Human immunodeficiency virus type-1 Tat protein induces nuclear factor (NF)-κB activation and oxidative stress in microglial cultures by independent mechanisms


Address correspondence and reprint requests to Dr Luisa Minghetti, Laboratory of Pathophysiology, Istituto Superiore di Sanità, Viale Regina Elena 299, 00161 Rome, Italy. E-mail:


We have extended our previous findings and shown that human immunodeficiency virus Tat protein, in addition to nitric oxide (NO), stimulated rat microglial cultures to release pro-inflammatory cytokine interleukin-1β and tumour necrosis factor-α in a nuclear factor (NF)-κB-dependent manner. At the same time, Tat stimulated the accumulation of free radicals, as indicated by the increased levels of isoprostane 8-epi-prostaglandin F (8-epi-PGF), a reliable marker of lipid peroxidation and oxidative stress, by a mechanism unrelated to NF-κB activation. The presence of free radical scavengers abrogated Tat-induced 8-epi-PGF accumulation without affecting NO and cytokine production. Consistently, Tat-induced IκBα degradation – an index of NF-κB activation – was not affected by free radical scavengers, but was prevented by an NF-κB-specific inhibitor. Our observations indicate that NF-κB plays a key role in Tat-dependent microglial activation, and that oxidative stress and NF-κB activation induced by Tat occur by independent mechanisms.

Abbreviations used

butylated hydroxytoluene


8-epi-prostaglandin F


HIV-associated dementia


human immunodeficiency virus type-1


inducible nitric oxide synthase


interleukin 1β


manganese (III) tetrakis 4-benzoic acid-porphyrin


nitric oxide


N-tosyl-l-phenylalanine chloromethyl ketone


nuclear factor-κB


tumour necrosis factor-α.

Human immunodeficiency virus type-1 (HIV-1) infection is known to affect the CNS causing severe cognitive and motor impairment in a significant proportion of patients (Kolson and Gonzalez-Scarano 2000). The histopathological signs of this neurological disorder, known as HIV-associated dementia (HAD), include pallor of myelin sheaths, gliosis, abnormalities of dendritic processes, neuronal apoptotic death and infiltration of inflammatory cells. As neurones are not infected, the neuronal abnormalities in HAD are likely to be the indirect result of infection of macrophage/microglial cells, the main cellular type productively infected by the virus in the brain. Increasing evidence suggests that diffusible viral and/or cellular gene products released from infected microglia may directly damage neurones or alter the integrity of glial functions required for neuronal survival. Among the putative viral toxins are the HIV-1 proteins gp120, gp41 and Tat (Nath and Geiger 1998).

The HIV-1 regulatory protein Tat is a potent transactivator of viral and cellular gene expression, produced in the early phase of infection and actively secreted into the extracellular environment, from where it can act in an autocrine or a paracrine manner (Chang et al. 1995). Elevated levels of Tat-mRNA have been detected in the brain of HAD patients (Nath and Geiger 1998), and in vitro and in vivo evidence indicates that Tat causes neuronal death, either directly (by mechanisms involving activation of glutamate receptors and calcium influx) or indirectly (possibly by stimulating the production of pro-inflammatory and/or neurotoxic factors from astrocytes and microglia) (Jones et al. 1998 and references therein; Nath et al. 1999).

We have previously shown that Tat stimulates inducible nitric oxide synthase (iNOS) expression and NO production in rat microglia cultures by activating the nuclear factor (NF)-κB (Polazzi et al. 1999). In the present study we investigated whether Tat could stimulate other functions typically associated with microglial activation, such as the production of free radicals and the pro-inflammatory cytokines interleukin-1β (IL-1β) and tumour necrosis factor-α (TNF-α). As a marker of free radical formation we measured the levels of isoprostane 8-epi-prostaglandin F (8-epi-PGF), one of the major compounds of the isoprostane-F family (Lawson et al. 1999). Isoprostanes are very stable molecules formed in situ in cell membranes by free radical catalysed peroxidation of arachidonic acid. Once synthesized, they are hydrolysed and released from membrane phospholipids and become easily detectable in biological fluids and cell culture supernatants.

We found that Tat induced IL-1β, TNF-α and NO synthesis by mechanisms strictly dependent on NF-κB activation. At the same time, Tat caused an accumulation of free radicals, which were, however, not necessary for NF-κB activation.

