 |
|
|
| |
Integrase/CCR5/TMC125 Resistance, New Drugs, GAG, Ultra-Deep Sequencing for Low-Level ART Drug Resistance, Abbreviated TDF/FTC PreP Therapy Effective in Monkeys. Barbados
|
| |
| |
Report by David Margolis, MD, University of North Carolina
The 16th International HIV Drug Resistance Workshop (june 12-16, 2007, Barbados) opened with an overview of HIV in the Caribbean from N. Adomakoh, the director of the Ladymeade Reference Unit, the comprehensive state HIV clinic for this 14-by-22 mile island country with universal healthcare. He reviewed the progress and shortcomings of responses to the HIV epidemic across the diversity of nations in the Caribbean, with wide variation in both the seroprevalence of HIV infection, and in the availability of resources and infrastructure to cope with the epidemic. Even in Barbados, a country with an adequate infrastructure and resources, it was ironic to learn that the stigma of HIV infection still prevents a substantial proportion of those who need HIV care from accessing it here.
The 16th HIV DRW covered a lot of ground, but this summary will focus primarily on presentations in three areas that will have immediate impact on clinical practice, or the directions of future clinical studies and drug development:
1. Resistance mutations developing in patients treated with integrase inhibitors
2. The impact of presence of CXCR4-using viral subpopulations on the use of CCR5 inhibitors
3. NNRTI mutations in patients failing therapy with TMC 125, a next-generation NNRTI
Resistance and using the newest drugs, part 1: Integrase inhibitors
The Merck integrase team expanded on the information presented at CROI on resistance to raltegravir (RAL; Abstr. 8). As Daria Hazuda was trapped in airport hell in Miami, Michael Miller of Merck stepped up to the podium to pinch-hit. Miller described drug resistance found in Protocol 005, a phase II study performed in patients with 3-class resistance.
Patients had received OBT (optimized background therapy) with 20, 400, or 600 mg twice a day of RAL. The cohort had advanced resistance to therapies, with a GSS score of 0 in 68% (i.e. no drugs in OBT were predicted to be fully potent). Of 133 patients analyzed, failure was seen in 38, and resistance mutations found in 35 of these.
As discussed at CROI, two primary "pathways" to RAL resistance were seen: a mutation at N155H, or a mutation at Q148 to Q148H, K, or R. The Q148 mutation was often accompanied by a mutation at G140S or A that appears to mitigate replication defects conferred by the Q148 mutation. 32 of the 35 patients had more than one RAL mutation. Not unexpectedly, the appearance of two or three RAL mutations conferred higher levels of resistance, and a greater ability of the mutant virus to replicate in culture ("fitness").
Again, as discussed at CROI, mutations were found in or around the integrase enzyme active site. The pattern of mutations seen did not relate to the dose the patient was assigned. In patients with viral rebound with the N155H virus, viremia was usually to a lower level than seen during previous episodes of treatment failure. N155H was often later replaced by Q148H, again suggesting that the Q148H mutation impairs viral replication to a lesser extent that the N155H mutation. No data was yet available as to whether failure was related to RAL levels achieved in patients. In initial studies, the 400 mg bid dose was selected when it was recognized that initial declines in viremia were no greater using a 600 mg bid dose than a 400 mg bid dose. However, as there is no increased toxicity, thus far, of the 600 mg dose a consideration for future research might be the testing of higher RAL doses in selected situations.
Information about resistance to the "other" integrase inhibitor, elvitegravir (EVG), in development by Gilead, was also presented at the DRW (Abstr. 9). The emerging data seems in some ways reminiscent of protease inhibitor resistance. In the laboratory, the initial mutations developed against EGV appear to distinguish it somewhat from RAL. Mutations at E92Q and T66I develop first, giving 33-fold and 15-fold resistance to EGV, respectively, and 6-fold and no resistance to RAL. Secondary mutations later develop in each of these two "resistance pathways;" only one of these increased resistance to RAL.
