HIV persistence, raltegravir/maraviroc intensification and immunology
David Margolis MD, University of North Carolina, for NATAP
Persistent HIV infection despite successful ART
Ultimately, we must either acquiesce to treat HIV patients with ART forever, or find ways to definitively prevent infection and to decisively eradicate infection. Tony Fauci, director of the NIHs infectious diseases institute highlighted these priorities in his keynote address. 
Notable at CROI this year were several exhaustive and demanding studies of persistent infection. These were emblematic of the types of studies that are required to make progress in understanding the persistence of HIV infection despite ART, and required if the field is ever going to make progress towards eradication or drug-free control of HIV infection. Both the investigators, and most especially the patient volunteers, are to be applauded for their efforts and contributions.
Mary Kearney co-workers at the Frederick Drug Resistance Program presented very important observations on viral evolution during the initiation of ART. 
When a patient initiates ART, suppression of viremia to <50 copies/mL takes many weeks to achieve. It has been unknown to what extent drug resistant populations of virus could develop or expand during this period of time. To carefully characterise genetic diversity and divergence in patients before and during suppressive ART, Kearney and colleagues studied samples from 10 HIV-positive patients taken from the time of ART initiation and thereafter for as long as 5 years. A total of 1300 HIV-1 sequences (gag-pro-pol from p6 to RT) sequences were obtained by single-genome sequencing (SGS). Diversity was measured by average pair-wise difference (APD), genetic variation over time was assessed by phylogenetic analyses and a test for panmixia (the state of maximal mixing of all possible genetic sequences), and sequence changes were characterized using a comparison program called Highlighter.
Before ART, circulating HIV-1 sequences had differences of 0.2 to 2.5% per site. This is consistent with the work of this group and others, showing a relatively homogeneous pool of sequences at a moment in time, with episodic evolution of the circulating swarm in response to immunological pressure. In this group, ART reduced plasma viral RNA levels to undetectable (<75 copies/mL) in all patients within 5 months of initiation. In 8 of 10 patients, phylogenetic analyses and measurements of intra-patient APD revealed no change in viral diversity or population structure between pre- and post-ART samples despite up to 4-log10 decreases in HIV-1 RNA levels. That is to say, that after a 2 log, and a 3 log decline of viral load the residual viral swarm did not change significantly, and no novel viruses were selected or emerged from underneath the majority population. This was still the case when viral sequences were recovered and studied when HIV RNA was <75 copies/mL.
In two patients, divergence of HIV-1 was eventually evident after prolonged ART, resulting in a shift from pre-therapy virus populations containing occasional unique viral variants within a population of a dominant species, to a population with frequent G to A mutations, stop codons, and shifts in CTL escape mutation profiles. In one of these patients, a variant found one day after initiating ART became a predominant plasma clone after fours years on suppressive therapy, similar to that described by Bailey et al. (J Virology 2006). In this patient the predominant plasma clone was significantly different (p<10-9) from the pre-therapy population by tests for panmixia and divergence (2.9%), resulting primarily from accumulation of G to A mutations.
These results suggest that both short- and long-lived cells are infected with diverse virus populations before therapy, but that viral replication is completely blocked by ART as no evolution of virus is observed during ART-induced decline of viremia. In some patients (as observed in the one patient described above) diverse replication-competent viral variants may decay with long-term suppressive therapy, leaving behind only defective proviruses produced by remaining clones of previously infected cells. This is an alternate or additional explanation for the predominant plasma virus clone described by Bailey, which is hypothesised by some to originate instead from an infected blood stem cell.
Steven Yukl exhaustively examined the HIV burden throughout the gut-associated lymphoid tissue (GALT).  He and his co-workers measured HIV DNA (cells that had been infected with HIV in the past) and HIV RNA (produced by infected cells that were potentially expressing HIV viral particles) in different area of the GI tract in patients on prolonged, suppressive ART. Eight HIV-positive patients on ART with CD4 counts >200 cells/mm3 (mean 480 cells/mm3) and plasma RNA viral load <40 copies/mL for 2.8 to 12 years were studied. Blood plasma, PBMC, and 710 endoscopic biopsies were taken from sites in the duodenum, terminal ileum, right colon, and rectum.
