Approximately one-third of adults and half of the children suffering from AIDS develop neurological sequelae culminating in severe cognitive impairments collectively described as AIDS dementia complex  (ADC). These neurological symptoms are accompanied by neuropathological changes which include widespread reactive astrocytosis, myelin pallor, infiltration by blood-derived macrophages, appearance of multinucleated giant cells and severe neuronal cell loss [2,3]. Mononuclear phagocytes (macrophages, microglia and multinucleated giant cells) are the chief cell types infected by HIV-1 in the brain and in contrast, infection of astrocytes is highly unusual in adults but occurs occasionally in children [4,5]. Viral infection of brain macrophages and microglia has been shown to be accompanied by neurodegeneration; however, this seems to occur without either direct infection of neurones or the involvement of an HIV-1-mediated autoimmune reaction .
Several neurotoxic factors have been suggested to play a central role in HIV-induced neurodegeneration , these may be classified as virotoxins or cellular toxins depending on whether they are derived from the viral or host genome. These factors include proteins derived from the viral genome such as structural envelope protein fragments (e.g. gp120 and gp41), viral regulatory proteins such as those encoded by tat, nef, and rev genes, as well as products produced indirectly as a result of macrophage infection such as eicosanoids, cytokines, quinolate, or nitric oxide . It has not clearly been established which of these factors are potentially important in triggering HIV-induced neurodegeneration, however, both gp120 and Tat proteins are released by infected cells and have been reported to have potent neurotoxic effects when added to cultures of isolated neurones [8-10] or when injected intracerebroventricularly into mice [11,12].
The nuclear regulatory protein Tat is the only potentially toxic HIV-induced factor, which is known to be actively and continuously secreted from unruptured chronically infected lymphoid cells [10,13]. Tat can be detected in the serum of HIV-infected patients  and both Tat mRNA and protein have been detected at elevated levels in the brains of patients with HIV encephalitis and dementia [15,16]. Injection of Tat protein into the brains of rats produces widespread and profound neurodegeneration . This degeneration appears to result from the triggering of an inflammatory cascade rather than by any direct actions of Tat itself as tissue destruction occurs several days after Tat injection . However, Tat has also been shown to directly excite dissociated cultured neurones at very low concentrations and to induce glutamate receptor-dependent intracellular calcium uptake, which could potentially be toxic . We reasoned that studying the neurotoxic actions of macrophage-derived Tat in vitro requires a system sufficiently representative of the brain to allow complex cellular interactions to occur, particularly as many potentially toxic factors probably interact with each other.
Organotypic slice cultures of brain tissue have recently been introduced and retain much of the organizational features of the intact brain including neuronal connectivity and a well preserved astrocyte neuronal interface . In the present study we have developed a new model involving the co-culture of a murine macrophage cell line RAW264.7  with organotypic hippocampal slice cultures. We have studied the effects of Tat protein in this system and provide direct evidence that the toxic effects of Tat are mediated via the production of a macrophage-derived diffusable factor.
Materials and methods
Organotypic hippocampal slice cultures and RAW264.7 co-cultures
Organotypic slice cultures of rat hippocampus were prepared from 8-10 day old Wistar rat pups and cultured by previously described methods . Sets of four organotypic hippocampal slice cultures were grown on a 0.22μm pore size membrane insert (Millicell CM, Millipore, Watford, UK) in six-well plates (Corning, Cambridge, MA, USA) and were maintained in vitro for 2 weeks before experiments. Co-culturing of organotypic slices was achieved by transferring the membrane inserts into six-well plates in which 2×106 murine RAW264.7 cells had been allowed to adhere to the plastic substrate for 1 h. This system ensured no direct physical contact between the macrophages and the slices. The tat-transfected (T21 and T31) and Nef-expressing RAW264.7 cell lines have been described previously [22,23].
