions (Roche).

Statistical Analysis

Comparison of medians between 2 groups was achieved by using the Mann-Whitney test (with a two-tail p value). Comparison of medians between more than 2 groups was done by using the Kruskall-Wallis test. The difference between the groups was considered significant when p < 0.05. On graphs, values are represented as means ± SEM.

RESULTS

Injection of SV(gp120) into the CP Induces a Lesion with Neuronal Loss

Examination of brains at 1, 2 and 4 weeks after injection of SV(gp120) into the CP showed partly hemorrhagic lesions (Fig. 1A). No lesions were seen in the CP of rats injected with SV(BUGT), a control vector, at any time point after the injection. H&E staining showed a progressive loss of tissue and cells (particularly neurons) in the area injected with SV(gp120) as well as at the periphery of it. Numerous cells with fragmented nuclei were observed 7 days after injection. Features suggestive of inflammation were seen 14 days after injection. Significant reduction in neurons was observed 28 and 84 days after injection of SV(gp120) into the CP. Loss of tissue led to the development of pseudo-cystic lesions at the later time points. No abnormalities were observed after injection of SV(BUGT) into the CP or in the contralateral side in SV(gp120) recipients (not shown) (Fig. 1B, first row). Sections stained with the neuronal marker NT showed a progressive loss of neurons not only in the area directly injected with SV(gp120) but also at the periphery; almost no NT cell loss was seen following injection of SV(BUGT) (Fig. 1B, second row). Iba1 identifies both quiescent and activated microglial cells. An appearance suggestive of reactive microglial cells was observed in numerous Iba-1-positive cells from day 14 to day 84 after SV(gp120) injection (insets), whereas 7 days after injection, reactive microglial cells coexisted with Iba-1-positive cells with more normal morphology (inset). Iba1-positive cells were increased to 84 days after injection of SV(gp120) into the CP with a peak 14 days after injection. No augmentation in the number of microglial cells was seen after injection of SV(BUGT) into the CP or on the contralateral side (not shown) (Fig. 1B, third row). The numbers of GFAP-positive astrocytes were also increased to 84 days after SV(gp120) injection; there was a peak at 2 weeks after injection (Fig. 1B, fourth row). Almost all remaining neurons stained by NT immunostained for NeuN, another neuronal marker, at the different time points (Fig. 1C). The area of the lesion caused by injection of SV(gp120) and measured on NR-stained sections and was between 12.9% and 20.6% of total CP area 1 and 12 weeks after injection; this was significantly higher than after injection of SV(BUGT) at the different time points (p = 0.0007) (Fig. 1D). Tissue damage measured on NR-stained sections extended several millimeters rostral and caudal to the injection site (range: 2.1 ± 0.3 to 3.2 ± 0.3 mm, with minimal differences between the different time points). The neuronal loss, expressed as a percentage of NT-positive cells compared to the contralateral side, was between 34.1% and 42.6% after injection of SV(gp120) in the CP and was significantly higher than after the injection of SV(BUGT) in the same area (p = 0.0009) (Fig. 1E). The number of Iba1-positive cells peaked 2 weeks after injection of SV(gp120) in the CP and remained significantly (p = 0.001) higher than after injection of SV(BUGT) (Fig. 1F) or compared to the uninjected contralateral side (p = 0.001) (not shown).

Figure 1.

