Report and Commentary from the Fifth International Workshop on HIV Persistence During Therapy

By Richard Jefferys

February 2012 – Inaugurated in 2003, the bi-annual International Workshop on HIV Persistence during Therapy (aka “the persistence workshop”) is the brainchild of researcher Alain Lafeuillade. The meeting presaged the recent explosion of interest in pursuing a cure for HIV infection, a pursuit many had considered quixotic until the case of Timothy Brown came to light in 2008. As has been extensively documented, Brown’s apparent cure resulted from a debilitating odyssey of treatments required for the grim diagnosis of acute myelogenous leukemia, enhanced with a mix of insight and good fortune on the part of his doctor Gero Hütter, who was able to provide a stem cell transplant from a donor lacking the major HIV co-receptor CCR5. The sea change wrought by this fortuitous “proof of concept” was much in evidence at the 2011 persistence workshop this past December; the tentative forays into basic science that were once emblematic of the field are now mixed together with more ambitious plans for advancing ideas into the clinic. Perhaps most strikingly, two large pharmaceutical companies—Gilead and Janssen/Tibotec—described their use of industrial scale screening to search for compounds that are active against latent HIV. This represents an unprecedented expansion of efforts once confined to under-resourced academic labs.

A number of online resources are available with information on presentations at the 2011 persistence workshop: Lafeuillade runs a website called the Reference Portal on HIV Reservoirs & Eradication Strategies which includes the official abstract book and late breaker supplement along with an expanding number of reports, video interviews and commentary. David Margolis from the University of North Carolina has written a comprehensive report for Jules Levin’s National AIDS Treatment Advocacy Project (NATAP) website. Jon Cohen also covered one the most notable presentations in the journal Science.

To try and briefly summarize the top-line stories that emerged from the 2011 meeting:

A triumvirate of researchers—Courtney Fletcher, Mario Stevenson and Tim Schacker—presented data suggesting that sporadic, very limited rounds of HIV replication may occur in some individuals on antiretroviral therapy (ART) due to poor penetration of certain drugs into the tissues. However, preliminary data were only available from a small number of participants (~4-5) so the implications are still uncertain. According to the clinicaltrials.gov entry for the study, it is now expanding from the original enrollment target of 12 to 40 so additional information should soon be forthcoming. Alain Lafeuillade has posted an interview with Mario Stevenson about the findings, and these presentations were the subject of Jon Cohen’s story in Science.
An Italian research group led by Andrea Savarino described a retrospective analysis involving 18 rhesus macaques infected with SIVmac251 that participated in various studies combining ART with drugs targeting the viral reservoir. The analysis found an association between the number of “anti-reservoir” drugs animals received and the likelihood of controlling SIV to undetectable levels after ART was interrupted; however only three macaques controlled SIV to this degree so the findings should be considered very preliminary. The workshop organizers issued a press release about the data suggesting that, for the first time, they show that anti-reservoir drugs may be able to contribute to what is now frequently referred to as a “functional cure” (control of viral load in the absence of ART). In an interview with Alain Lafeuillade, Savarino is careful to note that the findings require confirmation in human studies because they could relate to unknown factors specific to the three macaques that controlled SIV in the experiment.
David Margolis from the University of North Carolina presented the first ever data on the use of a histone deacetylase (HDAC) inhibitor named SAHA (aka vorinostat) in individuals with HIV. HDAC inhibitors are at the forefront of efforts to pharmaceutically lure HIV out of latency, so news from Margolis’s trial has been eagerly awaited. While very preliminary, and derived from just four participants, the results so far suggest that the approach is able to increase HIV expression by latently infected cells. It took Margolis many years to get the trial started due to concerns about the safety of HDAC inhibitors (which are used as cancer treatments and can cause significant toxicities) but no serious side effects have occurred to date. As Margolis stressed, much more work is needed before any conclusions can be drawn about the promise of the approach.
The burgeoning involvement of the pharmaceutical industry in cure-related research—represented by presentations from Romas Geleziunas from Gilead and Roger Sutmuller from Janssen/Tibotec—was widely viewed as important news because it has the potential to transform the drug discovery effort by increasing the number of compounds that are being screened by many orders of magnitude.

