Blood–brain barrier abnormalities caused by exposure to HIV-1 gp120 — Protection by gene delivery of antioxidant enzymes

Neurobiology of Disease 38 (2010) 313–325

Contents lists available at ScienceDirect

Neurobiology of Disease j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y n b d i

Blood–brain barrier abnormalities caused by exposure to HIV-1 gp120 — Protection by gene delivery of antioxidant enzymes Jean-Pierre Louboutin a,⁎, Beverly A.S. Reyes b, Lokesh Agrawal a, Christina R. Maxwell a, Elisabeth J. Van Bockstaele b, David S. Strayer a a b

Department of Pathology, Anatomy and Cell Biology, Thomas Jefferson University, Philadelphia, PA 19107, USA Department of Neurosurgery, Farber Institute for Neurosciences, Thomas Jefferson University, Philadelphia, PA 19107, USA

a r t i c l e

i n f o

Article history: Received 4 November 2009 Revised 13 January 2010 Accepted 15 February 2010 Available online 26 February 2010 Keywords: Blood–brain barrier Endothelial cells Gp120 HIV-1 Oxidative stress Gene therapy SV40

a b s t r a c t HIV-1 effects on the blood–brain barrier (BBB) structure and function are still poorly understood in animal models based on direct administration of recombinant HIV proteins. We therefore injected HIV-1 envelope glycoprotein, gp120, into rat caudate-putamens (CPs) and examined vascular integrity and function. Gp120 coimmunostained with endothelial cell marker, CD31. It induced apoptosis of endothelial cells in vitro and in vivo. BBB function was assessed by administering Evans Blue (EB) intravenously before injecting gp120. EB leaked near the site of gp120 administration. Within 1 h after intra-CP gp120 injection, structures positive for endothelial markers ICAM-1 and RECA-1 were greatly decreased. Vascular density assessed by laminin immunostaining remained decreased 1 month after gp120 injection. RECA-1-positive cells expressed hydroxynonenal, a marker of lipid peroxidation and rSV40-mediated gene delivery of antioxidant enzymes protected the BBB from gp120-related injury. Extravasated IgG accumulated following intra-CP SV(gp120) injection, an experimental model of continuing gp120 exposure. Thus: acute and chronic exposure to gp120 disrupts the BBB; gp120-mediated BBB abnormalities are related to lesions of brain microvessels; and gp120 is directly toxic to brain endothelial cells. © 2010 Elsevier Inc. All rights reserved.

Introduction The mechanism(s) by which Human Immunodeficiency Virus-1 (HIV-1) first enters the CNS remains obscure. However, loss of blood– brain barrier (BBB) integrity may be an important part of some of the tissue damage that accompanies HIV-1 infection of the brain, and may facilitate viral entry into the CNS (Annunziata, 2003). BBB maintains the immune-privileged nature of the CNS (Zozulya et al., 2007) and the combination of BBB compromise and elevated plasma viral load is associated with neurocognitive impairment and an increased risk for development of HIV-associated dementia (HAD) (Avison et al., 2004). Evidence of serum-protein leakage across the BBB has been demonstrated in the brains of HAD patients (Petito and Cash, 1992; Power et al., 1993). Absence or fragmentation of occludin and ZO-1, two important structural proteins of tight junctions, was demonstrated in brains of patients who died with HIV-1 encephalitis (HIVE), but in contrast, no significant changes were observed in tissues from

⁎ Corresponding author. Department of Pathology, Anatomy, and Cell Biology, Thomas Jefferson University, 1020 Locust Street, Room 255, Philadelphia, PA 19107, USA. Fax: + 1 215 503 1269. E-mail address: [email protected] (J.-P. Louboutin). Available online on ScienceDirect (www.sciencedirect.com). 0969-9961/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.nbd.2010.02.007

HIV-seronegative control patients and from HIV-1-infected patients without encephalitis (Dallasta et al., 1999). BBB abnormalities have been described in HIV-1 envelope glycoprotein gp120-transgenic mice (Finco et al., 1997; Toneatto et al., 1999), but in these models, gp120 is expressed in cell types different from the cells infected in humans, i.e., microglial cells. In vitro studies show that gp120 can be directly toxic to human endothelial cells, compromises blood–brain barrier integrity by reducing tight junction (occludin) protein expression and enhances monocyte migration across blood–brain barrier (Kanmogne et al., 2005, 2007). However, these results have been obtained in vitro, and studies of gp120 injection in vivo have not described abnormalities of the BBB so far (Bansal et al., 2000; Nosheny et al., 2004; Louboutin et al., 2009a). Experimental systems to study how gp120 and other HIV proteins affect the brain are limited to the acute effects of recombinant proteins in vitro or in vivo, or to SIV-infected monkeys. To circumvent these limitations, we have described an experimental rodent model in which HIV-1 Env gp120 is chronically expressed in the brain using a SV40-derived gene delivery vector, SV(gp120), inoculated in the CP (Louboutin et al., 2009b). Oxidative stress has been proposed as a key event in the pathophysiology of HAD (Sacktor et al., 2004) and in animals models of HAD (Agrawal et al., 2006, 2007). Moreover, HIV-1 gp120 and Tat induce oxidative stress in brain endothelial cells (Price et al., 2005). However, to our knowledge, there are no reports of a relationship

314

J.-P. Louboutin et al. / Neurobiology of Disease 38 (2010) 313–325

between oxidative damage and BBB disruption in animal models based on injection of HIV proteins. In this study, we investigated the cytotoxicity of gp120 for brain endothelial cells in vivo, gp120-related abnormalities of the BBB, and the long-term consequences of gp120 injection on vascular density. We also examined if BBB disturbances could be seen after injection of SV(gp120) into the CP. Finally, we studied if gp120-related BBB abnormalities were associated with oxidative stress and if gene transfer of oxidative enzymes could reduce BBB disturbances observed after acute and chronic exposure to gp120.

factor and heparin (Sigma). Cells were passaged twice after thawing, then were starved of serum and growth factors for 24 h. Cells were treated with serum free media (control) or 200 ng/ml gp120 in serum free media for 24 h, then fixed and permeabilized with cytoperm/ cytofix (BD Pharmingen). They were assayed for TUNEL positivity (Roche Diagnostics, Indianapolis, IN) (all cells were counted for morphometric study) and occludin immunoreactivity (Zymed) according to manufacturer's specifications.

Material and methods

The general principles for making recombinant, Tag-deleted, replication-defective SV40 viral vectors have been previously reported (Strayer, 1999). 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 (CMV-IEP) (McKee and Strayer, 2002). SOD1 and GPx1 transgenes were subcloned into pT7[RSVLTR], in which transgene expression is driven by the Rous Sarcoma Virus long terminal repeat (RSV-LTR). The cloned rSV40 genome was excised from its carrier plasmid, gelpurified and recircularized, 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 (Strayer et al., 1997). Crude virus stocks were prepared as cell lysates, then band-purified by discontinuous sucrose density gradient ultracentrifugation and titered by Q-PCR (Strayer et al., 2001). SV(human bilirubin-uridine 5′-diphosphate-glucuronosyltransferase) (BUGT), which was used here as negative control vector, with a non-toxic byproduct, has been reported (Sauter et al., 2000).