However in clinical samples, the story is somewhat different and suggestive of much more cross-resistance between EGV and RAL. This data was obtained from a Phase II clinical trial (GS-US-1830105). 278 patients were randomized to 20 mg, 50 mg or 125 mg once-daily doses of ritonavir-boosted EVG, or comparator-boosted PI (CPI/r), plus optimized NRTIs ąT-20. GSS was 0 in 50% of the patients, a cohort with advanced drug resistance. Darunavir and tipranavir were added in some patients at week 16 due to emerging data during the study. The late addition of a new protease inhibitor in some patients may influence the picture of emergence in EGV resistance from this study, but this is the best information we have at this time. Only resistance information from the highest EGV dose arm, 125 mg, was presented. Integrase genotyping and phenotyping for patients with viral failure (n=30) was performed using the experimental PhenoSense and GeneSeq INI assays (Monogram Biosciences). Resistance assays were available from 28 of the 30 patients.
The pre-clinical information from lab-based studies of resistance proved to be half-true in the clinic. The most common integrase mutations that developed were E92Q, E138K, Q148R/K/H, N155H, S147G and T66I/A/K. So mutations at E92, T66, and S147 seen after exposure to EGV in the lab were also seen in the clinic, but mutations at codons E138, Q148 and N155 seen after clinical exposure to RAL were also seen after clinical exposure to EGV. These mutations were seen in up to 50% of the samples, and other less frequent mutations were seen in up to 20% of the samples.
Significantly, at failure viral isolates showed a mean of 151-fold resistance to EGV, and a mean of 28-fold resistance to RAL. As we are not really sure yet how much resistance will translate to risk of drug failure with these new integrase inhibitors, one must for the time being assume that there is clinically cross-resistance between these two integrase inhibitors. And as one might expect, replication capacity declined to a median of 54% at viral failure. One might hope that if a patient fails RAL or EGV, this replication capacity effect might translate to some clinical benefit, as has been seen with NRTIs and PIs, but that has yet ot be proven.
It is clear that the licensure of HIV integrase inhibitors will be a watershed event for HIV therapeutics. It is also clear that clinical trials data is thus far insufficient to precisely, optimally guide the use of integrase inhibitors. And like all antiretrovirals, the optimal use of an integrase inhibitor will likely be different in the face of advanced drug resistance, intermediate levels of drug resistance, and in initial therapy. When integrase inhibitors are licensed information for their optimal use will not be simultaneously available for each clinical situation.
Conceptually, resistance to integrase inhibitors thus far appears somewhat similar to protease inhibitor resistance. A single point mutation near the enzyme active site develops, which allows enzyme function in the presence of inhibitor. The mutant enzyme functions at a lower efficiency. Secondary mutations later develop that allow improved enzyme function in the presence of inhibitor.
Clearly, resistance to integrase inhibitors can develop in some patients, nearly always in the setting of insufficiently potent background therapy. The clinical challenge to the use of integrase inhibitors arises as it is difficult to precisely judge the potency of background therapy required to insure success with integrase inhibitors. Certainly there are patients in clinical trials with background therapy with a GSS score of 0 who have been successfully suppressed on integrase inhibitors. Unfortunately, the very same sort of patients are also those that are most likely to fail therapy with integrase inhibitors.
The 14th HIV DRW, not unexpectedly, did not provide guidance for the use of integrase inhibitors at earlier stages of disease: in the face of intermediate levels of resistance, or during initial therapy. Some more information from RAL clinical trials will soon be presented at IAS in Sydney. Stay tuned.
Resistance and using the newest drugs, part 2: CCR5 receptor antagonists
The results of actual clinical experience in treating patients with the R5 inhibitor Maraviroc (MVC) was presented by a group from Pfizer (Abstr. 56). In the Phase III MOTIVATE 1 and 2 studies MVC was given in combination with optimized background therapy (OBT) to treatment-experienced patients. Detailed clonal analyses of plasma viral sequences were conducted on samples from 20 patients (16 on MVC and 4 on placebo) in whom CXCR4-using virus was detected by the Trofile assay during the study. Of these 20, 9 suffered early virological failure, 5 late failure, and 6 were still responding at week 24 of the study.