The workshop agenda was divided into discrete topic areas spread over three days. The first session addressed the subject of animal models, and was led by Jeff Lifson from the National Cancer Institute (NCI) at Frederick who has nearly two decades of experience studying SIV infection in rhesus macaques. Lifson delineated some of the considerations for developing an appropriate model for cure-related studies, which include mimicking the degree of viral suppression achieved with ART in humans and developing tools to comprehensively assess the impact of additional interventions on SIV reservoirs.

The models currently in use include:

Macaques infected with hybrid SIV/HIV viruses encoding HIV reverse transcriptase (SHIV-RT), treated with efavirenz, emtricitabine and tenofovir.
Macaques infected with SIVmac251 or SIVmac239 treated with multi-drug regimens (e.g. tenofovir, emtricitabine, raltegravir and ritonavir-boosted darunavir +/maraviroc).
Pigtailed macaques infected with SIV/17E-Fr & SIV/Delta B670 treated with tenofovir, integrase inhibitor, saquinavir, atazanavir (this model is primarily being used to assess issues relating to viral activity in the brain).

Lifson described a study conducted by his laboratory in which macaques were infected with the highly virulent challenge virus SIVmac239 and, after sixteen weeks, treated with a multi-drug antiretroviral regimen comprising an integrase inhibitor, tenofovir, emtricitabine, and ritonavir-boosted darunavir. Suppression of viral load to less than 30 copies/mL was eventually achieved, but Lifson noted that it took longer than is seen with HIV in humans. Like the vast majority of macaque studies, the experiment involved Indian rhesus macaques, and Lifson suggested that viral load control might be easier to achieve in Chinese rhesus macaques (this subspecies has been shown to restrict SIV replication somewhat better in the absence of ART). Lifson acknowledged that refinement of the SIV/macaque model for cure-related research is ongoing, and he cautioned against the premature adoption of any one approach as a standard. As an example of the pitfalls of premature standardization, he cited the HIV vaccine field’s mistake in adopting a SHIV89.6p challenge model that turned out to have essentially no relevance to human HIV infection.

One potentially important new technology that Lifson highlighted is called digital PCR, which is vastly superior to traditional quantitative PCR for measuring small quantities of nucleic acids in samples. Quantitiative PCR amplifies nucleic acid sequences from a single sample by inducing multiple rounds of copying of the original sequence, then back-calculates how many were originally present; these calculations can be imprecise for a number of reasons. Digital PCR divides a sample into many discrete “microfluidic” wells and then uses PCR to look for the nucleic acid sequence of interest in each well, providing a readout as to whether the sequence is absent (0) or present (1). The total amount of nucleic acid sequence present is then calculated based on the number of negative and positive wells, using an approach called a Poisson distribution. Digital PCR assays have only recently been commercialized and a number of laboratories are now busy using them to measure SIV and HIV levels in research studies.

The presentations following Lifson illustrated both the diversity of animal models in use, and the uncertainties associated with them. Andrea Savarino and Iart Luca Shyaj from the Istituto Superiore di Sanità in Rome provided an update on experiments conducted by their group involving macaques infected with SIVmac251. In a paper published in AIDS last year, Savarino and colleagues reported that the gold-based rheumatoid arthritis drug auranofin reduced the reservoir of SIV-infected cells in animals treated with combination ART. At the workshop, Savarino presented results of a retrospective analysis of 18 macaques (including those in the experiments reported in the paper) that have received various combinations of antiretrovirals and “anti-reservoir” drugs including auranofin and buthionine sulfoximine (BSO). The breakdown of the antiretroviral regimens employed was as follows:

ART: tenofovir, emtricabine, raltegravir.
Intensified ART (iART): tenofovir, emtricabine, raltegravir, ritonavir-boosted darunavir.
Mega-ART: tenofovir, emtricabine, raltegravir, ritonavir-boosted darunavir, maraviroc.

Three of the 18 macaques have controlled SIVmac251 to undetectable (<40 copies/ml) levels after interruption of all treatment for several months, and Savarino reported that there was a significant correlation between the number of “anti-reservoir” drugs received and this salutary outcome (for the purposes of this analysis, the CCR5 inhibitor maraviroc was counted as an anti-reservoir drug due to evidence that it reduced the amount of SIV DNA when added to intensified ART and preliminary results from a human study suggesting it may impact reservoirs). Some macaques also received the HDAC inhibitor SAHA, but an impact on the SIV reservoir could not be demonstrated.