Cell death in organotypic hippocampal cultures was assessed using the fluorescent exclusion dye propidium iodide (Sigma, Poole, UK) as described in detail elsewhere [24,25]. Cultures were placed in serum-free medium containing 75% minimal essential medium, 25% Hank‚s balanced salt solution, 5mg/ml glucose, 1mM glutamine and 5μg/ml propidium iodide for 20 min before imaging: propidium iodide fluorescence was visualized using a rhodamine filter set (488nm/515nm). The response of glutamate receptor agonists was studied by incubating slice cultures in the presence of 10μM N-methyl-D-aspartic acid (NMDA), 10μM agr;-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA), 3μM kainate, or 10mM glutamate. Images were captured using a frame grabber, visualized and stored on optical discs using the IMAGE 1.55 software (Wayne Rasband, National Institutes of Health).
Conditioned medium and recombinant Tat
Conditioned medium from tat-transfected RAW264.7 cells was produced by incubating cultures for 24h; medium was then removed, briefly centrifuged to pellet any non-adherent cells and placed in fresh six-well plates. Recombinant Tat HIV-1 IIIB (provided by J. Raina via the NIBSC AIDS Reagent Project) was added to cultures at a concentration of 10μg protein/ml.
Hippocampal slice cultures were fixed in 4% (w/v) paraformaldehyde in 0.1M phosphate buffer (pH7.4) at 4°C for 4 h and then either directly stained with Cresyl Fast Violet (CFV) or processed for immunocytochemical detection of specific astrocytic or neuronal markers. Astrocytic processes were visualized by glial fibrillary acidophilic protein (GFAP) immunocytochemistry using a polyclonal rabbit antibody (1:3000, kind gift from Dr Mepham, Southampton, UK). Neuronal processes were visualized using a monoclonal mouse antiserum against microtubule-associated protein (MAP) 2 (1:500; Sigma, Poole, UK). Sections were washed in Tris-buffered saline (TBS) pH 7.4 and blocked with 5% normal horse serum in TBS/0.1% Triton X-100 (TBSt) for 1 h at room temperature. Sections were then incubated overnight at 4°C with the primary antibody (1:3000 for GFAP; 1:500 for MAP2) in TBSt/1% normal horse serum with shaking. Following washing, slices were incubated for 1 h with a fluorescein isothiocyanate conjugated sheep anti- rabbit IgG (1:200; Serotech, Washington DC, USA) for GFAP and mounted on slides with PBS/glycerol medium. For (MAP2) staining, slices were incubated for 1 h with a rabbit anti-mouse IgG (1:200) in TBSt/1% normal horse serum then processed as described above. Sections were examined and images were captured and analysed using a Zeiss photomicroscope or a confocal laser-scanning microscope (Leica TCS 4D, Milton Keynes, UK).
We investigated the effects of RAW264.7 cells transfected with the HIV-1 tat gene (T21 and T31 Tat-expressing cell lines) on the viability of cultured hippocampal slices. Viability of organotypic hippocampal slice cultures was clearly affected by co-incubation with Tat-expressing cells as shown by the marked increase in propidium iodide fluorescence (Fig. 1c, e). Propidium iodide fluorescence appeared within the first 6-7 h of the incubation with Tat-expressing cells and distinct punctuate nuclear staining was clearly visible (Fig. 1c). Propidium iodide staining was diffusely located throughout the mainly astrocytic areas of the slice culture but was not localized within the principal neuronal cell layers. The propidium iodide staining increased in intensity over a period of 24 h becoming very extensive, now labelling neurone-rich dentate and pyramidal cell layers as well as the surrounding astrocytic areas (Fig. 1e). CFV staining confirmed the presence of widespread neurodegeneration 24 h after co-incubation with Tat-expressing cells (Fig. 1d) compared with controls (Fig. 1b). At higher magnification clear evidence of neurodegeneration was present in the form of pyknotic cell nuclei and a generally distorted cytoarchitecture throughout the extent of the slice culture (Fig. 1f).
