PERSISTENCE, RESERVOIRS AND ELIMINATION STRATEGIES
Reported by David Margolis, University of North Carolina
18th International HIV Drug Resistance Workshop (IHDRW)
June 9-12 2009 Ft Myers Florida
The 18th International HIV Drug Resistance workshop began with an examination of a different sort of drug resistance --- HIV that that persists despite successful ART. After an overview by this author that will be summarized by Mark Wainberg, the session addressed several aspects of this problem: proviral DNA persistence within long-lived cells, and persistent, low-level viremia that exists below the level of detection of clinical assays.
Laboratory and pre-clinical model systems that mimic the latent state of HIV infection in the most faithful means possible are needed to understand latency and develop therapeutic approaches to eliminate HIV infection. Cell lines, which are immortalized, transformed clones of primary cells are useful and easy to manipulate in the laboratory, but may not completely reflect latency within patients. Patient cells can be studied, but the rarity of latent infection makes some studies challenging, and other studies impossible.
Alberto Bosque, a Mathilde Krim fellow of the AmFAR and a member of the Planelles laboratory, described work recently published in Blood in which he developed a model of latency within infected primary peripheral blood CD4 cells obtained from uninfected volunteers (abstr. 2). Naive CD4+ T cells cells obtained from the peripheral blood of healthy donors are induced to undergo normal development ex vivo in the presence of a cocktail of cytokines and antibody co-stimulation of the T cell receptors CD3 and CD28. These cells are infected while in the activated state and while many cells die, some return to quiescence as memory cells. Infection of these ex vivo generated memory cells leads to latency with a high frequency (approximately 90% of the surviving cells). Surviving cells represent a polyclonal population of integrated HIV genomes. Bosque used these cells to examine viral expression from these "latent" cells in response to signaling via pathways reported to influence HIV-1 latency.
Bosque's found that viral reactivation following T cell receptor (TCR) signaling by anti-CD3 and anti-CD28 was completely blocked by drugs that blocked Lck kinase activity, a kinase that is required for one of the pathways of TCR signaling, and by cyclosporine, a drug that blocks that activation of the cellular factor NFAT. NFAT has previously been shown to be an important cellular activator that can induce HIV replication, but the cellular activator NFκB has classically been considered to be THE central cellular activator of HIV expression. However, in this model, NFκB is not required. This is not necessarily stunning, as older studies have shown that NFAT is an important activator of HIV expression, and HIV can replicate in the absence of NFκB binding sites. But this finding does highlight the NFAT pathway as an important alternative pathway of HIV reactivation, and perhaps a different target for therapeutic intervention.
Also, in contrast to studies performed directly in resting CD4+ T cells obtained from aviremic, ART-treated patients, histone deacetylase inhibitors had no ability to reactivate latent HIV-1 in primary memory cells. However, as previously reported in studies of patient cells by our laboratory and that of Roger Pomerantz, Bosque found that a combination of IL-2 and IL-7 is a potent inducer of latent HIV-1 independently of NFAT and NFκB. Therefore, reactivation via IL-2 and IL-7 represents an alternative pathway to T-cell receptor engagement and represents a novel therapeutic avenue for purging latent reservoirs. IL-2, and more recently IL-7, have been given safely to HIV-infected patients with the goal of improving immune reconstitution. IL-2, of course was recently proven by the ESPRIT and SILCAAT studies to increase CD4 counts but to have no beneficial effect on clinical outcomes. IL-7 has only been examined in one clinical trial so far, but was shown to improve CD4 cell counts. Another aspect of IL-7's effect, as Illustrated by Bosque and prior studies, is it potential to induce proliferation of central memory CD4 cells. As these are the cells that most frequently contain latent HIV, the important question is whether IL-7 will induce more expansion of the latent proviral reservoir via proliferation of cells carrying silent HIV, or more purging of those cells by inducing viral expression (and presumably cell death).