The complicated sequence of treatments and outcomes in the three macaques that have controlled viral load off ART can be roughly summarized as follows:

Macaque P252: ART, ART+auranofin, iART+auranofin, iART+SAHA, iART+auranofin, treatment interruption, viral load control to limit of detection, viral load rebound, Mega-ART, treatment interruption, viral load control, viral load rebound, viral load control, viral load rebound, Mega-ART, viral load control, viral load rebound, Mega-ART+BSO, viral load control (100+ days)
Macaque P157: ART, iART, Mega-ART+auranofin+BSO, treatment interruption, viral load rebound, viral load control (~60 days), viral load blip, viral load control (~50+ days)
Macaque P177: ART, iART, Mega-ART, Mega-ART+auranofin, treatment interruption, viral load rebound, Mega-ART, treatment interruption, viral load rebound, viral load control, viral load rebound, viral load control (~50+ days)

The data appear encouraging but there are potential caveats:

The model of SIVmac251 infection treated with combination ART (the drugs used in the study included tenofovir, emtricabine, raltegravir, ritonavir-boosted darunavir and maraviroc) is not well characterized, at least in terms of the published literature.
There were very few control animals, and the results are not from a single study but rather from multiple experiments, sometimes involving the same macaques being rolled over from prior experiments.
As can be seen from the sequence of events in the three controlling macaques, the treatments were complex and there was variability between animals in terms of exactly when different interventions were administered.

As Savarino stresses in his video interview with Alain Lafeuillade, human trials are now required to ascertain if the macaque results can be translated to HIV.

Paul Luciw from UC Davis presented results of an experiment in which macaques infected with SHIV-RT had prostratin and valproic acid added to long-term ART (efavirenz, emtricitabine and tenofovir) prior to an interruption. Luciw showed evidence of reduced viral RNA and DNA in tissues but when treatment was interrupted there was no significant difference in viral load rebound compared to macaques treated with ART alone. Daria Hazuda from Merck has included several of Luciw’s slides in her recent presentations on cure research so the main findings can be viewed online, however note that prostratin is only referenced as a “protein kinase C activator” and valproic acid as an “HDAC inhibitor”. Hazuda’s powerpoint can be downloaded online (bottom of the page, Luciw’s data is on slides 21-25). Luciw also mentioned that he repeated the experiment adding raltegravir to the ART regimen and in that case there was no additional viral RNA and DNA reduction in tissues resulting from the anti-reservoir drugs, but he was running out of time and was unable to give any details.

Victor Garcia-Martinez, University of North Carolina, gave an overview of a model at the smaller end of the mammalian scale: mice. As recently reported in several papers, mice can be equipped with “humanized” immune systems in order to facilitate HIV infection and the preclinical study of viral latency. Although there are a number of limitations to the system, it is likely to be a useful addition to the options for screening approaches prior to embarking on more costly macaque experiments.

In the subsequent basic science segment of the agenda, Jonathan Karn from Case Western Reserve University gave a presentation suggesting that the way HIV gets repressed into latency in cells may be somewhat different from the way that cellular genes get shut down, which could be a cause for optimism in terms of developing better ways to selectively de-repress latent HIV. Jerome Zack (UCLA) introduced the audience to a type of cellular particle called a vault, which consist of three proteins and a piece of RNA. Vaults were discovered by UCLA scientist Leonard Rome, and Zack is working to adapt them into drug-delivery nanoparticles for the purposes of delivering anti-reservoir compounds into cells. Zack presented some preliminary evidence that vaults can be engineered to deliver potential latency activators prostratin and bryostatin, he is also working with Paul Wender at Stanford to develop better analogs of these drugs to use. The goal is to come up with some lead vault-delivered anti-latency compounds to test in the humanized mouse model.

