HIV Pathogenesis, Innate Defense against HIV,
and HIV Cure at CROI 2014, Boston
By David Margolis MD
UNC Chapel Hill & the Collaboratory of AIDS Researchers for Eradication (CARE)
The annual Conference on Retroviruses and Opportunistic Infections (CROI) brings news of groundbreaking advances in some years, while other years report more incremental progress. This year's CROI, held again in Boston but suffering under the frigid grip of the polar vortex, reported advances in therapy for HCV and co-infection that are restructuring clinical practice, and exciting first news on long-acting antiretroviral therapy.
Innate Antiviral Immunity
Reflecting a dominant area of recent activity in the HIV pathogenesis field, much of the discussion and presentation at CROI touched on an aspect of innate immunity against HIV --- the ancient battle between retrovirus and host cell that reflects cycles of adaptation on an evolutionary scale of the host cell to block retroviral replication, and the viruses adaptations to evade such blockade. Understanding, and perhaps even manipulating these host and viral functions might someday lead to the ability to eradicate HIV infection, enable a durable ART-free remission of viremia and disease (so-called functional cure), or provide new ways to treat HIV infection.abstract
In this vein, Paul Bieniasz (Aaron Diamond AIDS Research Center, The Rockefeller University) opened CROI with the Bernard Fields Lecture in virology and viral pathogenesis (abstr. 17) with a discussion human innate immunity to the replication of retroviruses. We are equipped with a large and diverse set of genes the produce proteins that inhibit the replication of immunodeficiency viruses. The selection pressures exerted on humans by ancient viral infections (read: even before the evolution of homo sapiens), and on viruses by human cells have caused retroviruses to evolve at an unusually rapid pace --- that is in evolutionary time (read: millions of years). This is different that the evolution of ART drug resistance, which can happen between clinic visits.
Modern retroviruses like HIV and SIV have evolved to grow despite the restrictions imposed by antiviral cellular proteins. But these viruses can only grow in the specific cell types to which they have become adapted. Therefore it is difficult for viruses to "jump" species and infect a new host type. So monkey viruses cannot grow in humans (generally) and vice versa. This is a good thing, as it prevents the frequent spread of pathogens between species, but makes it difficult to development of primate models of HIV-1 infection for use in research.
Human genes that restrict or prevent retroviral replication are either constitutively expressed (that is the antiviral proteins are always being produced in the cells) or expressed after signaling from the antiviral cytokine interferon. These factors act in many ways to inhibit virus replication, and often affect invariant viral proteins or viral RNAs --- targets that the virus cannot easily alter or dispense with, so that host restriction is not easily escaped. The virus is then forced to evolve accessory proteins whose major function is the thwart a host restriction factor.
Bieniasz first discussed the host factor Tetherin, discovered primarily in the Bieniasz laboratory, that as the name suggests acts to link the HIV envelope to cell envelope, and prevent the release of fully formed viral particles from cell surface. In response HIV has evolved a protein called Vpu. Vpu binds one transmembrane domain of Tetherin, and blocks its linking function. Vpu can act on diverse enveloped viruses such as Moloney Leukemia virus and Marburg virus, a fact that illuminates how Vpu works, but is not clinically important. Vpu-like proteins are found in several retroviruses, but are specific to the virus of that species. For example, only some SIVs make a Vpu but other SIVs have evolved changes in their viral Nef protein to antagonize Tetherin restriction and allow SIV replication. Bieniasz described an elegant series of experiments in which by serial passage of virus between monkeys, SIV Vpu acquired ability to resist pigtail macaque tetherin, and a non-pathogenic form of SIV then caused lethal disease in macaques.
Bieniasz next discussed the ApoBEC3 family of host factors. These proteins bind foreign viral RNA, and after retroviral infection inhibit the reverse transcription of viral RNA to viral DNA, inducing very frequent and often lethal hypermutation in the viral genome that is copied and inserted in to the host genome. The restriction induced by the ApoBECs is the primary reason that more than 99% of all HIV DNA that can be detected in any form within a human cell is defective, and replication incompetent. Such HIV DNA is not a "reservoir" it is a boneyard of defective genetic material. In response to this potent threat to the ability of HIV to faithfully copy its genome, the virus has evolved the Vif protein to antagonize ApoBEC. Like Vpu, Vif is also species-specific. HIV-1 Vif neutralizes the APOBEC3 proteins by recruitment of an E3 ubiquitin ligase complex. This complex is cellular housekeeper that tags proteins for the trash heap, by poly-ubiquitination (adding multiple ubiquitin groups to protein tails), thereby assigning such proteins for transport to the cell proteasome, where they are chewed up and degraded.