Romerio (abstr. 85) presented studies from his laboratory of an alternate primary cell model of latency, some of which was recently published in J. Immunology. Romerio's system recapitulates CD4 cell activation by dendritic cells and antigen, are then infected in vitro with HIV-1 and returned to quiescence in the presence of interleukin-7. Infected, resting cells in this system lack expression
of activation markers, do not undergo cellular proliferation and do not sustain viral replication. Viral production is seen after cell stimulation. This system should contribute to the ability to study the mechanisms of latency, and test anti-latency therapy candidates.
Pepe Alcami from the Unidad de Inmunopatologia del SIDA at Instituto de Salud Carlos III, (abstr. 3) presented studies of a plant-derived phorbol ester, SJ23B. Of interest to the botanists among us, SJ23B is derived from a European plant related to the Samoan plant that yields prostratin, a phorbol ester with similar properties. For the chemists among us, this compound is a jatrophane diterpene, and induces HIV gene transcription and viral expression. This effect is seen at micromolar concentrations when tested in the Jurkat T cell line. Of interest, like some other compounds that signal through protein kinase C pathways, SJ23B simultaneously downregulates receptors required for HIV infection, CD4, CXCR4, and CCR5. This bimodal function makes SJ23B very interesting, as a major limitation to the strategy of purging latent HIV by inducing cell activation in concert with viral expression, is that one will spread more new infection than is purged. Further studies of SJ23B in human cells and and pre-clinical toxicity studies are planned. Their results will be eagerly awaited.
The presentations that followed dealt most with persistent HIV expression that can be detected by research assays in at least three-quarters of patients that are stably clinically suppressed on ART. First Sarah Palmer (Karolinska Institute, abstr. 4) presented a comparison of levels of residual viremia in "elite controllers" to patients suppressed by standard ART. As there are many definitions of what an elite controller (EC) is, for this work ECs were defined as patients with <50 copies/ml of plasma HIV-1 RNA for at least three measurements over a 12-month period. ECs were allowed to have occaisional blips, and some had episodic viremia of over 500 copies/ml.
The viral RNA levels of 84 elite controllers were compared with samples from 163 patients suppressed on therapy to <50 copies/ml for 4-333 weeks. Low-level, residual viremia was measured using the real-time RT-PCR assay with single copy sensitivity (SCA, single copy assay) that has become the standard in the field, first described by Palmer, Mellors, and Coffin at the Frederick NCI Drug Resistance Program in 2004. Viremia was detected in 85% of the samples from elite controllers (ranging 1-614 copies/ml) and from patients on suppressive therapy (ranging 1-43 copies/ml). Using a statistical test, the two groups had similar mean values of 0.5 log10 copies (3 copies; p = 0.95 based on Welch's ANOVA), but there was substantially variability in persistent viremia within each EC (P<0.0001 Levene's test for homogeneity). Although valid methodological complaints were voiced due to the differences in sample size and numbers of assays performed, it did seem intrinsically reasonable that immunological control of viremia would be less uniform than pharmacological control.
Mary Kearney of the Frederick DRP group then presented studies describing the viral genetics of this miniscule population of virus found leaking into the plasma despite stable ART suppression of viremia below the clinical level of detection of 50 copies/ml (abstr. 6). Kearney used single-genome sequencing (SGS) techniques to insure accurate characterization of this viral population. To perform SGS, plasma is diluted serially and PCR performed for HIV RNA. By doing PCR at the dilution level where most of the wells are negative, one can be assured that a positive well usually reflects the amplification of a single molecule (genome) of HIV RNA.
Using this technique a total of 1,105 HIV-1 sequences spanning the gag, protease, and polymerase genes were obtained by single genome sequencing (SGS) immediately prior to ART, after initiating ART, and after HIV-1 RNA decreased to <75 copies/ml. Diversity was measured in aligned sequences usinjg a statistical technique called average pairwise difference.