Injection of SV(gp120) into the caudate putamen (CP) induces a lesion with ongoing neuronal loss. (A) Frontal sections of the brain through the corpus callosum and the lateral ventricle showing a lesion in the CP area after injection of SV(gp120) (arrow). Injecting SV(BUGT) into the same area does not induce any lesion at any time point. (B) Cryostat sections of SV(gp120)- or SV(BUGT)-injected brains stained with hematoxylin and eosin (H&E) and different cell markers. (First row): H&E staining was normal after injection of SV(BUGT) into the CP as well on the contralateral side (not shown). There was a progressive loss of tissue and cells (particularly neurons) in the area injected with SV(gp120) and at the periphery. Numerous cells with fragmented nuclei were observed 7 days after injection (inset); other cells had normal shapes (arrows). Features suggestive of inflammation were seen 14 days after injection (arrows and inset). An important reduction in the number of cells with neuronal characteristics was observed 28 (with some cells remaining normal, arrows) and 84 days after injection of SV(gp120) into the CP. (Second row): Note the loss of NeuroTrace (NT)-positive cells, not only in the lesion area, but also at the periphery. The CP injected with SV(BUGT) is normal. (Third and fourth rows): Immunostaining for Iba1 (third row) and GFAP (fourth row), markers of quiescent and activated microglial cells and astrocytes, respectively, showed increased numbers of microglial cells and astrocytes until at least 84 days after injection of SV(gp120) into the CP; there was a peak 14 days after injection. Numerous Iba-1-positive cells exhibited an appearance suggestive of reactive microglia from d14 to d84 after SV(gp120) injection (insets), whereas 7 days after injection, reactive microglial cells coexisted with Iba-1-positive cells with a more normal morphology (inset). No augmentation in the number of microglial cells or astrocytes was seen after injection of SV(BUGT) into the CP or on the contralateral side (not shown). (C) Almost all remaining neurons stained by NeuroTrace (NT) immunostained for NeuN, another neuronal marker, at the different times. (D) Extent of the lesion (expressed as a percentage of the whole CP area) at different time points. (E) Neuron loss caused by SV(gp120) inoculation of the CP, as function of time post-injection (expressed as a percentage compared to the contralateral side). (F) Increase in the number of microglial cells, immunostained for Iba1 after injection of SV(gp120) into the CP (expressed as a number for the whole area). Bar: B = 120 μm, with 25 μm for inserts; C = 30 μm.

Transgene Expression of gp120 by SV(gp120) Injection into the CP

Expression of gp120 protein, which is produced as a result of transduction, was demonstrated by immunocytochemistry at different time points after injection. Gp120 was expressed in the cytoplasm. One month after SV(gp120) injection, gp120-positive cells were still present near the injection site of injection (Fig. 2A). A decrease in numbers of gp120-positive cells was seen at 12 weeks after the injections (Fig. 2B). For a positive control we used sections of brains harvested 4 days after injection of 500 ng/μl of the recombinant protein gp120 into the CP and immunostained them for gp120; they showed that neurons of the somatosensory cortex took up the protein, probably by retrograde transport (Fig. 2A). No gp120 was detected after injection of SV(BUGT) in the CP (Fig. 2A) or on the contralateral side of SV(gp120)-injected brains, nor was there any positive staining with an isotype-matched control primary antibody or by substitution of the antibody with PBS (not shown).

Figure 2.

Expression of gp120 in the caudate putamen (CP) after injection of SV(gp120) into the striatum. (A) Cryostat sections of brains injected with SV(gp120) or SV(BUGT) at the level of the CP immunostained for gp120. Expression of the transgene gp120 was tested in the CP at different time points (from day 7 [d7] to day 28 [d28]) after injection of SV(gp120) (third, fourth and fifth columns). Positive controls for gp120 immunostaining were sections of brains harvested after injection of 500 ng/μl of recombinant gp120 into the CP, immunostained for gp120; these showed that neurons of the somatosensory cortex took up the protein, probably by retrograde transport (first column). Control rats received SV(BUGT) instead and were analyzed in the same manner (second column). Insert: higher magnification of a cell immunostained for gp120 with a central nucleus stained by DAPI. (B) Numbers of cells expressing gp120 in the CP at the different time points studied were determined in consecutive sections. Transduction was assessed for each injected brain by serial cryo-sectioning of the whole brain (10 μm thick coronal sections), then by immunostaining. The numbers shown averages of results in the different sections examined. Note that there were no gp120-positive cells in the brains injected with SV(BUGT). (C) Western blotting on samples from the CP 2 weeks after injection of SV(gp120) or SV(BUGT) probed with antibody to gp120. GAPDH was used as a loading control. (D) Cryostat brain sections of SV(gp120)- injected rats at 1 week stained for the neuron marker NeuroTrace (NT), and immunostained for gp120. (E) Cryostat brain sections of SV(gp120)-injected animals at 4 weeks and immunostained for gp120 and Iba1, a marker of microglial cells. Nuclei are stained by DAPI. Note the cytoplasmic localization of gp120 in one Iba1-positive cell (arrow). Bar: A = 60 μm, insert: 20 μm; D = 30 μm; E = 25 μm.