Animals Female Sprague–Dawley rats (300–350 g) were purchased from Charles River Laboratories (Wilmington, MA). Protocols for injecting and euthanizing animals were approved by the Thomas Jefferson University Institutional Animal Care and Use Committee (IACUC), and are consistent with Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC) standards. Female Sprague– Dawley rats were used to maintain comparability to our prior studies on gp120: all of our previous experiments involving gp120 were performed in these rats (Agrawal et al., 2006; Louboutin et al., 2007a, b, 2009a,b). Experiments were done in female rats at similar points of their estrous cycle. The diet that the animals received was a standard commercial diet, regular powdered rodent diet without any component that might cause oxidative stress (such as high fat diet, or high manganese) and was not folate/methyl or iron deficient. Rats had free access to water and diet. Numbers of animals used in experiments are indicated in the “Experimental design” section. Antibodies Different primary antibodies were used: rabbit anti-occludin (1:100; IgG) (Zymed Laboratories, Inc., San Francisco, CA), mouse anti-rat Intracellular Adhesion Molecule-1 (ICAM-1) (1:100; IgG1, kappa) (Becton Dickinson Pharmingen, Franklin Lakes, NJ), goat antiPlatelet/Endothelial Cell Adhesion Molecule-1 (PECAM-1/CD31) (1:100; IgG), goat anti-ICAM-1 (1:100; IgG) (Santa Cruz, Santa Cruz, CA), mouse anti-rat Endothelial Cell Antigen-1 (RECA-1) (1:100, IgG1) (Serotec, Oxford, UK), rabbit anti-laminin (IgG; 1:100) (Sigma, Saint-Louis, MO), rabbit anti-N-acetyllysine-4-hydroxy-2-nonenal (HNE) (IgG; 1:50), a marker of lipid peroxidation (Calbiochem, La Jolla, CA). The monoclonal antibody against gp120 (monoclonal antibody to HIV-1V3, 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. Zolla-Pazner. This antibody reacts with HIV-1MN V3 epitope KRIHI (Gorny et al., 1991). Secondary antibodies were used at 1:100 dilution: Fluorescein IsoThioCyanate (FITC) and Tetramethyl Rhodamine IsoThioCyanate (TRITC)-conjugated goat anti-mouse, TRITC-conjugated goat antirabbit, FITC-conjugated sheep anti-rabbit, FITC-conjugated mouse anti-human IgG (γ-chain specific) (Sigma, Saint-Louis, MO), 7-amino4-methylcoumarin-3-acetic acid-NHS ester (AMCA)-, FITC- and TRITC-conjugated donkey anti-mouse, anti-rabbit and anti-goat antibodies (Jackson ImmunoResearch Laboratories, Inc., WestGrove, PA), FITC-conjugated goat anti-rat IgG H&L ((Fab)2 fragment; Abcam, Cambridge, MA). Human brain microvascular endothelial cells Human brain microvascular endothelial cells (HBMEC) were purchased from Sciencell (San Diego, CA). Cells were grown in F12/ Kaighns modification media (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum, 1% pen/strep, endothelial cell growth

Vector production

Experimental design Early consequences of gp120 administration into the CP In order to study gp120 immunolocalization, gp120-induced apoptosis, and microvessels immunostaining after gp120 administration in the CP, 1 µl saline containing 500 ng gp120 was injected stereotaxically into the CP of rats (n = 15). Brains were harvested at different times after the injection (30 min, 1 h, 1 day, with n = 5 for each time point, total = 15). Controls (n = 5 for each time point, total = 15) received only saline instead of gp120 in the CP; contralateral sides of the unilaterally injected brains were also used as controls. In order to test the specificity of the effects of gp120, 1 µl saline containing 500 ng rat IgG (Sigma) was injected into the CP as a control unrelated protein (n = 5 for each time point, total = 15). Recombinant HIV-1 BaL gp120 was obtained through the NIH AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH, Germantown, MD. Recombinant HIV-1 BaL gp120 was free of endotoxin (LPS), as was rat IgG, and the vehicle used (saline). Long-term consequences of gp120 administration on vascular density in the CP Long-term consequences of gp120 injection on vascular density in the CP were assessed after administration of 1 µl saline containing 500 ng gp120 injected stereotaxically into the CP of rats. Brains were harvested at different times after the injection (d2, d4, d7, d14, d28, with n = 5 for each time point, total = 25). Controls (n = 3 for each time point, total = 15) received only saline instead of gp120 in the CP; contralateral sides of the unilaterally injected brains were also used as controls. Brains were processed for immunostaining of laminin, a basement membrane protein, commonly used for evaluating vascular density in the brain (Paris et al., 2004; Marcon et al., 2009; Ling et al., 2009). Injection of SV(gp120) into the CP SV(gp120) was injected into the CP of Sprague–Dawley rats, whose brains were harvested 1, 2 and 4 weeks after injection (n = 6

J.-P. Louboutin et al. / Neurobiology of Disease 38 (2010) 313–325

for each time point, total = 18). Controls received saline or a control vector, SV(BUGT), instead of SV(gp120) in the CP (n = 3 for each time point for SV(BUGT), total = 9; n = 3 for each time point for saline, total = 9). Assessment of BBB disruption after gp120 injection BBB integrity was assessed by intravenous injection of 2% Evans Blue (EB, FW 960.81, Sigma) in 0.9% saline in the jugular vein 15 min before gp120, saline or control protein (rat IgG) injection. Then, 15 min later, gp120 was injected either: 1) stereotaxically into the right CP (500 ng gp120) (n = 10); 2) in the right middle carotid artery (CA) (2000 ng gp120) using a microcatheter (n = 10; or 3) stereotaxically into the right lateral ventricle (LV) (2000 ng gp120) (n = 10). Brains were removed 1 day later without prior fixation, half of the brains in each group were used for biochemical quantification of EB (n = 5), and half of the brains were used for morphological purpose (see below) (n = 5). The brains harvested for biochemical purpose were divided into ipsilateral and contralateral hemispheres, weighed, homogenized in 400 µl of N,N-dimethylformamide (Sigma), then centrifuged at 21,000 ×g for 30 min. Evans Blue was quantified using a spectrophotometer from the absorbance at 620 nm of each supernatant minus the background calculated from the baseline absorbance between 500 and 740 nm. Controls (n = 10 in each control group) received saline or rat IgG instead of gp120. Saline or rat IgG was used as the negative control, as was the contralateral side of the unilaterally injected brains. In addition, in order to study the relationship between the concentration of gp120 and BBB abnormalities, 1 µl saline containing either 100 ng or 250 ng gp120 was injected into the CP of rats whose brains were harvested without fixation 24 h after gp120 injection for EB concentration measurement (n = 5 rats for each concentration; total = 10 rats). EB injection was performed before gp120 administration as previously described. Challenge with gp120 and SV(gp120) after administration of SV(GPx1)/SV(SOD1) To study possible protection of the BBB from gp120-related injury by rSV40-mediated overexpression of SOD1 and GPx1, we first injected the CP of rats with SV(SOD1) (n = 5) and SV(GPx1) (n = 5). One month later, the CP in which SV(SOD1) or SV(GPx1) has been administered was injected with 500 ng gp120. Brains were harvested 1 day after injection of gp120 into the CP, and studied for BBB abnormalities (n = 5 for each vector). In all cases, controls received SV (BUGT) in the CP instead of SV(SOD1) and SV(GPx1) (n = 5) (total: n = 15 rats). We also evaluated the protection of the BBB from SV (gp120)-related insult: CPs of rats (n = 6 for each vector) were first injected with SV(SOD1) and SV(GPx1). One month later, the CPs previously injected with SV(GPx1) or SV(SOD1) were given SV (gp120). Brains were harvested 1 week after injection of SV(gp120) into the CP. Controls received SV(BUGT) in the CP instead of SV(GPx1) and SV(SOD1) (n = 5). In vivo injection of gp120 and SV(gp120) Rats were anesthetized with isofluorane UPS (BaxterHealthcare Corp., Deerfield, IL) (1.0 U isofluorane/1.5 l O2 per min) and placed in a stereotaxic apparatus (Stoelting Corp., Wood Dale, IL) for cranial surgery. Body temperature was maintained at 37 °C by 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 1 µl of gp120 (500 ng/µl), 1 µl of saline or 1 µl of saline containing rat IgG for the injection into the CP or 4 µl of gp120 (2000 ng) for the injection into the LV. The gp120-filled micropipettes were placed in the CP and the LV using coordinates obtained from the rat brain atlas of Paxinos and Watson (1986). For injection into the CP, a burr hole was placed + 0.48 mm anterior to bregma and −3.0 mm

