To determine if the tissue-specific bioavailability of the drugs was the cause for the selective loss of mucosal CD4 T cells, we evaluated the levels of viral DNA in subsets of CD4 T cells from the mucosa and compared them to those from peripheral blood. Naïve and memory CD4 T cells (discriminated based on the expression of CD28 and CD95 [21
]) were sorted and used in a highly quantitative PCR assay for SIV gag
DNA as a measure of the efficacy of the drugs to stop viral transcription. We previously demonstrated that this assay was highly sensitive for measuring viral infection (17
) in CD4 T cells.
Our results showed that the initiation of ART dramatically reduced the frequency of SIV gag
DNA in both mucosal and peripheral blood CD4 T cells of treated animals compared to that in untreated controls (Fig. ), suggesting that ART effectively limited infection in both tissues. At day 10 p.i., the average frequency of SIV gag
copies was ~8 × 103
memory CD4 T cells in the mucosa and ~16 × 103
memory CD4 T cells in peripheral blood of ART-treated animals. We previously showed that there were ~2 copies of SIV gag
/infected memory CD4 T cell at day 10 p.i. (17
). This suggests that at the peak of viral infection (day 10 p.i.), <5% of mucosal CD4 T cells and <10% of memory CD4 T cells in peripheral blood were infected in treated animals, whereas most of the memory CD4 T cells in both tissues were infected in untreated animals.
FIG. 3. Kinetics of cell-associated viral loads during early ART. Far fewer cells were infected in animals that received ART at day 7 p.i. The cell-associated SIV gag DNA levels in sorted memory CD4 T cells in peripheral blood (a) and rectal mucosa (b) were determined (more ...)
It is interesting that at day 7 p.i. there was a significant amount of virus in the plasma but few cells were infected. One reason for this discordance may be that initial infection targets highly activated memory CD4 T cells, which are more efficient at replicating the virus. Thus, even though very few of these cells are infected, they likely contribute to a significant level of early plasma viremia. After the initial phase, infection disseminates to encompass the resting memory CD4 T cells, which are not as efficient as the activated memory CD4 T cells at viral replication (14
). Additionally, it is possible that other cells, such as macrophages and dendritic cells, likely contribute to early plasma viremia. Numerous studies (7
) have shown that both of these cell types can be productively infected.
Our findings suggest that the higher level of memory CD4 T-cell preservation seen in peripheral blood of treated animals than in untreated animals was primarily due to the loss of fewer memory cells by direct viral infection. In contrast, the massive destruction of CD4 T cells in the rectal mucosa even when infection was controlled to very low levels indicates that most (>90%) of the mucosal CD4 T cells in treated animals were destroyed by mechanisms upstream of viral reverse transcription, such as SIV gp120 binding to CD4 T cells or viral entry.
Li et al. (14
) demonstrated that a majority of mucosal CD4 T cells were not productively infected and were likely killed at a high rate due to SIV gp120-mediated apoptosis. Additionally, Boirivant et al. (4
) showed that binding of HIV-1 gp120 accelerated the Fas-mediated apoptosis of human lamina propria T cells. Previous studies (6
) have shown that mucosal CD4 T cells express high levels of CCR5, making them highly susceptible to infection. Additionally, the recent demonstration (2
) that HIV can use the activated form of α4β7 receptor for high-affinity binding and entry into cells lends support to this hypothesis, as most mucosal CD4 T cells express high levels of α4β7 receptor on their surfaces (26
). Cummins et al. (9
) showed that high levels of α4β7 integrin expression on CD4 T cells exposed to an HIV-1 X4 strain were associated with bystander death in these cells. Vajdy et al. (26
) demonstrated that acute SIV infection was associated with a profound loss of CD4 T cells expressing α4β7 in the rectal mucosa.
We cannot rule out that viral RNA was present in the mucosal CD4 T cells due to viral entry, since our assay measures only viral DNA; however, our results suggest that either gp120 binding, viral entry, or bystander mechanisms, rather than productive viral infection, are sufficient to kill most of the mucosal CD4 T cells. It is highly unlikely that SIV-specific immune responses played a major role in the depletion of these mucosal CD4 T cells, as previous studies (24
) have shown that mucosal CD4 T cells are destroyed much earlier than the emergence of SIV-specific immune responses. Additionally, it is important to indicate that we did not sample rectal mucosa between days 10 and 63 p.i., and it is likely that the continuous viral replication in these tissues contributed to the ongoing attrition of CD4 T cells in the rectal mucosa, leading to their loss. Numerous studies have shown that low levels of viral replication continue in the mucosa even during continuous HAART (8
The plasma viral loads at day 7 p.i. were essentially similar in both the ART-treated and untreated animals (Fig. ), suggesting that by the time ART was initiated there were enough virions to interact with CD4 T cells resident in the mucosa. This likely explains why mucosal CD4 T cells in both groups displayed similar kinetics of depletion even though <5% of these cells were infected at day 10 p.i. in treated animals.
It is unlikely that the rapid and widespread dissemination of the virus across the mucosal tissue compartment immediately after intravenous challenge could have more readily exposed the mucosal CD4 T cells to the virus. A similar intravenous challenge was associated with the protection of mucosal CD4 T cells when ART was initiated within 24 h after challenge (15
). The protection observed by Lifson et al. (15
) was probably due to the rapid containment of early viral replication, thereby preventing the dissemination of SIV across the entire mucosal compartment. On the other hand, by day 7 p.i., viral replication was already in the ramp-up phase, with the virus disseminating rapidly across the mucosa.
Though mechanisms other than productive viral infection may explain the near-total loss of mucosal CD4 T cells during ART, a higher level of preservation of memory CD4 T cells (Fig. ) in peripheral blood of these animals than in the mucosa (Fig. ) suggests that mucosal CD4 T cells are much more susceptible to loss than cells in the periphery. Thus, although the plasma viral loads at day 10 p.i. were essentially similar between treated and untreated animals, fewer peripheral blood memory CD4 T cells were infected and lost in treated animals.
In conclusion, our data show that initiating ART prior to peak viremia has a selective effect on CD4 T cells in peripheral tissues relative to the rectal mucosa. Mucosal CD4 T cells are more easily lost following infection, and minimal viral interaction is probably sufficient for the loss of these cells. Most likely, the highly activated microenvironment in which the cells reside contributes to this process. On the other hand, the ability of peripheral blood CD4 memory T cells to survive even in the presence of high viremia suggests that peripheral memory CD4 T cells have a much higher threshold for loss than their mucosal counterparts. These findings may explain why HAART is associated with immediate and lasting repopulation in the periphery compared to the lack of or transient repopulation in mucosal tissues (1
); the continuous low level of viral replication in the mucosa (8
) is likely sufficient to destroy the repopulating cells.