Incubation of organotypic slice cultures with untransfected RAW264.7 cells or with RAW264.7 cells transfected with the HIV-1 nef gene (Table 1) did not result in any propidium iodide labelling or evidence of neurodegeneration on CFV-stained sections (Fig. 2b). RAW264.7 cells treated with lipopolysaccharide (LPS), 4β-phorbol 12-myristate 13-acetate (PMA) and interferon (IFN) γ were also without effect (Table 1). More surprisingly, incubation of the slice cultures with recombinant Tat protein (10μg/ml) in the absence of RAW264.7 cells was also without effect (Fig. 2a). However, co-incubation of untransfected RAW264.7 cells with recombinant Tat protein did result in widespread neurodegeneration similar to that seen with transfected cell lines (Fig. 2c). At high magnification, CFV-stained sections revealed clear signs of nuclear condensation and the formation of morphological features resembling apoptotic inclusion bodies (Fig. 2d, see also Fig. 4). We therefore propose that the incubation of RAW264.7 cells with Tat results in the production of a soluble toxic factor not typical of classically activated macrophages.
We next used confocal laser microscopy with immunofluorescence detection of the astrocytic marker GFAP and the neuronal marker MAP2 to characterize the effects of tat-transfected RAW264.7 cells on the cytoarchitecture of the hippocampal slice culture (Fig. 3). Using a double labelling technique, it was also possible to identify cells that had accumulated propidium iodide during the experiment. Untreated cultures revealed a clear meshwork of GFAP-positive astrocytic processes surrounding neurones, with no evidence of propidium iodide uptake (Fig. 3a). Twenty-four h after co- incubation with Tat-producing RAW264.7 cells, the astrocytic matrix was completely destroyed (Fig. 3e) and propidium iodide-positive nuclear staining was evident throughout the culture. At very high power, nuclear propidium iodide staining revealed clear signs of chromatin clumping and the formation of small inclusions similar to those described during apoptosis (Fig. 4). MAP2 staining of neuronal processes revealed clear healthy-looking neurones with well organized neuronal processes in untreated organotypic cultures (Fig. 3b). No propidium iodide staining was evident in these control cultures, however 24 h after incubation with Tat-producing RAW264.7 cells, intense propidium iodide staining was present in the neuronal cell fields (Fig. 3f). At this time point, MAP2 staining had almost completely disappeared with only a few neuronal processes still visible (Fig. 3f).
To characterize further the nature of the toxic effect slice cultures were incubated with conditioned medium from cultures of Tat-producing RAW264.7 cells. Although conditioned medium did not consistently result in propidium iodide labelling in all organotypic slice cultures, the cytoarchitecture of both GFAP- and MAP2- expressing cells was consistently found to be distorted significantly (Fig. 3c, d). Astrocytic processes appeared retracted and no longer tightly surrounded cells within principal neuronal cell layers (Fig. 3c), leaving large exclusion zones in the culture where no GFAP staining was evident. Similarly evidence of cytoplasmic and nuclear swelling in MAP2-expressing neurones indicated that these were also affected by the conditioned medium (Fig. 3d). These results suggest that conditioned medium from Tat producing RAW264.7 cells induced morphological changes in organotypic cultures but was not sufficient to reliably produce cell death. The pattern of damage seen with Tat-induced damage was not replicated by known glutamate receptor agonists (Fig. 5). Incubation of slice cultures with 10mM Glutamate, 10μM AMPA, 3μM kainate or 10μM NMDA produced typical regionally selective patterns of cell death. There was no evidence of glial cell death with any of the glutamate agonists tested, suggesting that toxicity of the Tat-induced factor is unlikely to be mediated via these receptors.
The results obtained by using a new model based on the co-culture of macrophages with organotypic hippocampal slice cultures, provides the first clear evidence that a soluble factor(s) may be responsible for the neurotoxicity produced by HIV-1 Tat protein in the brain, rather than Tat itself.
The model, which we have developed, has several advantages over existing in vitro models using cultures of dissociated neurones and glia. Organotypic cultures retain much of the organization of the brain regions of interest including intact neuronal connectivity and a well-defined glial neuronal matrix . Several studies have shown that the response of organotypic cultures to a number of excitotoxic insults is very similar to that seen in the intact brain [24-27]. The co-culture system used in the present study is well suited to the study of neurotoxicity in AIDS dementia because it allows the study of complex interactions between neurotoxic factors, which may be produced by several different kinds of cell. In addition to infected mononuclear cells, both astrocytes and neurones are known to secrete toxic factors, which may be important in HIV dementia. This model is therefore likely to be more representative of the situation with HIV infection in the brain than that seen with simpler models such as those based on isolated cell cultures. The fact that neuronal cultures and macrophages are separated by a semi-permeable membrane, and are never in physical contact, is also an important advantage, as it allows the study of soluble factors produced by stimulated macrophages.