Prior to ART genomes sequenced had an APD of 0.2-2.5% per site. As viral load dropped ART there was no change in diversity or population structure of circulating virus despite declines of viral load of up to 10,000-fold. This is interesting, as it suggests that with effective, successful therapy there is little chance for viral subpopulations to be selected. There has been long-standing concern, especially inpatients with high viral loads, "extra" therapy may be needed to prevent the evolution of resistance "on the way down." This observation would tend to allay that concern. Overall, the data suggest that virus populations established before therapy is initiated are fixed and stable during successful ART, and that active cycles of HIV-1 replication are completely blocked by ART.
A few years back the Siliciano lab described a few patients on ART with a "predominant plasma clone (PPC)." This was a single viral clonal population that persisted at low levels (ie <50 copiers/ml), despite ART, and could not be found in circulating T cells. The group speculated that this represented virus that might have entered a bone marrow stem cell, and was now being chronically produced at a low level.
Here Kearney reported that of many patients studied, the Frederick group was able to find a PPC in only one patient after 4 years on suppressive ART. The virus population containing the PPC after therapy was significantly different from the pretherapy virus population (2.9% different), however this PPC variant could be detected within the majority population prior to therapy, and during ART at very low frequency. Therefore the PPC seems not to be a unique virus that has emerged during ART, but a rare species found prior to ART. This does not rule out the possibility that the PPC is originating from a rare stem cell, but suggests that it is established prior to ART, during untreated viremia. However, it seems to appear in a minority of patients. The frequency, characteristics, and origin of this low-level viremia is not really of current clinical importance, but understanding it is critical to future efforts to eradicate infection or induce a "drug-free remission" of HIV infection.
Although viral species do not evolve as ART induces suppression of clinically detectable viremia, as previously discussed low-level persists despite ART. Maldarelli rounded out the Frederick group's presentations (abstr. 10) showing that intensified therapy does not reduce persistent HIV-1 viremia. 18 patients on NNRTI- or PI-based regimens, intensified therapy, and residual viremia measured by SCA. 13 participants underwent 4-week intensification with either efavirenz (n=2), lopinavir/r (n=2) or raltegravir (n=9); five participants underwent 8-week intensification with atazanavir. Median viral RNA levels prior to intensification were 1.7-2.7 copies/ml. During intensification median viremia ranged from 1.6 to 5.6 copies/ml. No significant decreases in viremia were noted. Following intensification, HIV-1 viremia levels were 1.1-4.8 copies/ml in the three studies. CD4+ T-cell counts remained stable with no significant changes during or following intensification.
At CROI in Montreal this year, Buzon and colleagues presented a provocative study from Spain (CROI 2009 abstr. 423a). In a group of 44 patients with <50 HIV-1 RNA copies/mL randomized to intensify their RT with raltegravir, they found an increased number for HIV DNA circles in some patients. As you know, HIV RNA is reverse transcribed into DNA copies. Most of the double-stranded copies that are produced are defective or incomplete, and can be degraded by cellular enzymes over time. A few complete copies can be packaged in pre-integration complexes with cellular proteins and viral integrase, enter the nucleus and become viral integrants. Some complete HIV DNA genomes fail to integrate and host repair enzymes knit the double strands into DNA circles. So HIV DNA in a cell can exist as linear DNA, circular DNA, or integrated DNA. Linear or circular DNA will be diluted if the cell divides, or as cellular nucleases act to chew up linear molecules. So in the absence of new infection of uninfected cells, a state thought to be achieved by current fully suppressive ART, no new HIV DNAs should be created.
The Buzon study suggested that intensification with RAL uncovered evidence for ongoing low-level de novo infection, that is viral entry and reverse transcription, although no evidence either for or against full rounds of viral replication (infection-integration-viral production-new infection). One part of their observation that was hard to explain was that this increase in 2-LTR DNA circles, when it was seen, was only seen at week 2, and returned to baseline after that.
Two studies presented at DRW sought evidence of a change in HIV circular DNA after ART intensification. A substudy of the ANRS 138 trial (abstr. 9) found no change in HIV-1 total DNA and 2-LTR circles after 24 weeks of raltegravir. In 30 patients whom maintained other ART and switched enfuvirtide (T-20) to raltegravir, HIV-1 DNA was measured at weeks 0 and 24. Episomal cDNA forms (2-LTR circles) were measured with primers that span the 2-LTR circle junction and total viral DNA forms were assayed using internal LTR primers. No change was seen from week 0 to week 24 in episomal HIV DNA circles or total HIV DNA. In 6 patients at week 0, circle were detected. In 4 different patients circles were detected at week 24.