Western blotting of brain samples harvested 2 weeks after injection of SV(gp120) demonstrated expression of gp120 (Fig. 2C). Gp120 was localized mainly within neurons in the first 2 weeks after injection (Fig. 2D), and in cells of neuronal and non-neuronal origin, such as microglia (Fig. 2E), at later time points (see below).

Injection of SV(gp120) into the CP Induces Apoptosis

Apoptotic cells were demonstrated by TUNEL assay at different times after injection of SV(gp120) into the CP. Rare apoptotic cells were seen after injection of SV(BUGT) (Fig. 3A) (p = 0.001, compared to SV(gp120)). No apoptotic cells were observed in the contralateral uninjected side (not shown). Three months after injection, some TUNEL staining colocalized with NT whereas other staining did not (Fig. 3B). By 4 weeks post-injection, the numbers of apoptotic cells decreased (Fig. 3C). At 2 weeks post-injection, some apoptotic cells were NT-positive, but some apoptotic cells were not NT-positive (Fig. 3D). Caspases are cysteine proteases often activated during apoptosis; activated caspase-3 has been described in HAD, as well in cortical or striatal cell cultures exposed to gp120 (40,41). At 2 weeks following SV(gp120) injection, TUNEL-positive cells also stained for the active form of caspase-3 (Fig. 3E). Figure 3E’ shows a NT-positive cell colocalizing with caspase-3 and TUNEL.

Figure 3.

Injection of SV(gp120) into the caudate putamen (CP) induces apoptosis. (A) Cryostat brain sections of SV(gp120)- or (control) SV(BUGT)-injected rats assayed for apoptosis by TUNEL at the indicated time points post-injection. (B) Left column: Cryostat sections of brains 12 weeks post-injection analyzed by TUNEL assay and stained with NeuroTrace (NT). Cells that are both TUNEL-positive and NT-positive are highlighted (arrows). Right column is a higher magnification of an area of the left column. (C) Number of TUNEL-positive cells in the CP after injection of SV(gp120) as a function of time post-injection. (D) Brains at 2 weeks post-injection assayed for TUNEL, stained for NT (NTB) and immunostained for gp120. Overlays demonstrate NT-, gp120- and TUNEL-positive cells. (E) Injected brains at 2 weeks analyzed by TUNEL, stained by NT and immunostained for caspase-3. Overlay shows colocalization of caspase-3 and TUNEL in neuronal cells. (E’) Panel shows on combined images an NT-positive cell colocalizing with caspase-3 and TUNEL (magnification of a field from Fig. 3E). Bar: A = 80 μm; B = 60 μm in left column and 20 μm in right column; C = 45 μm; D = 30 μm; E = 25 μm; E’ = 8 μm.

Activation of Microglial Cells after Injection of SV(gp120) into the CP

To assess microglial activation, we used an antibody against ED1/CD68, a marker of activated microglial cells. ED1/CD68-positive cells were observed at different time points after injection of SV(gp120) (Fig. 4A, B); some CD68-positive cells were also immunopositive for gp120 (not shown). Very few CD68-positive cells were seen in the area of inoculation when SV(BUGT) was substituted for SV(gp120) (Fig. 4A, B) (p = 0.0005 compared to SV(gp120)) or in the contralateral side (p = 0.0001 compared to SV(gp120). To assess the type of cells undergoing apoptosis, we enumerated TUNEL-positive cells that were either NT-positive or CD68-positive at different time points. TUNEL-positive cells that were not NT- or CD68-positive were considered to belong to another cell type. Apoptotic cells in the area of SV(gp120) injection were mainly of neuronal origin through 2 weeks post-injection. After that time point, microglia predominated (Fig. 4C). Because the numbers of cells varied with time after the injection of SV(gp120) (i.e. loss of neurons, increase of microglial cells), we measured the ratios of the number of TUNEL-positive cells to the number of cells for a given cell type. The results confirmed that the percentage of TUNEL-positive neuronal cells decreased after the first 2 weeks whereas the percentage of apoptotic CD68-positive cells increased with time (Fig. 4D). One month after gp120 injection, TUNEL staining was observed essentially in DAPI-positive nuclei of CD68-positive cells (Fig. 4E) (arrows). TUNEL-positive material was exceptionally seen outside of the nucleus in CD68-positive cells, suggesting that engulfment of apoptotic bodies by these cells occurred but was a rare event (Fig. 4F). Several CD68-positive cells are also immunostained for the proliferation marker PCNA at different time points after the injections, suggesting that CD68-positive cells were proliferating (Fig. 4G). No PCNA-positive cells were seen in the CP injected with saline or SV(BUGT) (not shown).