Shifting topics to the virological aspects of HIV persistence, Sarah Palmer from the Karolinska Institute reported results of an intensive evaluation of viral genetics pre-ART and on long-term ART (up to >12 yrs) in 12 people (seven treated at acute infection, five during chronic infection) to look for evidence of viral evolution that would be indicative of ongoing replication. No evidence suggestive of HIV replication was found in various CD4 subsets and other cell types in blood, lymph tissue, bone marrow and gut. Palmer noted that no hematopoetic progenitor cells (HPCs) containing HIV DNA could be found; occasional positive signals from HPC samples turned out to be due to low-level contamination with CD4 cells (this finding was recently echoed in a paper from Bob Siliciano’s group at Johns Hopkins).

Palmer drew attention to one case where a large amount of HIV DNA containing a huge deletion encompassing all of the protease gene was discovered. Since HIV can’t replicate without protease, this demonstrates that the division of CD4 T cells carrying integrated, non-functional proviral HIV DNA can contribute to what may appear to be an HIV reservoir by some measures (but really isn’t because the virus is defective). Mario Stevenson coined the term “junkyard DNA” for these non-functional proviruses, and it was quickly adopted at the workshop.

Tae-Wook Chun from the National Institute of Allergy and Infectious Diseases (NIAID) offered some data suggesting HDAC inhibitors may not be all they’re cracked up to be in terms of reversing HIV latency, in the hands of his lab they didn’t induce a significant amount of viral RNA from latently infected cells compared to prostratin (which is a potent activator generally considered too toxic for human use). Chun also said that the latently infected cells induced to produce viral RNA don’t seem to die (“we haven’t seen any evidence of cell death”), suggesting that induction using HDACs might have little effect in the absence of an immune response capable of killing the infected cell.

Day two of the persistence workshop featured the presentations from industry, with Romas Geleziunas from Gilead and Roger Sutmuller from Janssen/Tibotec talking back-to-back about the ongoing work at their companies.

Gilead is looking at both virus activators and immune modulators, with Romas Geleziunas seemingly already having taken on board what Tae-Wook Chun had suggested on the previous day: reactivating latent infection might not be enough to kill a cell, in which case immune mechanisms need to be induced to deliver the coup de grace. Geleziunas described Gilead’s high throughput primary cell screening assay, which is a modified version of an assay developed by Vincente Planelles and Alberto Bosque. So far they’ve identified three HDAC inhibitors from the Gilead drug library, named 001, 002 and 003. 001 is 10-fold more potent than SAHA but inhibits all classes of HDACs while 002 is of interest because, while less potent, it doesn’t score positive on the AMES test (the standard test used by regulators for assessing mutagenic potential). 001 and 003 were both AMES positive. Rats tolerated three weeks of 002 in a preliminary safety study. Geleziunas noted that HDAC inhibitors only activate a fraction of the virus expression seen with pan-activating CD4 T cell stimulation using CD3 and CD28 antibodies, raising the question of whether the HDAC inhibitors are only activating a proportion of the latently infected CD4 cells, or rather causing less virus expression per cell. This question remains to be resolved.

Massive high throughput screening of a Gilead library and a commercially available drug library has produced a 1% hit rate, identifying 89 compounds that could be grouped into 15 clusters based on their structures. One was a calcium pump inhibitor named thapsigargin, a “robust activator” of latency in cells from 6 out of 6 donors, however the compound is known to have carcinogenic potential. Another hit was a “broad spectrum nonspecific tyrosine kinase inhibitor” called tyrphostin A that worked on cells from 3/6 donors. Geleziunas explained that since they hadn’t expected to find kinase inhibitors, they then tried screening a library of those compounds and got a 20% hit rate. Preliminary studies involving these hits have revealed evidence of activity at low concentrations and dose responses. Next steps are to confirm activity with more selective kinase inhibitors and explore the signaling pathways that are causing these compounds to work.

Switching to the topic of bolstering immunity, Geleziunas said Gilead is looking at a TLR7 (toll-like receptor 7) agonist it has in development for hepatitis B. The product has been tested in chimps and woodchucks, where it has shown antiviral activity, dose-dependent induction of alpha interferon production and T cell and B cell activation. A small phase I human study has been safely conducted, also showing evidence of some T cell and B cell activation. Next step is to study the impact on HIV-infected cells and potentially test it in animal models in combination with HDAC inhibitors.