Finally discussed was TRIM 5 (TRIM stands for TRIpartite Motif). TRIM recognizes motifs within viral capsid proteins and interferes with the uncoating process as the virus enters a newly infected cell, therefore preventing successful reverse transcription.
This theme was continued in a subsequent abstract-driven session on host factors and HIV/SIV replication. Ned Landau's group from NYU discussed the retroviral restriction imposed by the deoxynucleoside triphosphohydrolase SAMHD1, which acts in myeloid cells (eg. Macrophages), but not lymphoid (eg. CD4+ T) cells ( abstract 45 http://www.croiwebcasts.org/console/player/22075?mediaType=slideVideo&). They found that the small molecule MLN4924 prevents cullin neddylation, part of the biochemical process used by the viral factor Vpx to fight against the antiviral factor SAMHD1. Their findings suggested that small molecule inhibitors such as MLN4924 could block Vpx effect, restore SAMHD1 restriction, and act as a novel antiviral for HIV replication in myeloid cells.
Kane and colleagues from the Bieniasz laboratory then described the activity of MX2, an interferon-induced inhibitor of HIV-1 ( abstract 46 http://www.croiwebcasts.org/console/player/22076?mediaType=slideVideo&). The group surveyed a number of cell lines with differing abilities to restrict HIV replication after interferon stimulation. They then compared gene expression profiles in these cell lines to discover genes whose expression correlated well with HIV restriction. This suggested that MX2 was a potential anti-HIV-1 protein. Subsequent biochemical and virological studies using MX2 and mutants of MX2 lacking functions such as GTPase activity, nuclear localization signal, or leucine rich effector domain revealed that MX2 inhibits HIV-1 nuclear import, or destabilizes nuclear HIV-1 DNA. Further mutations in the HIV-1 capsid protein that alter the nuclear import conferred resistance to MX2, suggesting that MX2 is an effector in the anti-HIV-1 activity of type I interferon, although there are other actions of interferon that inhibit the early steps of the HIV-1 replication cycle.
Amalio Telenti of the University of Lausanne ( abstract 79LB http://www.croiwebcasts.org/console/player/22156?mediaType=slideVideo&) identified the human IFI16 gene as an HIV restriction factor, using a genomics approach. Telenti performed a genome-wide screen for genes under positive selection pressure in primates, the assumption being that one of the positive selection pressure operating in primate evolution might be the need to control retroviral invasion of the human genome. A cell-based assay was used to determine the impact of candidate genes under positive selection during primate evolution on HIV-1 gene expression, virus production and virion infectivity. The group identified 30 gene candidates: many of them inhibited transcription or translation of HIV-1, but a few impaired viral infectivity. One of these, IFI16 was a known immune sensor of viral DNA, but also strongly reduced the production of infectious HIV-1 by interfering with viral gene activity and capsid protein processing.
Serra-Moreno from the Harvard Medical School primate facility in Southborough described the antiviral activity of BCA2 (Breast Cancer-Associated gene 2 or Rabring7). This gene encodes another ubiquitin ligase, but appears to have antiviral activity even in cells that lack tetherin. BCA2 promotes the ubiquitination and lysosomal degradation of HIV-1 Gag capsid protein, thereby, impairing virus assembly ( abstract 47 http://www.croiwebcasts.org/console/player/22077?mediaType=slideVideo&).
Two presentations then discussed primate retroviruses. Mirela D'arc from Universidade Federal do Rio de Janeiro and University of Montpellier presented genetic studies of SIVgor, an SIV in Western lowland gorillas that is related to HIV-1 groups O and P ( abstract 51 http://www.croiwebcasts.org/console/player/22081?mediaType=slideVideo&). By studying genetic material recovered from feces collected in the wild, the group found that gorilla populations from southwest Cameroon are the direct source of HIV-1 P and we show for the first time that SIV sequences closely related to HIV-1 O circulate also in wild gorilla populations. SIVgor was most likely transmitted to humans on two occasions, resulting in the establishment of HIV group O and P in humans.