Figure 4.

Microglial activation elicited by SV(gp120) injection into the caudate putamen (CP). (A) Brain sections were immunostained for CD68 at 7, 14 and 28 days after intra-CP inoculation of SV(gp120), or control, SV(BUGT). (B) Numbers of CD68-positive cells per section of CP after injection of SV(gp120) in the CP. (C) Distribution of TUNEL-positive cells of different lineages at different times after injection of SV(gp120) in the CP. TUNEL-positive cells are expressed as total numbers in the CP. (D) Repartition of TUNEL-positive cells of different lineages at different times after injection of SV(gp120) in the CP. TUNEL-positive cells are expressed as percentages related to the number of existing cell types in the CP. (E) One month after gp120 injection, TUNEL staining was primarily observed in nuclei (DAPI-positive) of CD68-positive cells (arrows). (F) TUNEL-positive material was exceptionally seen outside of the nucleus in CD68-positive cells, suggesting that engulfment of apoptotic bodies by CD68-positive cells was a rare event. Note that secondary antibodies were TRITC- and FITC-conjugated in E and F, respectively, and that TUNEL labels were fluorescein- and tetramethyl rhodamine-conjugated, respectively, in E and F. (G) Cryostat brain sections at 2 weeks after injection of SV(gp120) into the CP immunostained for CD68 and PCNA, a marker of cell proliferation. Several CD68-positive cells show immunostaining for PCNA, suggesting that they are proliferating. Insert shows a magnification of one area of the field. Bar: A = 150 μm; E = 15 μm; F = 4 μm; G = 80 μm, with 30 μm for insert.

Injection of SV(gp120) Induces Oxidative Damage

Since oxidative injury is seen in the brains of patients with HAD (42), we tested for oxidative damage following SV(gp120) injection. Lipid peroxidation was measured by calorimetric MDA assay (43) and was assessed by immunocytochemistry for HNE. HNE-positive neurons were seen in SV(gp120)-injected brains (Fig. 5A), and HNE-positive cells were detected by TUNEL assay (Fig. 5B), suggesting that, as in HAD, neuronal damage due to SV(gp120) is associated with oxidative stress. No HNE-positive cells were observed in the CP injected with SV(BUGT) or in the contralateral side (not shown). SV(gp120) elicited more MDA than did the control rSV40 (p = 0.0007) (Fig. 5C).

Figure 5.

Administration of SV(gp120) elicits oxidative stress. (A) Cryostat brain sections of 2 weeks after SV(gp120) injection immunostained for hydroxynonenal esters (HNE) and stained for NeuroTrace (NT) show colocalization of NT and HNE in some but not all HNE-positive cells (arrows). (B) HNE-positive cells may also be TUNEL-positive; all HNE-positive cells were not TUNEL-positive (arrows). (C) Increased brain MDA levels following injection of SV(gp120) assayed 1, 2 and 4 weeks post-injection. (D) Protein oxidation, demonstrated by immunostaining for dinitrophenol (DNP) in brain cryosections, 1 week after injection of SV(gp120) or SV(BUGT) in the CP. Insert: a DNP-positive cell (arrow) centered by a DAPI-positive nucleus. (E) Higher magnification of cryosections as in D, identifying some DNP-positive cells as neurons. Bar: A = 60 μm; B = 30 μm; D = 50 μm, with 20 μm for the insert; E = 20 μm.

Protein oxidation was also assessed by immunohistochemistry using antibodies against dinitrophenol (DNP) to identify oxidized proteins (44). DNP-positive cells were detected in brains injected with SV(gp120) mainly 1 and 2 weeks after SV(gp120) injection. DNP-positive cells were not seen in the control brains (Fig. 5D), nor when PBS or an isotype matched control primary antibody was used (not shown). DNP-positive cells were mainly NT-positive at the different time points assessed (Fig. 5E).

