The role of CD8+
T cells in the control of SIV replication in vivo awaits experimental confirmation. Here, we examined their importance by depleting these cells with an anti-CD8 mAb, OKT8F, and by following the subsequent virologic and immunologic changes. Although the extent of CD8+
T cell depletion was not examined in lymphoid tissues in our study, we found that depleting such T cells in peripheral blood led to a peak viral load increase of one to three orders of magnitude in five of six macaques studied. These results demonstrate that CD8+
T cells, at a minimum, must play a major role in the control of SIV replication during chronic infection. If OKT8F administration did not result in complete CD8+
T cell depletion in lymphoid tissues in our experiment, then the role of these cells in inhibiting viral replication in vivo would be even more significant than what was observed here. Thus, we concur with the conclusion drawn by Matano et al. (29
), who studied the effect of CD8 depletion in primary SHIV infection.
The depletion of CD8+
T cells by OKT8F was not unexpected; however, the speed at which CD8+
T cells disappeared from blood was surprising. Nevertheless, the kinetics of CD8 depletion observed in our experiments is consistent with previous depletion experiments done by others (28
). Likewise, the OKT4 antibody cleared CD4+
T cells from peripheral blood within 10 min of injection in rhesus monkeys (40
). The OKT3 antibody also depleted most of the peripheral T lymphocytes within 1 h after administration to humans (41
Several mechanisms may be responsible for the observed depletion of CD8+ T cells from the blood. First, the binding of OKT8F to CD8 molecules on cells may have led to complement-mediated cytolysis. Second, OKT8F may have only masked the epitope recognized by the staining anti-CD8 antibody, resulting in an artificially low number of CD8+ T cells. Third, there may have been an increased CD8 endocytosis after binding by OKT8F, thereby lowering surface CD8 expression without cellular depletion. Finally, there may have been a preferential trapping of antibody-bound CD8+ T cells in lymph tissues, leading to an apparent depletion of such cells in the circulation. To address the second and third possibilities above, we examined the difference in the number of CD3+ and CD4+ T cells. This difference declined in parallel with the number of CD8+ T cells measured directly, showing that there is a genuine depletion of CD8+ T cells in blood. Further experiments are necessary to examine the possibility of trapping of CD8 cells in lymphoid tissues after OKT8F injection. However, if the CD8+ T cells were trapped in the lymph nodes, where most of the viral replication takes place, they were probably nonfunctional given the substantial rise in plasma viremia. Irrespective of the underlying mechanism, the injection of OKT8F appears to have severely impacted the CD8+ T cells as to cause a dramatic increase in virus replication. However, it is still formally possible that the loss of control of viral replication is due partly to the elimination of NK cells that express CD8 on their surface.
Several different mechanisms could potentially be responsible for the increase in viral load. The removal of CD8+
CTLs could increase the life span of productively infected cells, leading to greater virion production. Alternatively, the depletion of CD8+
T cells may reduce the production of factors, such as β-chemokines, that block the entry of certain virions into target cells (26
), or of inhibitory factors that suppress viral transcription, thereby reducing virion output from productively infected cells (36
). To quantitatively evaluate these hypotheses, we used a standard model of HIV-1 dynamics (37
) adjusted to account for the eclipse phase of the viral life cycle (39, and our unpublished observations). For each monkey, a set of parameters was determined that corresponded to the animal being in a steady state with the measured baseline viral load and CD4+
T cell count. To mimic the loss of CTL killing, the death rate of productively infected T cells, δ
, was decreased. Even lowering δ
to 0 as the most extreme case did not cause the viral load to increase as rapidly as observed. Similarly, increasing the infection rate constant, k
, to mimic the loss of β-chemokines from CD8+
T cells, could not explain the rapid increase in viral load seen during the first 24 h (Fig. ). Finally, removing inhibitory factors on viral transcription (increasing p
) gives a better fit to the data. Thus, the rapid rise in plasma viremia over the first 2–3 d is more consistent with a higher virus production rate than a higher infection rate. Moreover, even a maximal decrease in δ
fails to give a good fit to the data, suggesting that decreased cytotoxic killing is not the primary mechanism responsible for the increased viremia. Overall, these conclusions on the mechanism of virus inhibition by CD8+
T cells should be viewed as preliminary and must be addressed in greater detail in future experiments.
The depletion of CD8+
T cells was accompanied by reduction in the CD4+
T cells in all animals receiving OKT8F. Several possible explanations might be considered. There may have been increased destruction of CD4+
T cells due to rapid rise in SIV replication after OKT8F infection. However, when the same antibody was given to an uninfected macaque (AT-03), a similar magnitude of depletion of CD4+
T cells was also observed. Therefore, SIV is unlikely to have contributed significantly to this phenomenon. In agreement with our findings, it was observed in a previous study that treatment with anti-CD8 antibody also resulted in lowering of B cells and CD4+
T cells (42
). It is worth noting that the injection of OKT3 also led to a depletion of non-T cells, such as B cells and NK cells, in humans (41
). Thus, another explanation is that OKT8F injection may be equivalent to a large dose of foreign antigen, activating numerous cell types, including CD4+
T cells that later undergo activation-induced apoptosis. Finally, we have noted that there are greater numbers of lymphocytes (up to 14%) that coexpress CD4 and CD8 in SIV-infected macaques (our unpublished observations). The elimination of such cells by OKT8F may contribute to the lowering of CD4+
T cell counts, although they are numerically insufficient to be the sole mechanism. Further studies are necessary to understand the indirect impact on the CD4+
T cell population.
In summary, we have demonstrated that CD8+ T cells do play a major role in controlling SIV in vivo. Therefore, we suggest that in the development of SIV or HIV vaccines, a stronger emphasis be placed on inducing specific CD8+ T cell responses.