8th Annual Retrovirus Conference
Viral Reservoirs (latency/persistance): blind men describing an elephant
Written for NATAP by David Margolis, MD, Associate Professor, University of Texas Southwestern Medical Center
--HIV & Resistance in CD4 memory cells
--Ongoing viral replication
--Is 5-drug therapy useful
--Reservoirs: mucosa, CNS, semen
This is a complicated topic & report but I will try to summarize it. I recommend reading the full report below. Robert Siliciano talked about two sources of persistance for HIV in the human body: ongoing viral replication & the latent reservoir in the resting memory CD4 cells. Because memory cells are designed to last a lifetime, eradication is unlikely with the current drugs. Siliciano reports it could take 73 years of therapy to eradicate this reservoir. He suggests that ongoing replication in these memory cd4s result in release of drug sensitive & drug resistant viruses stored in the memory CD4 cell. The drug resistant viruses are from stored resistant viruses from before HAART. He concluded that this suggests that despite stored resistance viral load suppression can be maintained. This is an important point and other studies suggest this. Despite resistance being present from previous therapy, if a person is fully adherent and doesn't get resistance to their current HAART regimen, they can stay undetectable for a long time. Andrew Phillips from London reported at the Resistance Workshop that if a person remains <50 copies/ml for the first two years of their therapy he estimates they ought to remain undetectable for 10 years if they are adherent.
In contrast, David Ho continues to say that ongoing replication results from incomplete viral suppression, which reseeds the reservoir. Ho reported using a potent 4-drug regimen (Kaletra+PMPA+efavirenz+3TC and found the initial decay in viral load was accelerated compared to less potent therapy. Previous data suggests, although this has never been proven, quicker initial decay in viral load may lead to more durable viral suppression. It's also Ho's feeling that this may be the way to prevent virus from storing in the reservoir of memory CD4s.
Finally, there are 3 reports below on reservoirs in mucosa, CNS and semen. It appears that HIV can persist in mucosa & semen despite undetectable HIV in blood. In the CSF, HIV may have similar or different resistant patterns than HIV in the blood. The CSF appears to be a compartment or reservoir where HIV replicates differently than in the blood. This is still not well understood. The good news is that despite these reservoirs, viral load cam remain undetectable if a person is fully adherent. It in fact appears that these reservoirs do not cause viral load rebound, so far 6 years out with HAART. Researcher's judgements I think are that these reservoirs ought not to lead to viral rebound but we need to follow patients on therapy for longer.
The Full Report from David Margolis
As well outlined by Robert Siliciano in a plenary lecture at the outset of the conference, reservoirs of HIV infection are an important obstacle to long-term effective therapy. Current models suggest there are 2 potential sources of persistent HIV infection in the face of HAART. The first is low-level continued replication that continues despite HAART. Even in patients who are rigidly adherent, ongoing replication may occur because of inadequate antiviral potency, or inadequate drug penetration to selected anatomic sites and reservoirs.
The second cause for HIV persistence is the reservoir of replication-competent proviral HIV DNA. This exists in two locations. The first is within the cytoplasm of resting CD4 cells, where HIV genomes may survive for several days to a few weeks before being degraded by cellular nucleases. More significant is the second reservoir of integrated proviral genomes persisting within the DNA of resting memory CD4 T cells.
As outlined by Dr. Siliciano, this reservoir in resting, memory CD4 T cells is vexing, as these cells are designed to last a long time, perhaps for the life of an individual. Longitudinal studies from his laboratory have found that the frequency of such infected, resting CD4 cells is about 1 in 1 million, and the half life of these cells on continuous HAART is 44 months, requiring 73 years of therapy for eradication. These studies now have follow-up of up to 5 years, and range of error of the measurements of decay is such that the true rate may actually be zero. The decay rate of this reservoir appears to be similar in the few children studied, but may be faster in some patients (such as those treated early after infection).
Dr. Siliciano discussed two mechanisms that could preserve this stable reservoir. In the first, proviral DNA would remain quiescent within resting cells, despite rare cell division of these cells. Cell division of resting memory CD4 cells has been proposed as a mechanism through which immunological memory is preserved. HIV would then remain as a silent passenger within these cells.
Alternatively, the reservoir could be refueled by low level replication of HIV. Dr. Siliciano discussed two type of studies which have suggested ongoing replication occurs. The first is the PCR detection of circular HIV DNA species called 2-LTR circles. 2-LTR circles are produced when genomic HIV-1 RNA is copied into DNA but does not integrate. It has been said that 2-LTR circles decay within few days, and so detection of 2-LTR circles is a marker of recent HIV-1 infection and replication.
Siliciano presented data suggesting that 2-LTR circles are stable, and that their decay reflects dilution due to cell division.
Siliciano then reported the findings of an ultrasensitive genotypic analysis of patients with <50 copies/ml of plasma HIV-1 RNA. Sequence evolution of proviral DNA within cells is the second type of study which has been presented as evidence of ongoing replication. Siliciano's ultra-sensitive assays find a continued release, for as long as 48 months, of archival drug-sensitive viruses devoid of new resistance mutations selected under HAART. Resistance mutations, when observed, reflect prior therapy with non-suppressive one or two drug regimens. Siliciano concluded that low-level viremia in patients on HAART with <50 copies/ml of HIV-1 RNA results primarily from release of archival, pre-HAART viruses rather than new, HAART-selected, partially resistant mutants. The results indicate that long-term suppression on HAART may be possible. An alternate, but not mutually exclusive, explanation of this data would be that these viral species originate in cell types or anatomic locations where antiretroviral drug levels are too low to exert selective pressure.
