Human immunodeficiency virus (HIV) infection is commonly associated with neurological disease that occurs in the apparent absence of extensive infection of brain cells by HIV, suggesting that indirect mechanisms account for neuropathogenesis in the CNS, perhaps including changes in the normal neuroprotective functions of astrocytes. To test this hypothesis, we examined the effect of the pro-inflammatory cytokine, tumor necrosis factor alpha (TNFalpha), produced by HIV-1-infected macrophages and microglia, on glutamate transport by primary human fetal astrocytes (PHFAs). A dose-dependent inhibition of high affinity glutamate uptake sites was observed 12-24 h after addition of exogenous recombinant human TNFalpha to PHFAs. This effect was specific since it was blocked by a neutralizing monoclonal antibody directed against TNFalpha. Furthermore, the inhibitory effect was reproduced by a monoclonal antibody that is an agonist at the 55-kDa TNF receptor. These results suggest that the neurotoxic effects of TNFalpha may be due in part to its ability to inhibit glutamate uptake by astrocytes, which in turn may result in excitotoxic concentrations of glutamate in synapses.

Tumor necrosis factor alpha (TNF-␣) is a candidate human immunodeficiency virus type 1-induced neurotoxin that contributes to the pathogenesis of AIDS dementia complex. We report here on the effects of exogenous TNF-␣ on SK-N-MC human neuroblastoma cells differentiated to a neuronal phenotype with retinoic acid. TNF-␣ caused a dose-dependent loss of viability and a corresponding increase in apoptosis in differentiated SK-N-MC cells but not in undifferentiated cultures. Importantly, intracellular signalling via TNF receptors, as measured by activation of the transcription factor NF-B, was unaltered by retinoic acid treatment. Finally, overexpression of bcl-2 or crmA conferred resistance to apoptosis mediated by TNF-␣, as did the addition of the antioxidant N-acetylcysteine. These results suggest that TNF-␣ induces apoptosis in neuronal cells by a pathway that involves formation of reactive oxygen intermediates and which can be blocked by specific genetic interventions.

HIV-1 infection of the brain results in chronic inflammation, contributing to the neuropathogenesis of HIV-1 associated neurologic disease. HIV-1-infected mononuclear phagocytes (MP) present in inflammatory infiltrates produce neurotoxins that mediate inflammation, dysfunction, and neuronal apoptosis. Neurologic disease is correlated with the relative number of MP in and around inflammatory infiltrates and not viral burden. It is unclear whether these cells also play a neuroprotective role. We show that the chemokine, fractalkine (FKN), is markedly up-regulated in neurons and neuropil in brain tissue from pediatric patients with HIV-1 encephalitis (HIVE) compared with those without HIVE, or that were HIV-1 seronegative. FKN receptors are expressed on both neurons and microglia in patients with HIVE. These receptors are localized to cytoplasmic structures which are characterized by a vesicular appearance in neurons which may be in cell-to-cell contact with MPs. FKN colocalizes with glutamate in these neurons. Similar findings are observed in brain tissue from an adult patient with HIVE. FKN is able to potently induce the migration of primary human monocytes across an endothelial cell/primary human fetal astrocyte trans-well bilayer, and is neuroprotective to cultured neurons when coadministered with either the HIV-1 neurotoxin platelet activating factor (PAF) or the regulatory HIV-1 gene product Tat. Thus focal inflammation in brain tissue with HIVE may up-regulate neuronal FKN levels, which in turn may be a neuroimmune modulator recruiting peripheral macrophages into the brain, and in a paracrine fashion protecting glutamatergic neurons.

Studies of interleukin 1 (IL-1) a and 13 have emphasized their functional similarities. IL-la and -P are encoded by ancestrally related genes that have diverged dramatically in primary sequence; however, only modest differences in the regulation or biological activity of IL-la and IL-11S have been documented. Here we show that mouse L cells transfected with murine pro-IL-la cDNA expressed biologically active, 33-kilodalton pro-IL-la, and that this pro molecule was neither processed to the 17-kilodalton mature form nor secreted. 5649The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Abstract. After skeletal muscle is denervated, fibroblasts near neuromuscular junctions proliferate more than fibroblasts distant from synaptic sites, and they accumulate adhesive molecules such as tenascin (Gatchalian, C. L., M. Schachner, and J. R. Sanes. 1989. J. Cell Biol. 108:1873-1890. This response could reflect signals that arise perisynaptically after denervation, preexisting differences between perisynaptic and extrasynaptic fibroblasts, or both. Here, we describe a line of transgenic mice in which patterns of transgene expression provide direct evidence for differences between perisynaptic and extrasynaptic fibroblasts in norreal muscle.Transgenic mice were generated using regulatory elements from a major histocompatibility complex (MHC) class I gene linked to the Escherichia coli/~-galactosidase (lacZ) gene. Expression of lacZ was detected histochemically. In each of eight lines, lacZ was detected in different subsets of cells, none of which included lymphocytes. In contrast, endogeous MHC is expressed in most tissues and at high levels in lymphocytes. Thus, the MHC gene sequences appeared inactive in the transgene, and lacZ expression was apparently controlled by genomic regulatory elements that were specific for the insertion site.In one line, cells close to the neuromuscular junction were lacZ positive in embryonic and young postnatal mice. Electron microscopy identified these cells as fibroblasts and Schwann cells associated with motor nerve terminals, as well as endoneurial fibroblasts, perineurial cells, and Schwann cells in the distal branches of motor nerves. No intramuscular cells >200 #m from synaptic sites were lacZ positive. These result indicate that there are molecular differences between perisynaptic and extrasynaptic fibroblasts even in normal muscle and that diverse perisynaptic cell types share a specific pattern of gene expression.