Recombinant HIV-1 Nef protein, but not Tat, gp120, and gp160, provoked leukocyte recruitment into the CNS in a rat model. The strong reduction of bioactivity by heat treatment of Nef, and the blocking effect of the mAb 2H12, which recognizes the carboxy-terminal amino acid (aa) residues 171–190 (but not of mAb 3E6, an anti-Nef Ab of the same isotype, which maps the aa sequence 168–175, as well as a mixture of mAbs to CD4) provided evidence for the specificity of the observed Nef effects. Using a modified Boyden chamber technique, Nef exhibited chemotactic activity on mononuclear cells in vitro. Coadministration of the anti-Nef mAb 2H12, as well as treatment of Nef by heat inhibited Nef-induced chemotaxis. Besides soluble Nef, chemotaxis was also induced by a Nef-expressing human astrocytoma cell line, but not by control cells. These data suggest a direct chemotactic activity of soluble Nef. The detection of elevated levels of IL-6, TNF-α, and IFN-γ in rat cerebrospinal fluid 6 h after intracisternal Nef injection hint at the additional involvement of indirect mechanisms in Nef-induced leukocyte migration into rat CNS. These data propose a mechanism by which HIV-1 Nef protein may be essential for AIDS neuropathogenesis, as a mediator of the recruitment of leukocytes that may serve as vehicles of the virus and perpetrators for disease through their production of neurotoxins.

Human immunodeficiency virus-1 invades the CNS early after viral exposure and persists (1). It is well established that mononuclear phagocytes (macrophages and microglia) are the predominant cell types carrying a replicating virus in brain specimens with AIDS-associated neuropathology (2). However, no or only isolated HIV-replicating macrophages/microglial cells could be detected by examination of brains of asymptomatic HIV-1 carriers, raising the question about the viral reservoir during latency (3). Recent investigations using post mortem tissues suggest that restricted HIV-1 infection occurs in astrocytes (4). In vitro infection of human fetal astrocytes, as well as astrocytoma cell lines, has also been demonstrated (5, 6, 7). HIV-1 establishes a persistent state of infection in astroglial cells in vitro, which is characterized by impairment of Rev function (8). During periods of viral persistence, the most abundant regulatory transcripts were those that coded for Nef and Rev (8, 9). Immunohistochemical staining of astrocytes persistently infected with HIV-1 further revealed expression of Nef proteins in these cells at times when little or no viral structural protein could be detected (10). In addition, recent studies have demonstrated that nef mRNA and protein were detectable in astrocytes in brain sections selected for extensive histopathology (10, 11), and (2) that Nef has structural and functional similarities to scorpion neurotoxin, suggesting a role in AIDS neuropathogenesis (12). HIV-1 Nef protein, together with Tat and Rev, is expressed early during virus infection (13). Nef expression is essential for the maintenance of high levels of viral replication and AIDS pathogenesis in SIV-infected monkeys (14). The mechanism(s) by which Nef has such an effect is (are) largely unknown.

Recently, a highly significant correlation between increased numbers of macrophages and HIV-associated dementia have been demonstrated (15), pointing at the role of macrophage infiltrates for the exacerbation of neuronal dysfunction. Several clinical and experimental studies have focused on the production of chemotactic factors during HIV infection (16, 17, 18, 19). In addition, HIV-1 Tat protein itself has been reported to enhance the chemotactic and invasive behavior of monocytes in vitro (20). However, pathogenic mechanisms responsible for the entry of blood-derived mononuclear cells into the CNS during HIV-1 infection remain speculative.

In this study, we investigated whether (1) recombinant HIV-1-Nef protein (compared with Tat, gp120, and gp160) can induce cellular trafficking into the CNS in vivo in a rat model, (2) and whether Nef is a chemoattractant for human blood-derived mononuclear and polymorphonuclear leukocytes (PMN)3 using a modified Boyden chamber assay

In total, 176 male Wistar rats were used in this study. The protocol of this study was approved by the Government of Upper Bavaria. To test acute effects of intrathecal injection of HIV-1 Nef protein, we used a well-characterized rat model of meningitis, which was previously described in detail (21, 22, 23). Rats (250–300 g) were anesthetized with 100 mg/kg thiopental (Byk Gulden, Konstanz, Germany), tracheotomized, and artificially ventilated with a small animal ventilator (Ap-10; K. Effenberger, Pfaffing, Germany). Mean arterial blood pressure (MABP) was measured through a cannula inserted into the right femoral artery using a pressure transducer (Statham P23; Viggo-Spectramed, Oxnard, CA). Arterial blood samples were drawn at the beginning and at the end of each experiment, as well as every 2 h, and were analyzed for arterial PaCO2, PaO2, and pH with a blood gas analyzer (IL1304; Instrumentation Laboratory, Kirchheim, Germany). The right femoral vein was cannulated for fluid substitution. Body temperature was monitored and maintained at 37.0 ± 0.5°C using a rectal thermometer-controlled heating pad.

Cerebrospinal fluid (CSF)3 samples were collected for white blood cell (WBC) count before and 6 h after intracisternal (i.c.) injection, as well as for the determination of concentrations of monocyte chemoattractant protein (MCP)-1, IL-6, TNF-α, and IFN-γ. Intracranial pressure (ICP) was recorded with a pressure transducer (Statham P23; Viggo-Spectramet, Oxnard, CA) connected to a cisterna magna catheter, placed through a burr hole in the occipital bone. For placement of the laser-Doppler probe, a 5 × 4-mm area of the right parietal bone was gently abraded to a transparent bone layer, while continuously rinsing with saline for cooling. The laser-Doppler probe (model BPM 403a; Vasamedics, St. Paul, MN) held by a micromanipulator was placed over an area free of large dural and pial vessels, as described previously (24). Cerebral blood flow (CBFLDF) was measured continuously by laser-Doppler flowmetry and was expressed as a percentage of a stable baseline of 100%. When stable baselines of CBFLDF and ICP were achieved for 30 min, 100 μl CSF was removed through the i.c. catheter. Then, 100 μl of the substances mentioned below (see Experimental groups section) were administered into the cisterna magna.

Since a predominance of PMN was found in early CSF samples (taken 6 h after Nef injection), additional experiments were done to examine whether a shift from polymorphonuclear to mononuclear leukocytes occurs later after Nef injection. For this purpose, rats were anesthetized with chloral hydrate i.p. (300 mg/kg). HIV-1 Nef protein, untreated or treated by heat, as well as PBS (100 μl each) was injected transcutaneously into the cisterna magna. Rats were put into cages and allowed to wake up. Then, 22 h after i.c. injection, rats were anesthetized with thiopental, tracheotomized, and artificially ventilated. A catheter was inserted into the cisterna magna for ICP monitoring and to determine CSF WBC counts and CSF concentrations of both MCP-1 and IFN-γ. The right femoral artery was canulated for monitoring of MABP and for blood gas and hematocrit analyses.

To investigate whether HIV-1 Nef protein can induce leukocyte infiltration in sites other than the CNS, rats were anesthetized, tracheotomized, and artificially ventilated as described above. A catheter was inserted into the right femoral artery for MABP monitoring. Using a disposable syringe with an integrated 29-gauge 0.5-inch needle, 100 μl of the substances mentioned below (see Experimental group sections) were injected either into the dorsal skin or the dorsal muscle. Then, 6 h later, rats were killed by exsanguination, tissue samples weighing 0.1–0.3 g were immediately removed and stored at −70°C until assayed for myeloperoxidase (MPO) activity.

MCP-1 and IFN-γ levels in the CSF were determined by commercially available ELISA systems (Laboserv; Biosource International, Camarillo, CA, and R&D Systems, Minneapolis, MN, respectively) according to the manufacturer’s instructions. The lower detection limit for MCP-1 and IFN-γ in CSF samples was 50 and 20 pg/ml, respectively. TNF-α was quantitated in a bioassay using TNF-sensitive L-M cells (25), and IL-6 was assessed using the IL-6-dependent B cell hybridoma 7TD1, as previously described (26). For both bioassays, a detection limit of 30 U/ml was found for the test samples.

Tissue MPO concentrations were determined as markers of leukocyte infiltration, according to Cuzzocrea et al. (27). Tissue samples were homogenized in 50 mM potassium phosphate buffer (pH 6.0) using an Ultra-Turrax (model TP18/10; Janke&Kunkel KG, Staufen, Germany) at 20,000 rpm for 1 min at 4°C. The homogenate from each sample was centrifuged (15 min at 14,000 rpm), and the resulting pellet was rehomogenized in 50 mM potassium phosphate buffer containing 0.5% hexadecyltrimethylammonium bromide, sonicated, freeze-thawed three times, and centrifuged for 15 min at 14,000 rpm. Then, the supernatants were assayed for MPO activity by monitoring the rate of H2O2-dependent oxidation of 0-dianisidine hydrochloride (0.167 mg/ml; Sigma Chemicals, Deisenhofen, Germany) at 460 nm at room temperature. One unit of MPO activity is defined as that degrading one micromole of peroxide per minute at RT.