The widespread degeneration seen in our model with tat-transfected cells is remarkable in its extent and resulted in almost total destruction of the slice culture within 24 h. These results are similar to those reported by Weeks  who found massive tissue destruction after injections of Tat directly into the striatum of the rat. Intracerebroventricular injection of Tat protein has been shown to produce a number of reactions including recruitment of inflammatory cells, gliosis and eventually neurodegeneration . Hence, a transient exposure to Tat appears to induce an inflammatory cascade resulting in cell death rather than directly inducing acute neurodegeneration. One of the most important findings in our study was that although the combination of Tat and RAW264.7 cells was neurotoxic, neither Tat nor untransfected RAW264.7 cells alone produced similar toxicity. Several reports in the literature suggest that Tat can have a number of direct effects on neurones. Femtomolar levels of Tat have been shown to directly depolarize hippocampal CA1 neurones and foetal human neurones by a tetanus toxoid-independent mechanism . Studies using dissociated cultures also suggest that both NMDA and non-NMDA glutamate receptors participate in direct neurotoxic effects of exogenously applied Tat via a calcium-dependent mechanism [28-30]. This has led to the suggestion that the neurotoxic effects of Tat may result from excessive glutamate-mediated depolarization and toxic intracellular calcium overload. As we found no toxic effects of Tat alone in the absence of macrophages we propose that Tat is more likely to require a macrophage-derived factor in order to mediate its neurotoxic effects. Unlike isolated neuronal cultures, organotypic cultures are extremely resistant to high extracellular glutamate concentrations, requiring millimolar levels before toxicity is evident . Thus, whereas Tat could directly depolarize neurones in an organotypic culture system, this would be unlikely to kill them because active glial uptake mechanisms should prevent parenchymal glutamate from rising to toxic levels. Glutamate receptor-mediated toxicity would result primarily in neuronal damage with minimal effects on astrocytes [26,27], which is not consistent with the pattern of propidium iodide staining we observed.
We were able to demonstrate that media conditioned by tat-transfected RAW264.7 cells was also able to illicit neurotoxic effects, albeit less profound than those observed in the co-culture experiments. Since recombinant Tat alone was not toxic to hippocampal slice cultures we suggest that Tat must induce the production by RAW264.7 cells of one or more soluble factors that have a direct or indirect neurotoxic effect. This toxic effect appeared to be specific to activation by Tat as activation of RAW264.7 cells with a variety of other stimulating agents did not produce any evidence of degeneration in the slice cultures. Our observations are strikingly similar to those reported using a complex culture system based on reaggregated neuronal/glial cultures of human brain tissue : these authors found that HIV-1 infected macrophages produced a potent soluble factor which was both gliotoxic and neurotoxic whheras stimulation of un-infected macrophages with LPS had no similar effects.
The reduced toxic effect of conditioned medium in our study may indicate that the Tat-induced factor is quite labile and hence is partially degraded when incubated with the slice cultures. Alternatively, the simultaneous presence of the slice cultures and Tat-producing macrophages may be required to express the full extent of this toxic effect, possibly involving some feedback between the two culture systems. This latter possibility could be related to the suggestion that amplificatory cellular responses involving astrocytes and neurones in addition to infected mononuclear cells may be responsible for the widespread neurodegeneration observed in ADC patients .