Injection of SV(gp120) in the CP of Nude Rats

To determine whether cytotoxic or other T lymphocytes were involved in the CNS lesions caused by SV(gp120), we injected SV(gp120) into the CP of athymic nude rats. These rats lack T-cells but have normal B and NK cells. SV(gp120) elicited the same type of lesions described above in nude rats with comparable numbers of TUNEL-positive cells (Fig. 6A, B) and microglial cells (Fig. 6 C, D) following the injection of SV(gp120). Likewise, injecting the control vector SV(BUGT) into the CP of nude rats was not different from injecting the control vector into the normal rats; significantly fewer TUNEL-positive cells (p = 0.0004) and CD68-positive cells (p = 0.0005) were seen in the control injected rats than were found following the injection of SV(gp120).

Figure 6.

SV(gp120)-induced apoptosis and microglial cell accumulation do not require T cell-mediated immunity. (A) Cryostat brain sections from nude rats that had been injected with SV(gp120) or (control) SV(BUGT) analyzed by TUNEL assay. (B) Numbers of TUNEL-positive cells in the CP after injection of SV(BUGT) or SV(gp120) in the striatum as a function of time post-injection. (C) Microglial cells accumulate in response to SV(gp120) in nude rats. Cryostat brain sections from injected nude rats immunostained for CD68. The cortex, far from the inoculation point was also examined. (D) Numbers of CD68-positive cells in the CP after injection of SV(BUGT) or SV(gp120) in the CP. Bar: A = 100 μm; C = 150 μm in 3 columns on the left and 100 μm in right column.

Blood-Brain Barrier Abnormalities following SV(gp120) Injection into the CP

Increased BBB permeability has been reported in brains of patients with HAD and in gp120-transgenic mice (45, 46). Immunohistochemical detection of serum IgG leakage from the blood vessels into the brain substance was used as a marker of vascular permeability in sections of rat brains injected with SV(gp120). Areas of IgG accumulation were observed 7 and 14 days after SV(gp120) injection into the CP of normal and nude rats (Fig. 7A). No IgG accumulation was seen in the contralateral side (not shown). The extent of such leakage was less at later time points. No IgG leakage was detected when SV(BUGT) was used instead of SV(gp120) (Fig. 7A). Results were similar in normal and nude rats. To study the relationship with IgG leakage and vascular breakdown, we immunostained brain microvessels for laminin, a basement membrane protein. At 7 days after injection of SV(BUGT) in the CP, laminin immunostaining was normal and no IgG leakage was observed. By contrast, after injection of SV(gp120) into the CP, few-laminin positive structures were seen, and those that remained were surrounded by areas of IgG accumulation (Fig. 7B). Abnormalities observed in the CP after injection of SV(gp120) are summarized in the Table.

Figure 7.

Intracerebral injection of SV(gp120) increases blood-brain barrier (BBB) permeability. (A) Cryostat brain sections of the caudate putamen (CP) from normal rats (upper panels) and nude rats (lower panels) injected with SV(gp120) or SV(BUGT) at the level of the striatum were immunostained for IgG to evaluate BBB leakage of plasma protein. (B) Seven days after injection of SV(gp120) into the CP, few-laminin positive structures were seen, and those remaining were surrounded by areas of IgG accumulation, whereas after injection of SV(BUGT), laminin immunostaining was normal and no IgG leakage was observed. Bar: A and B = 100 μm.

Table.

Parameters Studied Chronologically After Injection of SV(gp120) into the Caudate Putamen

	Day (d) post-injection
	d7	d14	d28	d84
gp120 positivity	+++	++++	++	++
Apoptosis	+++	++++	++	++
Neuron loss	++	+++	+++	++++
Microglial activation	+++	++++	++	+
Astrocyte proliferation	+++	++++	++	+
BBB abnorm alities	++++	+++	+++	++

The frequency of the abnormalities observed in the caudate putamen after injection of SV(gp120) was graded from low (+) to high (++++). BBB = blood-brain barrier;

Overexpression of Antioxidant Enzymes Protects against SV(gp120)-Induced Lesions

rSV40-delivered antioxidant enzymes Cu/Zn SOD1 and glutathione peroxidase (GPx1) strongly protect from the neurotoxicity of recombinant gp120 injected into the CP (21,23). To determine whether SOD1 or GPx1 could protect from SV(gp120)-induced injury, we administered SV(SOD1) and SV(GPx1) to the CP 1 month before SV(gp120), then assayed for apoptotic cells. Prior administration of SV(SOD1) or SV(GPx1) significantly protected the brains from the ongoing damage elicited by HIV-1 envelope glycoprotein (Fig. 8). Fewer apoptotic cells were observed when SV(SOD1) or SV(GPx1) (reduction by 84% and 71% compared to untreated rats, respectively) were administered 1 month before SV(gp120) was injected into the CP (p = 0.01) (Fig. 8A, B). Prior rSV40 delivery of these antioxidant enzymes also reduced SV(gp120)-induced lipid peroxidation (p = 0.003 for both SV(SOD1) and SV(GPx1)) (Fig. 8C), although the magnitude of decrease in MDA was less than the level of protection from apoptosis.