Walid Heneine from the CDC ( abstract 52LB http://www.croiwebcasts.org/console/player/22082?mediaType=slideVideo&) then reported the discovery of a new retrovirus in gorillas, Human T-lymphotropic Virus Type 4 (HTLV-4). As you may remember, HIV was for a time known as HLTV-3. Of the seven known species of human retroviruses, HTLV-4 had not been found in an animal thus far. Thus far a single case of HTLV-4 was reported in 2005 in a 48-yr-old male hunter from rural Cameroon who reported hunting monkeys, chimpanzees, gorillas, among others. Screening over 1000 wild and captive primates in Cameroon, genetic sequences related to a primate retrovirus similar to HTLV-4 were detected in six gorilla samples, with 99% nucleotide identity to HTLV-4 and to each other. Analysis suggested that HTLV-4 entered humans from a gorilla source 3,300 - 8,400 years ago. Heneine suggested that given the evidence that gorillas serve as a reservoir for the evolution and emergence of human infections, there was a need to management of contact with gorillas, develop assays for HTLV-4/STLV-4 detection, and monitor STLV-4 among gorillas and HTLV-4 among individuals exposed to gorilla populations.
These presentations were followed by a discussion of poster presentations on restriction factors lead by Paul Spearman of Emory. Ester Ballana of IrsiCaixa AIDS Research Institute ( abstract 188 http://www.croiwebcasts.org/console/player/22091?mediaType=slideVideo&), Barcelona showed in an elegant way that the viral factors that thwarted SAMHD1 restirction also reduced the activity of the thymidine analog HIV Reverse transcriptase inhibitors AZT and D4T. However, as these effects did not reduce the activity of the most-used NRTI drugs tenofovir or 3TC (and presumably FTC), these findings were largely of academic, mechanistic interest.
Refsland from the Harris lab at the University of Minnesota reported on the effect of natural human genetic variants in APOBEC3H on resistance to HIV-1 ( abstract 188A http://www.croiwebcasts.org/console/player/22093?mediaType=slideVideo&). As discussed, HIV Vif triggers the degradation of several members of the APOBEC3 family of DNA cytosine deaminases. Polymorphisms (natural variants) in human APOBEC3H gene encode proteins that differ widely in their expression levels, but the open question was whether these variants led to any difference in the potential of HIV to grow in cells, allowing for these difference to contribute to differing clinical outcomes. The group found that naturally occurring variants of APOBEC3H are expressed stably and capable of encapsidating and restricting HIV-1 replication in primary CD4+ T lymphocytes. Further, different naturally occurring Vif mutations allowed packaging of APOBECs into viral particles to varying degrees. This suggested the possibility that APOBEC polymorphisms, and the emergence of Vif variants in certain patients, could contribute to rapid disease progression or control of HIV infection.
Spragg and Emerman from University of Washington, Seattle ( abstract 82 http://www.croiwebcasts.org/console/player/22159?mediaType=slideVideo&) presented another example of parallel evolution of virus and host. They found multiple SAMHD1 haplotypes in African Green monkeys that were differently sensitive to restriction by the forms of Vpr found in various subtypes of SIV that infects these monkeys (SIVagm). The Vprs found in SIVagm could degrade the major forms of SAMHD1 found most commonly in their host AGM population, but were often incapable of degrading SAMHD1 found in other species of monkey.
Another plenary talk was dedicated to innate antiviral immunity, and focused on interferon-induced host factors. Michael Malim of King's College London reviewed the role of type-1 interferons in suppressing HIV-1 replication. Paradoxically, although interferon is produced at elevated levels during acute HIV-1 infection, and this is associated with long-term control of HIV infection, when interferon production is chronically elevated during chronic HIV infection, this is associated with increased viremia and poorer outcome. Data suggested that the clone or swarm of virus that initially establishes HIV infection, known as the transmitter/ founder virus, is less sensitive to interferon suppression when tested in laboratory culture replication assays. Interferon-stimulated genes can also be activated by signaling through Pattern Recognition Receptors, an evolutionarily conserved sensing system that detect foreign proteins and DNA. These receptors can turn on a cascade of cellular signals that induce an antiviral cellular state, such as cyclic GMP-AMP synthase, the STING pathway, and TREX and other cellular nucleases that might degrade foreign DNA. Interferon-stimulated genes also turn on the production of restriction factors discussed above, including TRIM 5a, ApoBEC, tetherin, SAMDH, and many other genes whose function and mechanisms of action have yet to be defined. Unfortunately, while these restriction factors are extremely efficient at blocking cross-species transmission of viruses, but human pathogenic viruses have adopted evasion strategies for the human defense system.