Using a sensitive research assay, plasma HIV RNA was detectable at very low levels in all patients (median 2.3 copies/mL). Unspliced HIV RNA was detectable in each gut site in the majority (63 to 88%) of patients using RT PCR, but levels were very low. So it is unclear if this HIV RNA represents tiny amounts of RNA produced by many cells, or large amounts of RNA produced by few cells. The latter would seem more likely, and consistent with the occasional ability to visualise HIV RNA by in situ hybridisation. Of course, the presence of HIV RNA does not guarantee that the cells involved are actually producing replication-competent viral particles, although for at least some of these cells that seems also likely to be the case.
Surprisingly, HIV DNA increased from the duodenum to the rectum, and the HIV DNA per CD4+ T cell was higher in all four anatomic regions relative to the PBMCs. Again, it is important to remember that the presence of HIV DNA means only that the virus has entered the involved cell at some time in the past, and does not prove that the cell is producing HIV RNA, or even competent viral particles. It does also not demonstrate that the involved cell was actually infected in the gut site from which the cell was recovered, as immune cells are frequently trafficking in tissues. Biopsies of the GALT, like any assay, are only snapshots in time. The terms persistence and reservoir are temporal and spatial ones, and these observations neither demonstrate that cells stably expressing virus persist in the GALT, or that the events that lead to persistent viral RNA production (and likely virion production) are occurring exclusively in the GALT. But we must measure what is possible to be measured.
Yukl and colleagues noted that the median unspliced HIV RNA was also higher in all gut sites compared to PBMCs, but oppositely was highest in the ileum and lowest in the rectum. Also surprising to the authors was the finding that cell activation markers were lower when HIV DNA levels in the gut tissue were higher. The authors concluded that The inverse relationship between HIV DNA and T cell activation in the gut and the paradoxically low levels of HIV expression in the large bowel suggest that different processes drive HIV persistence in the blood and gut.
This study is hopefully the first part of a longitudinal one, as such a study could tell us much of the processes that drive persistence. Persistence can be defined as:
- persistent infection without producer cell death,
- persistent rounds of infection with one cell passing new virus to another, or
- carriage of proviral DNA that can later express replication-competent HIV.
A conclusions from this study is only that cells that are expressing HIV RNA are found more frequently in the ileum. HIV DNA that predominates in the rectal tissue (also the tissue with the lowest levels of immune activation) could be, at least in part, a graveyard of proviral sequences that are incompetent for expression, in cells that were either infected there, or travelled there.
Ann Wiegand presented the results of an NIH intramural group study that found no reduction of persistent, low-level viremia as measured by the single-copy assay (SCA; Palmer J Clin. Micro. 2004) in treatment-experienced who intensified their ART with raltegravir (RAL). 
Eight participants had undergone an average of 4 suboptimal regimens with a prior history of virologic failure and genotypic resistance, but who had subsequently suppressed to <75 copies/mL were enrolled. Patients had baseline viral RNA levels determined during a 21-day period prior to RAL intensification, then weekly assays during a 30-day intensification period with raltegravir 400 mg twice daily, and then later additional sampling for 6 weeks after RAL was stopped. There was no significant change in viral RNA levels during intensification or afterwards. Some intensification studies have suggested a CD4 benefit in the absence of a change in viremia, but in this one mean CD4 cell numbers were not significantly different after 30 days of intensification.
Hatano reported the results of a similar study at UCSF in which 15 subjects with undetectable viral loads on ART for at least 1 year were randomised to add RAL 400mg twice daily, and 15 added placebo for 24 weeks. 
This was a group with a long disease history: the duration of HIV infection was 18 years, duration of HAART was 23 months, baseline CD4+ T cell count was 232 cells/mm3, and nadir CD4+ T cell count was 53 cells/mm3. Using a SCA with a lower limit of detection of <0.2 copies/mL, median baseline plasma HIV RNA level was 5.2 copies/mL; 9 subjects were below the limit of the assay. The proportion of subjects with undetectable plasma RNA levels at week 12 was not different across the 2 groups. Also, RAL intensification did not alter proviral DNA, cell associated HIV RNA, or CD8+ or CD4+ T cell activation in the blood. Further, 20 patients underwent GALT biopsies, and GALT CD8+ and CD4+ T cell activation was unaffected in the GALT by RAL. In addition, intensification did not significantly affect gag-specific responses in blood or GALT. However, higher levels of gag-specific IL2+INF-gamma CD4+ T cells and CD8+ T cells in GALT were associated with lower levels of cell associated RNA (rho = -0.52, p=0.02 for CD4+ T cell responses; rho = -0.53, p=0.04 for CD8+ T cell responses). This association held regardless of receipt of RAL or placebo, and it is hard to be sure if the association is a marker for patients with better control of tissue viral replication, or whether the responses themselves actually contribute to control of replication.