For each of the 20 patients, 192 individual sequences were examined from a baseline plasma sample (yes, that means 3840 sequences), and 48 sequences from on-treatment samples (960 of these), and phenotypically screened for tropism. In 14 patients, CXCR4-using viruses found in on-treatment samples were also found in the baseline sample. In 10 of the 14, CXCR4-using clones were present at a low frequency (16%), and at a frequency of >10% in the remaining 4 patients. These 4 patients had been designated to dual-mixed virus (using both R5 and X4) by tropism testing at entry.
However, in the 6 patients in whom CXCR4-using clones were not detected at baseline, CXCR4-using clones were found in on-treatment samples and these viruses were phylogenetically distinct from baseline and on-treatment CCR5-tropic clones, and contained between seven and 17 amino acid differences in the 35-amino acid V3 loop alone. This strongly suggested that the X4-using virus that developed on MVC therapy did not arise from serial mutation of the majority, circulating R5-using viral species, but more likely from very low-level X4 virus not detected in the 192 baseline samples studied (suggesting a frequency of < 5%).
Of interest, 4 patients on MVC and 2 on placebo continued to respond to OBT+MVC at week 24, despite the presence of X4 virus. In one patient who failed with X4 virus and stopped MVC, X4 virus disappeared and was replaced by R5 virus in the absence of continued MVC therapy.
The development of resistance to the R5 inhibitor vicriviroc (VCV) during a clinical trial of this inhibitor (ACTG 5211, abstr. 13) was reported. Coreceptor usage, VCV susceptibility, and envelope sequences of HIV-1 from 2 subjects with virological failure at week 16 of therapy were randomly selected from each of the 4 study arms for clonal sequence analysis (placebo, 5mg, 10mg, and 15mg of VCV). VCV susceptibility, co-receptor usage (using the Monogram Trofile assay), and envelope cloning and sequencing were performed on samples obtained at entry, confirmation of virological failure and week 24.
No consistent increase in resistance to VCV was observed in samples obtained from the eight randomly selected subjects through week 24 (maximum increase compared to control = 2.83-fold). HIV-1 variants from four of four subjects enrolled in the VCV 5 mg and 10 mg arms had changes in the HIV envelope V3 loop stem that became fixed in the population following VF; such changes were not found in samples from subjects in the placebo or 15 mg arms.
However, these changes did not correlate with VCV resistance.
In one subject, however, increased resistance to VCV did emerge in correlation with multiple changes in the V3 loop sequence. Tropism assays on samples from these same time points showed a modest capacity to infect CXCR4-expressing cells, although the virus remained predominantly R5. A sample obtained 5 months after stopping VCV showed R5 virus with return towards phenotypic susceptibility and baseline V3 loop sequences.
A central issue in the use of R5 inhibitors is the size of the viral population in a given patient that can use the alternate X4 co-receptor for entry, denoting intrinsic resistance to R5 inhibitors. Richard Harrigans presentation (Abstr. 149) resulted in a lively debate as to the possible utility of an initial screening test seeking a focused set of mutations in the V3 envelope sequence that might predict X4 receptor usage. Harrigan compared the ability of a variety of algorithms to analyze plasma viral sequences and predict X4 receptor usage. He concluded that current coreceptor prediction algorithms were not yet adequate for predicting HIV X4 coreceptor phenotype in clinical samples, and that currently assays such as Monograms Trofile were more sensitive. However, he proposed that such genotypes could be used as a screening tool, as they can be performed more cheaply than a Trofile assay. This may be true, for those that have the time to use the web-based analysis systems used by Harrigan. And certainly, genotypic predictors might improve over time.
However, given the greater complexity of receptor interactions in comparison to drug-binding site interactions, such a genetic database may take some time to develop. Consistent with this suggestion, Mori (Abstr. 10) reported that a wide variety of V3 sequence changes appeared in patients treated with Maraviroc in concert with a decrease in peak inhibition of these viruses in culture by MVC.