Iyer and colleagues ( abstract 78 http://www.croiwebcasts.org/console/player/22155?mediaType=slideVideo&) from the University of Pennsylvania presented studies to corroborate the claim that Transmitted/Founder (TF) HIV strains are relatively resistance to the antiviral effects of interferon α. They used matched pairs of TF molecular clones and consensus molecular clones that represented circulating viral species in untreated patients 6 months after primary infection. Samples were obtained from the CHAVI cohort, whose patients were diagnosed when viremic but seronegative, followed, and extensively characterized. The replication of these clones was assessed in human CD4+ T-lymphocytes in the presence and absence of IFNα. In 12 of 12 patient samples studied, TF viruses replicated more efficiently than the corresponding 6-mo viruses in the presence of IFNα. These findings show that interferon may exert significant selective pressure on HIV during acute infection, likely as this innate immune response is the primary one that is operational during acute infection. As adaptive anti-HIV immune responses come in to play later in infection, the circulating viral strains appear to prioritize evasion of those responses - which are likely more potent - over evasion of interferon.
However, Gilad Doish of the Gladstone Institute of Virology and Immunology in San Francisco demonstrated that there could be a dark side to the innate immune response to invading viruses. [CROI abstract 76 http://www.croiwebcasts.org/console/player/22153?mediaType=slideVideo&]. These findings have recently been published in Science and Nature. His studies in explanted tonsil tissue suggest that dying, HIV-infected CD4 cells release material that triggers an inflammatory form of cell death, pyroptosis, in bystander, uninfected CD4 cells. This may account for >95% of the CD4 T cell losses occurring in many lymphoid tissues like tonsil and spleen. Pyroptosis is triggered by caspase-1-mediated signaling, primed by high-level expression of pro-IL-1beta. Ironically, it is the IFI16 signaling pathway that triggers pyroptosis --- so in infected cells IFI16 may restrict replication, but in uninfected cells exposed to the products of HIV replication it can trigger cell death. Pyroptosis is efficiently blocked by VX-765, a small-molecule inhibitor of caspase-1 that has been shown to be safe in preliminary research studies in humans. Small-molecule inhibitors of caspase-1 could provide a form of "anti-AIDS" therapy that complements current treatment strategies by preventing CD4 cell death by pyroptosis.
Cure at CROI
ART and the beginning of its implementation around the world has significantly decreased HIV morbidity and has made it possible for HIV-infected individuals to live relatively healthy lives. This result, of course, ignores the tremendous challenges of delivering therapy to all who need it across the world. Nevertheless, for many decades in to the future, it would seem likely that a large population of patients throughout the world would require lifelong ART. While there may be some potential for long term ART to induce some toxicities, the primary reasons to seek therapies that could eradicate HIV infection --- a "cure for AIDS" --- or at least do induce a drug-free remission of disease and infectiousness --- a "functional cure" --- are to remove the medical and economic burden of lifelong therapy, reduce ongoing new infection, and to relieve the HIV-infected from the burden of stigma. Various presentations addressed aspects of this broad and aspirational topic.
Katherine Luzuriaga of the University of Massachusetts Medical School discussed the impact of early ART on HIV persistence in children ( abstract 53 http://www.croiwebcasts.org/console/player/22118?mediaType=audio&). As children infected at birth suffer higher levels of HIV viremia, earlier immune dysfunction, and earlier disease, ART initiated within three months of birth required to preserve immune function and markedly reduce HIV-1 related mortality. Such "early" ART yields a significant reduction in HIV DNA within cells, and also in true replication-competent reservoirs over the first year of therapy. Manysuch children have now been treated for more than 10 years with continued reductions in proviral DNA to extremely low levels, with the lack of recoverable replication-competent HIV-1. Such children (now adolescents) display an absence of detectable HIV-specific immune responses.
In the study PACTG 356, 15 children were treated within 2 weeks to 3 months after birth (mean 1.9 months) and achieved virological suppression by 16 weeks (median = 8 weeks). Those that maintained suppression have reversion of HIV antibody tests to negative by 15 months of life. These children have now been treated for up to 18 years, with very low levels of HIV PBMC DNA found (<10 copies/million PBMCs). There has been no evolution of virus, in the few patients in which replication-competent virus has been recovered, compared to the neonatal sample. The group now seeks to propose studies of vaccination in such children treated for years since birth, in preparation for a trial of therapy interruption in a highly monitored setting, in the hope that persistent virus has been cleared, or can be controlled.