315

lateral to the sagittal suture. Once centered, the micropipette was placed 6.0 mm ventral from the top of the brain. For LV injection, a burr hole was placed − 0.26 mm posterior to bregma and −1.0 mm lateral to the sagittal suture. Once centered, the micropipette was placed −3.4 mm ventral from the top of the brain. For injection of SV (gp120) into the CP, glass micropipettes were backfilled with 5 µl of SV(gp120) viral vector, which contains approximately 107 infectious units. The vector-filled micropipettes were placed in the CP using the same coordinates as previously described. Gp120 or SV(gp120) were injected using 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. Following surgery, animals were housed individually with free access to water and food. Brains were removed 1 day after injection without perfusion. Procedure for harvesting the tissue for morphology At different times after injection of gp120, or SV(gp120), into the CP, rats were anesthetized by intraperitoneal injection of sodium pentobarbital (Abbott Laboratories, North Chicago, IL) at 60 mg/kg and perfused transcardially through the ascending aorta with 10 ml heparinized saline followed by 1000 ml ice cold 4% paraformaldehyde (Electron Microscopy Sciences, Fort Washington, PA) in 0.1 M phosphate buffer (pH 7.4). Immediately following perfusion-fixation, the rat brains were dissected out, placed in 4% paraformaldehyde for 24 h, then in a 30% sucrose solution for 24 h, then frozen in methyl butane cooled in liquid nitrogen. The samples were cut on a cryostat (10 µm coronal sections). Brains of rats injected with EB for assessment of BBB disruption were frozen in cooled isopentane (− 70C) during 1 min and stored at −80 °C (n = 5 rats in each gp120-injected group – in CP, CA and LV – , total = 15 and n = 5 in each respective control group injected with saline or rat IgG instead of gp120, total = 10). To study Evans Blue extravasation on sections, 10-µm sections of unfixed frozen brains were cut on a cryostat and observed using a fluorescent microscope with a TRITC filter. Immunocytochemistry For immunofluorescence, coronal cryostat sections (10 µm thick) were processed for indirect immunofluorescence. Sections were first incubated for 60 min with 10% goat, or 10% donkey, serum in phosphate buffer saline (PBS; pH 7.4) to block non-specific binding. They were then incubated with antibodies diluted according to manufacturers' recommendations: 1 h with primary antibody, then 1 h with secondary antibody diluted 1:100, all at room temperature. Double immunofluorescence was performed as previously described (Rouger et al., 2001). Detection of BBB disturbances was assessed by employing a one-step immunohistochemical detection of IgG (molecular size: 130–140 kDa) in which sections were incubated during 1 h with the antibody (1:100), combined or not with immunostaining for laminin. All incubations were followed by extensive washing with PBS. To stain nuclei, we used mounting medium containing 4′,6diamidino-2-phenylindole (DAPI) (Vector Laboratories, Burlingame, CA). Specimens were finally examined under a Leica DMRBE microscope (Leica Microsystems, Wetzlar, Germany). Negative controls consisted of preincubation with PBS, substitution of nonimmune 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) and following a previously described protocol (Agrawal et al., 2006). To quantitate the TUNEL assay, TUNELpositive cells were expressed as a total number per CP measured in at

316

J.-P. Louboutin et al. / Neurobiology of Disease 38 (2010) 313–325

least 5 consecutive sections using the same computerized imaging system as described below (Morphometry section). The final number was an average of results measured in the different sections.

considered significant when p b 0.05. On graphs, values are represented as means ± s.e.m. Results

Staining of neurons using NeuroTrace In order to stain neurons, we used Neurotrace (NT), a fluorescent stain that has been used as a neuronal marker in numerous studies focusing on the characterization of neurons (Miura et al., 2003; Morinville et al., 2004; Nosheny et al., 2004; Wu et al., 2004; Bigini et al., 2006) and NT staining has been performed as previously reported (Agrawal et al., 2006; Louboutin et al., 2006, 2007a,b, 2009a,b). After rehydration in 0.1 M PBS, pH 7.4, sections were treated with PBS plus 0.1% Triton X-100 10 min, washed twice for 5 min in PBS then stained by NeuroTrace (NT) (InVitrogen, Carlsbad, CA) (1:100), for 20 min at room temperature. Sections were washed in PBS plus 0.1% Triton X-100 then ×2 with PBS, then let stand for 2 h at room temperature in PBS before being counterstained with DAPI. Combination NT then 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 stainings, 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 brain was assessed by neutral red (NR) staining of cryostat sections (10 µm thick). Briefly, cryostat sections were stained by NR for 5 min, then washed, dehydrated in alcohols and finally cleared in xylene. Morphometry ICAM-1-, RECA-1-, occludin- and laminin-positive structures were counted manually on the injected and uninjected sides in the CP of animals injected with either gp120, SV(gp120), SV(BUGT) or saline. The entire area of the CP per section was considered with at least 5 consecutive sections using a computerized imaging system (ImagePro Plus, MediaCybernetics, Bethesda, MD). In all cases, the final number was an average of results measured in the different sections and was expressed in the graphs as a number of the structures per microscopic field/area. TUNEL-positive cells were enumerated in a similar way, as well as gp120- and CD31-positive cells. To assess IgG accumulation in the CP after SV(gp120) administration into the same structure, we used the same computerized imaging system (Image-Pro Plus, MediaCybernetics, Bethesda, MD) to quantify the IgG-positive area on cryostat sections immunostained for IgG. Computer-assisted tracing of the perimeter of the striatal tissue positive for IgG surrounding the injection site as well as the whole CP was conducted to determine area measures. A ratio of the area of tissue positive for IgG compared to the whole CP area was determined for each section that was considered. A total of 20 sections (one section every 200 µm; 10 sections rostral and 10 sections caudal to the injection site) per animal were used. The final number was an average of results measured in the different sections. This procedure was similar to one previously described (Nosheny et al., 2004; Louboutin et al., 2009a). 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

Injection of gp120 into the CP leads to vascular damage and BBB leakiness We examined the effect of HIV-1 gp120 on BBB integrity by asking whether exposure of the brain to gp120 led to any alteration in BBB function. To do this, we injected rats with the vital dye, Evans Blue (EB) shortly before intra-CP gp120. One day after injection of gp120 into the CP, EB leakage was seen in the same structure. No such leakage was observed on the contralateral side or when saline or the control protein, rat IgG, was injected in the CP instead of gp120 (Fig. 1A). No significant leakage of EB was observed when gp120 was injected into the carotid artery (CA) or in the lateral ventricle (LV) (Fig. 1A). Compared to the side injected with gp120, quantitative measurement of EB revealed significantly lower levels of EB in the CP contralateral to the CP injected with gp120, as well as in CPs injected with saline or rat IgG (p b 0.01 for gp120-injected CP vs. uninjected CP and p b 0.01 for gp120-injected CP vs. saline- or rat IgG-injected CP; n = 5 for gp120- and for saline- or rat IgG-injected CPs). There was no difference between the two sides in the rats injected with gp120 into the CA or in the LV. Moreover, the concentration of EB in these groups was significantly lower than in the CP injected with gp120 (Fig. 1B). In summary, levels of EB were comparable without statistical difference: 1) in contralateral uninjected sides of the different groups studied; 2) after injection of saline or rat IgG into the CP; and as well as after injection of gp120 into the LV and CA. Compared to the uninjected side as well as to the saline- or rat IgG-injected CPs, significantly higher EB concentrations were measured after injection of gp120 into the CP (Fig. 1B). The concentration of EB was a function of the dose of gp120 injected into the CP (p b 0.01: 500 ng, 250 ng and 100 ng vs. saline and vs. saline containing rat IgG; p b 0.05: 100 ng vs. 250 ng and 500 ng vs. 250 ng) (Fig. 1C). EB extravasation was detected on cryostat sections as a red fluorescent signal, which was seen in the injected CP (Fig. 1D). No such fluorescence was seen when saline or rat IgG was injected in the CP instead of gp120 or in the contralateral side (not shown). Gp120 immunolocalization to brain microvessels To determine whether the observed vascular damage could reflect a direct effect of gp120 on blood vessels, we studied the localization of gp120 by immunocytochemistry following intra-CP administration. 30 min after injection into the CP, gp120 was immunolocalized both in the brain parenchyma (Fig. 2A, arrowheads), and close to or within the walls of small blood vessels walls (Figs. 2A, arrows, C). As we have previously reported, gp120 colocalized with neuronal markers (Neurotrace, NT) and is observed in the cytoplasm of neurons without any nuclear localization (Fig. 2B). It was also present in the walls of large ICAM-1 negative blood vessels (Figs. 2D,E), as well as in the walls of smaller vessels that were immunopositive for ICAM-1 (Figs. 2D,E). The HIV-1 envelope glycoprotein was identified in cells expressing the endothelial cell marker, CD31 (Figs. 2F,G). Vessel lumen was seen in ICAM-1- and CD31-positive structures (insets). Many ICAM-1- and CD31-positive structures were also positive for gp120. Among the different cell markers, 52.5 ± 5.9% of gp120-positive cells were neurons, 40.5 ± 4.4% were CD31-positive and 42.2 ± 4.6% of gp120-positive cells were ICAM-1-positive. 30.5 ± 3.7%, 41.2 ± 5.1%, and 38.8 ± 3.4% of respectively NT-, CD31-, and ICAM-1-positive structures were immunoreactive for gp120. Thus, in addition to being taken up by neurons, gp120 also binds to vascular structures. Immunodetection of gp120 was specific: no positivity was appreciated when saline or rat IgG was injected instead of gp120 (Figs. 2D,F), when