Vandenbroucke presented work done at Virco (Abstr. 139) that attempted to determine the detection limit of X4 viruses in an abundant R5-tropic virus population using mixtures of lab-derived HIV strains JRCSF (R5) and NL4.3 (X4) in a model system. They found that experimentally X4 minority populations could be detected at a level of 1% when the viral load was high (ca. 5 logs), but only when 20% of the viruses were X4 at a lower viral load (ca. 3 logs). They predicted that establishing cut-offs for X4 minority species will be difficult, and affected by many factors (input volume of patient sample, patient VL, detection limits, PCR variation).
Buontempo (Abstr. 122) studied how the affinity of viruses for the R5 receptor (how hard the virus sticks to the receptor) changes in the presence of R5 inhibitors in a laboratory model. As has been suggested in other studies, drug-resistant viruses had higher affinity for the drug-bound form of CCR5. That is, viruses learned how to evade drug by binding around the drug and still using R5 to enter cells. The authors suggested that high viral loads might therefore correlate with an increased chance of R5 inhibitor failure, a pretty safe bet as this is true with all previous antiretrovirals.
Other insights into resistance to CCR5 inhibitors were provided by Huang and a group of scientists at Schering Plough and Monogram (abst. 121). Resistance mutations to R5 inhibitors have frequently been studied within the critical V3 loop of the gp120 envelope glycoprotein, a domain of gp120 that reaches out to grasp the CCR5 receptor. Studies using the developmental R5 inhibitor SCH-C demonstrated that mutations outside the V3 loop could contribute to SCH-C resistance. As CCR5 antagonist-resistant variants can evolve using different pathways in addition to those involving primarily mutations in V3 loop, following the complexities of R5 inhibitor resistance may become extremely difficult.
A. Jekle and G. Heilek presented studies of a new CC5 entry inhibitor in development at Roche (Abstr. 11). The team first evolved a highly maraviroc- and vicriviroc-resistant virus by growing virus in the presence of maraviroc. The resultant virus was 8,000- and 12,000-fold resistant to maraviroc (MVC) and vicriviroc (VCV).
The team then tested the activity of the CCR5 small-molecule inhibitor RO1752 against this resistant virus, and found that RO1752 was still highly active. Binding studies suggested that MVC and RO1752 share the same binding site. However, RO1752 appears to bind more rapidly and in a different physical way to the site than MVC. Mutations that eliminated binding of MVC did not block binding of RO1752.
It would seem likely that the virus could find mutations that block RO1752 binding, or bound the R5 receptor despite the presence of RO1752, given enough time and selection. However, similar to the observations of antiviral synergy with some NRTIs, these observations raised the interesting possibility that different classes of R5 inhibitor might be developed, to which HIV would have a hard time developing simultaneous resistance.
Overall, there was considerable discussion and controversy about R5 inhibitors at the 14th DRW. It seemed likely that very low levels of X4 virus could be found in most, if not all patients, if one looked hard enough. A major point of controversy was whether or not tropism assays should therefore be held to a higher standard, and be expected to detect populations of less than 5%, a feat currently unachieved by other standard resistance tests.
Another major issue, was whether the emergence of X4 virus was a clinically different event when it occurred during the progression of untreated HIV viremia vs. when it occurs in the face of R5 inhibitor selection in a patient failing R5 inhibitor therapy. (that is, emergence of X4 before ART in patient vs on ART)
There has been a longstanding chicken-and-the-egg debate about X4 virus. Although its detection as the predominant viral species in the peripheral blood is clearly associated with a higher risk of progression to clinical AIDS, it has never been clear that X4 virus is the DRIVER of accelerated pathogenesis, or a MARKER of a collapsing immune system.
So far, with extremely limited experience, no cases have yet been reported in which a patient failed an R5 inhibitor, had X4 viremia which persisted, and was seen to then have rapid progression to clinical AIDS. But this is clearly the nightmare scenario that is feared.
So again, the use of R5 inhibitors is currently in an even more uncertain no-mans land than inhabited by integrase inhibitors. Clinical trials data will be insufficient to precisely, optimally guide the use of R5 inhibitors at the time of their licensure. It will be difficult to precisely judge the potency of background therapy required to insure success with R5 inhibitors. But certainly there is concern that patients with insufficient background therapy might be harmed by exposure to R5 inhibitors. Given the central importance of infection of cells via the R5 receptor to the initial pathogenesis of HIV infection, it is logical to hypothesize the such inhibitors would be of benefit in early disease, or as part of initial therapy. But such studies have yet to be performed, and will be challenged by the issue of X4 subpopulations.