Materials and methods


Recombinant Tat HIV-1 IIIB was purchased from NIBSC. Polyclonal anti-IκBα and anti-p65 were from Santa Cruz Biotechnology Inc. Specific ELISA for rat IL-1β and TNF-α were obtained from Endogen Inc., and 8-epi-PGF EIA was purchased from Cayman Chemical. Western blot enhanced chemiluminiscence (ECL) detection system and [3H]PGE2 (specific activity 164 Ci/mmol) were from Amersham International. Bicinchoninic acid (BCA) protein assay was from Pierce and manganese (III) tetrakis 4-benzoic acid-porphyrin (MnTBAP) was from Alexis Italia. All other chemicals, including specific antibody for prostaglandin E2, butylated hydroxytoluene (BHT) and N-tosyl-l-phenylalanine chloromethyl ketone (TPCK) were from Sigma-Aldrich.

Cell cultures

Microglial secondary cultures were prepared from 10- to 14-day mixed primary glial cultures obtained from the cerebral cortex of 1-day-old-rats (Minghetti et al. 1996). Animals were treated according to the directive of the Council of the European Communities N.86/609/CEE. Microglia was harvested by mild shaking from the mixed primary glial cultures, resuspended in 10% fetal calf serum containing basal Eagle's medium and plated at 1.25 × 105 cells/cm2 density. Cells were allowed to adhere for 20 min and then washed to remove non-adhering cells. After a 24-h incubation, the medium was replaced with fresh medium containing the substance(s) under study. For TPCK or radical scavenger treatment, cell cultures were incubated with the appropriate agents for 1.5 h before adding Tat. Cultures were > 98% positive for the macrophage marker ED1, and cell viability was > 95%, by Tripan Blue exclusion.

Determinations of microglial products

The supernatants of cells incubated for different intervals with or without 1 µg/mL Tat and/or other agents were collected, centrifuged and stored at − 80°C until tested.

The cytokine levels were assayed by specific ELISA for rat IL-1β and TNF-α (10–1000 pg/mL and 30–2500 pg/mL range of determination, respectively). The 8-epi-PGF level was measured by a specific EIA (detection limit: 2 pg/mL, see Greco et al. 1999). Nitric oxide was detected by measuring the level of nitrite, one of the end-products of NO oxidation, by Griess reaction (detection limit, 0.25 µm) and PGE2 by a specific radio-immunoassay (detection limit, 25 pg/mL; see Minghetti et al. 1996).

Western blot analysis

Cell lysates were prepared and protein concentrations measured as previously described (Minghetti et al. 1996). Equal quantities of proteins (25 µg) were separated by sodium dodecyl sulphate–polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes, which were then blocked with 10% non-fat milk, and incubated with anti-IκBα antibody (1 : 300 dilution) for 1 h at 25°C. After several washes the membranes were incubated with secondary antibody (anti-rabbit IgG conjugated with horseradish peroxidase, 1 : 5000) for 1 h at 25°C and visualized by the Amersham ECL system. The optical density of the bands (arbitrary units) was measured by a GS-700 Imaging Densitometer (Bio-Rad) and referred to the corresponding control samples (taken as 100%), run in the same gel.

Cell staining and immunofluorescence

Purified microglial cells grown for 24 h on coverslips were exposed to 1 µg/mL Tat for 30 and 60 min. Cells were fixed with cold methanol for 5 min at − 20°C. Cells were stained with the nuclear fluorochrome Hoechst 33258 for 20 min (to a final concentration of 5 µg/mL) and incubated for 20 min with 3% normal goat serum. After washing, single immunofluorescence staining was performed using polyclonal antiserum NF-κB p65 protein (1 : 200, overnight at 4°C) and fluorescein-conjugated goat IgG anti-rabbit IgG (1 : 50).

Statistical analysis

Data are expressed as mean ± SEM of (n) independent experiments, run in duplicate. Comparison of data between treatments was made by Student's t-test for paired samples. A two-tailed probability of less than 5% (p < 0.05) was taken as statistically significant.

Results and discussion

To study whether Tat elicited functions associated with microglial activation, we stimulated rat neonatal microglial cultures with 1 µg/mL of Tat for 24 h – conditions previously shown to evoke optimal NO synthesis (Polazzi et al. 1999).