Figure 8.

rSV40-delivered antioxidant enzymes glutathione peroxidase-1 [SV(GPx1)] and superoxide dismutase-1 [SV(SOD1)] protect against SV(gp120)-induced lesions. (A) TUNEL assay on cryosections of brains whose CP has been injected first with SV(GPx1), SV(SOD1), or (control vector) SV(BUGT). One month later the same sites were challenged with an injection of SV(gp120). One week thereafter, TUNEL-positive cells were visualized. (B) Protection from SV(gp120)-induced apoptosis by SOD1 and GPx1. TUNEL-positive cells were enumerated in the rats described in A (p = 0.01 for both SV(SOD1) and SV(GPx1) vs. control SV(BUGT). (C) Prior inoculation of SV(SOD1) or SV(GPx1) protects from SV(gp120)-induced lipid oxidation. Levels of malondialdehyde (MDA) were assayed in the CP that had been injected with SV(GPx1), or SV(SOD1), or (control, SV(BUGT)) prior to administration of SV(gp120) (p = 0.003 for both SV(SOD1) and SV(GPx1) vs. SV(BUGT)). Bar: A = 100 μm.

DISCUSSION

We describe a model system for HIV-1 Env-induced neurotoxicity in which injection of SV(gp120) into the CP of rats leads to chronic expression of gp120 in microglia and neurons, ongoing apoptosis of these cell types, neuronal loss, oxidative stress and increased BBB permeability. rSV40 vectors were used to deliver the gp120 gene into the brain, leading to the long-term expression of the transgene in neurons and microglial cells. In many respects, the consequences of rSV40-delivered gp120 expression in this system resemble the pathologic and biochemical alterations observed in neuroAIDS. Ongoing HIV-1 Env-induced apoptosis, especially neuronal apoptosis, in this system is associated with biochemical evidence of oxidative cellular injury, caspase activation, microglial cell accumulation, and increased vascular permeability.

rSV40-delivered expression of HIV-1 gp120 led to ongoing apoptosis of neurons and microglia. Because inflammatory cells disappear naturally by apoptosis, however, it is also possible that microglia/macrophage apoptosis observed in the present study represents to some extent a physiological phenomenon. Apoptosis caused by SV(gp120) was associated with caspase-3 activation, as has been described in HAD (40), and when cortical or striatal cell cultures were exposed to gp120 (40, 41).

The NT fluorescent stain binds to the Nissl substance of neuron cell bodies and exhibits bright fluorescence that is visible with filters appropriate for fluorescein or TRITC; it is considerably more sensitive than cresyl violet stain. NT has been used as a neuronal marker in numerous studies focusing on the characterization of neurons (34–36) and has been used to delineate the lineage of apoptotic cells in murine models of CNS HIV infection and of gp120 neurotoxicity (18, 25). It does not bind to astrocytes or microglial cells (36). In previous studies we found that NT staining correlated well with immunostaining using antibodies to other neuronal markers such as NeuN and MAP-2 (21, 23, 37, 38). We show here that almost all remaining neurons stained by NT immunostained for NeuN at the different times. Injection of SV(gp120) into the CP caused an increase in neuron loss with time.

We used antibodies against Iba1 and ED1/CD68 to stain microglial cells. An antibody against ED1 was chosen because ED1 is expressed in activated microglial cells whereas Iba-1 can be present in quiescent as well as in activated microglia. Even if ED1/CD68 is not entirely specific for activated microglial cells, antibodies against ED1/CD68 have been repeatedly used for recognizing activated microglial cells, particularly phagocytic microglia (47). Numbers of ED1/CD68-positive cells in normal brains are typically very low (47), whereas Iba1-positive cells are frequently observed in normal brains. The numbers of Iba1- and ED1/CD68-positive cells were both increased after injection of SV(gp120) into the CP, suggesting either microglial proliferation, or attraction of monocytes/macrophages to the lesion from the circulation. TUNEL staining was seen in the cytoplasm of a CD68-positive cell, suggesting that DNA can be phagocytosed as apoptosis occurs. In this regard, it should be mentioned that Gelbard et al described TUNEL-positive microglial cell cytoplasm in brains from children with HIV-1 encephalitis, postulating that microglia might phagocytose DNA in this condition (48).