Thirty-three different experimental groups were investigated. To examine whether HIV-1 Nef protein can induce CSF pleocytosis, as well as changes in CBFLDF and ICP, rats were injected i.c. with HIV-1 Nef protein (for preparation, see below) at four different dosages (groups 1–4): 1 ng (n = 3), 10 ng (n = 9), 100 ng (n = 15), and 1000 ng (n = 5). Rats injected i.c. either with heat-inactivated HIV-1 Nef protein at two different dosages (100 ng (n = 5) or 1000 ng (n = 3)), with mock solution (vehicle of Nef [vehNef] = vector plasmid without the nef gene purified according to the same procedure as recombinant Nef protein, n = 10), or with PBS (n = 10), served as negative controls (groups 5–8). To exclude that the observed effects of Nef depend on the preparation procedure, purified Nef protein (HIV-1LAV Nef) obtained from the National Institutes of Health (NIH) AIDS Research and Reagent Program (Rockville, MD; group 9, rats injected i.c. with 100 ng HIV-1LAV Nef (n = 8), compared with rats injected i.c. with 100 ng heat-inactivated HIV-1LAV Nef (group 10, n = 4)), as well as an additional independently produced recombinant Nef preparation (group 11, rats injected i.c. with 100 ng HisNefBru (n = 5), compared with rats injected i.c. with 100 ng heat-treated HisNefBru (group 12, n = 3)) was tested. To obtain information about the specifity of the observed effects of HIV-1 Nef proteins and about the regions of the protein involved, we examined: 1) whether two mAbs to Nef, both of the same Ig G1 subclass, (administered either individually or together), and a mixture of two mAbs to CD4 can modulate Nef activity, and 2) whether five synthetic Nef peptides by itself can mimic the observed effects of HIV-1 Nef proteins: 1) rats injected i.c. with 100 ng Nef and pretreated (30 min before Nef injection) i.c. with (a) a mixture of the mAbs 2H12 and 2E6 recognizing the Nef epitopes located at amino acid residues 171–190 and 168–175, respectively (5 μg of each mAb, n = 6, group 13), (b) a mixture of the mAbs MT310 and MT15 recognizing the first and second domain of CD4 (5 μg of each mAb, n = 4, group14), or (c) 5 μg of mAb 2H12 (n = 3) and 3E6 (n = 3), respectively (groups 15, 16); and 2) rats injected i.c. with the synthetic Nef peptide NP141–160 (100 ng, n = 5), NP171–190 (100 ng, n = 6), NP181–205 (100 ng, n = 5), NP160–174 (100 ng, n = 3), and NP181–191 (100 ng, n = 3), respectively (groups 17–21). Since in vitro studies have shown that Tat and gp120 can act as a chemoattractant for mononuclear leukocytes and T lymphocytes, respectively, (20, 28), we tested the activity of the following HIV-1 proteins in our acute rat model (groups 22–24): rats injected i.c. with 100 ng Tat (n = 5), gp120 (n = 8), and with 100 ng gp160 (n = 5). As positive controls for leukocyte recruitment into the CNS, additional rats were injected i.c. either with 100 μl 10−5 M fMLP (group 25, n = 4) or with 100 μg carrageenan type IV (group 26, n = 4; both from Sigma, Deisenhofen, Germany), well-known inducers of meningeal inflammation in rodents (29, 30). In CSF drawn 6 h after i.c. Nef injection, a predominance of PMN was observed, a well-known phenomenon in experimental in vivo models of inflammation (31). To test whether a switch from polymorphonuclear to mononuclear leukocytes occurs, we performed experiments for a period of 24 h in groups 27–29, rats injected with 100 ng Nef (n = 6), with 100 ng heat-inactivated Nef (n = 5), and with PBS (n = 5), respectively. To exclude that the observed effects of Nef depend on the preparation procedure, an additional independently produced Nef protein construct (group 30, rats injected i.c. with 100 ng HisNefBru (n = 5), compared with rats injected i.c. with 100 ng heat-treated HisNefBru (group 31, n = 3)) was tested. To investigate whether HIV-1 Nef protein can induce leukocyte infiltration in sites other than the CNS, 100 ng Nef (compared with 100 ng heat-inactivated Nef, PBS, and 100 μg carrageenan) were instilled either in rat dorsal skin (group 32, n = 4) or dorsal muscle (group 33, n = 4).

Veneous blood was drawn from healthy, drug-free volunteers and immediately mixed with preservative-free sodium heparin at a final concentration of 50 IU/ml. Briefly, PMN were isolated by sedimentation with a 1:1 solution containing 6% dextran and Histopaque 1.119 (Sigma), followed by density gradient centrifugation over Histopaque 1.077 (Sigma). After hypotonic lysis of remaining erythrocytes, PMN were washed twice in HBSS (Sigma) and resuspended in culture medium composed of RPMI 1640 (Sigma) and 1% FCS (Biochrom, Berlin, Germany). Viability, as tested by trypan blue dye exclusion, was >98%. By morphological criteria, the final cell preparation contained >99% PMN.

Mononuclear leukocytes were prepared by density gradient centrifugation over Histopaque 1.077 for 30 min and 900 × g at room temperature after 1/2 (v/v) dilutions in RPMI 1640. Mononuclear cells forming a cell layer above Histopaque 1.077 were collected and added to a fresh tube. After washing twice with RPMI 1640, PBMC were resuspended in RPMI 1640 containing 1% FCS, and viability was tested by trypan blue exclusion, which was always >95%. Nonspecific esterase staining of the mononuclear cell preparation showed the cell suspension to be composed of 15–25% monocytes and 75–85% lymphocytes.

The nef gene from the HIV-1 Bru-isolate, which was derived from permanently HIV-1-infected human astrocytoma cells (32), was cloned into the eukaryotic expression vector pSG5. U251 MG human astrocytoma cell lines were established cotransfected with constructs encoding nef and the neomycin-resistance gene (pSG-neo) (33). Single clones (astro-4/4.2) were isolated under selective conditions, and Nef expression was verified by immunoperoxidase staining and Western blot analysis. As control cell lines, clones of U251 MG cells (astro-pSGK1) were used, which were exclusively transfected with the pSG-neo construct. Both, astro-4/4.2 and astro-pSGK1 were cultured in a 5% CO2/95% air-humidified incubator at 37°C in RPMI 1640 medium supplemented with 10% FCS, 2 mM glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin (all from Sigma), and 200 μg/ml genticin (G418; Roche Molecular Biochemicals, Mannheim, Germany). For chemotaxis experiments, cells were seeded onto 24-well companion multiplates (Falcon; Becton Dickinson, Heidelberg, Germany) at a concentration of 106 cells/ml and allowed to adhere. Then, 4 h later, the cell culture medium was replaced by 500 μl RPMI 1640 containing 1% FCS. Then, polyethylene terephthalate membrane inserts (Falcon; Becton Dickinson) were placed into each well and filled with the PBMC suspension (4 × 106 PBMC/μl). After an incubation period of 4 h, the inserts were removed, and the membranes were processed as described below. In additional experiments, we examined whether the chemotactic activity exhibited by the Nef-expressing U251 MG cell clone astro-4/4.2 can be blocked by the mAb to Nef 2H12 given at three different concentrations (0.1, 1.0, and 10.0 μg/ml).

Leukocyte migration was assayed in vitro using a modified Boyden chamber technique. The initial experiments were performed in a 48-well microchamber (Neuroprobe, Bethesda, MD) technique, as previously described (34). Briefly, chemoattractants were diluted in RPMI 1640 with 1% FCS and placed in the lower well (25 μl). Fifty microliters of cell suspension at 1.1 × 106 PMN/ml or 4 × 106 PBMC/ml were added to the upper well of the chamber, which was separated from the lower well by polycarbonate, polyvinylpyrrolidone-free micropore filter (Nucleopore, Pleasanton, CA). The filter pore sizes for PMN and PBMC chemotaxis were 3 and 5 μm, respectively. Following incubation of the chamber (60 min for PMN and 90 min for PBMC migration) at 37°C in 5% CO2 in a humidified atmosphere, the filters were removed, after cleaning the upper side, they were fixed with methanol and stained with DiffQuik (Baxter Diagnostics AG, Dudingen, Switzerland). Chemotaxis was quantified by microscopic counting of the cells that migrated completely through the pores of the filter in 10 randomly chosen high power fields (×100). Cell migration was expressed as the mean number of leukocytes that migrated per field. FMLP was used as a positive control at concentrations of 10−6 M found to be optimal for the migration of PBMC and PMN.

Further sets of experiments were performed using 24-well companion multiplates. The lower compartment containing the chemoattractants (500 μl) was separated by a polyethylene terephthalate membrane insert from the upper compartment, which was filled with the cell suspension (350 μl of 1.1 × 106 PMN and 4 × 106 PBMC/μl, respectively). Apart from the different chamber assembly, there were no differences in the experimental setup.

Recombinant HIV-1MN gp160 was kindly provided by Prof. F. Dorner (Immuno AG, Wien, Austria). Recombinant HIV-1SF2 gp120, recombinant HIV-1 Tat, and recombinant HIV-1LAV Nef were obtained from the NIH AIDS Research and Reference Program (contributed by Dr. K. Steimer, Dr. J. Brady, and Division of Acquired Immunodeficiency Syndrome (DAIDS), National Institute of Allergy and Infectious Diseases, respectively). mAbs to Nef 2H12 and 3E6, which recognize the amino acid residues 171–190 and 168–175, respectively (35), were a generous gift from Prof. K. Krohn (Department of Biomedical Sciences, Tampere, Finland). Anti-CD4 Abs M-T15 and M-T310, specific for the first and second domain of CD4, respectively, were kindly provided by Prof. E.P. Rieber (Institute for Immunology, Munich, Germany). Synthetic Nef peptides (residues 141–160, 171–190, and 181–205) were a generous gift from Prof. B. Wahren (Swedish Institute for Infectious Disease Control, Stockholm, Sweden). The Nef peptides NP160–174 and NP181–191 were purchased from German-American Institute for Applied Biomedical Research and Intracel (Berlin, Germany).