Meanwhile the overarching goals of Gilead’s program continue to be:

Conduct more high throughput screening.
Uncover novel mechanisms (e.g. as may happen as a result of the identification of kinase inhibitors).
Discover new chemical entities (NCEs).

Roger Sutmuller from Janssen/Tibotec then described his company’s efforts, which have not been discussed publicly before. He outlined the basic goal of discovering safe and effective compounds to reactivate latent HIV i.e. those that cause little or no cell activation and ideally have the potential to be combined. Unlike Gilead, Tibotec starts with a Jurkat cell line assay to identify compounds, after which they have a preplanned set of steps involving evaluation of:

Toxicity/immune stimulation.
Virus reactivation in primary T cell assays.
Virus reactivation in latently infected cells from HIV+ individuals ex vivo.
Medicinal chemistry selection of lead compounds.
Testing in a humanized mouse model developed by Roberto Speck.
Testing of the pathways involved in drug activity eg using microarrays, HIV mutants with various signaling elements disabled, short-interfering RNAs etc.

Using the Jurkat cell line assay, a fairly staggering 35,000 compounds have been screened to date, and the next step is to screen an incredibly staggering 480,000 compounds from a Johnson & Johnson “diversity library.” Of those screened to date, 800 HDAC inhibitors have popped out (a 20% hit rate), 25 protein kinase C agonists (the family prostratin belongs to) and 600 unknowns that can be grouped into 11 different “chemotypes.”

Sutmuller went on to describe their in-house primary T cell assay, which involves fresh cells expanded in the lab and infected with an HIV encoding green fluorescent protein (GFP). Cells are rested to create latency and then drug activity is measured based on the extent to which the cells light up green. They’re using this assay to screen medium sized libraries; it can handle about 2,000 compounds per week. He showed some data from one compound “229,” which induced virus at about half the level of pan-stimulator PMA, and worked even better in combination with SAHA. The next step is to study these and other compounds in Roberto Speck’s humanized mouse model, which involves 3TC and TDF given in food pellets and a long-acting version of TMC 278 that is delivered by weekly injection. They have seen good viral suppression and can recover latently infected CD4 cells using this system.

Among the other highlights from day two, Una O’Doherty from the University of Pennsylvania showed that CD8 T cells from elite controllers can kill what appear to be latently infected CD4 cells because they express the HIV Gag protein, just with much slower kinetics than seen with activated CD4 cells (and without causing spreading infection). O’Doherty suggested that perhaps this means latently infected CD4 cells aren’t as invisible to the immune system as has been thought, which provoked some controversy because—as she happily acknowledged—it is not yet known whether the same holds true for latently infected CD4 cells from individuals on ART.

In an effort to hone in on which elements of the Berlin patient’s treatment were necessary to achieving the apparent cure of HIV infection, John Mellors (University of Pittsburgh) presented an analysis of ten people who had undergone myeloablative chemotherapy and autologous stem cell transplants for lymphoma. None of these individuals were cured of HIV infection, leading Mellors to conclude that in the case of Timothy Brown, the CCR5-negative transplant was important, possibly along with the graft-versus-host disease Brown experienced. In the Q&A afterwards, workshop attendee Mike McCune from UCSF suggested that total body irradiation (TBI) might also have played a role.

Santiago Moreno (Hospital Ramon Y Cajal, Madrid, Spain) presented some preliminary evidence that the CCR5 inhibitor maraviroc may activate a protein complex named NF-kappaB when the drug binds to the CCR5 receptor. Because NF-kappaB activation can stimulate latent HIV, Moreno suggested that maraviroc might have anti-reservoir activity, as was previously suggested by a small uncontrolled pilot study conducted by Moreno’s laboratory and reported at a symposium prior to the 2010 International AIDS Conference in Vienna. However, results from a randomized trial of ART intensification with maraviroc were debuted at the persistence workshop by Maria Puertas, and this study was unable to document any additional declines in HIV reservoirs associated with receipt of the drug (HIV DNA levels fell by ~8-fold in both arms).