This effort was, of course, spurred by the Mississippi baby, reported by Persaud at CROI last year. A short update of the course of the child was provided ( abstract 75LB http://www.croiwebcasts.org/console/player/22149?mediaType=slideVideo&). As everyone on planet Earth with a television knows, this was a child born to an untreated HIV+ mother. Therapy was started at 31 hours of life but viremia was detected for several days thereafter. The child was treated for many weeks, lost, then returned to care off ART without detectable HIV. Now 41 months old, the child has been off ART for 23 months. Plasma HIV RNA remains <20 c/ml (no target detected), HIV antibody tests show seronegativity, and no HIV-specific CTL responses can be detected by research assays. Single-copy HIV assays show less than 1 copy/ml from months 26 to 40, with a single positive test at 1 c/ml at 24 months that is likely a false positive. 17 to 22 million cells have been cultured without viral recovery. Low levels of HIV DNA have been detected in PBMCs (3.4 copies/million cells) at 40 months. HIV DNA sequencing, which might rule out laboratory contamination or validate that the sequence detected were similar to those seen in early viremia, but this has not been possible due to low levels of HIV DNA.
The IMPAACT group is seeking other similar cases, and reported the discovery of a second infant with high-risk exposure to HIV who started ART at four hours of age. The mother took no antiretrovirals during pregnancy and had a viral load of 138,811 copies near delivery.
HIV infection was confirmed by positive peripheral blood HIV DNA PCR at four hours of life and HIV RNA of 217 copies/mL at 36 hours of age. The "Long Beach baby" began treatment immediately with zidovudine, lamivudine, and nevirapine. At 2 weeks, the child was switched from nevirapine to lopinavir/ritonavir, and therapy has continued for 9 months. HIV RNA was detected in CSF at 32 copies/mL on the 6th day of life. Plasma HIV RNA was undetectable on the 11th day, and remained undetectable through age 8 months. HIV DNA testing was negative by the 6th day, and remained undetectable at day 47 and 67.
Replication-competent HIV was not recovered from resting CD4+ T cells at 1 and 3 months of age; but HIV DNA was detected in culture-negative wells at one month but not age three months. At age 3 months, HIV antibody is indeterminate (western blotreactivity to gp160). CD4+ T cell percentages remained normal for age.
The very important distinctions from the Mississippi baby to draw here are two-fold, First, the Long Beach child was treated hours earlier, with much lesser evidence of persistent viremia at any substantial level. In a way this case is much like post-exposure prophylaxis. In primate experiments, infection has been blocked in SIV-exposed infant macaques treated with 2-drug ART up to 48 hours after birth. Secondly, this child remains on therapy, which has never been interrupted. Nevertheless, even if such cases are irrelevant for the cure of infected adults, they are critical wake-up calls for efforts to define safe and effective methods for the immediate treatment of HIV-exposed newborns with fully effective ART.
Two other studies touted the benefit of immediate ART in acute/early infection, towards the goal of immune preservation and the restriction of reservoir size (a potential future benefit, rather than a current one). Schuetz ( abstract 77 http://www.croiwebcasts.org/console/player/22154?mediaType=slideVideo&) found that early ART initiation in the USAMC-AFRIMS Thai cohort study prevents the appearance of biomarkers said to correlate with disruption of the mucosal barrier, and that this was associated with lower levels of T cell activation. In 38 subjects in Fiebig stages I to III (acute infection) were enrolled and underwent sigmoid biopsy at baseline, 6 and 24 months after initiation of ART. 10 HIV uninfected and 5 ART-naive HIV+ controls were studied. HIV+ patients treated in Fiebig I/II exhibit levels of Th17 cells and CD8 T cell activation similar to HIV- subjects and maintain those levels under ART. This may prevent the loss of Th17 cells and subsequent CD8 T cell activation.
