A complex relationship exists between HIV and its cellular targets. The lethal effect of HIV on circulating CD4(+) helper T lymphocytes parallels the degree of the infected individual's immunodeficiency and ultimately the transition to AIDS and death. However, as with other members of the Lentivirus family of retroviruses, the ubiquitous, mobile macrophage is also a prime target for HIV infection, and apparently, in most instances, is the initial infected cell, since most people are infected with a CCR5 chemokine-tropic virus. Unlike the lymphocyte, the macrophage is apparently a more stable viral host, capable of a long infected life as an HIV reservoir and a chronic source of infectious virus. Published in vitro studies have indicated that whereas lymphocytes replicate HIV solely on their plasma membrane, macrophages have been envisaged to predominantly replicate HIV within cytoplasmic vacuoles, and thus have been likened to a "Trojan horse," when it comes to the immune system. Recent studies have revealed an ingenious way by which the cultured monocyte-derived macrophage (MDM) replicates HIV and releases it into the medium. The key macrophage organelle appears to be what is alternatively referred to as the "late endosome" (LE) or the "multivesicular body" (MVB), which have a short and a long history, respectively. Proof of the association is that chemically, LE/MVB and their vesicles possess several pathopneumonic membrane markers (e.g., CD63) that are found on released HIV particles. The hypothesis is that HIV usurps this vesicle-forming mechanism and employs it for its own replication. Release of the intravacuolar virus from the cell is hypothesized to occur by a process referred to as exocytosis, resulting from the fusion of virus-laden LE/MVB with the plasma membrane of the macrophage. Interestingly, LE/MVB are also involved in the infection stage of MDM by HIV. Close review of the literature reveals that along with the Golgi, which contributes to the formation of LE/MVB, the MVB was first identified as a site of HIV replication by macrophages many years ago, but the full implication of this observation was not appreciated at the time. As in many other areas of HIV research, what has been totally lacking is an in vivo confirmation of the in vitro phenomenon. Herein, the ultrastructure of HIV interaction with cells in vitro and in vivo is explored. It is shown that while HIV is regularly found in LE/MVB in vitro, it is infrequently the case in vivo. Therefore, the results challenge the "Trojan horse" concept.