^{+}T Cell Reservoir for HIV in Patients on HAART

^{1}

^{1}

^{1}

^{2}

^{3}

^{1}

^{4}

^{*}

ARS, COW, and RFS conceived and designed the experiments and wrote the paper. ARS, JDS, and TPB performed the experiments. ARS, COW, and RFS analyzed the data.

The authors have declared that no competing interests exist.

Whereas cells productively infected with human immunodeficiency virus type 1 (HIV-1) decay rapidly in the setting of highly active antiretroviral therapy (HAART), latently infected resting CD4^{+} T cells decay very slowly, persisting for the lifetime of the patient and thus forming a stable reservoir for HIV-1. It has been suggested that the stability of the latent reservoir is due to low-level viral replication that continuously replenishes the reservoir despite HAART. Here, we offer the first quantitative study to our knowledge of inflow of newly infected cells into the latent reservoir due to viral replication in the setting of HAART. We make use of a previous observation that in some patients on HAART, the residual viremia is dominated by a predominant plasma clone (PPC) of HIV-1 not found in the latent reservoir. The unique sequence of the PPC serves as a functional label for new entries into the reservoir. We employ a simple mathematical model for the dynamics of the latent reservoir to constrain the inflow rate to between 0 and as few as 70 cells per day. The magnitude of the maximum daily inflow rate is small compared to the size of the latent reservoir, and therefore any inflow that occurs in patients on HAART is unlikely to significantly influence the decay rate of the reservoir. These results suggest that the stability of the latent reservoir is unlikely to arise from ongoing replication during HAART. Thus, intensification of standard HAART regimens should have minimal effects on the decay of the latent reservoir.

The discovery of a stable latent reservoir for human immunodeficiency virus type 1 (HIV-1) [^{+} T cells uncovered a major obstacle to curing HIV-1 infection and revealed limitations of previous analytical predictions concerning eradication [^{+} T cells and possibly other viral reservoirs, as well as some degree of active viral replication that continues despite HAART [

The mechanism underlying the stability of the latent reservoir remains unclear. Some investigators have argued that residual viral replication continuously reseeds the latent reservoir [^{+} T cells through de novo infection of cells that then enter the reservoir. The other major explanation for the persistence of the latent reservoir is that the stability arises from the intrinsic dynamic properties of the latently infected cells. Because the reservoir consists of resting memory T cells [

Curing HIV-1 infection will require elimination of the latent reservoir. It is therefore critical to understand which of these potential mechanisms are responsible for its stability. As a step in this direction, we have used mathematical modeling to understand the dynamics of the reservoir. Mathematical models have proven useful for analysis of several aspects of HIV-1 infection, including measurement of the turnover of different T cell subsets [

The procedures for obtaining and analyzing the sequences used in this study have been described in detail elsewhere [

To allow consistent amplification and sequencing of the small number of viral genomes present in the plasma of patients with viral loads below 50 copies/ml, plasma virus was first pelleted by ultracentrifugation, and then analyzed by limiting dilution reverse transcriptase (RT)-PCR, cloning, and sequencing using a previously described ultrasensitive genotyping method [^{+} T cell reservoir were analyzed by a novel limiting dilution PCR assay [^{+} T cells were purified from peripheral blood mononuclear cells by magnetic bead depletion as previously described [

As described previously, sequence analysis was carried out using techniques designed to avoid PCR resampling and PCR error [

The number of latently infected cells carrying replication-competent virus was quantified as previously described [^{+} T cells was determined by quantitative real-time PCR [

To evaluate replenishment quantitatively, we used a simple mathematical model to represent the dynamics of the latent reservoir in patients on HAART who had suppression of viremia to <50 copies/ml and whose residual viremia was largely composed of a PPC. Because only a small number of plasma virus sequences can be obtained from a given blood sample when the viral load is below 50 copies/ml, patients underwent intensive (every other day) plasma sampling over a 3-mo period, and data from this period of intensive sampling were pooled. Consistent with our sequencing data, we assumed that there were no latently infected cells containing PPC at the beginning of the 3-mo sampling period. Otherwise, we did not make any assumptions about the origin of the PPC. We make the conservative assumption that the PPC first appears at the beginning of the observation period (time ^{+} T cells [_{e}_{e}_{1}_{2}_{in}_{out}