Okoye of the Oregon National Primate Research Center ( abstract 136LB http://www.croiwebcasts.org/console/player/22273?mediaType=slideVideo&) presented a analogous study of the effect of early ART in acute SIV Infection. Rhesus macaques were infected intravenously with SIVmac239 treated with tenofovir, emtricitabine, dolutegravir and darunavir boosted with ritonavir starting at day 7 (group A), day 10 (group B) or day 42 (group C) after viral exposure. Early ART, when initiated prior to peak virus replication, limited systemic virus dissemination and seeding of the reservoir in peripheral and extra lymphoid mucosal compartments. A delay as short as three days during the "hyperacute" phase can result in over a log higher tissue-based reservoir size, once therapy is started and maintained. Central memory T cells were found to be the major cell population containing SIV DNA in the early ART setting. Tissue biopsies showed persistence of cell-associated HIV RNA on ART, which persisted for longer times if ART was delayed. Early ART also decreased levels of cell-associated HIV DNA, with larger decreases seen with earlier ART. Virus was undetectable by coculture given the numbers of cells that could be collected, but using "adoptive transfer" of cells to an uninfected animal as few as 60 million cells transferred infection, but 3 million did not.
Victor Garcia-Martinez of UNC Chapel Hill ( abstract 56 http://www.croiwebcasts.org/console/player/22121?mediaType=slideVideo&) presented in impressive study of the leading alternative model to SIV in primates for HIV research, the humanized mouse models (nb. This author is a colleague and co-author with Dr. Garcia). In this model system, human hematopoietic cells into are transplanted in to immunodeficient mice. Garcia discussed the "BLT" model that co-implants bone marrow and a human thymic organoid grown from liver and thymic tissue (hence BLT), yielding bona fide educated human T cells in the context of HLA. The resulting mice can be infected with HIV by any route, and treated with standard ART to suppress peripheral viremia. Standard assays to quantitate latent infection in resting CD4+ T cells can be performed. Most recently, Garcia demonstrated the efficacy of envelope-targeted immunotoxins to kill HIV infected cells that were continuing to express residual viral RNAs and proteins despite full ART suppression. The model appears to be a flexible platform for the in vivo evaluation of novel approaches for the study of HIV latency, reservoirs and eradication strategies.
Several presentations described cell populations carrying persistent infection despite ART. Joe Wong of UCSF ( abstract 137 http://www.croiwebcasts.org/console/player/22274?mediaType=slideVideo&) described the distribution of cellular HIV DNA and RNA in tissues and blood in naive (Nv), central- (CM), transitional- (TM) and effector- (EM) memory populations in 8 patients durably suppressed on ART. 3 patients were treated in early (but not acute) HIV infection (Feibig VI) and 5 in chronic infection. The highest HIV DNA levels were in EM cells in rectum, but only about 30% higher then TM/CM. HIV DNA levels in TM/CM cells in rectum (which could not be separated from this site) were also higher than in CM or TM and EM in blood. Most of these results were similar to those of Chomont in 2009. HIV expression per infected cell (expressed as ratio of HIV RNA to DNA) was lower in all subsets in rectum compared to blood, perhaps as in that activated environment most cells expressing HIV RNA are cleared. Wong concluded that the local host environment affects HIV expression and how latently infected cells take up residence. The impact of potentially curative interventions on HIV persistence is expected to be different across anatomic sites. This may be true, but as most lymphoid immune cells engage in periodic trafficking, this is not necessarily the case.
Wagner of the University of Washington, Seattle ( abstract 138 http://www.croiwebcasts.org/console/player/22275?mediaType=slideVideo&) studied viral integration sites and envelope sequences to test the hypothesis that proliferation of HIV-infected cells helps sustain the HIV DNA reservoir during ART. The group found HIV integrants preferentially present at the human BACH2 gene, as previously reported. Viral envelope sequences from identical integrations sites are identical, suggesting that this was a clone of proliferating cells. As the preferred integration sites were enriched for cancer-associated genes, Wagner hypothesized that the HIV integrants were affecting host cell proliferation, leading to copying of the HIV genome. The major weakness of this study, however, was the lack of any demonstration that the expanding clones identified were actually replication-competent, as through proliferative effects one might expect an over-representation of defective, hypermutated genomes.
Lichterfeld from the Massachusetts General Hospital presented his recently published work on HIV latency within CD4+ T memory stem cells, a recent subclassification of CD4+ memory cells ( abstract 54 http://www.croiwebcasts.org/console/player/22119?mediaType=audio&). These are not true hematopoetic stem cells, with the ability to differentiate in to any blood cell, but thought to be the earliest and most immature developmental stage of memory T cells, and recently described in humans, mice and non-human primates. Tscm compose 2-4% of CD4 T cells. Lichterfeld reported studies of 9 patients, treated with ART for a median of 8 years, and found that the decline of HIV DNA within the Tscm population was less than that in standard central memory T cells.