In a separate study from France (abstr. 82), 25 antiretroviral-experienced patients receiving a RAL-containing regimen were studied. The 2-LTR DNA were quantified in peripheral blood mononuclear cells (PBMCs) before starting RAL (month 0) and in follow-up samples up to month 12. With decline of plasma HIV RNA (in the group, no therapy failures were mentioned), the mean amount of 2-LTR DNA increased 0.14 ±1.07 log10 copies/106 PBMC at month and 0.95 ±1.37 log10 copies/106 PBMC at month 12. But of the 25 patients, circles were repeatedly undetectable in 15, increased in 7, decreased in 2, and in one followed the pattern reported by Buzon of an early increase followed by a return to baseline.
Overall one is left with the impression that we are still uncertain if measuring circular HIV DNA as a measure of low-level infection is a worthwhile, reliable assay. Along these lines, one poster (abstr. 87) by Lafeuillade suggested that cell-associated HIV-1 RNA obtained from rectal biopsies might be a marker for studies of HIV persistence and low-level replication. In 23 HIV-1-infected patients (16
on ART and 7 naive) fibre-optic rectosigmoidoscopy was performed and biopsies obtained a mean distance of 20 ±2 cm from the anal margin. Lymphoid cells were assayed for intracellular
HIV-1 RNA levels (extraction with Qiamp Blood RNA [Qiagen], DNAse RNAse Free treatment, then Cobas Ampliprep/Cobas Taqman HIV1 [Roche]), and the last for proviral HIV-1 DNA following Qiamp Blood DNA (Qiagen) extraction. Gut-associated viral load was measurable using these standardized, commercially available kits.
Finally, on the immunological side of the equation, two studies examined the question of immune activation in either ECs or aviremic, ART-treated patients. Hunt at UCSF (abstr. 5) hypothesized that HIV-specific CD4+ T cells in ECs, while supporting the cytotoxic HIV-specific CD8+ T-cells that are likely to play a key role in the special status of ECs, might also serve as target cells contributing to viral persistence. In 31 HIV-infected ECs with plasma HIV RNA levels <75 copies/ml in the absence
of antiretroviral therapy, 82% had detectable plasma HIV RNA levels by an alternate, more sensitive HIV RNA test called the transcription-mediated amplification assay (TMA; GenProbe). The percent of activated CD8+ T-cells correlated with plasma HIV RNA levels (rho 0.32; P=0.048). However, proviral cellular DNA and cell-associated HIV RNA were correlated with the % of activated CD4+ T-cells and % of HIV-specific CD4+ cells. Hunt concluded that his findings suggest that activated and/or HIV-specific CD4+ T-cells are major contributors to viral persistence in ECs. To me this seemed a bit circular, as these cells presumably are required to be activated and function to identify and kill infected cells, and maintain the fortunate few EC patients in their elite controlling state.
Conversely, the Frederick group (abstr. 8) showed that markers of cellular immune activation do not correlate with levels of residual viremia in patients on long-term suppressive antiretroviral therapy. 33 subjects initiating ART and maintaining HIV-RNA<75 copies/ml for ≥3 years without viral blips (≥2 consecutive HIV-1 RNA≥500 copies/ml) were studied. After 7 years on ART, activated CD8+ T-cells less frequent than prior to ART, and not significantly different than that of healthy uninfected controls. A marginal difference in CD8+HLA-DR+ T-cells (P=0.020) was not significant after correction for multiple comparisons. Residual viremia was determined by SCA in 18 patients and no significant correlation was found between the level of residual viremia and immune activation markers.