Soluble gp120 induces apoptosis in many cell types including cardiomyocytes (49), lymphocytes (50) and neurons (51, 52). At high concentrations, HIV gp120 may be directly neurotoxic (53). Among other effects, gp120 binds neuron cell membrane co-receptors (CCR3, CCR5 and CXCR4) and elicits apoptosis that is signaled via these G-protein coupled pathways (54, 55). When we previously injected recombinant gp120 directly into the CP (21, 23), gp120 was detected in the CP by immunocytochemistry only at early time points after injection (from 1 hour to 2 days), and then disappeared. Apoptosis peaked one day after the injection of recombinant gp120 protein, and oxidative stress was seen at early post-injection time points as well. By 4 days after the injection of gp120, ongoing apoptosis and oxidative damage were no longer detectable in the CP. In contrast, gp120 was still present 3 months after inoculation of SV(gp120), and apoptotic cells were still seen. Oxidative stress continued to be observed at least one month after injection of SV(gp120). Direct injection of recombinant gp120 induces acute lesions, which are reproducible but are unlike the chronic injury observed in AIDS. For this reason, we tried to develop an in vivo model of chronic HIV-1 Env-induced neurotoxicity that recapitulated some of the chronicity and pathogenetic features of HIVE not generally seen when recombinant HIV-1 proteins are inoculated. It should be pointed out that although the gp120 used in these studies was from the X4-tropic strain HIV-1 NL4-3, the pathology and biochemical changes are indistinguishable from those caused by gp120 from R5-tropic HIV-1 (21, 23).

The lesions caused by SV(gp120) were virtually identical in nude rats and their non-nude counterparts. Neuron apoptosis, vascular leakage and accumulation of microglial cells were all similar in both types of rats. Thus, effective T cell- mediated immune responsiveness does not appear to play a role in these CNS lesions elicited by gp120 expression. This conclusion may be significant for understanding the pathogenesis of the CNS lesions seen in HIVE.

Injection of SV(gp120) in the CP of rats induced oxidative stress, as assessed by increased MDA levels in the brain, HNE production and immunolocalization of DNP indicative of protein oxidation in neuronal cells. Oxidative stress can play a role in HIV-1-associated dementia in several ways. Oxidation of polyunsaturated fatty acids results in production of multiple aldehydes with different carbon chains including hydroxynonenal esters (HNE) (56) that can mediate oxidative stress-induced apoptosis of cultured neurons (57) and damage neurons and cause cognitive dysfunction in vivo (58). HNE-positive neurons, as well as increases in ceramide, have been demonstrated in the brains of patients with active progressive HIV-1 encephalitis (42, 59). It has also previously been shown that HIV-1 gp120 can cause lipid peroxidation (60).

Accumulation of extravasated IgG in the CP suggested that SV(gp120) increased BBB permeability. Abnormalities in the BBB are important in mediating some of the tissue damage that accompanies HIV-1 infection of the brain, as well as facilitating viral entry into the CNS (45). Evidence of serum-protein leakage across the BBB has been demonstrated in the brains of patients with HAD (46, 61), and accumulation of serum proteins in subcortical neurons and glia has been observed more frequently in HIV-1-positive patients with dementia than in those with no cognitive impairment (62). BBB compromise is associated with neurocognitive impairment and elevated plasma viral load in the presence of BBB compromise may increase the risk for development of HAD (63). Absence or fragmentation of occludin and ZO-1, 2 important structural proteins of tight junctions, was demonstrated in brains of patients who died with HIVE, but in contrast, no significant changes were observed in tissues from HIV-seronegative control patients and from HIV-1-infected patients without encephalitis (64). It is reported that gp120 leads to extravasation of albumin, increased numbers of vessels immunostained for ICAM-1 and VCAM-1, as well as immunoreactivity for substance P at the endothelial cell surface were demonstrated in gp120-transgenic mice (65, 66).