In a session on acute HIV infection, Marty Markowitz from Aaron Diamond AIDS Research Center presented 96-week results from a 3-drug vs. 5-drug treatment study, showing essentially no significant differences in a variety of reservoir and immunological measures in blood and gut. There was a slight reduction in cell-associated HIV RNA levels at week 96 in the 5-drug group but Markowitz felt this was unlikely to be meaningful. Jintanat Ananworanich (HIV Netherlands Australia Thailand Research Collaboration) described a study involving treatment of people with very, very early HIV infection, in which 60 people have so far been enrolled, with an average time from screening to enrollment of just 3 days. This would not seem like much time for someone to process the news that they have become HIV infected and make a decision to enter a trial involving a multiple treatments and sampling from the peripheral blood, CNS and GI tract, but Ananworanich said “acceptance rates are quite high.” Participants were in Fiebig stages I through IV at enrollment which designate:

I: within 5 days of infection (34% of participants)
II: 10 days (9%)
III: 13 days (48%)
IV: 19 days (9%)

24-week results indicated significantly smaller HIV reservoirs in blood and gut samples of participants enrolled at Fiebig stage I vs. III or IV, with total and integrated HIV DNA being undetectable in a proportion of the earliest-treated individuals.

The very last presentations of day two involved the tag team of Timothy Schacker (University of Minnesota), Courtney Fletcher (University of Nebraska) and Mario Stevenson (University of Miami) outlining very preliminary results from their small study of viral replication in anatomical and cellular reservoirs. A total of 12 individuals are enrolled, ART naive at baseline but then treated (mostly with TDF, FTC and ritonavir boosted atazanavir) and analyzed regularly up to six months. Not all individuals have data available yet, and the number of individuals from whom data were reported varied between the different presenters. Courtney Fletcher looked at drug levels in nine people, finding that some drugs (particularly atazanavir, FTC and efavirenz) may not reach adequate levels in lymph nodes and gut. Mario Stevenson then showed that in some study participants, 2-LTR circles increased in lymph tissue after starting ART, in one case along with a rise in proviral DNA. In one other individual, levels of both 2-LTR circles and proviral DNA went down. Stevenson stated: “this does not necessarily denote ongoing replication” but proposed an alternative model in which a population of long-lived cells can generate virions that infect one more cell and that’s it – just one cycle of replication, in other words. He stated this would not lead to viral evolution but could replenish the latent reservoir. In the Q&A, John Coffin from the NCI got up to the microphone and noted that since latency is a rare event in infected cells, and since Stevenson was saying these were single-cycle rounds of infection, the number of times latency would be created is not known, and may well not be often enough replenish the reservoir.

Timothy Shacker closed out the talks with a description of his efforts to correlate Fletcher’s and Stevenson’s results with measurements of viral RNA on the follicular dendritic cell (FDC) network in lymph tissue (using in situ hybridization). Schacker created 3D graphs for several participants that included 2-LTR circle levels, DNA levels, levels of viral RNA on FDCs and, lastly, drug levels. There appeared to be correlations between the various measures, but how many people had evidence of ongoing HIV replication cycles was unclear. Schacker noted that there was a significant inverse correlation between levels of FTC diphosphate in lymph tissue and viral RNA on FDCs. Additional results from the expanded version of this study are needed in order to understand if this is a broadly applicable phenomenon, and whether poor tissue penetration of antiretrovirals represents an under-appreciated obstacle to curing HIV infection.

The major news on day three of the workshop was the presentation by David Margolis (University of North Carolina) of very preliminary results from the phase I/II study of the HDAC inhibitor vorinostat (SAHA). The trial has a complicated schema, largely due to the safety concerns of the FDA regarding the drug, which scores positive on the AMES mutagenic test (a red flag for regulators even though the significance is not fully understood).

The first step of the protocol involved screening potential participants to assess whether vorinostat could reactivate latent HIV from their CD4 T cells ex vivo. Thirteen individuals had ~4 billion lymphocytes extracted by leukopheresis, then sorted into discrete pools of 1 million purified resting CD4 cells each (ending up with 24-36 pools per participant). These pools were exposed to either vorinostat or no drug, and a mean level of HIV RNA per million cells (and a standard deviation) was calculated for each person (the assay used can measure down to 10 copies per million cells). Margolis noted that the statistical approach used to calculate the mean RNA levels is robust but complicated, and a paper explaining it is currently in press at an unnamed statistics journal.