RESISTANCE TO ENTRY INHIBITORS
John Moore of Cornell summarized issues around drugs that inhibit HIV replication by the blockade of the use of human CCR5 receptor by the HIV envelope "entry claw" (abstr. 11). Moore summarized the mechanisms of action of, and the development of drug resistance to, HIV entry inhibitors such as maraviroc (MVC), vicriviroc (VCV), and others in clinical development. In brief, CCR5 receptor antagonist resistance mechanisms consist of 1) more efficient use of receptor to out-compete drug, 2) binding "around" drug to the receptor, and 3) switching to use another receptor, eg. CXCR4.
In the lab, Moore took an R5-using virus from a patient who was known to later have developed X4-tropism. By serially growing the virus in human blood cells in the presence of an R5 inhibitor, AD101, Moore selected an escape mutant that was resistant to AD101. However, the mutant was still sensitive to RANTES, the human chemokine that naturally binds and signals through the CCR5 receptor, and is a natural anti-HIV factor, as well as sensitive to another R5 inhibitor. So the virus had first developed the ability to out-compete AD101 for occupancy and use of the R5 receptor. This change came about by a single amino acid change in the HIV envelope sequence. As this virus was grown in culture, 3 additional mutations were selected in the presence of AD101. Each change added about 10-fold resistance to AD101. In similar experiments done by Mike Westby at Pfizer, using MVC as the selecting drug, similar changes emerged in the V3 loop of the gp120 envelope. However, using VCV as a selecting drug, one resistant isolate developed changes in the gp41 fusion peptide, not the gp120 site.
Moore emphasized that CCR5 inhibitors act by an allosteric mechanism, binding CCR5 and stabilizing the envelope in a altered conformation that is not recognized by the virus. Resistant viruses can bind either the altered or native receptor, so give the "plateau effect" seen in the Phenosense assay. This plateau is not seen in traditional assays done with PBMCs.
Moore then explained that the HIV envelope glycoprotein gp120 interacts primarily with two domains in the third variable domain of CCR5, a V3 loop and region called the V3 tip. In the presence of inhibitor in the V3 loop alters the geometry which virus must bind. The development of resistance alters the shape of the HIV V3 region, exposing a new epitope that is sensitive to specific monoclonal neutralizing antibodies. This change makes the virus more neutralization sensitive, but drug-resistant. So some drug-resistance escape mutants in vivo may become sensitive to humoral immunity (antibodies). As passive antibody therapy for HIV infection is under development, that is bimonthly or monthly doses of synthetic antibodies designed to block HIV entry, this fact might be of use in future patients who have developed resistance to CCR5 inhibitors.
Finally, Moore reviewed the issue of the escape of virus from CCR5 receptors by "tropism shift," that is a switch from using CCR5 to the use of CXCR4. In the MOTIVATE studies that led to the licensure of MVC, ca. 50% of the clinical failures of therapy were correlated with the emergence of X4-using virus; in ACTG 5211, a study of VCV, ca. 35% of failures were from X4 emergence. Later studies showed that these viruses largely emerged from pre-existing resistant populations that were not detected at screening.
Following this plenary, Eric Arts of Case Western (abstr. 12) discussed resistance to PSC-RANTES, a synthetic form of the chemokine under study for use as a microbicide. PSC-RANTES
interacts with the second extracellular loop of CCR5 and effectively downregulates surface CCR5 expression. However, Arts showed that PSC-RANTES inhibits via competitive inhibition of HIV binding to CCR5, and that receptor downregulation contributed little to the antiviral effect of PSC-RANTES. However, PSC-RANTES resistance but not MVC resistance, increased HIV fitness. PSC-RANTES has been shown to be effective in primate model experiments of vaginal SIV infection and PSC-RANTES protection, but that when the microbicide fails a PSC-RANTES-resistant SHIV clone is established in the infected animal.