In a companion piece to this study, Yukl and colleagues also presented a study of the effect raltegravir intensification on HIV RNA and T cell activation in the GALT in patients of suppressive ART. 
In this study, 7 HIV-positive men with viral load <40 copies/mL for 3 to 12 years and a CD4 count >200, underwent a variety of 12-week intensifications: raltegravir (n = 4), raltegravir and efavirenz (n = 2), or raltegravir and darunavir/ritonavir (n = 1). Again, GALT biopsies were done in 4 sites (duodenum, ileum, colon, and rectum) at entry and week 12.
As before, HIV RNA was detectable in plasma and PBMCs. HIV DNA was detectable in all GALT sites, and RNA detected in the GALT in most patients. Intensification resulted in no consistent change in HIV RNA in the plasma, peripheral blood mononuclear cells, or gut. There was a trend (in 5 of 7) towards decreased unspliced HIV RNA per 106 CD4+ T cells in the ileum, and similarly a trend towards decreased activation of CD4+ and CD8+ T cells in all GALT sites. The authors concluded that intensification reduced HIV RNA, reduced immune activation, and increased CD4+ T cells in the ileum, suggesting that the ileum may support ongoing productive infection in some patients on ART, even if the contribution to plasma RNA is not discernible. However, like their other study, this study should be followed-up with longitudinal evaluations. Most of the confidence intervals for the effects were wide, and not all the effects were concordant in each part of the GI tract.
Javier Martinez-Picado from Barcelona presented a third raltegravir intensification study actually initially presented last year at CROI and now published in Nature Medicine (March 2010). 
69 patients with <50 HIV RNA copies/mL for >1 year were randomised to intensify their ART with raltegravir (n = 45), or to continue their HAART (n = 24) for 48 weeks. As discussed last year, raltegravir intensification resulted in no change or difference in total HIV DNA. This year it was also shown that SCA viremia was unchanged. However raltegravir intensification of a 3-drug suppressive ART regimen resulted in a specific and transient increase in 2-LTR circle DNA in a significant percentage (29%) of ART-suppressed patients, primarily those on therapy that included a PI rather than an NNRTI. This was the same data as shown on a poster at CROI in 2009, but all the individual patient data points were shown. There was no doubt that in the subgroup of patients in the study who had an increase in circles, most followed a similar pattern of increase at week 2 and/or 4, then a decrease to baseline. It should be noted that the levels in most of the 69 patients in the study were <1 copy of DNA circle per million cells, and that in some of the few patients with LTR circles the values were very low and increased little (eg. increase from <1 copies to 6 copies, then decreased to 2 copies) and in others was more variable (eg. 40 copies at baseline, increased to 70 copies at week 2, then <1 copy at week 4).
Interestingly, patients who were 2-LTR+ showed higher levels of immune activation (HLA-DR+CD38+, CD38+CD45RO+, HLA DR+CD45RO+ in CD8 T-cells) at baseline and a decline after intensification. These immune changes appeared quite modest but real, but it was unclear what mechanism induced them. Remember, levels of viremia (by SCA) did not change with intensification, and so the effect of RAL did not reduce the amount of circulating viral antigen – the typical mechanism by which we believe that ART dampens immune activation. The group interpreted their findings as evidence of ongoing replication, but it is important to be picky about terminology here. Replication means full rounds of the HIV life cycle, which on ART would generally be expected to induce drug resistance. Although this could be happening at a level that is somehow too low to select for drug resistance, there is no evidence for full rounds of replication here. There is evidence for expression of virus (as measured by SCA in some patients), and evidence that RAL had its effect by blocking integration and increasing 2-LTR circles. It would be interesting to know how the individual SCA levels related to the individual 2-LTR circle levels, but this was not shown.
de Laugerre reported on another RAL intensification study for the EASIER-ANRS 138 study team. 