Transfection of Escherichia coli strain K12 with plasmid pTG1166 containing HIV LAV-1 Bru Nef under the control of a thermoinducible promoter (kindly provided by Transgène, Strasbourg, France) and induction of recombinant Nef protein expression was performed as described (36). Lysis of the bacterial pellets and purification of recombinant Nef protein was described in detail by Kohleisen et al. (37). Briefly, Nef-containing pellets were solubilized with 8 M urea and 1% 2-ME. Supernatants collected by centrifugation were subjected to affinity chromatography on heparin-sepharose (Pharmacia, Freiburg, Germany). A total of 10 mM sodium phosphate buffer containing 50 mM NaCl was used for equilibration and elution of bound Nef protein was performed with 200 mM NaCl by a step-wise salt gradient. Fractions were assayed for purity of Nef-protein by SDS-PAGE on a 12% gel and subsequent silver staining. Western blot analysis was done using Nef-specific mono- and polyclonal Abs, as previously described (32). Enzymatic reactions of alkaline phosphatase- or HRP-conjugated secondary Abs were detected with colorigenic substrates (NBT/BCIP) or chemiluminescence (ECL; Amersham-Buchler, Braunschweig, Germany). Western blot analysis and silver staining after SDS-PAGE required 100 ng/ml and 400 ng/ml of recombinant Nef protein for the detection of clearly visible bands of the protein, respectively.

To exclude that the observed effects of Nef depend on the preparation procedure, an additional independently produced recombinant Nef (HisNefBru) was tested. The Nef-gene derived from HIV-1 Bru was cloned into the pQE30 vector (Qiagen, Hilden, Germany), and expressed as recombinant fusion-proteins containing six histidine residues and a factor Xa cleavage site at the N terminus of Nef. Nef was purified from bacterial pellet under nondenaturing conditions using metal chelate affinity chromatography (Ni-NTA) and subsequent chromatography on heparin-sepharose (37). In this preparation, the concentration of recombinant protein was determined by the bicinchoninic acid assay (Pierce, Rockford, IL).

Protein concentration of purified recombinant Nef protein was determined by gel filtration (PC3.2/10, Pharmacia Freiburg, Germany) and subsequent peak area determination at 214 nm, as already described (37). At least five concentrations of a BSA (Pierce) were used for standardization. Protein concentrations of mAbs used for blocking studies were determined according to the same procedure. Additionally, Nef protein preparations were determined by bicinchoninic acid reaction at 570 nm (Pierce). Identical concentrations for recombinant Nef were obtained by this method compared with peak area determination at 214 nm.

We have identified several factors that may reduce the bioactivity (amounts) of purified Nef protein, including maintaining solutions at room temperature, repeated freeze-thaw cycles, and also prolonged storage (beyond 3 mo), even at −80°C.

The principal statistical test was one-way ANOVA and Student-Newman-Keuls test. Differences were considered significant at p < 0.05. Data are expressed as mean ± SEM. For analysis of CSF cytokine/chemokine concentrations, data were compared using Kruskal-Wallis H test. Spearman-rho correlation analysis was used to evaluate the relationship between the Nef dosage used and the observed CSF leukocyte recruitment (response).

Intracisternal injection of recombinant HIV-1 Nef protein induced a significant increase of CSF WBC counts (Fig. 1 A), irrespective of the Nef preparation investigated. Statistical analysis revealed a significant correlation between the concentrations of Nef protein injected i.c. and CSF leukocyte counts (Spearman-rho correlation: r = 0.877; p < 0.01; Kendall-tau-b: r = 0.739; p < 0.01), indicating a clear (sigmoid) dose-response relationship.

FIGURE 1.

A, HIV-1 Nef protein induced a significant increase in CSF leukocyte count, irrespective of the protein preparation investigated (Nef preparation described above (Nef), compared with HIV-1LAV Nef from DAIDS, National Institute of Allergy and Infectious Diseases). ∗, p < 0.05, compared with rats injected i.c. either with PBS, the mock solution of Nef (mock), 1 ng Nef, or 10 ng Nef. As positive controls, rats were injected (i.c. either with 100 μl 10−5 M fMLP or with 100 μg carrageenan type IV (n = 4). Both fMLP and carrageenan induced a significant increase in CSF leukocyte counts. B, Inactivation of the protein by heat (60°C for 20 min) and pretreatment with a mixture of two mAbs to Nef (mAbsNef, 2H12, 3E6) prevented CSF pleocytosis. In contrast, a mixture of two anti-CD4 mAbs had no effect. , p < 0.05, compared with rats injected i.c. with 100 ng Nef with or without anti-CD4 mAbs. Data are expressed as means ± SE.

FIGURE 1.

A, HIV-1 Nef protein induced a significant increase in CSF leukocyte count, irrespective of the protein preparation investigated (Nef preparation described above (Nef), compared with HIV-1LAV Nef from DAIDS, National Institute of Allergy and Infectious Diseases). ∗, p < 0.05, compared with rats injected i.c. either with PBS, the mock solution of Nef (mock), 1 ng Nef, or 10 ng Nef. As positive controls, rats were injected (i.c. either with 100 μl 10−5 M fMLP or with 100 μg carrageenan type IV (n = 4). Both fMLP and carrageenan induced a significant increase in CSF leukocyte counts. B, Inactivation of the protein by heat (60°C for 20 min) and pretreatment with a mixture of two mAbs to Nef (mAbsNef, 2H12, 3E6) prevented CSF pleocytosis. In contrast, a mixture of two anti-CD4 mAbs had no effect. , p < 0.05, compared with rats injected i.c. with 100 ng Nef with or without anti-CD4 mAbs. Data are expressed as means ± SE.

Close modal

Both inactivation of the protein by heat (60°C for 20 min) and pretreatment with a mixture of two mAbs to Nef (2H12 and 3E6) strongly inhibited CSF pleocytosis (Fig. 1,B). Pretreatment with the mAb 2H12, which map to the amino acid residues 171–190 (but not with mAb 3E6, an anti-Nef Ab of the same isotype [IgG G1], which maps the aa sequence 168–175), prevented Nef-induced CSF pleocytosis (Fig. 2). However, i.c. injection of different synthetic Nef peptides selected from this epitope (amino acid residues 141–160, 160–174, 171–190, 181–191, and 186–205) were ineffective. Furthermore, a mixture of two anti-CD4 mAbs (M-T15 and M-T319) produced no effect on Nef-induced CSF pleocytosis. When injecting the mock solution of Nef (vector plasmid without the nef gene purified according to the same procedure as recombinant Nef protein) or PBS into the cisterna magna of rats, no CSF pleocytosis could be detected (p < 0.05, compared with rats injected with 100 ng and 1000 ng Nef, respectively; Fig. 1 B). Likewise, the i.c. instillation of other HIV-1 proteins, including recombinant gp160, gp120, and Tat (each in a dosage of 100 ng), did not cause an increase in CSF WBC counts (19 ± 10, 39 ± 5, and 102 ± 81, respectively, NS).

FIGURE 2.

Pretreatment with the mAb 2H12 to Nef that map to the amino acid residues 171–190 (but not with mAb 3E6) prevented Nef-induced CSF pleocytosis. Intracisternal injection of different synthetic Nef peptides (aa residues 141–160, 160–174, 171–190, 181–191, and 186–205) were ineffective. ∗, p < 0.05, compared with rats injected i.c. with 100 ng Nef irrespective whether rats were untreated or pretreated either with mAbNef 3E6. Data are expressed as means ± SE.

FIGURE 2.

Pretreatment with the mAb 2H12 to Nef that map to the amino acid residues 171–190 (but not with mAb 3E6) prevented Nef-induced CSF pleocytosis. Intracisternal injection of different synthetic Nef peptides (aa residues 141–160, 160–174, 171–190, 181–191, and 186–205) were ineffective. ∗, p < 0.05, compared with rats injected i.c. with 100 ng Nef irrespective whether rats were untreated or pretreated either with mAbNef 3E6. Data are expressed as means ± SE.

Close modal

Apart from CSF pleocytosis, HIV-1 Nef protein (100 ng) also increased CBF, the effect being inhibited by anti-Nef Abs (123.0 ± 3.1% vs 100.4 ± 6.6% in anti-Nef Ab treated, Nef-injected rats and 99.8 ± 6.1% in PBS controls; p < 0.05). Furthermore, a slight increase of intracranial pressure was noted after HIV-1 Nef injection (4.6 ± 0.9 mm Hg vs 2.3 ± 0.6 mm Hg in anti-Nef Ab treated, Nef-injected rats and 2.7 ± 0.7 mm Hg in PBS controls; NS). Neither the mock solution of Nef nor PBS nor the other above-mentioned HIV-1 proteins influenced these parameters (data not shown).