J.-P. Louboutin et al. / Neurobiology of Disease 38 (2010) 313–325

317

Fig. 1. A. Injection of gp120 (2000 ng) into the carotid artery (CA) or in the lateral ventricle (LV) did not induce any extravasation of EB previously injected intravenously, compared to the injection of 500 ng gp120 into the CP. B. These data were confirmed by spectrophotometry measurements. The results represent the specific absorbance of Evans Blue at 620 nm. As a control for perfusion efficiency, the absorbance value from contralateral uninjected hemispheres were subtracted from that of hemispheres injected with gp120. p b 0.01 for gp120-injected CP vs. uninjected CP and p b 0.01 for gp120-injected CP vs. saline- or vs. IgG-injected CP; n = 5 for gp120- and saline- or rat IgG-injected CPs. Levels of EB were comparable without statistical difference: 1) in contralateral uninjected sides in the different groups studied; 2) after injection of saline or rat IgG into the CP as well as after injection of gp120 into the LV and CA. C. The concentration of EB was a function of the dose of gp120 injected into the CP (p b 0.01: 500 ng, 250 ng and 100 ng vs. saline and vs. saline containing rat IgG; p b 0.05: 100 ng vs. 250 ng and 500 ng vs. 250 ng). D. EB extravasation was detected on cryostat sections as a red fluorescent signal in the injected CP. Bar in C: 40 µm.

the anti-gp120 primary antibody was replaced by non-immune isotype-matched control antibody or when the primary antibody was omitted (data not shown). Gp120 induces apoptosis of brain endothelial cells in vitro and in vivo To understand the consequences of gp120 localization in cerebral blood vessels and the consequent vascular damage, we asked whether gp120 could induce apoptosis of endothelial cells. Cultured human brain microvascular endothelial cells (HBMEC) were treated for 24 h with serum free medium alone (control) or with gp120, then assayed for apoptosis by TUNEL. Cellular immunopositivity for occludin (a tight junction protein in brain endothelial cells) was used to ascertain that the cells were endothelial cells. The dose of 200 ng/ml was chosen because previous experiments showed that significantly fewer apoptotic cells were observed with lower concentrations of gp120 (not shown). Numerous TUNEL-positive endothelial cells were seen after treatment with gp120 200 ng/ml, while almost no apoptosis was seen in control cultures (p b 0.001) (Fig. 3A). TUNEL-staining was localized in the nucleus as shown by the merged image between occludin, DAPI, and TUNEL. No TUNEL-staining was seen in the

cytoplasm. In gp120-treated cultures, almost all occludin-positive cells were TUNEL-positive (99.7 ± 8.8%), while 32.4 ± 4.3% of TUNELpositive cells were not occludin-positive, suggesting either a degree of heterogeneity in the culture or that occludin expression might be altered after gp120 treatment. Control with buffer only and no enzyme showed no TUNEL-staining (not shown). Similar studies were performed in vivo. Rats were injected with 500 ng HIV-1 BaL gp120 in 1 µl saline (or, control, saline or rat IgG only), stereotaxically into the CP. Over the course of the next 24 h, increasing numbers of CD31positive cells underwent apoptosis (Fig. 3B). Sections of brains injected with 500 ng gp120 were immunostained for CD31 and gp120, then processed for TUNEL assay. Numerous CD31- and gp120immunoreactive cells were positive for TUNEL (Fig. 3C). However, some CD31-immunoreactive cells were TUNEL-positive, but not gp120-positive (Fig. 3C, insert) and some CD31- and gp120-positive cells were not TUNEL-positive (Fig. 3C, arrow). TUNEL assays processed with buffer and without enzyme showed no TUNELstaining (Fig. 3D). Quantitation of TUNEL assay gave the following results: 42.8 ± 4.5, 65.1 ± 7.2, 315.4 ± 35.5 TUNEL-positive cells/CP were detected respectively 1, 6, and 24 h after gp120 injection. No apoptotic cells were observed after injection of saline, as previously

318

J.-P. Louboutin et al. / Neurobiology of Disease 38 (2010) 313–325

Fig. 2. Cryostat sections of CP injected with 500 ng gp120 30 min earlier and immunostained for gp120 and NT, a neuronal marker (B), ICAM-1, a marker of microvessels (D, E), and CD31, a marker of endothelial cells (F, G). Nuclei were stained by 4′,6-diamidino-2-phenylindole (DAPI). Gp120 was observed not only in parenchyma (A, arrowheads) but in the walls of vessels (A, C, arrows) and colocalized with NT (B), as well as with ICAM-1 (D, E) and CD31 (F, G). In C, the vessel lumen is indicated by dashed lines. In D and F, note the lumen in ICAM-1- and CD31-positive structures (insets). No immunoreactivity for gp120 was observed when saline (D) or rat IgG (F) was injected into the CP. Note that in Fig. 2D, gp120 is also present in the walls of larger vessels which are not stained by ICAM-1 (arrows). In D, cryostat sections have been incubated with anti-ICAM-1 and anti-gp120 primary antibodies with respectively TRITC- and FITC-conjugated secondary antibodies, while in E secondary antibodies were conjugated with FITC and TRITC. G is a higher magnification of field of the right column of F indicated by *. Bar in A: 120 µm, B: 50 µm, C: 60 μm, D: 120 µm, insert: 60 µm, E: 20 µm, F: 120 µm, insert: 60 µm, G: 20 µm.

reported (Agrawal et al., 2006), neither after injection of rat IgG into the CP. Taken together, these results suggest that gp120 induces apoptosis of endothelial cells, both in vitro and in vivo. Gp120 causes a reduction in the number of brain microvessels ICAM-1 and laminin have been used in the present study as markers of brain microvessels. Injection of gp120 into one CP led to rapid loss of ICAM-1- (Figs. 4A,B) and laminin- (Figs. 4C,D) positive structures, when compared to the uninjected contralateral side (p b 0.01 and p b 0.01 respectively) 1 h after injection of gp120 into the CP. Control animals, receiving unilateral intra-CP saline or rat IgG (not shown), showed approximately equal numbers of ICAM-1 and laminin-positive structures on the contralateral (uninjected) as on the injected sides (p N 0.05). Vessel lumens were visualized within the

ICAM-1-positive structures (insets). Similarly, there was a significant diminution of RECA-1-positive structures 1 h and 1 day after injection of gp120 into the CP (p b 0.01 for both times) (Figs. 4E,F). Fewer occludin-positive structures were detected in CPs injected with gp120, while no such diminution was seen in the contralateral uninjected sides (Fig. 4G). The reduction in the number of occludinpositive structures after gp120 injection compared to the uninjected side was highly significant (p b 0.01) (Fig. 4H). Long-term consequences of gp120 injection on vascular density in the CP Long-term effects of gp120 administration on brain vessels were studied by immunostaining for laminin, an extracellular matrix protein that is a key component of vascular basement membranes, at different time points after gp120 administration. Very few laminin-

J.-P. Louboutin et al. / Neurobiology of Disease 38 (2010) 313–325

319

Fig. 3. A. Human brain endothelial cells immunostained for occludin and incubated with 200 ng/ml gp120 became apoptotic. The nucleus was stained by 4′,6-diamidino-2phenylindole (DAPI). Numerous TUNEL-positive cells were observed after treatment with gp120, while almost no apoptotic cells were seen in cultures that were not incubated with gp120 (p b 0.001 serum free media (control) vs. 200 ng/ml gp120 in serum free media for 24 h). TUNEL-staining was localized in the nucleus stained by DAPI. No TUNEL-staining was seen in the cytoplasm. Almost all occludin-positive cells were TUNEL-positive for 200 ng/ml gp120 (99.7 ± 8.8.3%), while 32.4 ± 0.3% of TUNEL-positive cells were not occludinpositive, suggesting that occludin staining might be altered after gp120 treatment. Control with buffer only and no enzyme showed no TUNEL-staining (not shown). B. Cryostat section of CP injected with 500 ng gp120 1 h earlier immunostained for CD31 and processed for TUNEL assay. Some of the endothelial cells, immunostained for CD31, are apoptotic, TUNEL-positive. C. Cryostat section of CP injected with 500 ng gp120 1 h earlier immunostained for CD31, gp120, and processed for TUNEL assay. Numerous CD31- and gp120immunoreactive cells were positive for TUNEL. However, some CD31-immunoreactive cells were TUNEL-positive, but not gp120-positive (insert). Some CD31- and gp120-positive cells were not TUNEL-positive (arrow). *: a vessel-like structure positive for CD31. D. TUNEL assay of cryostat section of CP injected with 500 ng gp120 1 h earlier processed with buffer and without enzyme showed no TUNEL-staining. Bar in A: 15 µm; B: 60 µm; C, D: 80 µm.