Lichterfeld also reported various post hoc analyses of the study of the HDAC inhibitor panobinostat performed in Denmark. Unfortunately he displayed a key piece of data incorrectly, presenting changes in cell-associated HIV RNA that were actually changes in histone acetylation, and so it is best to look at the poster by Rasmussen. In this trial of 15 patients (poster abstract 438LB http://croiconference.org/sites/all/ abstracts/438LB.pdf) received oral panobinostat 20 mg 3 times/week every other week for 8 weeks. The drug was generally well-tolerated. The primary endpoint was cell-associated unspliced HIV-RNA (CA-US RNA) was analyzed in total CD4+ T cells using semi-nested qPCR. Levels of CA-US RNA increased significantly during panobinostat treatment (p<0.0001) with significant increases on time points on-treatment as compared to baseline. The median maximal fold-increase in CA-US RNA was 3.5 (range 2.1-14.4). Levels of CA-US RNA remained elevated 4 weeks post-panobinostat (fold-increase 1.60; 95% CI: 1.17-2.19; p=0.003). Overall, this pattern of response looks very similar to that shown by Lewin in her study of 14 days of daily vorinostat (as yet unpublished, and unfortunately not currently available on the old CROI website). Similar to the Lewin study, and the single dose study of vorinostat by Archin et al (Nature 2012) an increase of CA-RNA on the first day of dosing is apparent. However, like the Lewin study, the fold increase of CA-RNA on following says is variable, not always associated with dosing, and strangely persists for many weeks off drug. The level of total HIV DNA/million CD4+ T cells decreased transiently from baseline to day 14, but returned to baseline levels. There was no change from baseline in integrated HIV DNA. In contrast, our multiple-dose vorinostat study did not show a repeated induction of CA-RNA ( abstract 435 lb http://croiconference.org/sites/all/abstracts/435LB.pdf and J Infect Diseases 2014). In this writer's opinion, although the ability of HDAC inhibitors to disrupt latency has now been demonstrated in 3 studies, it is still not clear how best to administer multiple doses of HDAC inhibitor to achieve the desired effect of tolerable, safe, and effective disruption of HIV latency.
This situation was made more uncertain by two subsequent presentations. Sharon Lewin of Monash University, Melbourne ( abstract 140 http://www.croiwebcasts.org/console/player/22277?mediaType=slideVideo&) presented an analysis of the effect of vorinostat on host gene expression in samples from 9 of the 20 patients who received daily doses of vorinostat for 14 days. Gene expression was analyzed at baseline and two time points on vorinostat (day 1 and 14) and post-cessation of vorinostat (day 84). Lewin stated that 2 hours after the 7th dose the same gene changes were seen after 1st dose, and that there was persistent differential gene expression at day 14 and 84. Lewin felt that the effects of vorinostat on chromatin largely occur within the 1st day after the first dose of drug and that after 14 days of continuous dosing, there are compensatory mechanisms associated with transcriptional repression and cell survival. During her talk she highlighted gene expression changes that persisted at day 84, 70 days after the end of dosing, and expressed concern about long-term toxicities. As circulating peripheral blood cells are short-lived cells, and any blood cells that had been exposed to the last dose of vorinostat are unlikely to be extant 70 days later, it seems that further analysis is clearly indicated.
However, even if HDAC inhibitors can be safely administered, Greg Laird ( abstract 139 http://www.croiwebcasts.org/console/player/22276?mediaType=slideVideo&) questioned if they could ever be clinically effective on their own. Laird and colleagues in the Siliciano laboratory evaluated the effects of candidate latency reversing agents (LRAs) on resting CD4+ T cells from HIV-1 infected patients on suppressive ART. They made two key changes in the assays they used to perform this study. First they tested LRAs for their ability to induce viral outgrowth from latently infected patient cells. However, when they treated purified rCD4s from patients on ART with individual LRAs for 18 hours and then cultured them to recover HIV, they used MOLT-4/CCR5 cell line cells as targets rather than activated (allogeneic) primary CD8-depleted PBMCs from an HIV-negative donor. This was done for a laudable reason, to remove the potential cross-stimulation from the donors' cells to the patient's cells. MOLT cells lack the HLA antigen to give this stimulation. However, MOLT cells (as all cell lines) are significantly less efficient at replicating HIV. So although this new assay represents the activity of the LRA tested in a more "purist" assay, it may also be an assay that only detects the activity of very strong LRAs. Disulfiram, JQ1, bryostatin-1 and the HDAC inhibitors vorinostat, romidepsin, and panobinostat were tested at clinically relevant concentrations, and in fact in this system, none of these LRAs induced viral outgrowth from cells from any patient tested.