Analysis of CSF differential cell count revealed a predominance of PMN at 6 h after i.c. injection of Nef (in two randomly selected rats: 93% and 86% PMN, respectively). Then, 24 h after Nef application, a switch from polymorphonuclear to mononuclear cells was observed in the investigated CSF samples (in two randomly selected rats: 53% and 49% lymphocytes, respectively, and 13% and 32% monocytes, respectively). At this time point, leukocyte counts in CSF samples of rats injected i.c. with HIV-1 Nef protein were still significantly higher than in rats injected i.c. with heat-treated Nef and PBS, respectively (see Table II).

Table II.

Effect of HIV-1 Nef protein on CSF leukocyte count, as well as CSF IFN-γ and MCP-1 concentration 24 h after i.c. injection in an experimental rat modela

Groups (no. of rats)CSF Leukocyte Count (cells/μl)CSF IFN-γ (no. of rats with detectable IFN-γ levels (>20 pg/ml)bCSF MCP-1 (ng/ml)
PBS (6) 62 ± 39c 0 /5 1.6 ± 0.3 
100 ng Nef (6) 695 ± 162 1 /5 1.2 ± 0.2 
100 ng Nefheat-treated (5) 226 ± 54c 0 /5 1.2 ± 0.4 
Groups (no. of rats)CSF Leukocyte Count (cells/μl)CSF IFN-γ (no. of rats with detectable IFN-γ levels (>20 pg/ml)bCSF MCP-1 (ng/ml)
PBS (6) 62 ± 39c 0 /5 1.6 ± 0.3 
100 ng Nef (6) 695 ± 162 1 /5 1.2 ± 0.2 
100 ng Nefheat-treated (5) 226 ± 54c 0 /5 1.2 ± 0.4 
a

Twenty-four hours after i.c. injection, HIV-1 Nef protein caused a significant increase in CSF leukocyte count (but not in CSF MCP-1 and IFN-γ concentrations).

b

In one rat, CSF IFN-γ concentration was above the detection limit (20.2 pg/ml). Data are expressed as means ± SE.

c

, p < 0.05, compared with rats injected i.c. with 100 ng HIV-1 Nef protein.

In additional experiments, we tested the chemotactic activity of soluble Nef protein in sites other than the CNS, namely the dorsal skin and the dorsal muscle. A significant increase in tissue MPO activity, a marker for leukocyte infiltration, was observed after both s.c. and i.m. injection of Nef, compared with PBS and heat-inactivated Nef, suggesting that the Nef-induced chemotaxis is not CNS-specific (Fig. 3).

FIGURE 3.

Soluble Nef protein acts as a chemoattractant in sites other than the CNS, namely in the dorsal skin (A) and the dorsal muscle (B). Tissue MPO levels were determined as markers of leukocyte infiltration. PBS and heat-treated Nef (100 ng) were used as negative controls. Carrageenan (100 μg) served as a positive control. In these experiments, we observed a 7-fold increase in tissue MPO activity after s.c. injection of Nef, and a 2-fold increase after i.m. Nef application, compared with both negative controls. ∗, p < 0.05, compared with rats injected with 100 ng HIV-1 Nef protein. Data are expressed as means ± SE.

FIGURE 3.

Soluble Nef protein acts as a chemoattractant in sites other than the CNS, namely in the dorsal skin (A) and the dorsal muscle (B). Tissue MPO levels were determined as markers of leukocyte infiltration. PBS and heat-treated Nef (100 ng) were used as negative controls. Carrageenan (100 μg) served as a positive control. In these experiments, we observed a 7-fold increase in tissue MPO activity after s.c. injection of Nef, and a 2-fold increase after i.m. Nef application, compared with both negative controls. ∗, p < 0.05, compared with rats injected with 100 ng HIV-1 Nef protein. Data are expressed as means ± SE.

Close modal

Since both polymorphonuclear and mononuclear cells could be detected in CSF samples of rats injected i.c. with Nef, dependent on the time of investigation, we examined whether HIV-1 Nef protein by itself exhibits chemotactic activity on both cell types using a modified Boyden chamber technique.

Using two different chemotaxis chamber assemblies (see Materials and Methods), soluble HIV-1 Nef protein was found to act as a chemoattractant on both cell types (Table III). A maximum chemotactic response of PMN and PBMC was attained with 100 ng/ml and 10 ng/ml HIV-1 Nef protein, respectively. Similar to our in vivo investigation, coadministration of the mAb 2H12 (but not of the mAb 3E6), as well as treatment of HIV-1 Nef protein by heat inhibited Nef-induced chemotaxis. In contrast to Nef, the synthetic HIV-1-Nef peptides did not act as chemoattractants for both PMN and PBMC (data not shown).

Table III.

Chemotactic activity of HIV-1 Nef protein on human polymorphonuclear and mononuclear leukocytes as measured by modified Boyden chamber techniquesa

GroupsPolymorphonuclear LeukocytesMononuclear Leukocytes
24-well multiplate chemotaxis assemblyb48-well microchamber chemotaxis assayc24-well multiplate chemotaxis assemblyc48-well microchamber chemotaxis assayd
Cell no.%Cell no.%Cell no.%Cell no.%
Control 5.9 ± 2.9ef 100 6.5 ± 2.4 100 6.2 ± 1.0eg 100 5.7 100 
0.1% BSA 6.1 ± 2.1ef 103 ND  6.3 ± 1.1eg 102 ND  
100 ng Nefheat-treated 5.3 ± 0.8ef 90 5.7c 88 6.6 ± 0.7eg 106 ND  
1 ng Nef 7.4 ± 1.1ef 125 10.5 ± 3.8 162 9.2 ± 1.2 148 11.25 197 
10 ng Nef 10.2 ± 1.4 173 12.6 ± 2.6 194 21.3 ± 5.2 344 15.6 274 
100 ng Nef 15.2 ± 2.1 258 16.8 ± 0.2 258 14.6 ± 1.9 235 140 
1000 ng Nef ND  10.1 ± 2.2 155 ND  6.4 112 
mAbNef 2H12+ 100 ng Nef 8.9 ± 1.1ef 151 ND  6.9 ± 0.3eg 111 ND  
mAbNef 3E6+ 100 ng Nef 14.1 ± 1.8 239 ND  16.9 ± 0.6 273 ND  
fMLP (1 μM) 56.5 ± 3.9 958 65.2 ± 13.3 1003 49.8 ± 2.7 803 26.8 470 
GroupsPolymorphonuclear LeukocytesMononuclear Leukocytes
24-well multiplate chemotaxis assemblyb48-well microchamber chemotaxis assayc24-well multiplate chemotaxis assemblyc48-well microchamber chemotaxis assayd
Cell no.%Cell no.%Cell no.%Cell no.%
Control 5.9 ± 2.9ef 100 6.5 ± 2.4 100 6.2 ± 1.0eg 100 5.7 100 
0.1% BSA 6.1 ± 2.1ef 103 ND  6.3 ± 1.1eg 102 ND  
100 ng Nefheat-treated 5.3 ± 0.8ef 90 5.7c 88 6.6 ± 0.7eg 106 ND  
1 ng Nef 7.4 ± 1.1ef 125 10.5 ± 3.8 162 9.2 ± 1.2 148 11.25 197 
10 ng Nef 10.2 ± 1.4 173 12.6 ± 2.6 194 21.3 ± 5.2 344 15.6 274 
100 ng Nef 15.2 ± 2.1 258 16.8 ± 0.2 258 14.6 ± 1.9 235 140 
1000 ng Nef ND  10.1 ± 2.2 155 ND  6.4 112 
mAbNef 2H12+ 100 ng Nef 8.9 ± 1.1ef 151 ND  6.9 ± 0.3eg 111 ND  
mAbNef 3E6+ 100 ng Nef 14.1 ± 1.8 239 ND  16.9 ± 0.6 273 ND  
fMLP (1 μM) 56.5 ± 3.9 958 65.2 ± 13.3 1003 49.8 ± 2.7 803 26.8 470 
a

Different concentrations of HIV-1 Nef protein (with or without mAbs to Nef) were supplied to the lower compartment of each chemotaxis chamber. A suspension of either polymorphonuclear (PMN) or mononuclear (PBMC) leukocytes, in control medium, was placed into the upper compartment. The chambers were incubated 60 min for PMN and 90 min for PBMC migration. The results are means ± SE (percent of control values) of migrated cells per field in 10 randomly chosen high-power fields (×100).

b

The experiments were performed three times.

c

The experiments were performed twice.

d

The experiment was performed once.

e

, p < 0.05, compared with 100 ng/ml HIV-1 Nef protein.

f

, p < 0.05, compared with 100 ng/ml HIV-1 Nef protein with the mAbNef 3E6.

g

, p < 0.05, compared with 10 ng/ml HIV-1 Nef protein.