Presence of a PPC in the Plasma and Resting CD4^{+} T Cells of Patients on HAART

We extend this simple model of reservoir dynamics so that we consider _{1}(_{2}(

Using _{e}

The fraction of latently infected cells at time _{PPC}

To be conservative, the initial number of latently infected cells containing PPC DNA (_{1}(0)) was set to zero for each patient. Because blood samples from time _{2}(0)) was determined by extrapolation from experimental measurements by limiting dilution PCR at various time points for each patient (

Initial Latent Reservoir Size, _{2}(0), under Different Reservoir Half-Life Assumptions for Patients on HAART

We used a maximum likelihood approach in order to find the value of _{in}_{out}_{in}_{in}_{in}

We report analyses of patients that yielded informative results for _{in}_{in}_{out}L

All simulations and calculations were performed with MATLAB version 7.2.0.232 (

In a previous report, we described a population of HIV-1-infected individuals on HAART who had suppression of viremia to less than 50 copies/ml for an average of 34 mo [^{+} T cells (from our previous work, we know that these sequences are a reasonable surrogate for rescuable virus in the same population of cells [^{+} T cells at baseline (^{+} T cells, entry of a substantial number of these plasma viruses into the latent reservoir at later time points could be readily detected. Therefore, we were able to employ this unique plasma virus population as a functional label for measuring the rate of replenishment for the latent reservoir in the setting of HAART.

(A and B) Results are shown for two patients, pt. 148 (A) and pt. 154 (B). The heterogeneity of the latent reservoir in comparison to the homogeneity of the plasma virus in these patients is represented with pie charts in which distinct genotypes are indicated in different colors. The PPC for each patient is shown in red. Intensive sampling of the plasma virus was carried out by sampling three times per week over a 3- to 4-mo period as indicated by the thin vertical marks on the time line. For pt. 154, additional samples of plasma virus (small circles) were obtained before and on several occasions after the period of intensive sampling. These document the persistence of the PPC for a minimum of approximately 900 d. Sampling of proviruses in resting CD4^{+} T cells was carried out before and on multiple occasions after the period of intensive plasma sampling. With one exception (*), all of the cellular sequences remained distinct from the PPC. The numbers below each circle represent the number of independent sequences analyzed. For pt. 154, the plasma samples after study day 900 were analyzed by RT-PCR of the

In order to find the _{in}_{in}_{out}_{out}L_{in}_{in}_{out}_{out}_{in}/L

(A) Regime 1; _{out}L_{in}

(B) Regime 2; _{out}L_{in}

(C) Regime 3; _{out}L_{in}_{out}L_{in}

For each time point when cellular sequences were obtained, the most likely _{in}_{in}_{in}_{in}_{in}_{in}

For the case when _{out}_{out}^{−1}, which reflects the previously reported 44-mo half-life of the latent reservoir. Initially, we assumed that the PPC was present for only the 90-d intensive sampling period despite the fact that two of the three patients maintained the PPC beyond this time. Under these assumptions, our maximum likelihood analysis found a most likely reservoir inflow rate of 0 cells/day, with an upper 95% confidence bound of 151 cells per day in pt. 135 (_{in}_{in}_{in}

Estimation of Patient-Specific, Maximum Reservoir Inflow Rates

Because the true decay rate of the latent reservoir remains controversial, we repeated the above analyses for _{out}^{−1} (_{in}_{in}_{in}_{in}

In our analysis above, we used an approximation of _{out}_{in}_{in}_{out}_{in}_{in}_{in}_{in}_{in}_{out}_{in}

It is also helpful to view the maximum absolute flow rate into the reservoir in the context of the overall reservoir size in each patient. We approximate the percentage of the total reservoir that daily reservoir inflow represents by normalizing the daily reservoir inflow (the upper bound of the 95% confidence intervals for each _{in}_{in}_{1}(0) + _{2}(0))) (_{in}_{in}

Patient-Specific Maximum Daily Reservoir Inflow as a Percent of the Total Latent Reservoir Size

We also consider whether our predicted maximum daily inflow of cells into the latent reservoir during HAART reflects a reduction from the predicted pre-HAART inflow. We have previously described how long each patient in this study had consistently suppressed viremia (79 mo for pt. 135, 35 mo for pt. 148, and 17 mo for pt. 154 from _{out}^{−1}), we are able to back calculate the most recent size of each patient's pre-therapy latent reservoir, _{in}^{+} T cell reservoir. These fold reductions are several orders of magnitude larger if we assume a 6-mo half-life. These results indicate that HAART can drastically reduce the flow of cells into the resting CD4^{+} T cell reservoir.