Huang from Virologics (abstr. 13) outllined the genetic pathways traveled by HIV in the tropism switch from CCR% use to CXCR4 use for cell entry. The group characterized coreceptor utilization along with envelope clone sequences derived from 8 patient's virus populations, in patients that were known to have evolved the ability to use CXCR4 in the absence of antiretroviral drug pressure. The results were convincing, but surprisingly heterogeneous. It appeared that HIV can take different genetic pathways to emergence of X4 use. Viruses that are dual tropic (can use X4 or R5) or mixed (a swarm of distinct X4 and R5 viruses) can appear, disappear, and reappear over serial sampling. V3 envelope mutations that are thought to primarily drive co-receptor selection may appear in different order in different patients. CCR5 to CXCR4 transitions appeared to be incremental, not sudden.
Harrigan from Vancouver (abstr. 15) then reported efforts to accurately predict viral tropism using genotype sequence algorithms on data obtained from viral genomes detected in stored samples from patients in the MVC MOTIVATE study. In the MOTIVATE-1 study, MVC was compared to placebo in patients on optimized background therapy. Patients were screened prior to entry into the study with the standard Monogram Trofile, a less-sensitive test of viral tropism that is no longer in use.
For this study, genotype was determined and tropism inferred from the genotype sequence by using the algorithms geno2pheno (g2P; 5% false-positive rate) or PSSM (-2.96 cutoff). Preliminary genotypes and Trofile results were available from 1,230 samples (of these 553 were non-R5 by Trofile).
Compared with Trofile, genotype had sensitivities of 63% and 56% and specificities of 91% and 90% for detecting non-R5 virus using g2P and PSSM, respectively. However, short-term viral load decreases were similar regardless of the assay used. Harrigan concluded that despite poor sensitivity of standard genotyping and algorithm prediction of the presence of non-R5 HIV when compared to standard Trofile, early virological reductions in this treatment-experienced population were similar regardless of the assay used.
One obvious shortcoming of this comparison is that it could not evaluate the effect of the newer Trofile of improved sensitivity. At this point many clinicians using Trofile testing to guide MVC use might prefer that a direct comparative study be performed; others without access to Trofile testing might feel comfortable using MVC based on genotypic information. While it seems obvious that at some point sufficient genetic information will be available to predict tropism with great precision from a genotype, it is also obvious that day has not arrived. Clinical outcome is a sum of many factors. In HIV therapy one of these factors is the correct use of a test to guide the selection of the best drug options for an individual patient. In this presentation, the position put forward is that genotypic interpretation of viral tropism is sufficient to guide the use of MVC, because a successful clinical outcome, herein measured as week 8 or 24 viral load decline in a selected subgroup of clinical trial patients, occurred equally in groups of patients said to have R5 virus by an old Trofile and a genotypic interpretation. Whether this means that in clinical practice such interpretations are "equal" or "close enough" to the clinical results garnered using the improved Trofile test is the current subject of debate and controversy.
RESISTANCE TO INTEGRASE INHIBITORS
Although perhaps of interest only to resistance nerds, Jay Grobler of Merck provided one of the premier mechanistic insights of the meeting (abstr. 25). Although the mutations in integrase that confer drug resistance, and the "pathways" through which these mutations evolve (ie. which mutation comes first, and which follow) has been well described, the true molecular mechanism of resistance to integrase inhibitors has not been elucidated.
When integrase inhibitor resistance mutations were first described to encircle the enzyme's active site, many (the author included) assumed that integrase inhibitor resistance was like protease inhibitor resistance ---- a battle for access to the enzyme active site. However, using time-of-drug-addition and drug washout experiments the residence time of drug on integrase/DNA complexes was measured. The HIV integrase enzyme was assembled on immobilized, synthetic DNA that mimicked viral genome strands, and equilibrated with radiolabeled drug. Dissociation rates were determined by counting label bound to the complex in the presence of a large excess of unlabeled compound.