There were not obvious technical differences in the implementation of the assays used, but the French group reported No evolution of HIV-1 total DNA and 2-LTR circles after 48 weeks of raltegravir-containing therapy in patients with controlled viremia a self-evident title and conclusion. However, the French group did not make measurements at the 2-4 week timepoint after intensification, the time when increased levels of circles were seen by the Spanish.
Several studies examined the effect of the CCR5 antagonist and HIV entry inhibitor maraviroc, as an alternate intensification agent in place of raltegravir. Wilkin reported the results of ACTG 5256 in which patients with suboptimal CD4 recovery after ART added MVC for 24 weeks. 
This was a single-arm pilot trial that enrolled patients with a CD4 count <250 celss/mm3, whose CD4 counts were stable despite continued ART such that their calculated CD4 slope was between -20 and +20 cells/mm3/year, despite at least 2 years of undetectable plasma HIV-1 RNA. 34 subjects enrolled in this study. The median baseline CD4 count was 153 cells/mm3 and despite maraviroc intensification a CD4 count increase of > 20 cells/mm3 was not seen. The mean increase in CD4 count to was 11 cells/mm3. Despite this lack of enhanced CD4 recovery, modest decreases in activation (%CD38+ cells, or % HLA-DR+/CD38+ cells) were seen.
Evering and colleagues enrolled patients infected with CCR5-tropic HIV-1 and treated with ART during acute, early infection. 
Subjects received ART for an average of 4 years prior to study entry. 4 patients intensified with maraviroc for 24 weeks; 2 patients intensified their NRTI for 12 weeks, followed by crossover to maraviroc for 12 weeks. Phlebotomy and flexible sigmoidoscopy with mucosal biopsies were performed at entry, weeks 12 and 24. HIV RNA was generally undetectable in tissue. In contrast to other studies, levels of immune activation and CD4+ T cell depletion in the GALT persist despite maraviroc intensification. And in the small sample thus far, no statistically significant effect of intensification of ART with MVC on a variety of immunologic and virologic parameters in the GALT is seen.
Carolina Guti*ez presented a difficult study for the group from Ram* Cajal in Madrid. 
Again, stably suppressed patients intensified ART with maraviroc, this time for 12 weeks. In this study latently infected resting CD4 T cells were quantified using a limiting dilution co-culture assay. Residual viremia was measured by quantitative real-time RT-PCR assay (Single Copy Assay, SCA, threshold: 0.3 copies/mL), episomal 2-LTRs DNA in peripheral blood mononuclear cells was measured, and activation measured (HLA-DR and CD38 on CD4 and CD8 cells).
In nine patients studied so far, at baseline, the reservoir could be quantified in 6 patients (mean 2.04 infectious units per million (IUPM)). After 12 weeks of maraviroc intensification, all patients maintained viral load <50 copies/mL and a decrease in the latent reservoir was observed in 5 patients, while no decrease was found in one (mean 0.08 IUPM, P =0.048 compared to baseline). SCA increased in several patients, but not >50 copies/mL, and episomal 2-LTRs DNA were undetectable in the 9 patients at baseline and became detectable in 4 of them at week 12. HLA-DR and CD38 decrease 2.9% on CD4s (p=0.003) and not significantly on CD8s.
So unexpectedly, there was an apparent depletion of the resting cell reservoir, with a paradoxical increase in SCA, 2-LTR circles, but a decrease in activation. These findings appear discordant, and are hard to reconcile. However, most of the effects seen are of marginal magnitude, and so if possible longer-term study might help to clear up these confusing results.
Is there a summary?
Overall, intensification of durable, suppressive ART appears to have no effect on low-level plasma viremia. In some but not all studies modest effects can be measured such as small reductions of HIV RNA in tissue, or HIV DNA species in cells, or of cellular activation markers in tissue or circulating cells, and in one study resting CD4 cell infection frequency. It does appear that drug intensification is perturbing some equilibrium in the virus-infected patient, but the mechanism is not clear. I would propose that RAL and MVC are exerting a subtle, uncharacterized, non-virological effect resulting in the cellular changes measured. Repeated measures, and interventions might help clarify the mechanism of these phenomenon.
Tired T cells
Michael Lederman and his Bad Boys of Cleveland looked for correlates to understand why a large minority (ca. 25%) of patients who are successfully treated with ART are not able to enjoy an increase of CD4+ T cells into the normal range. 