Additionally, we examined whether astocytoma cells (U251 MG) expressing the HIV-1 Nef protein exhibit chemotactic activity on mononuclear cells in a modified Boyden chamber assembly. A >2-fold increase in PBMC migration was observed when the Nef-expressing U251 MG cell clone astro-4/4.2 was seeded in the lower compartment of the Boyden chamber, compared with the control U251 MG cell clone pSGK1 (23.3 ± 2.0 (n = 9) vs 8.6 ± 1.1 (n = 9) cells/microscopic field in controls; p < 0.05). Coadministration of the mAb to Nef 2H12 at a concentration of 10 μg/ml (but not of 1.0 and 0.1 μg/ml) significantly reduced the chemoattractant activity exerted by the Nef-expressing U251 MG cell clone astro-4/4.2 (14.4 ± 1.0;∗, 21.3 ± 0.4, and 23.3 ± 1.3 cells/microscopic field; n = 3 for each group;∗ p < 0.05)

Our in vitro experiments provided evidence for a direct chemotactic activity of the HIV-1 Nef protein, but did not exclude the existence of (an) additional indirect mechanism(s) involved in Nef-induced leukocyte recruitment in vivo. Therefore, we determined CSF concentrations of soluble host factors that may be involved in HIV-1 Nef protein-induced CSF pleocytosis, namely MCP-1, IL-6, TNF-α, and IFN-γ. Increased levels of these cytokines/chemokines have been detected in both CSF samples and brain sections of HIV-1-infected patients or animal models of AIDS (19, 38, 39, 40). In cell culture experiments, HIV-1 Nef protein was also observed to induce the production of IL-6 and IFN-γ (41, 42). At 6 h, IFN-γ was detectable in the CSF samples of rats injected with HIV-1 Nef protein, but not of rats injected with heat-treated Nef (Table I). At 24 h, IFN-γ concentrations were below the detection limit in both groups investigated (Table II). Neither at 6 h nor at 24 h, was there a significant difference in CSF MCP-1 levels between rats injected with Nef and rats injected with heat-treated Nef (Tables I and II). In addition, at 6 h after i.c. injection, HIV-1 Nef induced a slight, but significant increase in both CSF TNF-α and CSF IL-6 concentrations, as compared with controls (Table I).

Table I.

Effect of HIV-1 Nef protein on CSF IL-6, TNF-α, IFN-γ, and MCP-1 concentrations 6 h after i.c. injection in an experimental rat modela

GroupsCSF IL-6 (U/ml)CSF TNF-α (U/ml)CSF IFN-γ (pg/ml)CSF MCP-1 (ng/ml)
PBS 4576 ± 4048 (6)b ND (6) ND (5) 2.5 ± 0.7 (10) 
100 ng Nef 115,183 ± 60,171 (6) 50.5 ± 12.9 (6)cd 56.1 ± 9.0 (5) 3.1 ± 0.7 (10) 
100 ng Nefheat-treated 8033 ± 3652 (6) ND (6) ND (5) 2.9 ± 0.8 (5) 
GroupsCSF IL-6 (U/ml)CSF TNF-α (U/ml)CSF IFN-γ (pg/ml)CSF MCP-1 (ng/ml)
PBS 4576 ± 4048 (6)b ND (6) ND (5) 2.5 ± 0.7 (10) 
100 ng Nef 115,183 ± 60,171 (6) 50.5 ± 12.9 (6)cd 56.1 ± 9.0 (5) 3.1 ± 0.7 (10) 
100 ng Nefheat-treated 8033 ± 3652 (6) ND (6) ND (5) 2.9 ± 0.8 (5) 
a

Intracisternal application of HIV-1 Nef protein (100 ng) induced a significant increase in CSF concentrations of IL-6, TNF-α, and IFN-γ, but not of MCP-1 at 6 h after i.c. application.

b

No. of rats investigated.

c

CSF TNF-α concentration was below the detection limit (30 μ/ml) in two out of six rats. Data are expressed as means ± SE.

d

, p < 0.05, compared with rats injected i.c. with 100 ng HIV-1 Nef protein.

The experiments described in this report reveal a novel property of HIV-1 Nef protein. We demonstrated that soluble Nef protein of HIV-1 provoked leukocyte recruitment into CNS in a rat model. The absence of bioactivity of heat-treated Nef, the blocking effect of the mAb 2H12 that recognizes the aa sequence 171–190 of the Nef protein (but not of mAb 3E6, an anti-Nef Ab of the same isotype, directed to the aa sequence 168–175, as well as of mAbs against CD4), and the lack of activity of the other investigated HIV-proteins (gp120, gp160, Tat) provided important evidence of the specificity of the observed effects of soluble Nef Ag. Experiments using two different mAbs to Nef and Nef peptides demonstrated that: 1) Nef-induced CSF pleocytosis was blocked almost completely by mAb 2H12, which maps to the carboxy-terminal domain of Nef (amino acid residues 171–190); and 2) that the i.c. application of the corresponding Nef peptide (amino acid residue 171–190) was ineffective. These data suggest that the carboxy-terminal region of Nef is necessary, but not sufficient for Nef-induced leukocyte recruitment into rat CNS, and that Nef is not involved as a linear epitope. Similar to our study, Fujinaga et al. (43) reported that the carboxy terminus and a certain protein conformation are necessary for the ability of extracellular Nef to activate HIV-1 from latent to productive infection in a CD4+ T cell line and in PBMC derived from asymptomatic HIV carriers. In addition, epitope-mapping studies have shown that the carboxy-terminal region of Nef was predominantly exposed at the cellular surface of HIV-1-infected T cell lines and PBMC (44). The carboxy-terminal Nef domain was also observed: 1) to elicit the highest immunogenic response in chimpanzees and in rodents (45), 2) to play an important role in syncytium formation between HIV-1-infected and CD4+-uninfected T cells (44), and 3) to specifically induce cytolysis of CD4+ T cells (46).

Using a modified Boyden chamber technique to test HIV-1 Nef protein for direct chemotactic activity, we found that Nef by itself can induce migration of both polymorphonuclear and mononuclear leukocytes. These data suggest an active role of HIV-1 Nef protein in the recruitment of leukocytes into rat CNS. Besides HIV-1 Nef protein itself, immunomodulatory host factors may be involved in Nef-induced leukocyte migration into the subarachnoid space in our rat model. Experimental and clinical studies have identified a variety of cytokines/chemokines in both CSF samples and brain sections of HIV-1-infected patients or animal models of AIDS that could be involved in leukocyte recruitment to the brain (16, 17, 19, 38, 40). Recently, Sasseville et al. (17) demonstrated elevated immunohistochemical expression of chemokines (MIP-1α and -β, RANTES, MCP-3, and IP-10) in encephalitic brains of SIV-infected rhesus macaques. An increased production of IL-1β, IL- 6, and TNF-α was also observed in a SCID mouse model of HIV-1 encephalitis (40). Moreover, elevated levels of IL-6 were identified in the CSF of 16 out of 38 HIV-1-infected patients (16). Brain extracts from patients with mild HIV encephalitis showed elevated levels of TNF-α in deep gray matter, while, in more severe cases, elevated TNF-α levels were also found within cortical white and cortical gray matter (19). Immunocytochemical analysis of frozen specimens taken at autopsy from the cerebral cortex and white matter of HIV-seropositive individuals frequently showed positive staining of endothelial cells for IL-1β and IFN-γ and less frequently for TNF-α and IL-6 (38). In addition, Sozzani et al. (47) found markedly elevated levels of MCP-1 in CSF but not in blood of HIV-infected patients with CMV encephalitis. CSF MCP-1 concentrations were also found to be significantly higher in patients with than in those without HIV encephalitis (39). Combined, these studies illustrate the complex nature of the host factors that may also be involved in leukocyte trafficking into the brain during HIV-1 infection.

In this study, we examined whether HIV-1 Nef protein, aside from its chemotactic activity, alters rat CSF concentrations of MCP-1, the most potent monocyte chemoattractant in the brain parenchyma (48), as well as of IL-6, TNF-α, and IFN-γ. Whereas Nef injection had no effect on CSF MCP-1 levels, a slight, but significant increase in CSF concentrations of TNF-α, IL-6, and INF-γ could be detected 6 h after intrathecal Nef application. In accordance with our study, Chirmule et al. (41) have reported that HIV-1 Nef was able to induce IL-6 production and secretion in human lymphocytes. Furthermore, soluble HIV-1 Nef protein was found to stimulate IFN-γ production by Th1 cells (49). Since IL-6 was previously shown to play a positive role in local inflammatory reactions by amplifying leukocyte recruitment and function (e.g., oxidative burst) (50), and IFN-γ was recently demonstrated to be involved in the switch from polymorphonuclear to mononuclear cell infiltration (51), these data suggest that the release of soluble host factors in response to Nef may be involved in Nef-induced leukocyte recruitment and/or in the switch from polymorphonuclear to mononuclear cell infiltration as observed in vivo.