In time-of-addition and washout studies, raltegravir effectively blocked wild-type HIV-1 replication
when present during a relatively short window of time following the completion of reverse transcription. In
dissociation kinetic studies, raltegravir exhibited a long residence time on the integrase/DNA complexes that is comparable to or exceeds the half-life of the pre-integration complex in the cell. The clinically important integrase mutation N155H increased the rate of raltegravir dissociation by 10-fold and allowed the inhibitory effects of the compound to be washed out in viral replication assays. The second-generation InSTI MK2048 (no longer in clinical development), which better inhibits raltegravir-resistant variants, exhibited a substantially longer residence time on integrase than raltegravir. While a similar increase in dissociation rate was observed with MK-2048 with integrase containing the N155H mutation, the half-life of MK2048 on the mutant is nearly that of raltegravir on the wild-type enzyme.
So HIV develops InSTI resistance by evolving the ability of the integrase enzyme to escape from the grasp of the inhibitor. Inhibitors that can hold on tighter and longer to the viral enzme will be more "resistant to resistance" and in the absence of resistance may be also be able to be taken once a day or even less frequently.
Francois Clavel (abstr. 26) then presented an elegant assay to simultaneously assess HIV integrase inhibitor resistance and HIV replicative fitness in one assay, the "selective advantage profile." Previously, studies have shown that Integrase inhibitor resistant HIV strains are less able to replicate robustly in culture ("less fit"). So Clavel performs a comparative replication assay of drug-sensitive wild-type HIV to drug-resistant HIV ("fitness assay") in the presence of as range of integrase inhibitor drug concentrations. Clavel's studies showed why HIV first selects the Q155H mutations first, followed later by the dual mutations including the Q148HKR mutation, as 155H replicates best a low concentrations of RAL.
Finally Thibaut presented preliminary data (abstr. 27) describing the activity of a new class of integrase inhibitors, the quinoline family, that inhibit integrase binding. The results showed that a quinoline drug, QNL111, inhibited the binding of HIV-1 integrase to its substrate, whereas RAL did not. Moreover, QNL111 remained active on viruses carrying the main resistance mutations selected by RAL. The QNL111 and RAL were felt to be synergistic in culture against wild-type and single-mutant virus strains. No synergy was seen against double-mutant strains.
INSIGHTS AND ADVANCES BY STUDY OF THE STRUCTURE OF HIV RNA
Kevin Weeks, University of North Carolina at Chapel Hill discussed new technology to analyze the structure of an entire HIV-1 RNA genome, and the potential for this to lead to drugs that bind HIV RNA (abstr. 28). Any RNA constrained by its hydroxyl (-OH). This constrained RNA site can be reacted with an ester, and acetylated. Weeks used this selective 2' hydroxyl acylation, and then performed reverse transcription using a pool of cDNA primers. Extended RT products would stop at site of acylation. So the population of cDNA product sequences can be linked on the RNA genomes to create a map of where acylated modifications lie. This reveals which RNA residues are available to be acylated (constrained) and which are available (free, unstructured). The more frequent a particular cDNA is, the more constrained that RNA was.
This sort of analysis can be automated and rapid. Weeks calls it SHAPE analysis. When the HIV NL4-3 strain, a standard laboratory virus, is analyzed this was the structure is rapidly revealed. Domains of the HIV RNA that are known to be regulatory, TAR, RRE, etc, were apparent. But many new structured regions were apparent, and revealed many new structural regulatory region of HIV RNA.
For example, in the POL poly protein that encodes protease, RT, RNase, and integrase, contains 15 linker areas, of which 14 are highly structured. Mutations of these linker regions greatly decreased the fitness of virus in competitive viral culture assays. The hypothesis is that as RNA is translated into protein, the ribosome must pause at structured regions of RNA, giving the newly formed protein the chance to fold correctly into the final protein before the ribosome proceeds to produce the next protein encoding within the polyprotein RNA.
The Weeks lab also treated virions with AT-2, a zinc ion chelator that steals zinc from proteins. In the case of HIV, this leads to a loss of any structure that is coordinated by zinc ions. By following changes induced in SHAPE by AT-2, regions that are structurally coordinated by zinc are revealed. Such analysis has given a new picture of the RNA packaging site in the virion capsid. It is expected that these insights will greatly illuminate how HIV replicates, and may lead to new therapies directed against the structures of HIV RNA itself.