Four of the thirteen people screened showed a statistically significant upregulation of HIV RNA expression in this analysis and were therefore recruited into the next step of the trial. A 200mg dose of vorinostat was given first for safety, followed by a 400mg dose to study pharmacokinetics and for analyses of histone acetylation and acetylation of the p21 gene (in other words, analyses of the effects of the drug on cellular genetic machinery and not HIV). The pharmacokinetic data mirrored that reported in cancer studies and cellular acetylation (both total and p21 gene) was maximal by 8 hours then trended down by 24 hours.

A final 400mg dose of vorinostat was then administered with leukopheresis performed 4-6 hours afterward based on the pharmacokinetic data indicating this would be around the time of maximum activity. No grade 1 or greater toxicities were seen, and HIV RNA expression increased compared to baseline in all four individuals by a mean of 4.4-fold (range: 3-6.6 fold). HIV RNA in peripheral blood was also assessed using a single copy assay but no change was detected, perhaps not surprisingly given that this was a single dose study.

Margolis was obviously very encouraged by the data and stated that they had successfully “demonstrated induction of full length HIV RNA expression within a window of time after a single vorinostat exposure.” He concluded that obstacles to HIV RNA expression can overcome “at least in some cells.” But he stressed that many questions remain, including:

Is there an equal effect from multiple doses or does it become attenuated?
How much drug exposure is needed?
Should drug be administered continuously or pulsed?
Will toxicities emerge?
What number of cells is needed to measure relatively rare reactivation events?
Does RNA expression lead to virion production or clearance of the infected cell?
Are additional inducers needed?
Are additional interventions needed to clear the latently cells that have been induced to express HIV RNA?

The final session of the meeting was on functional cures. Dishearteningly, the crowd of attendees thinned noticeably but the first presenter, Paula Cannon from UCLA, was undeterred. “This is the first time people are going to be talking about functional cures,” she opened sunnily. “I know you’re all very obsessed with the reservoirs but we don’t really care about the reservoir – if there’s a little bit of virus left in the body, so what?” Cannon highlighted three key goals for those pursuing a functional cure:

Reducing the pool of HIV target cells and thereby reducing the harmful immune activation and inflammation that is central to pathogenesis.
Creating HIV-resistant HIV-specific CD4 T cells.
Taking advantage of HIV as a selection agent to drive the expansion of resistant cells.

Cannon went on to review the Sangamo zinc finger nuclease (ZFN) approach to deleting CCR5 and the work conducted by her laboratory to adapt it to modify hematopoietic stem cells (HSC). Cannon noted that efficacy was demonstrated in a published experiment in which humanized mice were engrafted with the CCR5-deleted stem cells and challenged with HIV. Work is now underway to advance the approach into HIV-positive people who need stem cell transplants as treatment for lymphoma, in collaboration with John Zaia and David DeGusto from City of Hope who have previous experience of studying gene-modified HSCs in this setting. Cannon explained that preparation for the trial has involved switching from relatively easy-to-use HSCs obtained from fetal cord blood to rather more uncooperative adult stem cells. These cells are called mobilized peripheral blood stem precursor cells (mPSCs) and sampling involves giving G-CSF for four days then conducting apheresis to extract white blood cells, followed by ex vivo purification of CD34+ cells. This procedure has now been performed on 13 donors, obtaining 42 billion white blood cells of which around 0.5% were CD34+ cells; Cannon estimates that around 1% of the CD34+ cells are “true” stem cells. These mPSCs are now being used in mouse studies to address a number of issues prior to human testing.

One such experiment assessed whether pre-existing immunity to adenovirus might be problematic, because an adenovirus vector is used to deliver the zinc finger nuclease into the mPSCs. Mice were given a high titer of anti-adenovirus antibodies prior to delivery of the mPSCs and, encouragingly, no difference was seen in the extent of engraftment compared to controls given phosphate buffered saline (PBS).

Next steps include large scale tumorigenicity studies in “NOD scid gamma” (NSG) mice and evaluation of modified mPSC under “maximizing” conditions to test the upper limit of on and off target effects (there is some evidence that ZFNs can disrupt genes other than the CCR5 target, particularly a similar region of the CCR2 gene). Mice given the maximized mPSCs will be kept for many months and extensively analyzed for safety.