Several in vitro studies have provided evidence that HIV-1 Tat protein can act as a chemoattractant on human monocytes and dendritic cells (20, 52), presumably via binding to the β chemokine receptors CCR-2 and CCR-3 (53), as well as to the VEGFR-1/Flt-1 receptor (54). To our surprise, recombinant HIV-1 Tat protein did not exert chemotactic activity in our in vivo rat model. There are several possible explanations for the discrepancy between the above-mentioned studies and our study: 1) the experimental setup used (in vitro chemotaxis assemblies vs in vivo model, where the substances are injected into the CSF, and leukocyte migration through the blood-brain barrier was monitored), and 2) the species investigated (human vs rat). For comparison of the chemotactic potency of different HIV-1 proteins, CSF leukocyte counts were determined at 6 h after injection. At this time point, PMN were the predominant cell type in the CSF of rats injected either with HIV-1 Nef, fMLP, or carrageenan. HIV-1 Tat protein, however, was reported to mimic β-chemokine features, thus serving to recruit mononuclear cells (53). Data on the role of HIV-1 gp120 in leukocyte recruitment are more complex. On the one hand, Weissman et al. (28) have reported that recombinant gp120-mediated signal transduction through CCR5 receptors induces chemotaxis of T lymphocyte. On the other, Wang et al. (55) have observed that monocytes preincubated with gp120 exhibit markedly reduced chemotactic response to either chemokines or fMLP. An inhibitory effect of gp120 was also demonstrated on both fMLP- and IL-1-induced chemotactic responses of human granulocytes, as well as monocytes (56). In accordance with the latter studies, recombinant HIV-1 gp120 did not induce leukocyte recruitment in our study. In previous animal studies, where recombinant HIV-1 gp120 was injected either intracerebrally (57) or intracerebroventricularily (58), or where gp120 was expressed in astrocytes of transgenic mice (59), leukocyte recruitment into the CNS was not examined.

HIV-1 Nef protein is predominantly localized inside infected cells. Unmyristoylated Nef is present mainly in the cytoplasmic cell fraction, whereas myristoylation is necessary for the association of Nef with cytoplasmic membrane structures (60). However, several recent studies suggest that HIV-1 Nef protein may also be present extracellularly. The HIV-1 Nef protein is highly immunogenic (61). Nef-specific Abs and CTL have been detected in HIV-1-seropositive individuals (61, 62). This Ab response may occur as a result of lysis of Nef-expressing HIV-1-infected cells or a specific export of Nef protein from infected cells (63). In addition, in vitro studies have shown that HIV-infected cells can release Nef protein in vesicles (64, 65). Furthermore, HIV-1 Nef protein derived from yeast cells have been found in the extracellular medium during stress (66). Using an Ag capture ELISA, soluble Nef Ag was detected in 21 sera out of 32 of HIV-1-seropositive individuals, but not in 28 healthy volunteers (67). Moreover, several in vitro studies have suggested that extracellular Nef can bind to a cellular surface receptor. For example, Torres and Johnson (42) have reported binding of Nef to MHC class II molecules. Immunohistochemical investigations of brain sections from HIV-seropositive individuals have shown that perivascular cells, consisting primarily of macrophages with some CD4+ and CD8+ T cells and rare B cells, were consistently MHC class II-positive (38). In addition, Okada et al. (68) have described that soluble Nef protein possesses binding affinity to the cell surface of a wide range of human blood cell lines, including CD4+ T, CD8+ T, and B lymphocytes, macrophages, and neutrophils. A 25-kDa protein responsible for Nef binding has been also identified on the surface of various murine cell lines, including T and B lymphocytes and macrophages, and the cell binding site of Nef was found to be located in the C-terminal domain (69). Similarly, Brigino et al. (70) have shown that IL-10 induction by extracellular Nef involves the calcium/calmodulin signal transduction pathway, suggesting that Nef may bind to a surface receptor or to calcium channels. Combined, these studies hint at the existence of both extracellular Nef and binding of Nef to cellular surface receptors.

The CNS, or at least the CSF, is frequently exposed to HIV-1 early in the course of systemic infection (1). Evidence for early virus invasion into the CNS includes detection of early local immune response (including intrathecal Ab production) and more direct identification of the virus and its components in the CSF of HIV-infected patients (71, 72). Similarly, experimental inoculation of SIV-1 into rhesus macaques revealed that infected cells invade brain tissue within 2 wk of systemic virus challenge (4). Recent studies support the idea that viral invasion into the brain is mainly mediated through cell-associated HIV-1 in monocytes and T cells that traffic across the blood-brain barrier (73, 74). The numbers of immunocompetent macrophages and microglia in the brain have been shown to be a more relevant predictor of neurologic disease than the absolute level of virus production (15). However, the exact mechanisms responsible for the recruitment of leukocytes to the CNS remain speculative. The data presented in this report suggest that HIV-Nef could facilitate transendothelial passage and direct leukocyte migration into the CNS. The major cells for HIV-1 replication of the CNS are macrophages and microglia (75). HIV-1 also establishes a persistent infection in astroglial cells (76, 77) that can be accompanied by an expression of Nef protein (10, 11). The expression of Nef in astrocytes may be of special significance in regard to the chemotactic property of Nef because astrocytes are in close contact to the endothelial cells and thus may be able to form a gradient of nef protein in the vessel wall. In addition, experimental studies have provided evidence for an extravascular localization HIV-1 Nef protein and for the ability of Nef to bind to cell surface receptors (42, 61, 64, 65, 67), two further arguments for a possible involvement of HIV-1 Nef protein in leukocyte recruitment into brain in HIV-infected patients. Thus, our data propose a mechanism by which HIV-1 Nef protein may be essential for the maintenance of a cellular reservoir for HIV replication in the CNS and for AIDS neuropathogenesis by acting as a mediator of the recruitment of leukocytes.

We thank Ms. A. Heng, Dr. Robert Paul, and Stefan Lorenzl from the Department of Neurology, Ms. E. Götz from the Institute of Molecular Virology for technical support, and Dr. M. Wick from the Department of Clinical Chemistry for performing the cerebrospinal fluid leukocyte differential counts. We thank Dr. C. Eckerskorn from the Department of Proteinanalytics (Max-Plank-Institute, Martinsried) for evaluating Nef determination by silver staining and SDS-PAGE. We thank Prof. K. Krohn (Tampere, Finland) for support with mAbs, Prof. B. Wahren (Stockholm, Sweden) for support with synthetic Nef peptides, and Prof. L. Gürtler (Max von Pettenkofer-Institute, Munich) for clinical cooperation. We thank Mrs D. Förth and Ms J. Benson for manuscript preparation.

1

This work was supported by grants from the Deutsche Forschungsgemeinschaft (Sonderforschungsbereich 464, project B3 (to H.-W.P.), Sonderforschungsbereich 464, project B2 (to V.E.)) and the Swiss National Science Foundation (Grant 31-42900.95 (to A.F.)).

3

Abbreviations used in this paper: PMN, polymorphonuclear leukocyte; MABP, mean arterial blood pressure; CSF, cerebrospinal fluid; WBC, white blood cell; i.c., intracisternal; MCP, monocyte chemoattractant protein; ICP, intracranial pressure; CBFLDF, cerebral blood flow measured by laser-Doppler flowmetry; MPO, myeloperoxidase.

