The central nervous system manifestations of AIDS were originally thought to consist solely of white matter lesions, but recent evidence has shown that a substantial degree of neuronal loss can also occur. This review presents evidence for HIV-related toxic factors that may account at least in part for this newly-recognized neuronal injury. One potential neurotoxin is the HIV-1 envelope glycoprotein gp-120 or a fragment of this molecule. This coat protein is shed by the virus and potentially released from HIV-infected immune cells. In tissue culture experiments on rodent neurons, gp120 produces an early rise in intracellular calcium concentration and, subsequently, delayed-onset neurotoxicity. In addition, HIV-infected macrophages or microglia release as yet undefined toxic factor(s) that kill rodent, chick, and human neurons in vitro. It is as yet unknown if one of these macrophage toxic factors might represent a gp120 fragment, or alternatively, if gp120, in the absence of HIV-1 infection, might be capable of activating macrophages to release these toxic factors). In at least some neuronal cell types, gp120-induced neurotoxicity can be prevented by antagonists of L-type voltage-dependent calcium channels or by antagonists of W-methyl-D-aspartate (NMDA, a subtype of glutamate receptor). Degradation of endogenous glutamate also protects neurons from gp120-related neuronal injury, suggesting that gp120 and glutamate are both necessary for neuronal cell death as synergistic effectors. Antagonists acting at the other types of glutamate receptors
(non-NMDA antagonists) are ineffective in affording protection from gp120. Interestingly, NMDA, but not non-NMDA, antagonists also block the lethal effects of the macrophage toxic factor(s). The similar profile of pharmacological protection may possibly reflect the fact that at least one of the macrophage toxic factors is related to gp120, as suggested above. However, molecular-sieving and protease-digestion experiments suggest that the macrophage toxic factor(s) does not appear to be intact gp120, although a gp120 fragment remains a possibility. Alternatively, it is plausible that macrophages secrete several unrelated neurotoxic factors. Astrocytes may also be important in mediating HIV-related neurotoxicity. For example, in some neuronal cultures gp120-induced toxicity can be prevented by vasoactive intestinal polypeptide (VIP) or by a five amino acid substance with sequence homology, peptide T. VIP has been found to act on astrocytes to increase oscillations in intracellular calcium and to release factors necessary for normal neuronal outgrowth and survival. These results raise the possibility that gp120 may compete with endogenous VIP for a receptor, most likely on astrocytes, that is important for neuronal function. In summary, toxic factor(s) from HIV-infected human monocytoid cells may lead to neuronal damage in vitro. K is as yet unknown if these factors include a gp120 fragment or if gp120 may trigger the release of these neurotoxic factors. Based upon in vitro studies, calcium channel antagonists or NMDA antagonists may represent promising forms of pharmacological intervention to protect neurons from HIV-related injury. In the brains of AIDS patients, neuronal injury may be mediated by several separate pathways that most likely originate from toxins released by HIV-infected macrophages. Alternatively, there may be an intricate web of neurotoxic factors interacting with macrophages/microglia, astrocytes, and neurons; this complex may be amenable to pharmacotherapy because of common final pathways of attack involving growth factors, NMDA receptors, and deleteriously high levels of intracellular calcium ions.