Following Paula Cannon, Carl June from the University of Pennsylvania gave an update on the use of the same technology to modify CD4 T cells that are extracted from individuals with HIV using apheresis, expanded and modified in the laboratory, and reinfused into the same individual. Previous presentations of data from these phase I trials has generated considerable excitement, because the proportion of modified CD4 T cells persisting in the blood and gut of participants far exceeds the extremely modest levels obtained with prior gene therapies. Significant CD4 T cell count increases have also been documented up to nine months of follow up. Unusually, CD4:CD8 ratios have also improved from an average of 0.5 at baseline to 1.5 at last analysis; this type of improvement is rarely observed as a result of ART, and may have implications in terms of improving long-term health because inverted CD4:CD8 ratios are a well-documented risk factor for illness in the HIV-uninfected elderly.

Most intriguing, however, is a trial involving a 12-week analytical treatment interruption (ATI). Data is now available from six individuals who have undergone the ATI and while all experienced a viral load rebound, levels began falling prior to the reinitiation of ART, which June noted was not the case in a prior gene therapy study involving an ATI (an evaluation of a candidate named VRX496). One notable individual controlled viral load to below the level of detection (<50 copies/mL) before ART was restarted. This person turned out to be heterozygous for the delta32 CCR5 deletion, which means that the ZFNs could work more efficiently because only one CCR5 gene in each cell had to be disrupted in order for CCR5 expression to be completely abrogated (instead of two as is normally the case). Importantly, June found a significant correlation between the proportion of modified CD4 T cells and viral load control during the ATI. This suggests that an antiretroviral effect is achievable with the approach, and that the potency of the effect may be boosted if the proportion of modified cells can be increased. In the Q&A period, June was asked if he had assessed whether gene-modified HIV-specific CD4 T cells may have contributed the viral load results; he replied that HIV-specific CD4 T cell responses have not yet been analyzed in the ATI trial.

The last two talks in the final workshop session addressed the development of methods that attempt to specifically target latent HIV and excise it from the DNA of infected cells (or damage the provirus in order to render it non-functional). On paper, at least, these approaches sound very appealing but it was clear that significant hurdles remain. Jan van Lunzen (University Medical Centre Hamburg-Eppendorf) discussed the modification of an enzyme called Cre recombinase to target HIV DNA. The modified version, dubbed Tre recombinase, has successfully excised proviral DNA from cells in vitro and work is now underway to study how it might be delivered. Next steps involve studies in humanized mice using a lentiviral vector to deliver the Tre recombinase to CD34+ stem cells; the vector is designed to be “self-inactivating” in cells that do not contain HIV DNA. As an aside, Jan van Lunzen also mentioned a patient of his who started ART during early infection, was treated for 5 years, then stopped 6 years ago, had a small viral load blip and has been undetectable ever since. HIV RNA cannot be found in blood, gut or CNS. According to van Lunzen, the individual has a “very strong HIV-specific CD4 response,” and he highlighted the case as being similar to Christine Rouzioux’s report of five individuals treated very early who have controlled viral load to undetectable levels off ART for an average of around five years. These case reports may bode well for prospects for a functional cure, van Lunzen suggested.

Keith Jerome from the Fred Hutchinson Cancer Research Center recounted the efforts of his group to employ different enzymes, endonucleases, to target latent HIV. The idea in this case is to induce mutations in the HIV provirus in order to render it non-functional. Some success has been achieved in vitro but considerable challenges remain in terms of improving the efficiency of targeting and developing delivery methods that might be able to get the endonucleases to where they are needed. Jerome’s work is now being supported by a Martin Delaney Collaboratory grant from NIH.

The last word at the 2011 persistence workshop was given to Nobel laureate Françoise Barré-Sinoussi, who outlined the International AIDS Society’s development of a Global Scientific Strategy “Towards an HIV Cure” and encouraged audience members to attend an IAS symposium on the subject that will take place in Washington DC immediately ahead of the 2012 International AIDS Conference. Barré-Sinoussi also stressed the importance of the work and the need to continue the momentum which has placed curing HIV infection back at the top of the research agenda.

The 6th International Workshop on HIV Persistence, Reservoirs & Eradication Strategies is scheduled for 2013 in Miami.