1
Michaels, J., L. R. Sharer, L. G. Epstein.
1988
. Human immunodeficiency virus type 1 (HIV-1) infection of the nervous system: a review.
Immunodefic. Rev.
1
:
71
2
Wiley, C. A., M. Baldwin, C. L. Achim.
1996
. Expression of HIV regulatory and structural mRNA in the central nervous system.
AIDS
10
:
843
3
Gray, F., F. Scaravilli, I. Everall, F. Chretien, S. An, D. Boche, H. Adle-Biassette, L. Wingertsmann, M. Durigon, B. Hurtrel, et al
1996
. Neuropathology of early HIV-1 infection.
Brain Pathol.
6
:
1
4
Sharer, L. R., Y. Saito, A. da-Cunha, P. C. Ung, H. A. Gelbard, L. G. Epstein, B. M. Blumberg.
1996
. In situ amplification and detection of HIV-1 DNA in fixed pediatric AIDS brain tissue.
Hum. Pathol.
27
:
614
5
Brack-Werner, R., A. Kleinschmidt, A. Ludvigsen, W. Mellert, M. Neumann, R. Herrmann, M. C. Khim, A. Burny, N. Muller-Lantzsch, D. Stavrou, et al
1992
. Infection of human brain cells by HIV-1: restricted virus production in chronically infected human glial cell lines.
AIDS
6
:
273
6
Cheng-Mayer, C., J. T. Rutka, M. L. Rosenblum, T. McHugh, D. P. Stites, J. A. Levy.
1987
. Human immunodeficiency virus can productively infect cultured human glial cells.
Proc. Natl. Acad. Sci. USA
84
:
3526
7
Dewhurst, S., K. Sakai, J. Bresser, M. Stevenson, M. J. Evinger-Hodges, D. J. Volsky.
1987
. Persistent productive infection of human glial cells by human immunodeficiency virus (HIV) and by infectious molecular clones of HIV.
J. Virol.
61
:
3774
8
Neumann, M., B. K. Felber, A. Kleinschmidt, B. Froese, V. Erfle, G. N. Pavlakis, R. Brack-Werner.
1995
. Restriction of human immunodeficiency virus type 1 production in a human astrocytoma cell line is associated with a cellular block in Rev function.
J. Virol.
69
:
2159
9
Tornatore, C., K. Meyers, W. Atwood, K. Conant, E. Major.
1994
. Temporal patterns of human immunodeficiency virus type 1 transcripts in human fetal astrocytes.
J. Virol.
68
:
93
10
Tornatore, C., R. Chandra, J. R. Berger, E. O. Major.
1994
. HIV-1 infection of subcortical astrocytes in the pediatric central nervous system.
Neurology
44
:
481
11
Saito, Y., L. R. Sharer, L. G. Epstein, J. Michaels, M. Mintz, M. Louder, K. Golding, T. A. Cvetkovich, B. M. Blumberg.
1994
. Overexpression of nef as a marker for restricted HIV-1 infection of astrocytes in postmortem pediatric central nervous tissues.
Neurology
44
:
474
12
Werner, T., S. Ferroni, T. Saermark, R. Brack-Werner, R. B. Banati, R. Mager, L. Steinaa, G. W. Kreutzberg, V. Erfle.
1991
. HIV-1 Nef protein exhibits structural and functional similarity to scorpion peptides interacting with K+ channels.
AIDS
5
:
1301
13
Klotman, M. E., S. Kim, A. Buchbinder, A. DeRossi, D. Baltimore, F. Wong-Staal.
1991
. Kinetics of expression of multiply spliced RNA in early human immunodeficiency virus type 1 infection of lymphocytes and monocytes.
Proc. Natl. Acad. Sci. USA
88
:
5011
14
Kestler, H. W., III, D. J. Ringler, K. Mori, D. L. Panicali, P. K. Sehgal, M. D. Daniel, R. C. Desrosiers.
1991
. Importance of the nef gene for maintenance of high virus loads and for development of AIDS.
Cell
65
:
651
15
Glass, J. D., H. Fedor, S. L. Wesselingh, J. C. McArthur.
1995
. Immunocytochemical quantitation of human immunodeficiency virus in the brain: correlations with dementia.
Ann. Neurol.
38
:
755
16
Gallo, P., K. Frei, C. Rordorf, J. Lazdins, B. Tavolato, A. Fontana.
1989
. Human immunodeficiency virus type 1 (HIV-1) infection of the central nervous system: an evaluation of cytokines in cerebrospinal fluid.
J. Neuroimmunol.
23
:
109
17
Sasseville, V. G., M. M. Smith, C. R. Mackay, D. R. Pauley, K. G. Mansfield, D. J. Ringler, A. A. Lackner.
1996
. Chemokine expression in simian immunodeficiency virus-induced AIDS encephalitis.
Am. J. Pathol.
149
:
1459
18
Schmidtmayerova, H., H. S. Nottet, G. Nuovo, T. Raabe, C. R. Flanagan, L. Dubrovsky, H. E. Gendelman, A. Cerami, M. Bukrinsky, B. Sherry.
1996
. Human immunodeficiency virus type 1 infection alters chemokine β peptide expression in human monocytes: implications for recruitment of leukocytes into brain and lymph nodes.
Proc. Natl. Acad. Sci. USA
93
:
700
19
Achim, C. L., M. P. Heyes, C. A. Wiley.
1993
. Quantitation of human immunodeficiency virus, immune activation factors, and quinolinic acid in AIDS brains.
J. Clin. Invest.
91
:
2769
20
Lafrenie, R. M., L. M. Wahl, J. S. Epstein, I. K. Hewlett, K. M. Yamada, S. Dhawan.
1996
. HIV-1-Tat protein promotes chemotaxis and invasive behavior by monocytes.
J. Immunol.
157
:
974
21
Pfister, H. W., K. Frei, B. Ottnad, U. Koedel, A. Tomasz, A. Fontana.
1992
. Transforming growth factor β2 inhibits cerebrovascular changes and brain edema formation in the tumor necrosis factor α-independent early phase of experimental pneumococcal meningitis.
J. Exp. Med.
176
:
265
22
Koedel, U., A. Bernatowicz, R. Paul, K. Frei, A. Fontana, H.-W. Pfister.
1995
. Experimental pneumococcal meningitis: cerebrovascular alterations, brain edema, and meningeal inflammation are linked to the production of nitric oxide.
Ann. Neurol.
37
:
313
23
Koedel, U., A. Bernatowicz, K. Frei, A. Fontana, H.-W. Pfister.
1996
. Systemically (but not intrathecally) administered IL-10 attenuates pathophysiologic alterations in experimental pneumococcal meningitis.
J. Immunol.
157
:
5185
24
Koedel, U., H.-W. Pfister.
1997
. Protective effect of the antioxidant N-acetyl-L-cysteine in pneumococcal meningitis in the rat.
Neurosci. Lett.
225
:
33
25
Frei, K., D. Nadal, A. Fontana.
1990
. Intracerebral synthesis of tumor necrosis factor-α and interleukin-6 in infectious meningitis.
Ann. NY Acad. Sci.
594
:
326
26
Frei, K., U. V. Malipiero, T. P. Leist, R. M. Zinkernagel, M. E. Schwab, A. Fontana.
1989
. On the cellular source and function of interleukin-6 produced in the central nervous system in viral diseases.
Eur. J. Immunol.
19
:
689
27
Cuzzocrea, S., A. Tailor, B. Zingarelli, A. L. Salzman, R. J. Flower, C. Szabo, M. Perretti.
1997
. Lipocortin 1 protects against splanchnic artery occlusion and reperfusion injury by affecting neutrophil migration.
J. Immunol.
159
:
5089
28
Weissman, D., R. L. Rabin, J. Arthos, A. Rubbert, M. Dybul, R. Swofford, S. Venkatesan, J. M. Farber, A. S. Fauci.
1997
. Macrophage-tropic HIV and SIV envelope proteins induce a signal through the CCR5 chemokine receptor.
Nature
389
:
981
29
Täuber, M. G., U. Borschberg, M. A. Sande.
1988
. Influence of granulocytes on brain edema, intracranial pressure, and cerebrospinal fluid concentrations of lactate and protein in experimental meningitis.
J. Infect. Dis.
157
:
456
30
Gamache, D. A., J. T. Povlishock, E. F. Ellis.
1986
. Carrageenan-induced brain inflammation: characterization of the model.
J. Neurosurg.
65
:
679
31
Adams, D. H., A. R. Lloyd.
1997
. Chemokines: leukocyte recruitment and activation cytokines.
Lancet
349
:
490
32
Kohleisen, B., M. Neumann, R. Herrmann, R. Brack-Werner, K. J. Krohn, V. Ovod, A. Ranki, V. Erfle.
1992
. Cellular localization of Nef expressed in persistently HIV-1- infected low-producer astrocytes.
AIDS
6
:
1427
33
Kohleisen, B., G. Sutter, B. Koettgen, E. Shumay, and V. Erfle. 1996. Stable expression of HIV-1 Nef in a human astrocytoma cell line. In ECEAR First European Conference on Experimental AIDS Research, Cannes.
34
Seebach, J., D. Bartholdi, K. Frei, K. S. Spanaus, E. Ferrero, U. Widmer, S. Isenmann, R. M. Strieter, M. Schwab, H. Pfister, et al
1995
. Experimental Listeriameningoencephalitis: macrophage inflammatory protein-1α and -2 are produced intrathecally and mediate chemotactic activity in cerebrospinal fluid of infected mice.
J. Immunol.
155
:
4367
35
Ovod, V., A. Lagerstedt, A. Ranki, F. O. Gombert, R. Spohn, M. Tahtinen, G. Jung, K. J. Krohn.
1992
. Immunological variation and immunohistochemical localization of HIV-1 Nef demonstrated with monoclonal antibodies.
AIDS
6
:
25
36
Guy, B., M. P. Kieny, Y. Riviere, C. Le-Peuch, K. Dott, M. Girard, L. Montagnier, J. P. Lecocq.
1987
. HIV F/3′ orf encodes a phosphorylated GTP-binding protein resembling an oncogene product.
Nature
330
:
266
37
Kohleisen, B., K. Gaedigk-Nitschko, M. Ohlmann, E. Götz, H. Ostolaza, F. M. Goni, V. Erfle.
1996
. Heparin-binding capacity of the HIV-1 NEF-protein allows one- step purification and biochemical characterization.
J. Virol. Methods
60
:
89
38
Tyor, W. R., J. D. Glass, J. W. Griffin, P. S. Becker, J. C. McArthur, L. Bezman, D. E. Griffin.
1992
