The authors are in the Department of AIDSand Retroviruses, Institut Pasteur, 28rue du Dr. Roux, 75724 Paris Cedex, France. that programmed cell death induction is related to AIDS pathogenesis (4). Chimpanzees that have been infected with HIV-1 are productively infected but do not develop disease and do not show any immune deficiency, and simian immunodeficiency vi-rus (SIV)-infected macaque rhesus monkeys show a rapid CD4 cell decline leading to AIDS-like disease. Programmed cell death did not occur in T cells from HIV-1-infected chimpanzees and, by contrast, cell death by apoptosis was observed upon activation of T cells from SIV-infected macaques. The ab-sence of apoptosis in HIV-infected chimpanzees suggests that T cell programmed cell death is not an obligatory consequence of HIV infection, but reflects a more complex array of interactions between HIV and the immune system.

Is there in vivo evidence that binding of CD4 receptors on lymphocytes by gpl20 oc-curs in AIDS patients? The site of the CD4 receptor (D1 domain), which bindsto gpl 20, is often masked in AIDS patients, presumably by the gpl20 itself (13). Viral particles can shed freegpl20, which could bind rapidly to CD4 receptors on lymphocytes or could be complexed by specific antibodies. Upon treatment ofpatients with AZT, a transient unmasking of CD4 occurs. Recognizing the importance of apoptosis in AIDSpathogenesis may have dramatic consequences for the conception ofnew strat-egies of research and treatment for combating the disease. In order to prevent apoptosis, it is essential to dissect the subtle interaction of molecules involved in the process. Par-ticularly important will be the identification of novel surface markers for an early diagno-sis of apoptosis. The biochemistry also needs to be clarified. The role of oxidative stress and mitochondrial alterations should beana-lyzed. Antioxidants such as N-acetyl-cystein, vitamins C and E, and superoxide dismutase should be included in clinical tri-als in combination with antiviral therapy. The masking of CD4 receptor by gpl20 may be used as a surrogate marker for the clinical evaluation ofantiviraIs, as suggested by the rapid, although transient, unmasking of this receptor by AZT. Soluble CD4 could prevent the binding of gpl20 to the CD4 receptor. As a viral inhibitor, soluble CD4 has been rather disappointing, but it could have a beneficial effect on the course of the disease by preventing the abnormalsignaling of the CD4 lymphocytes due to the binding of the viral glycoprotein. We suggest there-fore that the pharmaceutical industryreenter this neglected field and make sufficient amounts ofsoluble CD4 for clinical trials. Finally, it would be very important to identify the superantigens, if any, that activate CD4+ and CD8+ Tcells. Because of the exist-ence ofsuperantigens in animal retroviruses (14), it has been proposed that HIV encodes