Besides NO formation, measured as nitrite accumulation (Fig. 1a; Polazzi et al. 1999), Tat elicited IL-1β and TNF-α synthesis (Figs 1b and c). The pretreatment with the NF-κB inhibitor TPCK prevented the synthesis of all three products (Figs 1a–c), suggesting that their induction was strictly dependent on NF-κB activation. At the same time, Tat stimulated a 3.1 ± 0.5-fold increase of 8-epi-PGF basal level (p < 0.0025, n = 8), indicating an enhanced free radical production and lipid peroxidation. As previously discussed (Greco et al. 1999), in vivo formation of 8-epi-PGF appears to be exclusively caused by a non-enzymatic free radical catalysed process. However, in vitro evidence suggests that this compound could in part derive from cyclooxygenase, the enzyme responsible for prostaglandin synthesis. In our system 8-epi-PGF formation was very unlikely to be attributable to cyclooxygenase activity, as in five independent experiments (not shown) Tat failed to stimulate a significant production of prostaglandin E2, one of the major prostaglandins released by activated microglia (Minghetti et al. 1996). At variance with NO and the two cytokines, 8-epi-PGF accumulation was not affected by the presence of TPCK, suggesting that Tat triggered free radical formation by mechanism(s) not dependent on NF-κB activation (Fig. 1d).

Figure 1.

 Cytokine and free radical production in microglial cultured treated with Tat and the NF-κB inhibitor TPKC. Cells were treated for 24 h with or without 1 µg/mL Tat. Where indicated, the inhibitor TPCK was added to the cultures 1.5 h before Tat. Supernatants were collected and analysed for their content of nitrite (a), IL-1β (b), TNF-α (c) and 8-epi-PGF (d) (n = 4). *p < 0.025, **p < 0.01 and ***p < 0.0025 compared with control cultures.

In agreement with our previous data on NO, lower concentrations of Tat (100 ng/mL) did not significantly stimulate the synthesis of IL-1β, TNF-α and 8-epi-PGF (not shown). The effect of 1 µg/mL Tat was specific, as it was abolished by heat-inactivation (not shown), and largely prevented (67 ± 6% of inhibition of nitrite accumulation) by a specific anti-Tat antibody, as previously reported (Polazzi et al. 1999).

As NF-κB is sensitive to the cellular redox state (Christman et al. 2000), we investigated whether Tat-dependent NF-κB activation was related to free radical formation. If this were the case, prevention of Tat-induced free radical formation should result in a decreased activation of NF-κB. We found that pretreatment of microglial cultures with the antioxidant BHT (1–10 µm) inhibited in a dose-dependent way the accumulation of 8-epi-PGF stimulated by Tat (Fig. 2). Conversely, 10 µm BHT did not affect the production of IL-1β and had only a minor, though significant, inhibitory activity on NO and TNF-α synthesis (19 ± 8% and 15 ± 2% of inhibition, n = 4 and n = 3, respectively; Table 1). Similar results were obtained by pretreating microglial cultures with the specific oxygen radical scavenger MnTBAP, which at 1 µm concentration inhibited Tat-induced 8-epi-PGF levels (50 ± 22% inhibition, n = 3) without affecting NO synthesis (3.0 ± 0.4 µm and 2.9 ± 0.6 µm in the presence of Tat alone or with 1 µm MnTBAP, respectively, n = 3). It was not possible to obtain the complete abrogation of 8-epi-PGF formation by MnTBAP, as higher doses exhibited some toxic effects.

Figure 2.

 Effects of antioxidant BHT on Tat-elicited 8-epi-PGF accumulation. Microglial cells were incubated with or without 1 or 10 µm BHT for 1.5 h before the addition of 1 µg/mL Tat. Supernatants were collected after 24 h and analysed for 8-epi-PGF accumulation (n = 4). p < 0.025 compared with Tat-treated cultures.

In resting cells NF-κB is retained in the cytoplasm in a latent form by binding to a family of inhibitory molecules (IκBs), of which IκBα is the most intensively studied (Christman et al. 2000). The rapid phosphorylation and degradation of IκBα allows the active complex p65/p50 of NF-κB to be released, translocate to the nucleus and transactivate target genes. The ability of Tat to stimulate NF-κB activation was firstly assessed by monitoring the nuclear translocation of p65 by immunocytochemistry. Unstimulated control cultures showed a disperse cytoplasmatic staining (Fig. 3a), whereas after 30 min (not shown) and 60 min (Fig. 3c) of exposure to 1 µg/mL Tat a significant proportion of cells exhibited intense staining localized to the nucleus. To prove directly the lack of correlation between oxidative stress and NF-κB activation in Tat-treated microglial cells, we analysed by western blot the expression of IκBα. As shown in Fig. 4, after 60 min of stimulation, 1 µg/mL Tat induced a significant reduction of IκBα level (40 ± 7% reduction of IκBα expression in control cultures). Such a reduction was partially reversed by the presence of 10 µm TPCK (16 ± 4% of reduction compared with control), but remained unaffected by 10 µm BHT (39 ± 3% of reduction). The basal IκBα level was not modified by TPCK or BHT alone (not shown).