Recent studies have demonstrated the ability of Tat to induce activation of human monocytes and to stimulate their expression of a range of soluble factors including matrix metalloproteinase-9, transforming growth factor-β1, tumour necrosis factor (TNF)-agr;, interleukin (IL)-1β, IL-6 and IL-8 [33-35]. In addition, Lafrenie et al. showed that Tat treated monocytes displayed enhanced monocyte-endothelial cell adhesion and promoted profound disruption of endothelial cell monolayers, which could be mechanistically related to the profound cytoarchitectural disruption that we observed in the organotypic slice cultures [34,35]. As expression of TNF-agr;, IL-1 and IL-6 is induced in Tat-stimulated monocytes  and as all three cytokines have been implicated in Tat-induced neurotoxicity [12,36,37] one or more may be responsible for the neurotoxic damage observed in the current study. Recent work in our laboratory has revealed that TNF-agr; is not toxic when added on its own in organotypic slice cultures but that ir modulates toxicity induced by other excitotoxins significantly . Moreover, both PMA and LPS induce expression of TNF-agr; by RAW264.7 cells [39,40] yet neither PMA- nor LPS-treated RAW264.7 cells had neurotoxic effects when co-cultured with hippocampal slices. Increased production of TNF-agr; could act indirectly to potentiate damage; however, the primary damage must be triggered by other factors induced by Tat. It is also unlikely that either IL1-β or IL-6 are directly responsible for the primary damage as expression of both cytokines by RAW264.7 cells is induced co-ordinately with TNF-agr; in response to LPS [39-41]. It has been suggested that macrophages stimulated with cytokines or infected by HIV-1 may release other toxic substances such as quinolinic acid [42-44] thus killing neurones indirectly. Quinolinic acid is known to be selectively neurotoxic and does not cause the death of astrocytes because these contain enzymes which convert quinolinic acid to non-toxic metabolites [7,45]. A selectively neurotoxic agent such as quinolinate would result in a different pattern of propidium iodide staining than the one observed in our experiments and should not result in any astrocytic damage.
Our observations suggest that the toxicity observed with Tat-transfected cells or macrophages incubated with Tat may be primarily astrocytic. Propidium iodide images from early incubation times show widespread degeneration throughout regions that are not particularly rich in neurones but that contain mainly astrocytes (e.g., 7 h, Fig. 1c). Only at later time points (24 h) do the CA1 and CA3 regions, which are rich in pyramidal neurons, become affected (Fig. 1e). It therefore seems possible that the neurotoxicity produced by Tat could be secondary to a loss of astrocytic function, as neurones would be unlikely to survive long in the absence of glial support mechanisms. High resolution confocal microscopy of the propidium iodide signal revealed clear signs of nuclear fragmentation reminiscent of that described during apoptosis (see Fig. 4). This pattern of nuclear propidium iodide staining is never seen in organotypic cultures with classical necrotic insults such as incubation with glutamate receptor agonists (see Fig. 5) or ischaemic insults and suggests, at the very least, that the mode of death in this case is distinct from classical necrosis.
In summary, our work suggests that the HIV-1 regulatory protein Tat can induce the production of soluble toxic factor(s) from macrophages which is sufficient, if studied in a complex in-vitro system, to cause almost complete destruction of brain tissue. HIV-1 Tat has been found extracellularly and could therefore stimulate production of the neurotoxic factor by uninfected macrophages as well as by infected cells [12,15]. It is clear that if such mechanisms are operational in the brains of individuals with AIDS this could result in the widespread neurodegeneration characteristic of patients suffering ADC. We propose that the co-culture model described in this study can be used to identify HIV-1 induced macrophage-derived neurotoxic factors and to study the mechanisms by which these factors induce cell death in the brain.
The authors thank R. Alston and S. Bottoms for their help with confocal imaging.