positive structures were detected 2 and 4 days after gp120 inoculation. Numbers of laminin-positive structures increased thereafter, through d28, particularly at the edge of the lesion. However, 1) laminin-positive structures were less numerous at all times in the gp120-injected CPs than in CPs injected with saline, or in the contralateral sides, and 2) these vessels were often of smaller caliber than in control CPs (Fig. 5A). Even at 1 month post-gp120 injection, vascular density remained significantly lower than in control CPs (p b 0.01) (Figs. 5B,C). Vascular density was unchanged with time in the contralateral uninjected side or in CPs injected with saline. Chronic exposure to gp120 leads to BBB abnormalities To assess the effects of long-term exposure to gp120 on the BBB, we employed a model of more protracted exposure to HIV-1 envelope

gp120 (Louboutin et al., 2009b). In CNS, SV40-derived vectors transduce microglia and macrophages. They also transduce neurons. They do not transduce astrocytes or oligodendrocytes detectably (Agrawal et al., 2006; Louboutin et al., 2007a, 2009b). Injection of SV(gp120) allows for continued gp120 production by transduced cells for weeks after the vector is administered. Thus, SV(gp120) was injected into the rat CP unilaterally and animals were then studied over time for BBB injury. Immunohistochemical detection of IgG leakage from the serum into the brain substance was used as a marker of vascular permeability. Areas of IgG accumulation were observed 7 days after SV(gp120) injection into the CP (Fig. 6A). No significant IgG leakage was detected when a control vector, SV(BUGT), was used instead of SV(gp120), nor was it seen on the contralateral side (Fig. 6A). The area positive for IgG was significantly higher when SV(gp120) was injected into the CP instead of SV(BUGT) (24.7 ± 3.1% vs. 1.2 ± 0.3% respectively; p b 0.001).

320

J.-P. Louboutin et al. / Neurobiology of Disease 38 (2010) 313–325

Fig. 4. After the injection of 500 ng gp120 into the CP, fewer ICAM-1-, laminin-, RECA-1-, and occludin-positive structures were observed in the injected sides on cryostat sections of brains harvested 1 h after the injection (A, C, E, and G respectively). Brains injected with saline showed no difference between injected and uninjected sides. A, C: Sections immunostained for ICAM-1 and laminin show a reduction in the number of ICAM-1- and laminin-positive structures in the CPs injected with 500 ng gp120 compared to the uninjected contralateral side as well as to the CPs injected with saline. Note the lumen in ICAM-1-positive structures (inset). B, D: Graphs showing the decrease of ICAM-1- and laminin-positive structures in the CP injected 1 h earlier with 500 ng gp120. For ICAM-1 and laminin immunostaining, p b 0.01 for gp120-injected CP vs. uninjected CP and for gp120-injected CP vs. saline-injected CP. E: Fewer RECA-1-positive structures were seen after injection of gp120 into the CP. F: Graph showing the decrease of RECA-1-positive structures in the CP injected with 500 ng gp120. (p b 0.01 for gp120-injected CP vs. uninjected CP at 1 h and 1 day post-injection). G: Immunostaining for occludin showed fewer occludin-positive structures in the CPs injected with gp120, while no reduction was seen in the contralateral uninjected side. H: Graph showing a significant reduction in the number of occludin-positive structures after gp120 injection compared to the uninjected side (p b 0.01 gp120 injected CP vs. uninjected CP). Bar in A, C: 90 µm; E: 120 µm; G: 90 µm.

To assess vascular breakdown in this setting, we studied lamininpositive structures. One week after administration, SV(gp120) greatly reduced laminin-positive structures, particularly in areas of IgG

accumulation (Fig. 6B). This decrease (31.4± 3.8 laminin-positive structures/unit CP area) was highly significant compared to the CPs injected with SV(BUGT) (65.7 ± 5.8/unit area) (p b 0.02). With

J.-P. Louboutin et al. / Neurobiology of Disease 38 (2010) 313–325

321

Fig. 5. Long-term consequences of gp120 injection on vascular density in the CP. CPs were injected with 500 ng gp120 and brains were harvested at different times later. A. Cryostat sections (10 mm thick) of the CP from rats injected with gp120, or saline. Few laminin-positive structures were detected 2 and 4 days after gp120 inoculation. Numbers of lamininpositive structures increased thereafter, through d28, particularly at the edge of the lesion. Laminin-positive structures were however less numerous at all times in the gp120injected CPs than in CPs injected with saline or on the contralateral side, and these vessels were often of smaller caliber than in control CPs (Fig. 6A, insert). In all photographs, the site of injection (and consequently the lesion) is indicated by *. Even 1 month after gp120 injection, the vascular density remained significantly lower than in control CPs (p b 0.01) (Figs. 5B,C). Bar: 50 mm in A.

continued exposure (2 and 4 weeks), number of laminin-positive structures remained as low as they were at 1 week (not shown). By contrast, laminin immunostaining was normal, and no IgG leakage was observed after intra-CP injection of SV(BUGT).

In vivo exposure to gp120 induces oxidative stress in endothelial cells Gp120 has been reported to activate reactive oxygen species (ROS), leading to oxidation of cellular lipids and proteins. We have used lipid peroxidation by ROS as assayed by formation of hydroxynonenal (HNE) (Sacktor et al., 2004), as an index of gp120-induced oxidative injury to cellular macromolecules. To identify cell populations susceptible to gp120-induced lipid oxidation, we performed double immunostaining for cell markers and HNE. One hour after administering recombinant HIV-1 Bal gp120 into the CP, we identified HNE-positive cells in brain vessels. These cells expressed the endothelial cell marker RECA-1 (Fig. 7A), suggesting that gp120 rapidly induces lipid peroxidation in endothelial cells. Similarly, HNEpositive cells were seen in the CP after inoculation of SV(gp120). Several HNE-positive cells expressed RECA-1 (Fig. 7B). No HNEpositive cells were detected after injection of SV(BUGT).

Protection of the BBB from acute and chronic gp120-related injury by rSV40-mediated gene delivery of SOD1 and GPx1 Gene transfer of antioxidant enzymes using recombinant SV40derived vectors has been effective in mitigating neuronal injury and apoptosis due to HIV-1 gp120 and Tat (Louboutin et al., 2007b, 2009a; Agrawal et al., 2007). We therefore tested whether rSV40-delivered Cu/Zn superoxide dismutase (SOD1) or glutathione peroxidase (GPx1) could protect the BBB from gp120-related damage. SV(SOD1) or SV(GPx1) were injected stereotaxically into one CP. Subsequent intra-CP challenge was accomplished using either recombinant gp120 (acute exposure) or SV(gp120) (chronic exposure). Fewer HNEpositive cells were observed in the challenged CPs that received prior gene transfer of antioxidant enzymes, suggesting that lipid peroxidation was decreased by overexpression of antioxidant enzymes (p b 0.01 SV(SOD1) + gp120 and SV(GPx1) + gp120 vs. SV(BUGT) + gp120) (Fig. 8A). We also tested the effect of overexpression of antioxidant enzymes on BBB leakage of blood constituents into the brain, following acute exposure to recombinant gp120 or chronic exposure induced by SV(gp120) injection (Louboutin et al., 2009b). Prior gene transfer of antioxidant enzymes decreased the levels of EB (Fig. 8B) or the area of extravasated IgG (Fig. 8C) that followed respectively intra-CP gp120 or

322

J.-P. Louboutin et al. / Neurobiology of Disease 38 (2010) 313–325

Fig. 6. Injection of SV(gp120) into the CP increases blood–brain barrier (BBB) permeability. A. Cryostat sections (10 µm thick) of the CP from normal rats injected with SV(gp120) or a control vector, SV(BUGT), at the level of the striatum were immunostained for IgG to evaluate leakage of plasma protein through the blood–brain barrier (BBB). B. Seven days after injection of SV(gp120) into the CP, significantly less laminin-positive structures were seen (p b 0.02), particularly in the areas of IgG accumulation, while after injection of SV(BUGT), laminin immunostaining was normal and no IgG leakage was observed. Bar: 100 µm in A, B.

intra-CP SV(gp120) injection, suggesting that reactive oxygen species (ROS) are central to BBB dysfunction caused by both acute and chronic exposure to HIV-1 gp120, and that gp120-induced increased BBB permeability can be alleviated by gene transfer of antioxidant enzymes.