Patients with advanced drug resistance are likely to benefit for therapy that combines several new agents, such as R5 inhibitors, integrase inhibitors, or enfuvirtide. What is needed are algorithms to precisely calibrate the potency of therapy with these new agents and current therapeutics, and match it to the potency needed in each patient to ensure stable suppression of viremia. Until such metrics are available, treatment decisions will be a challenge.
Resistance and the next generation of NNRTIs: TMC 125
Resistance correlates for TMC125 (etravirine) were presented by the Tibotec group in an analysis culled from the Duet 1 and 2 studies. Efficacy data for these studies is due to be presented at IAS in Sydney next month. This study provided patients failing therapy with darunavir, other optimized background therapy, and the new NNRTI TMC125 or placebo. Resistance in 406 of these patients who did not use enfuvirtide as a new drug, and did not fail for "non-drug" reasons (eg. toxicity, nonadherence, etc) was analyzed to give the "purest" view of mutational drug resistance to TMC 125.
The group identified resistance-associated mutations (RAMs) in several ways. 14 mutations are identified in IAS-USA drug resistance guidelines for other NNRTIs, but the number of these mutations carried by a patient did not very well predict response or lack of response to TMC 125. Response was defined as the attainment of <50 copies of HIV RNA at 24 weeks in the subgroup of patients (n = 52) who did not have RAMs on genotype at entry to the studies. The number of a larger group of 44 NNRTI RAMs, consisting of mutations suggested to confer resistance in prior studies of TMC 125 and other NNRTIs was also found to be poorly predictive of response to TMC 125. However, a subgroup of 13 of these RAMs, if present at baseline, did predict lack of response to TMC 125. In patients with 3 or more of these mutations, attainment of <50 copies/ml of HIV RNA was seen in about a third less patients. With 5 or more of such mutations, suppression was 10 times less likely. These mutations do not include K103N, but do include Y181C, G190A, and other familiar and unfamiliar NNRTI mutations (90I, 98G, 101E or P, 106I, 179D or F, 181 I or V, 190S). It would seem a new algorithm or pocket card is needed to guide the clinician when TMC 125 is used. And as has been said in the past, prolonged failure while on an NNRTI is to be avoided, as there is no good evidence of a "fitness benefit" of failure and accumulation of NNRTI mutations, and now it is clear that acquiring an array of NNRTI mutations will seriously compromise TMC 125, and perhaps other new NNRTIs.
Working on better versions of old drugs
Despite the excitement that swirls around antiretrovirals (ARVs) with new mechanisms of action such as integrase inhibitors or CCR5 inhibitors, considerable effort continues to improve and expand the "old" classes of ARVs. Little about therapeutics directed at completely new targets was discussed at the 16th DRW. Perhaps this is because the development path for improving on current classes of drugs is better defined, less risky, and perhaps ultimately cheaper. Or perhaps there is nothing really new in the pipeline that is ready for primetime yet.
K. Klumpp (abstr. 28) presented the efforts of Roche to develop a novel non-nucleoside reverse transcriptase inhibitor (NNRTI). RO-0335 was developed based on crystal structure modeling of NNRTIs in the active site binding pocket of reverse transcriptase. The new drug was designed to sit in the binding pocket in a way that was shifted away from sites in RT where resistance mutations are know to obstruct drug binding. RO-0335 was found to be 5-10-fold more active in laboratory tests against virus isolates that were resistant to efavirenz (Sustiva) or etravirine (TMC-125). When tested against a panel of viruses from patients who failed therapy in the TORO studies of enfuvirtide, containing a mean of 3 NNRTI mutations, nearly 90% were inhibited by RO-0335. Resistance can be developed to RO-0335 after about 12 weeks in culture, and when seen is associated with an unusual triple mutation cluster (V106A/F227C/M230L; abstract 33). However, as the drug was found to be poorly bioavailable, a phosphonate prodrug was developed. This drug, R1206, appears to have a long half-life, good bioavailability, an acceptable animal safety profile, and is soon entering human phase I studies.