1. Everall IP, Lanthos P. The Neuropathology of HIV: A review of the first 10 years. Int Rev Psychiatr
2. Everall IP, Luthert PJ, Lanthos P. Neuronal loss in frontal cortex in HIV infection. Lancet
3. Masliah E, Ge N, Achim CL, DeTeresa R, Wiley CA. Patterns of neurodegeneration. NeuroAIDS
4. Tornatore C, Chandra R, Berger JR, Major EO. HIV infection of subcortical astrocytes in the paediatric CNS. Neurology
5. Bagasra O, Lavi E, Bobroski L, et al.Cellular reservoirs of HIV
-1 in the CNS of infected individuals identification by the combination of in situ polymerase chain reaction and immunohistochemistry. AIDS
6. Wiley C, Shrier RD, Nelson JA, Lampert PW, Oldstone MBA. Cellular localization of human immodeficiency virus infection within the brains of acquired immunodeficiency syndrome patients. Proc Natl Acad Sci USA
7. Nath A, Geiger JD. Neurobiological aspects of human immunodeficiency virus infection: neurotoxic mechanisms. Prog Neurobiol
8. Magnuson DSK, Knudsen BE, Geiger JD, Brownstone RM, Nath A. Human immunodeficiency virus type 1 Tat activates non-N-methyl-D-aspartate excitatory amino acid receptors and causes neurotoxicity. Ann Neurol
9. Wu P, Price P, Du B, Hatch WC, Terwilliger EF. Direct cytotoxicity of HIV-1 envelope protein gp120 on human NT neurons. Neuroreport
10. Cheng J, Nath A, Knudsen B, et al.Neuronal excitatory properties of human immunodeficiency virus type 1 TAT protein.Neuroscience
11. Sabatier JM, Vives E, Mabrouk K, et al.Evidence for neurotoxic activity of tat from human immunodeficiency virus type 1. J Virol
12. Philippon V, Vellutini C, Gambarelli D, et al.The basic domain of the lentiviral Tat Protein is responsible for damages in mouse brain: Involvement of cytokines. Virology
13. Ensoli B, Buonaguro L, Barilari G. Release uptake and effects of extracellular HIV-1 Tat protein on cell growth in viral transactivation. J Virol
14. Westendorp OP, Frank R, Ochsenbauer C, et al. Sensitisation of T-cells to CD95-mediated apoptosis by HIV-1 Tat and GP120. Nature
1995, 1995, 375
15. Wesselingh SL, Power C, Glass JD, et al.Intracerebral cytokine messenger RNA expression in aquired immunodeficiency syndrome dementia. Ann Neurol
16. Hofman FM, Dohadwala MM, Wright AD, Hinton DR, Walker SM. Exogenous Tat protein activated central nervous system-derived endothelial cells. J Neuroimmunol
17. Weeks BS, Lieberman DM, Johnson B, et al.Neurotoxicity of the human immunodeficiency virus type 1 tat transactivator to PC12 cells requires the Tat amino acid 49-58 basic domain. J Neurosci Res
18. Jones M, Olafson K, Del Bigio MR, Peeling J, Nath A. Intraventricular injection of human immunodeficiency virus type 1 (HIV-1) tat protein causes inflammation, gliosis, apoptosis, and ventricular enlargement. J Neuropathol Exp Neurol
19. Gahwiler BH, Capogna M, Debanne D, McKinney RA, Thompson SM. Organotypic slice cultures: a technique has come of age. Trends Neurosci
20. Raschke WC, Baird S, Ralph R, Nakoinz I. Functional macrophage cell lines transformed by Abelson leukemia virus.
21. Stoppini L, Buchs PA, Muller D. A simple method for organotypic cultures of the nervous system. J Neurosci Methods
22. Barton CH, Biggs TE, Mee TR, Mann DA. The HIV-1 regulatory protein Tat inhibits interferon-induced i-Nos activity in a murine macrophage cell line. J Gen Virol
23. Cooke SJ, Coates K, Barton CH, et al.Regulated expression-vectors demonstrate cell-type-specific sensitivity to HIV-1 Nef-induced cytostasis. J Gen Virol
24. Vornov JJ, Tasker R, Coyle JT. Direct observation of the agonist specific regional vulnerability to glutamate. NMDA and kainate neurotoxicity in organotypic hippocampal cultures. Exp Neurol
25. Pringle AK, Iannoti F, Wilde GJC, Chad JE, Seeley PJ, Sundstrom LE. Neuroprotection by both NMDA and non-NMDA receptor antagonists in in vitro ischemia. Brain Res