Discussion HIV encephalopathy covers a range of HIV-related CNS dysfunction (Mattson, et al., 2005; McArthur et al., 2005; Nath and Sacktor, 2006; Ances and Ellis, 2007). HIV-1 can affect CNS function through a variety of mechanisms. Neurons themselves are rarely infected by HIV-1, and neuronal damage is felt to be mainly indirect. HIV-1 infects resident microglia, periventricular macrophages and some astrocytes, leading to increased production of cytokines and chemokines (Kaul and Lipton, 1999). Virus shedding or cytopathic effects in virus-infected cells result in release of gp120 protein and peptide products into the serum as well as the brain (Jones et al., 2000). Within the brain, infected microglia and macrophages release HIV-1 proteins, several of which are neurotoxins, including the envelope (Env) glycoprotein gp120 used here. Soluble gp120 may bind chemokine receptors such as CXCR4 and CCR5 (Kaul et al., 2001), and can induce apoptosis in a wide variety of cells including lymphocytes, cardiomyocytes and neurons (Meucci et al., 1998; Kaul and Lipton, 1999; Kaul et al., 2001; Nosheny et al., 2004; Garden et al., 2004; Agrawal et al., 2006). In vitro gp120 is cytotoxic to human umbilical vein endothelial cells (Huang et al., 1999), as well as those from the brain and lung (Kanmogne et al., 2005). Human brain endothelial cells express CCR5 and CXCR4. Compounds that block those chemokine receptors decreased [Ca2+]i release induced in endothelial cells by gp120 from corresponding X4- or R5-tropic HIV-1 isolates (Kanmogne et al.,

2007). However, the effects of gp120 on endothelial cells in vivo have not been reported so far. HIV-1 in the circulation can cross the BBB either as free virus or within infected immune cells (Banks et al., 2001). HIV-1 infected cells can shed gp120 into the circulation and gp120 can cross the BBB by way of a pathway related to adsorptive endocytosis (Banks et al., 2005). There are regional variations in the permeability of the BBB to gp120 which can contribute to the inhomogeneous distribution of HIV-1 within the CNS. These variations can be explained by regional differences not only in blood vessel density, but also in types and concentrations of glycoproteins on the luminal surface of brain endothelial cells, and in the permeability of the BBB to gp120. For example, gp120 uptake, which relates to endothelial cell internalization, is greater in the cerebellum, cortex and midbrain, and less in the striatum and hypothalamus. On the other hand, transport of gp120 across the BBB is highest in hypothalamus, hippocampus and ponsmedulla, and is lowest in striatum (Banks et al., 1999). It has been shown that circulating gp120 is cytotoxic and may damage the BBB (Cioni and Annunziata, 2002), but those data were obtained using an in vitro model of BBB. In the present report, we did not observe any BBB abnormality in the striatum after injecting gp120 either into the carotid artery or the lateral ventricle. This might be due to a lower dose of gp120 compared to the one used in the in vitro experiments. Our findings might also have been different had gp120 administration been repeated over a period of several days. Alternatively, the differences between our results and the in vitro studies might also be explained by the low intake/transport of gp120 across the BBB towards the striatum, which is discussed above (Banks et al., 1999, 2001, 2005). We report here that gp120 injection into the CP greatly increases vascular permeability. We reported previously that injection of gp120 into the CP induced oxidative stress and that antioxidant enzyme gene

J.-P. Louboutin et al. / Neurobiology of Disease 38 (2010) 313–325

323

Fig. 7. Gp120 injection induces oxidative stress in endothelial cells. A. Cryostat sections (10 µm thick) of the CP from normal rats injected 1 h earlier with 500 ng gp120 or saline, at the level of the striatum were immunostained for HNE, a marker of lipid peroxidation, and RECA-1 a marker of endothelial cells. Nuclei were stained with DAPI. No HNE immunostaining was seen in brains injected with saline. HNE immunoreactivity was observed in the CPs injected with gp120 and colocalized with RECA-1. B. Similarly, HNE-positive cells were detected in the CP after inoculation of SV(gp120). Several HNE-positive cells expressed RECA-1. No HNE-positive cells were detected after injection of SV(BUGT) (Fig. 7B). Dashed line represents vessel walls and * represents the lumen of the vessel. Bar: 50 µm in A, B.

delivery was able to protect from HIV-1 gp120- and Tat-induced neuronal apoptosis (Agrawal et al., 2006, 2007; Louboutin et al., 2007b). It has been reported that oxygen radicals were involved in the pathogenesis of HIVE (Sacktor et al., 2004). More particularly, HIV-1 gp120 and Tat induce oxidative stress in brain endothelial cells (Price et al., 2005) and antioxidant interventions limit gp120-mediated damage (Price et al., 2006). Gp120-induced apoptosis in HUVEC cells is mediated by caspases (Ullrich et al., 2000). We confirm these in vitro findings with our present results in vivo. Several reports demonstrated that oxidative stress promotes in vivo BBB disruption (Chan et al., 1991). Our results suggest that such might be also the case in gp120-related injury: reduction in acute or chronic gp120induced oxidative stress by gene delivery of antioxidant enzymes was associated with decreased numbers of HNE-positive cells and protected the BBB from gp120-induced disruption. In the present study, gp120 injection reduced numbers of ICAM-1and laminin-positive structures. Laminin and ICAM-1 have been used in the present study as markers of brain microvessels. An increase of ICAM-1 expression by human endothelial cells cell has been demonstrated in vitro after exposure to gp120 (Ren et al., 2002; Stins et al., 2003). However, the degree of upregulation of ICAM-1 differed among the various human brain microvascular endothelial cell isolates (Stins et al., 2003) and these conditions were different from our model, in which an increase of ICAM-1 might not have been observed considering the acute cytotoxic effect of gp120. HIV-1 gp120-transgenic mice developed by Finco et al. were characterized

by mRNA expression of gp120 in the brain and in the thymus and serum circulating gp120. In this model, extravasation of albumin was associated with increased numbers of vessels immunostained for Intracellular Adhesion Molecule-1 (ICAM-1) and Vascular Cell Adhesion Molecule-1 (VCAM-1), as well as immunoreactivity for substance P at the endothelial cell surface (Finco et al., 1997; Toneatto et al., 1999). Circulating gp120 altered the BBB permeability in these gp120-transgenic mice, but these results were observed using an in vitro model of BBB (Cioni and Annunziata, 2002). In other gp120transgenic mice, gp120 was expressed chiefly in astrocytes (Toggas et al., 1994) or neurons (Michaud et al., 2001), considering the promoter used, while in humans HIV-1 infects mainly microglial cells. With injection of recombinant gp120, exposure is more acute and might reflect the direct action of gp120 on microvessels. Immunostaining for RECA-1, a marker of rat endothelial cells and microvessels, showed a reduction in RECA-1-positive structures that paralleled the ones of ICAM-1-positive microvessels. Similarly, a significant decrease in vascular density assessed by immunostaining for laminin was observed at early times and persisted through 1 month after gp120 injection, suggesting that even the acute effects of gp120 maybe transient, there are also longer term consequences of gp120 exposure on the vascular density. Occludin is present in the membrane of brain microvascular endothelial cells when these cells are confluent, in monolayer (Zhong et al., 2008). In the present case, in which HBMEC were not confluent, occludin was not expected to localize at the plasma membrane level.

324

J.-P. Louboutin et al. / Neurobiology of Disease 38 (2010) 313–325

Fig. 8. Gene delivery of antioxidant enzymes mitigates gp120-related BBB disturbances. A. Prior gene delivery of antioxidant enzymes into the CP before injection of gp120 decreased the number of HNE-positive cells when compared with injection of the vector control SV(BUGT) (p b 0.01 SV(SOD1) + gp120 and SV(GPx1) + gp120 vs. SV(BUGT) + gp120). B. Gene delivery of antioxidant enzymes by SV40-derived vectors in the CP 1 month before the injection of gp120 into the same structure mitigated the extent of the BBB breakdown after injection of gp120 into the CP, as demonstrated by spectrophotometry measurements of Evans Blue (EB) (absorbance at 620 nm). As a control for perfusion efficiency, the absorbance value from contralateral uninjected hemispheres were subtracted from that of hemispheres injected with gp120 (p b 0.01 SV(SOD1) + gp120 and SV(GPx1) + gp120 vs. SV(BUGT) + gp120). C. The extent of leakage of IgG in the CP following the injection of SV(gp120) in the same structure was mitigated by prior gene delivery of antioxidant enzymes. SV(BUGT) was used as a control vector (p b 0.01 SV(SOD1) + gp120 and SV(GPx1) + gp120 vs. SV(BUGT) + gp120). Bar: 80 µm in A.