A poster from Pat Bucy's laboratory (#385) was consistent with the findings of Siliciano's ultrasensitive genotype studies. The Bucy group studied viral DNA sequences and absolute vDNA quantity in PBMC obtained from two individual patients. Each had been treated with nevirapine (NVP) for several months in a monotherapy trial. NVP resistance mutations were quickly selected in the plasma pool, and more slowly in the proviral DNA pool. The subjects were then treated with HAART and HIV RNA suppressed below the limits of detection. The total amount of vDNA declined significantly, but the proportion of NVP resistance mutations decreased to about 30% of the pool of sequences that remained after suppression of active viral replication. Bucy concluded that the active quasispecies appear to contribute to the stable, latently infected pool of cells, but slowly, and that the latent pool reflect primarily an archive of viruses from early in infection (i.e. wild-type, drug-sensitive HIV).
Impotent HAART: low level replication despite therapy
David Ho (symposia #17) discussed his group's studies that point to incomplete viral suppression as the primary cause of HIV persistence. His group has previously reported measurements of decay rates of the latent reservoir, significantly higher that those measured by Siliciano, that are inversely proportional to the degree of the residual HIV-1 replication. Ho contends that low level replication continuously reseed the latent reservoir, and so that the apparent rate of decay of this reservoir might be zero.
In this lecture Ho outlined studies presented in detail in poster #383. Using a 4-drug combination, nineteen patients were treated with lopinavir/ritonavir, efavirenz, tenofovir, and lamivudine for a mean of 45 days (range 7-70). This combination was expected to be more potent than "standard" 3- or 4-drug HAART. All patients were either drug-naïve or had drug-sensitive HIV-1, and the baseline viral loads and CD4 levels were comparable in the 3 groups. The initial decay rate of plasma viremia was compared.
Frequent measurements showed that the first phase of viral decline with the highly potent regimen was faster than that seen with "standard" drug regimens.
The mean log change in viral load from day 0 to day 7 was 1.59+0.12, compared to 1.14+0.10 for the other 4-drug regimen and 1.32+0.13 for the 3-drug regimen. While this may not seem significant for those used to assessing decreases in viral load by half log increments, this 0.3-0.45 log difference is quite substantial as it acccrues over 1 week. Correspondingly, the decay slope for this phase was -0.57+0.06/d, compared to -0.41+0.03/d and -0.45+0.04/ d.
These measurements allowed Dr. Ho to calculate the relative potency of the regimens. If the new 4-drug regimen was said to be 100% potent, the saquinavir, ritonavir, zidovudine, and lamivudine regimen would be72% potent, and nelfinavir, zidovudine, and lamivudine 78% potent. Dr. Ho suggested that substantial improvements in drug potency should be sought, and might "drain" the latent reservoir of HIV infection. Longitudinal studies of the kinetics of the infected resting cell pool in these patients will be of great interest.
(Comments from Jules Levin: even if the 5-drug regimen doesn't drain reservoirs, it could be more effective in reaching <50 copies/ml more quickly. Data suggests reaching <50 more quickly results in more durable suppression).
That special hidden place: anatomic reservoirs
Mucosa. Several posters presentations discussed the detection of HIV in anatomical compartments. A poster by Elliot (#388) measured HIV DNA and RNA in the GI mucosal tissue. 9 subjects receiving HAART for >2 years and having undetectable plasma HIV-1 RNA <50 copies/ml for at least 3 months were studied. HIV RNA and DNA were extracted from recto-sigmoid biopsies, and quantified using a sensitive (3 copies/reaction) PCR assay. 3 subjects had blips of plasma HIV RNA <600 copies. Mucosal HIV-1 DNA was detectable in 9/9 subjects at all time points with no evidence of decay over 15 months. This is not surprising in comparison with previous studies of tissue. Of note, however, mucosal HIV-1 RNA was undetectable at all time points in 4/9 subjects, intermittently detectable in 2/9 and present at all time points in 3/9. It was unclear, however, whether RNA detected was recovered from mucosal epithelial cells themselves, or from trafficing PBMCs.
Seminal fluid & cells. The Pomerantz group (poster #391) reported findings in a different compartment. They studied PBMCs and seminal cells from a cohort of HIV-1-positive individuals on HAART with <50 copies/ml of viral RNA in plasma. Ultra-sensitive assays have previously detected HIV RNA in seminal fluid and cells. Viral out-growth assays were performed, as well as quantitative PCR to detect 2-LTR circles. Viral out-growth from PBMCs in vitro directly correlated with in vivo 2-LTR circular DNA levels in PBMCS in the thirty patients that were evaluated. Surprisingly, 2-LTR circular DNA was not detected in any semen sample, even when replication-competent HIV-1 virus had been recovered from semen cells by viral outgrowth co-culture assays. This suggested ongoing, low-level replication at levels of <50 copies/ml in PBMCs, but persistence of replication-competent HIV without ongoing replication in seminal cells.
Central Nervous System. Ronald Ellis (Symposia #19) discussed HIV replication and persistence in the CNS. Both concordant and discordant evolution of viral genetic sequences has been documented in the CSF when compared to plasma. Envelope sequences may diverge in the CSF due to sequence requirements for entry and replication into the CNS. Polymerase sequences may lack drug resistance mutations when isolated from the CSF, as antiretrovirals may penetrate the CSF poorly. Particularly in advanced AIDS, HIV may replicate autonomously in the CNS, independent of reinfection by HIV in the blood. The level of HIV RNA in the CSF correlates less and less with the level of plasma HIV RNA as disease progresses, but the cause of this is unclear (poster #616). The efficacy of antiretrovirals in the CSF may be best reflected by a ratio of CSF conc/IC50 for a given drug. Measuring the IC50 of HIV in the CSF of individual patients may be clinically useful in predicting CSF viral response (poster #614).