Gradual shutdown of virus production resulting in latency is the norm during the chronic phase of human immunodeficiency virus replication and differential rates and mechanisms of shutdown are determined by viral sequences

Most CD4(+) lymphocytes in lymph nodes of both asymptomatic HIV-1-infected individuals and AIDS patients are nonproductively or latently infected. It is not clear how these cells come about because infection of resting lymphocytes results in abortive infection and infection of activated lymphocytes results in productive infection. The frequency and mechanisms underlying nonproductive or latent HIV infections of normal CD4(+) lymphocytes largely remain unexplored, and because HIV latency has principally been studied in latently infected cell clones of established cell lines, it is not even dear how often this type of infection occurs in cell lines. We demonstrate herein that chronic HIV replication in populations of normal phytohemagglutinin-stimulated peripheral blood CD4(+)-enriched lymphocytes, as well as an established T-cell line (GEM), gradually shuts down in the vast majority of cells. The nonproducing cells in these cultures still harbored HIV provirus, and HIV could be reactivated in CEM cells by treatment with phorbol ester, showing that this was latent infection. Thus, HIV's life cycle should probably be considered as consisting of two phases: an acute exponential rise in production of virus progeny which levels at some peak, followed by a gradual decline of progeny production during the chronic phase leading to viral latency. Temporal analyses of the steady-state levels of viral mRNAs in populations of chronically infected CEM cells as virus production declined revealed the two mechanisms of HIV latency which have previously been described in the OM-10.1 and U1 or ACH-2 latently infected cell clones (i.e., apparent overall shutdown of HIV transcription and ''blocked early-stage latency'' involving enhanced splicing of viral pre-mRNAs). However, which mechanism was employed, as well as the rate of shutdown, depended on the virus strain. (C) 1996 Academic Press, Inc.