Subcellular localization of occludin in brain microvascular endothelial cells can also be influenced by stress (Colgan et al., 2007) and coculture with HIV-1 infected monocytes (Persidsky et al., 2006). As exposure to HIV-1 in HIV/AIDS patients is protracted, we tested whether BBB injury could be seen in a more chronic gp120 exposure model. Intra-CP inoculation of SV(gp120) leads to protracted gp120 production and continuing gp120-induced injury, including ongoing apoptosis of these cell types, neuronal loss, oxidative stress and increased permeability of the BBB (Louboutin et al., 2009b). Observed accumulation of extravasated IgG in the CP suggested that injection of SV(gp120) increased BBB permeability. The consequences of rSV40delivered gp120 expression thus resemble many of the pathologic and biochemical alterations observed in neuroAIDS. Ongoing HIV-1 Envinduced apoptosis, especially neuronal apoptosis, in this system is associated with biochemical evidence of oxidative cellular injury, caspase activation, microglial cell accumulation (Louboutin et al., 2009b), and increased vascular permeability. We demonstrated here that HIV-1 glycoprotein gp120 can induce BBB alterations and the present results suggest that HIV proteins can bind to microvessel endothelial cells, and cause degradation of vascular basement membrane and vascular tight junctions without requiring intact HIV-1 virions. Oxidative stress is associated with these events. Whether these proteins act alone or either in combination with vasoactive molecules (prostaglandins, nitric oxide, substance P, etc.) (Annunziata, 2003), or associated activation of matrix metalloproteinases remains to be established. The answers to these

questions may help better understand HIV-1 neuropathogenesis and facilitate identification of new therapeutic approaches for neuroAIDS. Acknowledgment This work was supported by NIH grants MH69122, MH70287 and AI48244. References Agrawal, L., Louboutin, J.P., Reyes, B.A.S., van Bockstaele, E.J., Strayer, D.S., 2006. Antioxidant enzyme gene delivery to protect from HIV-1 gp120-induced neuronal apoptosis. Gene Ther. 13, 1645–1656. Agrawal, L., Louboutin, J.P., Strayer, D.S., 2007. Preventing HIV-1 Tat-induced neuronal apoptosis using antioxidant enzymes: mechanistic and therapeutic implications. Virology 363, 462–472. Ances, B.M., Ellis, R.J., 2007. Dementia and neurocognitive disorders due to HIV-1 infection. Semin. Neurol. 27, 86–92. Annunziata, P., 2003. Blood–brain barrier changes during invasion of the central nervous system by HIV-1. Old and new insights into the mechanism. J. Neurol. 250, 901–906. Avison, M.J., Nath, A., Greene-Avison, R., Schmitt, F.A., Bales, R.A., Ethisham, A., Greenberg, R.N., Berger, J.R., 2004. Neuroimaging correlates of HIV-associated BBB compromise. J. Neuroimmunol. 157, 140–146. Banks, W.A., Ibrahimi, F., Farr, S.A., Flood, J.F., Morley, J.E., 1999. Effects of wheatgerm agglutinin and aging on the regional brain uptake of HIV-1 gp120. Life Sci. 65, 81–89. Banks, W.A., Freed, E.O., Wolf, K.M., Robinson, S.M., Frankko, M., Kumar, V.B., 2001. Transport of human immunodeficiency virus type 1 pseudoviruses across the blood–brain barrier: role of envelope proteins and adsorptive endocytosis. J. Virol. 75, 4681–4691.

J.-P. Louboutin et al. / Neurobiology of Disease 38 (2010) 313–325 Banks, W.A., Robinson, S.M., Nath, A., 2005. Permeability of the blood–brain barrier to HIV-1 Tat. Exp. Neurol. 193, 218–227. Bansal, A.K., Mactutus, C.F., Nath, A., Maragos, W., Hauser, K.F., Booze, R.M., 2000. Neurotoxicity of HIV-1 proteins gp120 and Tat in the rat striatum. Brain Res. 879, 42–49. Bigini, P., Gardoni, F., Barbera, S., Cagnotto, A., Fugamalli, E., Longhi, A., Corsi, M.M., DiLuca, M., Mennini, T., 2006. Expression of AMPA and NMDA receptor subunits in the cervical spinal cord of wobbler mice. BMC Neurosci. 7, 71. Chan, P.H., Yang, G.Y., Carlson, E., Epstein, C.J., 1991. Cold-induced brain edema and infarction are reduced in transgenic mice overexpressing CuZn-superoxide dismutase. Ann. Neurol. 29, 482–486. Cioni, C., Annunziata, P., 2002. Circulating gp120 alters the blood–brain barrier permeability in HIV-1 gp120 transgenic mice. Neurosci. Lett. 330, 299–301. Colgan, O.C., Ferguson, G., Collins, N.T., Murphy, R.P., Meade, G., Cahill, P.A., Cummins, P.M., 2007. Regulation of bovine brain microvascular tight junction assembly and barrier function by laminar shear stress. Am. J. Physiol. Heart Circ. Physiol. 292, H3190–H3197. Dallasta, L.M., Pisarov, L.A., Esplen, J.E., Werley, J.V., Moses, A.V., Nelson, J.A., Achim, C.L., 1999. Blood–brain barrier tight junction disruption in human immunodeficiency virus-1 encephalitis. Am. J. Pathol. 155, 1915–1927. Finco, O., Nuti, S., De Magistris, M.T., Mangiavacchi, L., Aiuti, A., Forte, P., Fantoni, A., van der Putten, H., Abrignani, S., 1997. Induction of CD4+ T cell depletion in mice doubly transgenic for HIV gp120 and human CD4. Eur. J. Immunol. 27, 1319–1324. Garden, G.A., Guo, W., Jayadev, S., Tun, C., Balcaitis, S., Choi, J., Montine, T.J., Moller, T., Morrison, R.S., 2004. HIV associated neurodegeneration requires p53 in neurons and microglia. FASEB J. 18, 1141–1143. Gorny, M.K., Xu, J.Y., Gianakakos, V., Karwowska, S., Williams, C., Sheppard, H.W., Hanson, C.V., Zolla-Pazner, S., 1991. Production of site-selected neutralizing human monoclonal antibodies against the third variable domain of the human immunodeficiency virus type 1 envelope glycoprotein. Proc. Natl. Acad. Sci. U. S. A. 88, 3238–3242. Huang, M.B., Hunter, M., Bond, V.C., 1999. Effect of extracellular human immunodeficiency virus type 1 glycoprotein 120 on primary human vascular endothelium cell cultures. AIDS Res. Hum. Retroviruses 15, 1265–1277. Jones, M.V., Bell, J.E., Nath, A., 2000. Immunolocalization of HIV envelope gp120 in HIV encephalitis with dementia. AIDS 14, 2709–2713. Kanmogne, G.D., Primeaux, C., Grammas, P., 2005. HIV-1 gp120 proteins alter tight junction protein expression and brain endothelial cell permeability: implications for the pathogenesis of HIV-associated dementia. J. Neuropath. Exp. Neurol. 64, 498–505. Kanmogne, G.D., Schall, K., Leibhart, J., Knipe, B., Gendelman, H.E., Persidsky, Y., 2007. HIV-1 gp120 compromises blood–brain barrier integrity and enhance monocyte migration across blood–brain barrier: implication for viral neuropathogenesis. J. Cereb. Blood Flow Metab. 27, 123–134. Kaul, M., Lipton, S.A., 1999. Chemokines and activated macrophages in HIV gp120induced neuronal apoptosis. Proc. Natl. Acad. Sci. U. S. A. 96, 8212–8216. Kaul, M., Garden, G.A., Lipton, S.A., 2001. Pathways to neuronal injury and apoptosis in HIV-associated dementia. Nature 410, 988–994. Ling, L., Zeng, J., Pei, Z., Cheung, R.T., Hou, Q., Xing, S., Zhang, S., 2009. Neurogenesis and angiogenesis within the ipsilateral thalamus with secondary damage after focal cortical infarction in hypertensive rats. J. Cereb. Blood Flow Metab. 29, 1538–1546. Louboutin, J.P., Liu, B., Reyes, B.A.S., van Bockstaele, E.J., Strayer, D.S., 2006. Rat bone marrow progenitor cells transduced in situ by rSV40 vectors differentiate into multiple central nervous system lineages. Stem Cells 24, 2801–2809. Louboutin, J.P., Reyes, B.A.S., Agrawal, L., van Bockstaele, E.J., Strayer, D.S., 2007a. Strategies for CNS-directed gene delivery: in vivo gene transfer to the brain using SV40-derived vectors. Gene Ther. 14, 939–949. Louboutin, J.P., Agrawal, L., Reyes, B.A.S., van Bockstaele, E.J., Strayer, D.S., 2007b. Protecting neurons from HIV-1 gp120-induced oxidant stress using both localized intracerebral and generalized intraventricular administration of antioxidant enzymes delivered by SV40-derived vectors. Gene Ther. 14, 1650–1661. Louboutin, J.P., Agrawal, L., Reyes, B.A.S., van Bockstaele, E.J., Strayer, D.S., 2009a. HIV-1 gp120 neurotoxicity proximally and at a distance from the point of exposure: protection by rSV40 delivery of antioxidant enzyme. Neurobiol. Dis. 34, 462–476. Louboutin, J.P., Agrawal, L., Reyes, B.A.S., van Bockstaele, E.J., Strayer, D.S., 2009b. An experimental model for HIV-1 lesions in the brain using envelope glycoprotein gp120 expression delivered by SV40 vectors. J. Neuropath. Exp. Neurol. 68, 456–473. Marcon, J., Gagliardi, B., Balosso, S., Maroso, M., Noe, F., Morin, M., Lerner-Natoli, M., Vezzani, A., Ravizza, T., 2009. Age-dependent vascular changes induced by status epilepticus in rat forebrain: implications for epileptogenesis. Neurobiol. Dis. 34, 121–132. Mattson, M.P., Haughey, N.J., Nath, A., 2005. Cell death in HIV dementia. Cell Death Diff. 12, 893–904. McArthur, J.C., Brew, B.J., Nath, A., 2005. Neurological complications of HIV infection. Lancet Neurol. 4, 543–555. McKee, H.J., Strayer, D.S., 2002. Immune responses against SIV envelope glycoprotein, using recombinant SV40 as a vaccine delivery vector. Vaccine 20, 3613–3625.