The lesion induced by the injection of SV(gp120) into the CP was partly hemorrhagic whereas we did not observe hemorrhagic lesions in previous studies when we injected either saline or other SV40-derived vectors into the CP (23, 38). On the other side, direct injection of recombinant gp120 protein into the CP was associated with hemorrhage, the extent of the lesion being dependent on the concentration of gp120 injected (Supplementary Fig. 1). Hemorrhage has been considered to be a rare feature of HIV-related brain damage, occurring mainly in children. Recently, however, subarachnoid or multiple intra-cerebral hemorrhages associated with diffuse vasculopathy have been described in adults (67–72). It has also been reported that the introduction of HAART in children with HIV-associated multiple intracranial aneurysms resulted in the non-progression, or resolution, of the aneurysms (73 74). Postmortem histology in some patients showed intimal hyperplasia, fibrosis and thickened, beaded internal elastic lamina with fragmentation, suggesting that the focus of the insult in the intracranial vasculopathy might be the intima (71). The mechanism of development of vascular abnormalities in HIV-infected patients remains unknown. Vascular changes might be due to any number of factors, including membrane changes in the endothelial cells in response to cytokines, or direct toxic effects of a viral protein on the endothelial cells. In this respect, we have found that recombinant gp120 protein alone can induce apoptosis of brain microvascular endothelial cells, breakdown of the BBB (unpublished observation), and hemorrhage (Supplementary Fig. 1). It is, therefore, not surprising to observe hemorrhage associated with the disruption of the BBB after injection of gp120 into the CP (Supplementary Fig. 1). The hemorrhagic lesions disappear 7 days after injection of recombinant gp120 into the CP, whereas they still can be observed more than one month after inoculation of SV(gp120) into the same structure; this underscores our contention that SV(gp120)-related damage is more chronic.

A potential therapeutic strategy for the treatment of HAD might be to limit oxidative stress-induced neurotoxicity. Gene transfer of antioxidant enzymes has been achieved in a gene therapy perspective in numerous models of neurological disorders using diverse viral vectors: adenovirus (75), lentiviruses (76), herpes simplex (77). We previously showed that SV40-vectors mediated overexpression of antioxidant enzymes in vitro and in vivo protects against gp120- and Tat-related neuronal apoptosis (21–23). Since the lesions caused by recombinant gp120 were very brief in duration and thus different from the ongoing ones in this report, our present findings emphasize the role of antioxidant molecules in the neuroprotection against gp120 in a chronic model of HAD. They also indirectly reinforce the putative link between oxidative stress and genesis of the lesions, because fewer apoptotic cells were achieved when a reduction of the MDA levels was obtained after gene transfer of antioxidant enzymes. It is important to note that the protection from apoptosis afforded by SV(SOD1) and SV(GPx1) was substantially greater than the percentage decrease in MDA. These findings have potential implications for therapies of other neurological diseases in which protein, lipid and DNA oxidation are thought to play key roles.

One of the potential uses of this experimental system is to facilitate development of experimental therapeutics. Interesting information could also be provided by neuropsychological tests (e.g., water maze testing), as well as behavioral assays focused on the basal ganglia (rotational behavior (39); behavioral testing of the sensitivity to psychostimulants (78). SV(gp120) inoculation into the CNS does not recapitulate perfectly all aspects of neuroAIDS pathogenesis. However, its reproducibility, relative accessibility and low cost, as well as its chronicity and susceptibility to inhibition by antioxidant gene delivery, all suggest that this model may lend itself readily to screening and analysis of experimental neuroAIDS therapeutics.

Supplementary Material

1

Supplementary Material 1. Injection of gp120 in the CP induces a lesion with tissue loss. A. Photographs of frontal sections of the brain through the corpus callosum and the lateral ventricle showing a lesion in the caudate putamen (CP) area at different time points (from 6 hours to 14 days-d14) after injection of gp120 in the striatum. Tissue loss in the CP was dependent of the dose of gp120. Injecting saline into the same area induced a minimal lesion at the different time points considered. B. Schematic distribution of the lesion area in brains whose CP had been injected 2 days earlier with gp120. The extent of tissue damage was dose-dependent.

Supplementary Material 2: Color version of Figure 2

Supplementary Material 3: Color version of Figure 6

Supplementary Material 4: Color version of Figure 7

Supplementary Material 5: Color version of Figure 8