During the progress of an HIV-infected patient to AIDS, the CD4+ T4cells progressively diminish in number so that eventually the immune response is compromised and opportunistic infections ensue. During the late stages of infection when disease appears, CD8+ cells also decrease in number. As the course of infection progresses, the HIV population switches from non-syncytium-inducing types that use CCR5 as a co-receptor to syncytium-inducing variants that use CXCR4 as their co-receptor. CD8+ T cells have been reported to express low levels of CD4 antigen when they are activated and appear to be infected in small numbers by HIV in the later stages of disease but this is probably not enough to explain the loss of CD8+ cells.

There have been reports of HIV infection of CD8 T cells isolated from AIDS patients. Clearly, it is difficult to exclude contamination of these CD8+ cells by a small number of CD4+ cells and even if the CD8 cells had acquired a productive infection, it may have occurred via the low levels of CD4 antigen that these cells may express. The use of clones of CD8+ cells isolated from HIV-infected patients, however, allowed investigators to rule out the possibility of CD4+ cell contamination and to demonstrate that apparently CD4 antigen-negative CD8+ T cells can not only be infected by HIV but can also replicate the virus. It is possible that these cells become infected by a CD4 antigen-independent pathway and virus particles isolated from CD8+ T cells can re-infect uninfected CD8+ T cells independently of CD4 antigen and CCR5 or CXCR4.

Thus, at late stages of HIV infection, mutants with altered tropism may arise that can directly infect CD8 cells. This may result from the intense selective pressure on the constantly mutating virus population as a result of CD4 cell depletion and may be part of the explanation of the reduction in CD8+ cell numbers. Since HIV infection is primarily controlled by CD8+ T cells, this could be important in the immunocompromization of the infected patients.