26. Pringle AK, Benham CD, Sim L, Kennedy J, Iannoti F, Sundstrom LE. Selective N-type calcium channel antagonist omega conotoxin MVIIA is neuroprotective against hypoxic neurodegeneration in organotypic hippocampal-slice cultures. Stroke
27. Wilde GJC, Pringle AK, Wright P, Iannotti F. Differential vulnerability of the CA1 and CA3 subfields of the hippocampus to superoxide and hydroxyl radicals in vitro. J Neurochem
28. Magnuson DSK, Knudsen BE, Geiger JD, Brownstone RM, Nath A. Human immunodeficiency virus type 1 Tat activates Non
-N-Methyl-D-Aspartate excitatory amino acid receptors and causes neurotoxicity. Ann Neurol
29. Strijbos PJLM, Zamani MR, Rothwell NJ, Arbuthnott G, Harkiss G. Neurotoxic mechanisms of transactivating protein Tat of Maedi
-Visna virus. Neurosci Lett
30. Nath A, Psooy K, Martin C, et al.Identification of a human immunodeficiency virus type 1 Tat epitope that is neuroexcitatory and neurotoxic. J Virol
31. Pulliam L, Herndier BG, Tang NM, McGrath MS. Human immunodeficiency virus
-infected macrophages produce soluble factors that cause histological and neurochemical alterations in cultured human brains. J Clin Invest
32. Lipton SA. Neuronal injury associated with HIV-1: Approaches to treatment. Annu Rev Pharmacol Toxicol
33. Gibellini D, Zauli G, Re MC, et al.Recombinant human immunodeficiency virus type-1 (HIV-1) Tat protein sequentially up-regulates IL-6 and TGF-β mRNA expression and protein synthesis in peripheral blood monocytes. Br J Haematol
34. Lafrenie RM, Wahl LM, Epstein JS, Hewlett IK, Yamada KM, Dhawan S. HIV-1 Tat modulates the function of monocytes and alters their interactions with microvessel endothelial cells. J Immunol
35. Lafrenie RM, Wahl LM, Epstein JS, Yamada KM, Dhawan S. Activation of monocytes by HIV-Tat treatment is mediated by cytokine expression. J Immunol
36. Gelbard HA, Dzenco KA, Diloreto D, Delcerro C, Delcerro M, Epstein LG. Neurotoxic effects of tumor necrosis factor alpha in primary human neuronal cultures are mediated by activation of the glutamate AMPA receptor subtype: Implication for AIDS neuropathogenesis. Dev Neurosci
37. New DR, Maggirwar SB, Epstein LG, Dewhurst S, Gelbard HA. HIV-1 Tat induces neuronal death via tumor necrosis factor-alpha and activation of non-N-methyl-D-aspartate receptors by NfkappaB-independent mechanism.J Biol Chem
38. Wilde GJC, Pringle AK, Sundstrom, LE, Iannotti F. The protective effects of tumor necrosis factor in an in-vitro model of transient global ischaemia. J Neurol Neurosurg Psychiatry
39. Flatch R, Schade FU. Peritoneal macrophages from endotoxin-tolerant mice produce an inhibitor of tumor necrosis factor alpha synthesis and protect against endotoxin shock. J Endotoxin Res
40. Williams JG, Bernstein S, Prager M. Effect of melatonin on activated macrophage TNF, IL6 and reactive oxygen intermediates. Shock
41. Godambe SA, Chaplin DD, Takova T, Read LM. Bellone CJ. A novel cis-acting element required for lipopolysaccharide-induced transcription of the murine interleukin-1-βgene.Mol Cell Biol
42. Espey MG, Chernyshev JF, Reinhard Jr, JF, Namboodiri MAA, Colton CA. Activated human microglia produce the excitotoxin quinolinic acid. NeuroReport
43. Pemberton LA, Kerr SJ, Smythe G, Brew BJ. Quinolinic acid production by macrophages stimulated with IFN-γ, TNF-agr; and IFN-agr;. J Interferon Cytokine Res
44. Kerr SJ, Armati PJ, Guillemin GJ, Brew BJ. Chronic exposure of human neurons to quinolinic acid results in neuronal changes consistent with AIDS dementia complex. AIDS
45. Noack H, Lindenau J, Rothe F, Asayama K, Wolf G. Differential expression of superoxide dismutase isoforms in neuronal and glial compartments in the course of excitotoxically mediated neurodegeneration: relation to oxidative and nitrergic stress. Glia