Figure 3.

 Effect of Tat on nuclear traslocation of NF-κB-p65 protein in microglia. Immunofluorescence labelling of microglia with polyclonal anti-NF-κB-p65 (a and c) and nuclear staining Hoechst 33258 (b and d) of untreated (a and b) and Tat treated cultures (c and d). Arrows indicate cells presenting more intense nuclear labelling.

Figure 4.

 Effects of antioxidant and NF-κB inhibitor agents on IκBα level in Tat-activated microglia. Cells were incubated with or without 10 µm BHT or TPCK for 1.5 h before the addition of 1 µg/mL Tat. Cell lysates were prepared and proteins analysed as described in the Materials and methods section. (a) One representative experiment of n = 5 is shown. (b) The intensity of the bands was measured by densitometry and the value (arbitrary units) for IκBα levels in control cultures taken as 100% (n = 5). *p < 0.025 compared with control; **p < 0.025 compared with control and Tat-treated cultures; ***p < 0.005 compared with control.

In conclusion, our observations indicate that NO, IL-1β and TNF-α synthesis are induced by Tat through activation of NF-κB, adding further evidence to the key role of this transcription factor in microglial activation (Polazzi et al. 1999; Patrizio et al. 2001; Visentin et al. 2001). Furthermore, NF-κB activation is not correlated with the concomitant Tat-induced oxidative stress.

An induction of cytokines by Tat has been recently reported in astrocytes and microglia (Nath et al. 1999; Sheng et al. 2000), but the mechanisms involved in such induction were not investigated. At variance with the present work, Sheng et al. (2000) reported that Tat did not stimulate superoxide production. This discrepancy is probably attributable to the different doses of Tat, as well as to the method used to monitor the occurrence of radical formation in the two studies: the measurement of 8-epi-PGF accumulation allows a time-integrated quantification of free radicals and lipid peroxidation throughout the 24-h period of Tat treatment, whereas the ferricytochrome c reduction assay used by Sheng et al. (2000) relies on the ability of microglia to form free radicals over a period of 90 min in the presence of the inducer, phorbol myristate acetate, following Tat-treatment.

Consistently with our finding, Tat was shown to induce pro-oxidative conditions by decreasing the expression of Mn-dependent superoxide dismutase in HeLA cells (Flores et al. 1993), and to increase NADPH oxidase activity in human peripheral blood mononuclear cells (Lachgar et al. 1999).

Our data do not support a causal role of oxidative stress in Tat-induced NF-κB activation, in apparent contrast with other studies in other cell systems in which Tat acted through a change in the cellular redox state (Demarchi et al. 1996). However, the link between oxidative stress and NF-κB has been recently challenged and it has been proposed that in most cases the role of oxidative stress is facilitatory rather than causal, depending on specific cell types (Bowie and O'Neill 2000). Because NF-κB activation requires high concentrations of Tat (1 µg/mL) and is not dependent on extracellular Ca2+– as EGTA up to 100 µm did not prevent NO or TNF-α synthesis (not shown) – we suggest that Tat enters the cells and activates NF-κB, bypassing membrane receptor binding, as previously discussed (Polazzi et al. 1999). This hypothesis would be consistent with the known ability of Tat to form protein complexes with transcriptional regulators such as p300/CREB-binding protein (Marzio et al. 1998), or to activate the RNA-dependent protein kinase R, resulting in IκBα phosphorylation and degradation (Demarchi et al. 1999).


The authors thank Immunodiagnostics and NIBSC Centralized Facility for AIDS Reagents, supported by EU Programme EVA (contract BMH4 97/2515) and the UK Medical Research Council, for providing recombinant Tat HIV-1 IIIB. We also thank Dr Paqualina Bovenzi (Tema Ricerche SRL, Bologna, Italy) for continuos technical support. This work was funded by Project on AIDS of the Italian Ministry of Health (grant n. 30 C/H to GL).