Just in case you thought the last protease inhibitor had been invented, Vacca from Merck (Abstr 15) unveiled PL-100. In the lab, PL-100 selects for a novel pattern of mutations (K45R, M46I, T80I and P81S) in the protease gene. This virus was only 5-fold resistant to PL-100, and took almost a year to select in the lab. Phase 1 human studies are beginning.
Gilead also presented studies (abstr. 16) of a new PI, called GS-8374. Extensive drug selection against lopinavir, atazanavir, and darunavir did not select a virus resistant to GS-8374. Selection with GS-8374 for 5 months induced only Gag mutations and approximately fivefold reduced susceptibility to GS-8374. Further extension of the selection to 11 months resulted in protease mutation R41K together with additional mutations in Gag and 15-fold reduced susceptibility to GS-8374, with minimal cross-resistance to other PIs.
Protease resistance due to changes outside of the protease enzyme:
Dierynck presented an analysis (Abstr. 21) of the role of HIV gag cleavage site mutations in resistance to darunavir, using isolates from the POWER studies. Mutations in the gag polyprotein that is cleaved by protease have been for some time speculated to contribute to protease inhibitor resistance. One way to think about this is to consider that changes in the Gag target of protease allow a crippled protease enzyme that carries many resistance mutations to continue to function, cleaving this more receptive Gag target and allowing replication to proceed. Dierynck showed that while rare (found in only 6% of patient) Gag cleavage site mutations were associated with decreased response to darunavir. Practically speaking, it remains to be seen whether Gag cleavage site mutations are a cause of PI resistance or solely an effect of protease resistance mutations.
Combination Pre-exposure Prophylaxis (PreP) in a monkey model
Garci_a-Lerma from the CDC (Abstr. 85) presented an exciting study of antiretroviral pre-exposure prophylaxis (PreP) with tenofovir (TFV) and emtricitabine (FTC). Daily PreP with TFV and FTC fully protected macaques from rectal transmission of a hybrid HIV-SIV virus called SHIV. While this virus is not identical to HIV, it is an accepted model.
The researchers tested if PreP given around the time of virus exposure can be as protective as daily PreP. Monkeys were doses IV FTC/TFV 2 hrs. before and 24 hrs. after virus exposure, another other group of animals received FTC/TFV only 2 hrs. before virus exposure. All animals were exposed rectally once weekly for up to 14 weeks with a low dose of R5-using SHIV.
21 untreated macaque monkeys were challenged with virus without PreP. 16 of 21 were infected during the first four challenges, four were infected between exposures 8 and 12, and only one remained uninfected after 14 exposures. All the animals receiving PrEP before and after virus exposure were protected. Single-dose PreP was effective but did not protect all the animals.
Looking "Ultra-Deeply" for resistance
Ultra Deep sequencing is a new, expensive, high-tech method to simultaneously examine thousands of DNA or RNA molecules. Kozal (abstr. 134) presented the use of this technology to detect low levels of resistant HIV in patients participating in the FIRST study. The study was carried out between 1999 and 2002 in 1397 antiretroviral-naive people. See Jules Levins and Mark Mascolinis two separate detailed reports on the NATAP web site (www.natap.org) for in-depth information.
In brief, standard sequencing detected PI mutations in 2.3% of the patients, NNRTI mutations in 6.6% of the patient, and NRTI mutations in 5.8% of the patients. Ultra Deep sequencing found more patients with mutations: 4.7%, 15.1%, and 14.3%, respectively. Importantly, detection of NRTI or NNRTI mutations by Ultra Deep was significantly associated with virological failure during the study.
Ultra Deep will be an important research tool, as it should lead to a better understanding of which factors lead tiny populations of drug-resistant viruses (likely present in many patients) to suffer failure of therapy. Currently, it is too costly to employ as a routine clinical assay. And these findings need validation in a more modern study. However, this is an important FIRST step. (note from Jules: Roche recently purchased this test)
|
| |
|
|
|
|
|