The Cleveland failure project, perhaps an accidental reference to place called by some the mistake on the Lake (sorry, couldnt resist), studied local patients who has been on ART for more than 2 years, and had viremia suppressed to <50 copies/mL for at least 2 years. Patients within incomplete immune restoration were those that had CD4 <350 cells/mm3 and they were compared to those with >500 cells/mm3 (ignoring those in the middle, 350-500, for clarity of analysis).
Ledermans group studied 61 patients who met the failure criteria, identified 168 who met the criteria for success, and studied 20 of these lucky patients. In multivariate analysis, only nadir CD4 was independently associated with immune failure (p <0.001). Failures were more predominantly male (82% vs. 70% for immune restoration patients), and slightly older (ca. 4 years older on average). Of note CD8 cells were 2-fold higher in those with immune restoration. Of the subpopulations of CD4+ cellsnaive cells, central memory cells, and effector memory cellsall were lower in the patients without full CD4 cell restoration. So failure was not specific to a CD4 cell subpopulation.
The activation markers CD38 and HLA-DR on both CD4 cells and CD8 cells tended to be somewhat higher in failure patients. Ki67, a marker of recently proliferating/cycling cells was higher in CD4, especially central and effector memory cells, but not in CD8 cells. This was one of the sentinel findings of the study, which Lederman entitled: Immune failure after suppressive ART: high level CD4 and CD8 T cell activation but only memory CD4 cells are cycling. This finding suggested that failure to fully recover the CD4 cell count might be related to a greater need to expand the CD4 cell pool to replace cells, and an inability to completely accomplish that task, despite apparently equivalent viral suppression. This finding is consistent with the observation that nadir CD4 was most strongly associated with immunological failure, and adds another reason to consider ART sooner rather than later. An additional useful tidbit was the suggestion that a lower CD8 cell count might portend poorer CD4 recovery, and if validated might also be used as a trigger to initiate ART earlier, or implement an immunotherapy strategy to boost CD4 cells (if we can find one that really works).
This report is an edited version of a longer report on the NATAP website.
Unless stated otherwise, all references are to the Programme and Abstracts of the 17th Conference on Retroviruses and Opportunistic Infections. 16-10 February 2010, San Francisco. All oral abstracts are available as webcasts.
1. Tony Fauci lecture. Pre-meeting plenary. Webcast: The HIV/AIDS Research Agenda. Tuesday 5.30pm.
2. Kearney M et al. The genetic diversity of HIV-1 in plasma persists despite suppression with ART. Oral abstract 98.
3. Yuki S et al. Differences in HIV burden throughout the gut of patients on suppressive ART: implications for HIV persistence. Oral abs 97.
4. Wiegand A el al. Raltegravir intensification does not reduce persistent HIV-1 viremia in treatment-experienced patients. Oral abs 280.
5. Hatano H et al. Raltegravir intensification in ARV-treated patients exhibiting a suboptimal CD4+ T cell response. Oral abstract 101LB.
6. Yukl S et a. Effect of raltegravir-containing intensification on HIV burden and T cell activation in the gut of HIV+ adults on suppressive ART. Poster abstract 279.
7. Buzon MJ et al. HIV-1 replication and immune dynamics are impacted by raltegravir intensification of HAART-suppressed Patients. Oral abstract 100LB.
8. de Laugerre C et al. No evolution of HIV-1 total DNA and 2-LTR circles after 48 weeks of raltegravir-containing therapy in patients with controlled viremia: a sub-study of the randomized EASIER-ANRS 138 trial. Poster abstract 281.
9. Wilkin T et al. Maraviroc intensification for suboptimal CD4+ cell response despite sustained virologic suppression: ACTG 5256. Poster abstract 285.
10. Evering T et al. The antiviral and immunological effects of intensification of suppressive ART with maraviroc, a CCR5 antagonist. Poster abstract 283.
11. Guti豲ez C et al. Effect of the intensification with a CCR5 antagonist on the decay of the HIV-1 latent reservoir and residual viremia. Poster abstract 284.
12. Lederman M et al. Immune failure after suppressive ART: high level CD4 and CD8 T cell activation but only memory CD4 cells are cycling. Oral abstract 47LB.