325

Meucci, O., Fatatis, A., Simen, A.A., Bushell, T.J., Gray, P.W., Miller, R.J., 1998. Chemokines regulate hippocampal neuronal signaling and gp120 neurotoxicity. Proc. Natl. Acad. Sci. U. S. A. 95, 14500–14505. Michaud, J., Fajardo, R., Charron, G., Sauvageau, A., Berrada, F., Ramia, D., Dilhuydy, H., Robitaille, Y., Kessous-Elbaz, A., 2001. Neuropathology of NFHgp160 transgenic mice expressing HIV-1 env protein in neurons. J. Neuropathol. Exp. Neurol. 60, 574–587. Miura, Y., Misawa, N., Kawano, Y., Okada, H., Inagaki, Y., Yamamoto, N., Ito, M., Yagita, H., Okumura, K., Mizusawa, H., Koyanagi, Y., 2003. Tumor necrosis factor-related apoptosis-inducing ligand induces neuronal death in a murine model of HIV central nervous system infection. Proc. Natl. Acad. Sci. U. S. A. 100, 2777–2782. Morinville, A., Cahill, C.M., Aibak, H., Rymar, V.V., Pradhan, A., Hoffert, C., Mennicken, F., Stroh, T., Sadikot, A.F., O'Donnell, D., Clarke, P.B., Henry, J.L., Vincent, J.P., Beaudet, A., 2004. Morphine-induced changes in delta opioid receptor trafficking are linked to somatosensory processing in the rat spinal cord. J. Neurosci. 24, 5549–5559. Nath, A., Sacktor, N., 2006. Influence of highly active antiretroviral therapy on persistence of HIV in the central nervous system. Curr. Opin. Neurol. 19, 358–361. Nosheny, R.L., Bachis, A., Acquas, E., Mocchetti, I., 2004. Human immunodeficiency virus type 1 glycoprotein gp120 reduces the levels of brain-derived neurotrophic factor in vivo: potential implication for neuronal cell death. Eur. J. Neurosci. 20, 2857–2864. Paris, D., Patel, N., Delledonne, A., Quadros, A., Smeed, R., Mullan, M., 2004. Impaired angiogenesis in a transgenic mouse model of cerebral amyloidosis. Neurosci. Lett. 366, 80–85. Paxinos, G., Watson, C., 1986. The Rat Brain in Stereotaxic Coordinates2nd ed. Academic Press, New York, NY. Persidsky, Y., Heilman, D., Haorah, J., Zelivyanskaya, M., Persidsky, R., Weber, G.A., Shimokawa, H., Kaibuchi, K., Ikezu, T., 2006. Rho-mediated regulation of tight junctions during monocyte migration across the blood–brain barrier in HIV-1 encephalitis (HIVE). Blood 107, 4770–4780. Petito, C.K., Cash, K.S., 1992. Blood–brain barrier abnormalities in the acquired immunodeficiency syndrome: immunohistochemical localization of serum proteins in postmortem brain. Ann. Neurol. 32, 658–666. Power, C., Kong, P.A., Crawford, T.O., Wesselingh, S., Glass, J.D., McArthur, J.C., Trapp, B.D., 1993. Cerebral white matter changes in acquired immunodeficiency syndrome dementia: alterations of the blood–brain barrier. Ann. Neurol. 34, 339–350. Price, T.O., Ercal, N., Nakaoke, R., Banks, W.A., 2005. HIV-1 viral proteins gp120 and Tat induce oxidative stress in brain endothelial cells. Brain Res. 1045, 57–63. Price, T.O., Uras, F., Banks, W.A., Ercal, N., 2006. A novel antioxidant N-acetylcysteine amide prevents gp120- and Tat-induced oxidative stress in brain endothelial cells. Exp. Neurol. 201, 193–202. Ren, Z., Yao, Q., Chen, C., 2002. HIV-1 envelope glycoprotein 120 increases intercellular adhesion molecule-1 expression by human endothelial cells. Lab. Invest. 82, 245–255. Rouger, K., Louboutin, J.P., Villanova, M., Cherel, Y., Fardeau, M., 2001. X-linked vacuolated myopathy: TNF-alpha and IFN-gamma expression in muscle fibers with MHC class I on sarcolemma. Am. J. Pathol. 158, 355–359. Sacktor, N., Haughey, N., Cutler, R., Tamara, A., Turchan, J., Pardo, C., Vargas, D., Nath, A., 2004. Novel markers of oxidative stress in actively progressive HIV dementia. J. Neuroimmunol. 157, 176–184. Sauter, B.V., Parashar, B., Chowdhury, N.R., Kadakol, A., Ilan, Y., Singh, H., Milano, J., Strayer, D.S., Chowdhury, J.R., 2000. A replication-deficient rSV40 mediates liverdirected gene transfer and a long-term amelioration of jaundice in Gunn rats. Gastroenterology 119, 1348–1357. Stins, M.F., Pearce, D., Di Cello, F., Erdreich-Epstein, A., Pardo, C.A., Sik Kim, K., 2003. Induction of intercellular adhesion molecule-1 on human brain endothelial cells by HIV-1 gp120: role of CD4 and chemokine coreceptors. Lab. Invest. 83, 1787–1798. Strayer, D.S., 1999. Gene therapy using SV40-derived vectors: what does the future hold? J. Cell. Physiol. 181, 375–384. Strayer, D.S., Kondo, R., Milano, J., Duan, L.X., 1997. Use of SV40-based vectors to transduce foreign genes to normal human peripheral blood mononuclear cells. Gene Ther. 4, 219–225. Strayer, D.S., Lamothe, M., Wei, D., Milano, J., Kondo, R., 2001. Generation of recombinant SV40 vectors for gene transfer. SV40 protocols. In: Raptis, L. (Ed.), Methods in Molecular Biology, vol. 165. Humana Press, Totowa, NJ, pp. 103–117. Toggas, S.M., Masliah, E., Rockenstein, E.M., Rall, G.F., Abraham, C.R., Mucke, L., 1994. Central nervous system damage produced by expression of the HIV-1 coat protein gp120 in transgenic mice. Nature 367, 188–193. Toneatto, S., Finco, O., van der Putten, H., Abrignani, S., Annunziata, P., 1999. Evidence of blood–brain barrier alteration and activation in HIV-1 gp120 transgenic mice. AIDS 13, 2343–2348. Ullrich, C.K., Groopman, J.E., Ganju, R.K., 2000. HIV-1 gp120- and gp160-induced apoptosis in cultured endothelial cells is mediated by caspases. Blood 96, 1436–1442. Wu, C.C., Reilly, J.F., Young, W.G., Morrison, J.H., Bloom, F.E., 2004. High throughput morphometric analysis of individual neurons. Cereb. Cortex 14, 543–554. Zhong, Y., Smart, E.J., Weksler, B., Couraud, P.O., Hennig, B., Toborek, M., 2008. Caveolin-1 regulates human immunodeficiency virus-1 Tat-induced alterations of tight junction protein expression via modulation of the Ras signaling. J. Neurosci. 28, 7788–7796. Zozulya, A.L., Reinke, E., Baiu, D.C., Karman, J., Sandor, M., Fabry, Z., 2007. Dendritic cell transmigration through brain microvessel endothelium is regulated by MIP-1α chemokine and matrix metalloproteinases. J. Immunol. 178, 520–529.