Next Laird evaluated the ability of LRAs to induce increases in intracellular HIV-1 mRNA in patient resting CD4 cells. For this they used a PCR primer that amplifies sequence at the very (3' poly A) end of an HIV RNA transcript, so that any signal measured had a better chance of originating from a full-length, authentic HIV RNA that would be packaged in to a virion. They criticized the standard primers used by others (Archin et al., Nature 2012), theorizing that RNA detected by these primers could preferentially reflect RNA transcripts that would not result in full-length, RNAs packaged in virions. Again, when they tested the ability of LRA to induce viral transcripts using the poly A primer, as before they found that vorinostat, romidepsin, panobinostat, disulfiram, and JQ1 all failed to increase intracellular HIV-1 mRNA, and only the PKC agonist bryostatin-1 caused significant increases, though the effect was only seen in some patients. As before, the goal of detecting only authentic transcripts is a worthy one, but the ability of the poly A primer to detect rare transcripts is unclear. This is an important area for further work, to clarify and optimize the best assays systems to be used to discover and test LRAs.
In other work from the Siliciano laboratory, Kai Deng ( abstract 142 http://www.croiwebcasts.org/console/player/22279?mediaType=slideVideo&) found that CTL escape variants dominate in the latent reservoirs of chronic phrase-treated patients, but not acute phase-treated patients. After reactivation, outgrowth viruses still carry the same CTL escape variants, insensitive to the relevant epitope-specific CTL clones, but sensitive to clearance by broad spectrum CTLs in culture. The researchers concluded that broad spectrum or more specifically, subdominant CTL responses could be required for the elimination of viral latent reservoir.
Macatangay from the Mellors group at University of Pittsburgh reported such an attempt to boost immunity via a dendritic cell-based HIV therapeutic vaccine ( abstract 143 http://www.croiwebcasts.org/console/player/22280?mediaType=slideVideo&). To their surprise vaccination with an autologous DC-HIV vaccine not only failed to prevent viral rebound after analytic ART interruption (well, not so surprising), but increased residual viremia a mean of 14-fold in 40% of subjects despite continuous ART. The increase in residual viremia was associated with increases in T cell activation, and higher levels of IL10 and IL12.. Therapeutic vaccination may increase HIV-1 replication or expression from latent reservoirs.
Finally several presentations discussed various aspects of marrow transplant research seeking to cure HIV infection. Tim Henrich ( abstract 144LB http://www.croiwebcasts.org/console/player/22281?mediaType=slideVideo&) provided further information on the two Boston patients widely hoped to be cured when announced at the IAS meeting last summer, but announced at the Persistence workshop in December to have both suffered rebound viremia.
Neither patient received a marrow transplant that was intrinsically resistant to HIV infection due to the absence of a function CCR5 coreceptor, as did the Berlin patient, Timothy Brown. Both patients had graft-vs-host-disease (GVHD), controlled with intermittent or no immunosuppression. The first patient (patient A) was studied 1.8 years after marrow transplant, prior to ART interruption, and culture of 150 million cells was negative. No HIV DNA was detected in 26 million PBMCs. This patient had viral rebound 12 weeks after interruption, with symptoms of acute infection.
2.6 years after marrow transplant, Patient B had < 2.4 HIV DNA copies/million cells, and no virus recovered from cell cultures. Although the patient was HIV RNA and DNA negative just prior to rebound, at day 220 viremia returned with symptoms of acute infection.
This suggests that allogeneic HSCT may lead to loss of detectable HIV-1 from blood and gut tissue, but a rare cell can retain replication-competent HIV, and apparently survive without inducing rebound via reactivation for several months.
In another presentation ( abstract 418) Henrich found that cytotoxic chemotherapy could not always be relied upon to deplete reservoirs of persistent HIV infection, although one patient had undetectable cell-associated DNA and RNA on salvage chemotherapy for T cell lymphoma. The impact of cytotoxic chemotherapy on the HIV reservoir may vary depending on specific oncologic, chemotherapeutic and viral disease factors.