. Cytokine expression in the brain during the acquired immunodeficiency syndrome.
Ann. Neurol.
31
:
349
39
Cinque, P., L. Vago, M. Mengozzi, V. Torri, D. Ceresa, E. Vicenzi, P. Transidico, A. Vagani, S. Sozzani, A. Mantovani, A. Lazzarin, G. Poli.
1998
. Elevated cerebrospinal fluid levels of monocyte chemotactic protein-1 correlate with HIV-1 encephalitis and local viral replication.
AIDS
12
:
1327
40
Persidsky, Y., J. Limoges, R. McComb, P. Bock, T. Baldwin, W. Tyor, A. Patil, H. S. Nottet, L. Epstein, H. Gelbard, et al
1996
. Human immunodeficiency virus encephalitis in SCID mice [see comments].
Am. J. Pathol.
149
:
1027
41
Chirmule, N., N. Oyaizu, C. Saxinger, S. Pahwa.
1994
. Nef protein of HIV-1 has B-cell stimulatory activity.
AIDS
8
:
733
42
Torres, B. A., H. M. Johnson.
1994
. Identification of an HIV-1 Nef peptide that binds to HLA class II antigens.
Biochem Biophys Res Commun.
200
:
1059
43
Fujinaga, K., Q. Zhong, T. Nakaya, M. Kameoka, T. Meguro, K. Yamada, K. Ikuta.
1995
. Extracellular Nef protein regulates productive HIV-1 infection from latency.
J. Immunol.
155
:
5289
44
Otake, K., Y. Fujii, T. Nakaya, Y. Nishino, Q. Zhong, K. Fujinaga, M. Kameoka, K. Ohki, K. Ikuta.
1994
. The carboxyl-terminal region of HIV-1 Nef protein is a cell surface domain that can interact with CD4+ T cells.
J. Immunol.
153
:
5826
45
Estaquier, J., C. Boutillon, J. C. Ameisen, M. H. Gras, A. Delanoye, J. P. Lecocq, A. Dixson, A. Tartar, A. Capron, C. Auriault.
1992
. Determination of B-cell epitopes of nef HIV-I protein: immunogenicity related to their structure.
Mol. Immunol.
29
:
1337
46
Fujii, Y., K. Otake, M. Tashiro, A. Adachi.
1996
. Human immunodeficiency virus type 1 Nef protein on the cell surface is cytocidal for human CD4+ T cells.
FEBS Lett.
393
:
105
47
Sozzani, S., M. Introna, S. Bernasconi, N. Polentarutti, P. Cinque, G. Poli, A. Sica, A. Mantovani.
1997
. MCP-1 and CCR2 in HIV infection: regulation of agonist and receptor expression.
J. Leukocyte Biol.
62
:
30
48
Bell, M. D., D. D. Taub, V. H. Perry.
1996
. Overriding the brain’s intrinsic resistance to leukocyte recruitment with intraparenchymal injections of recombinant chemokines.
Neuroscience
74
:
283
49
Torres, B. A., T. Tanabe, H. M. Johnson.
1999
. Characterization of Nef-induced CD4 T cell proliferation.
Biochem. Biophys. Res. Commun.
225
:
54
50
Romano, M., M. Sironi, C. Toniatti, N. Polentarutti, P. Fruscella, P. Ghezzi, R. Faggioni, W. Luini, V. van-Hinsbergh, S. Sozzani, et al
1997
. Role of IL-6 and its soluble receptor in induction of chemokines and leukocyte recruitment.
Immunity
6
:
315
51
Musso, T., R. Badolato, D. L. Longo, G. L. Gusella, L. Varesio.
1995
. Leukemia inhibitory factor induces interleukin-8 and monocyte chemotactic and activating factor in human monocytes: differential regulation by interferon-γ.
Blood
86
:
1961
52
Benelli, R., R. Mortarini, A. Anichini, D. Giunciuglio, D. M. Noonan, S. Montalti, C. Tacchetti, A. Albini.
1998
. Monocyte-derived dendritic cells and monocytes migrate to HIV-Tat RGD and basic peptides.
AIDS
12
:
261
53
Albini, A., S. Ferrini, R. Benelli, S. Sforzini, D. Giunciuglio, M. G. Aluigi, A. E. Proudfoot, S. Alouani, T. N. Wells, G. Mariani, et al
1998
. HIV-1 Tat protein mimicry of chemokines.
Proc. Natl. Acad. Sci. USA
95
:
13153
54
Mitola, S., S. Sozzani, W. Luini, L. Primo, A. Borsatti, H. Weich, F. Bussolino.
1997
. Tat-human immunodeficiency virus-1 induces human monocyte chemotaxis by activation of vascular endothelial growth factor receptor-1.
Blood
90
:
1365
55
Wang, J. M., H. Ueda, O. M. Howard, M. C. Grimm, O. Chertov, X. Gong, W. Gong, J. H. Resau, C. C. Broder, G. Evans, et al
1998
. HIV-1 envelope gp120 inhibits the monocyte response to chemokines through CD4 signal-dependent chemokine receptor down-regulation.
J. Immunol.
161
:
4309
56
Stefano, G. B., E. M. Smith, P. Cadet, J. Hughes.
1993
. HIV gp120 alteration of DAMA and IL-1 α induced chemotaxic responses in human and invertebrate immunocytes.
J. Neuroimmunol.
43
:
177
57
Glowa, J. R., L. V. Panlilio, D. E. Brenneman, I. Gozes, M. Fridkin, J. M. Hill.
1992
. Learning impairment following intracerebral administration of the HIV envelope protein gp120 or a VIP antagonist.
Brain Res.
570
:
49
58
Kimes, A. S., E. D. London, G. Szabo, L. Raymon, B. Tabakoff.
1991
. Reduction of cerebral glucose utilization by the HIV envelope glycoprotein gp-120.
Exp. Neurol.
112
:
224
59
Toggas, S. M., E. Masliah, E. M. Rockenstein, G. F. Rall, C. R. Abraham, L. Mucke.
1994
. Central nervous system damage produced by expression of the HIV-1 coat protein gp120 in transgenic mice.
Nature
367
:
188
60
Harris, M..
1995
. The role of myristoylation in the interactions between human immunodeficiency virus type I Nef and cellular proteins.
Biochem. Soc. Trans.
23
:
557
61
Bahraoui, E., M. Yagello, J. N. Billaud, J. M. Sabatier, B. Guy, E. Muchmore, M. Girard, J. C. Gluckman.
1990
. Immunogenicity of the human immunodeficiency virus (HIV) recombinant nef gene product. Mapping of T-cell and B-cell epitopes in immunized chimpanzees.
AIDS Res. Hum. Retroviruses
6
:
1087
62
Ameisen, J. C., B. Guy, S. Chamaret, M. Loche, Y. Mouton, J. L. Neyrinck, J. Khalife, C. Leprevost, G. Beaucaire, C. Boutillon, et al
1989
. Antibodies to the nef protein and to nef peptides in HIV-1-infected seronegative individuals.
AIDS Res. Hum. Retroviruses
5
:
279
63
Ratner, L., T. M. Niederman.
1995
. Nef.
Curr. Top. Microbiol. Immunol.
193
:
169
64
Guy, B., Y. Riviere, K. Dott, A. Regnault, M. P. Kieny.
1990
. Mutational analysis of the HIV nef protein.
Virology
176
:
413
65
Kienzle, N., M. Bachmann, W. E. Muller, L. N. Muller.
1992
. Expression and cellular localization of the Nef protein from human immunodeficiency virus-1 in stably transfected B-cells.
Arch. Virol.
124
:
123
66
Macreadie, I. G., L. A. Castelli, A. Lucantoni, A. A. Azad.
1995
. Stress- and sequence-dependent release into the culture medium of HIV-1 Nef produced in Saccharomyces cerevisiae.
Gene
162
:
239
67
Fujii, Y., K. Otake, M. Tashiro, A. Adachi.
1996
. Soluble Nef antigen of HIV-1 is cytotoxic for human CD4+ T cells.
FEBS Lett.
393
:
93
68
Okada, H., R. Takei, M. Tashiro.
1997
. HIV-1 Nef protein-induced apoptotic cytolysis of a broad spectrum of uninfected human blood cells independently of CD95(Fas).
FEBS Lett.
414
:
603
69
Okada, H., S. Morikawa, M. Tashiro.
1998
. HIV-1 Nef binding protein expressed on the surface of murine blood cells.
Med. Microbiol. Immunol.
186
:
201
70
Brigino, E., S. Haraguchi, A. Koutsonikolis, G. J. Cianciolo, U. Owens, R. A. Good, N. K. Day.
1997
. Interleukin 10 is induced by recombinant HIV-1 Nef protein involving the calcium/calmodulin-dependent phosphodiesterase signal transduction pathway.
Proc. Natl. Acad. Sci. USA
94
:
3178
71
Elovaara, I., I. Seppala, E. Poutiainen, J. Suni, S. L. Valle.
1988
. Intrathecal humoral immunologic response in neurologically symptomatic and asymptomatic patients with human immunodeficiency virus infection.
Neurology
38
:
1451
72
McArthur, J. C., E. Sipos, D. R. Cornblath, D. Welch, M. Chupp, D. E. Griffin, R. T. Johnson.
1989
. Identification of mononuclear cells in CSF of patients with HIV infection.
Neurology
39
:
66
73
Epstein, L. G., H. E. Gendelman.
1993
. Human immunodeficiency virus type 1 infection of the nervous system: pathogenetic mechanisms.
Ann. Neurol.
33
:
429
74
Nottet, H. S., Y. Persidsky, V. G. Sasseville, A. N. Nukuna, P. Bock, Q. H. Zhai, L. R. Sharer, R. D. McComb, S. Swindells, C. Soderland, H. E. Gendelman.
1996
. Mechanisms for the transendothelial migration of HIV-1-infected monocytes into brain.
J. Immunol.
156
:
1284
75
Takahashi, K., S. L. Wesselingh, D. E. Griffin, J. C. McArthur, R. T. Johnson, J. D. Glass.
1996
. Localization of HIV-1 in human brain using polymerase chain reaction in situ hybridization and immunocytochemistry.
Ann. Neurol.
39
:
705
76
Ward, J. M., T. J. O’Leary, G. B. Baskin, R. Benveniste, C. A. Harris, P. L. Nara, R. H. Rhodes.
1987
. Immunohistochemical localization of human and simian immunodeficiency viral antigens in fixed tissue sections.
Am. J. Pathol.
127
:
199
77
Tornatore, C., A. Nath, K. Amemiya, E. O. Major.
1991
. Persistent human immunodeficiency virus type 1 infection in human fetal glial cells reactivated by T-cell factor(s) or by the cytokines tumor necrosis factor α and interleukin-1 β.
J. Virol.
65
:
6094