The latent reservoir for HIV-1 in resting CD4^{+} T cells is the primary known barrier to eradication of HIV-1 infection. Therefore, eradication of HIV-1 infection depends on successful purging of the latent reservoir from an infected individual. Unfortunately, experimental evidence has shown that the latent reservoir is highly stable. Whereas the half-lives of other types of infected cells in the setting of HAART are on the order of days to weeks, the half-life of the latent reservoir has been reported to be on the order of months to years [

The basis for the stability of the latent reservoir is controversial. Some reports have shown an increase in the reservoir decay rate with an intensified HAART regimen [^{+} T cells, which are inherently long-lived cells, some have hypothesized that the latent reservoir's longevity stems from its intrinsic stability [^{+} T cells. It is difficult to find direct, experimental support for either argument, however, due to the lack of a readily accessible experimental model.

Opportunities do arise when patient-derived data may be used to gain unique insights into the latent reservoir. In this study, we offer a quantitative glimpse into the replenishment of the latent reservoir in the setting of HAART. Although several studies have suggested that reservoir replenishment might occur during HAART [^{+} T cells and subsequently following that label. In this study we take advantage of a previously reported phenomenon where a unique, patient-specific viral sequence (PPC) dominated the residual plasma virus but could not be readily found in the patient's activated or resting CD4^{+} T cells [

Our model was constructed with as few assumptions as possible regarding the nature of reservoir dynamics. Our model does rely on the assumption that the PPC is replication competent, or at least capable of infecting and integrating into the genome of a CD4^{+} T cell. Our previous study strongly suggested that each patient's PPC was not different than other patient-specific plasma virus sequences in functionality (by direct examination of ^{+} T cells [^{+} T cells is due to infection of that cell after initiation of HAART. Clearly this is not the only possibility, as a PPC sequence may have entered a resting CD4^{+} T cell before HAART was initiated.

By applying this approach to data derived from three independent HIV-1-infected individuals on HAART who exhibited a PPC, we have been able to quite conservatively constrain the replenishment rate of the latent reservoir to be at most on the order of 100 cells carrying replication-competent virus per day. Given that the average size of the latent reservoir is approximately one million cells [^{+} patient on HAART. Our results predict a substantial reduction in the reservoir inflow in the setting of HAART compared to pre-HAART levels. While we could not demonstrate a drastic reduction in reservoir inflow for pt. 148 due to the limited number of available sequences, we were able to show in other patients that HAART reduces the reservoir inflow by at least 10- to 20-fold from pre-HAART levels. Given pt. 154′s treatment history of frequent blips suggestive of low-level viral replication, that we are able to predict a ∼20-fold reduction in reservoir replenishment suggests that HAART would reduce the reservoir replenishment rate by even more in patients like pt. 135 and pt. 148, who exhibit no signs of potential viral replication. Subsequent analyses and procurement of additional sequence data may allow us to reduce the maximum replenishment rate of the reservoir even further. Our present results, however, do not establish whether or not there is any replenishment of the latent reservoir by low-level viral replication in the setting of HAART. Our analysis uses patient-derived data to conservatively constrain the

Of the three patients in our study, we detected the PPC in the resting CD4^{+} T cell compartment of one. We believe that the infrequent detection of PPC in resting CD4^{+} T cells of our study participants reflects the following: 1) few resting CD4^{+} T cells contain PPCs, and 2) replenishment of the latent reservoir in the setting of HAART must be slow, which is consistent with our results. Furthermore, while we can only infer a maximum daily inflow into the latent reservoir (since we cannot sequence the entire latent reservoir, but rather only a sample of the latent reservoir at each time point), the actual replenishment rate may easily be much lower than our calculated maximum replenishment rate and may even be zero. Finally, because the data have constrained the maximum reservoir inflow rate to be small compared to the total reservoir size, it may be that the flow of new cells into the reservoir does not significantly affect the decay rate of the latent reservoir in these patients (regime 1 versus regimes 2 and 3 described above).

The finding that the daily inflow into the reservoir is small compared to the overall reservoir size suggests that the decay of the reservoir in our patients (who have all been on HAART for several years) is more likely determined by _{out}_{in}_{in}_{out}

The results of our analysis are important on a practical level. It has been suggested that intensification of HAART may stop residual viral replication in the setting of standard HAART and increase the decay rate of the latent reservoir [

(56 KB DOC)

(21 KB DOC)

highly active antiretroviral therapy

human immunodeficiency virus

predominant plasma clone

reverse transcriptase