Butting Heads With Resistance
The XV International HIV Drug Resistance Workshop, held June 13-17, 2006, at Sitges, Spain, was notable for its inclusion of data on resistance to new antiretroviral and anti-hepatitis drugs, as well as new information on structured treatment interruptions (STIs), the potential benefit of genotypic testing prior to initiation of antiretroviral therapy (ART), and resistance in nonsubtype B viruses.
Johan Vingerhoets (Tibotec) compared the virologic response of patients whose virus had the NNRTI-resistance mutation Y181C, which is associated with reduced TMC-125 susceptibility, with those whose viruses did not have Y181C (but often had K103N alone) [Abstract 17].1 In an unadjusted analysis, the 23 patients with Y181C had a 0.86-log reduction in plasma HIV-1 levels compared with a 1.7-log reduction in those without Y181C. Although this difference was statistically significant, the results were confounded by the fact that viruses with Y181C generally had more NNRTI-resistance mutations than those without Y181C.
This study supports the idea that TMC-125 will be a useful option in patients who have failed a previous NNRTI-containing regimen, and that it will be particularly active in those having viruses with just K103N. It also provides a rationale for not continuing currently approved NNRTIs in patients who have developed NNRTI resistance, because continued treatment will likely select for additional NNRTI-resistance mutations without providing virologic benefit.
At last year's XIV International HIV Drug Resistance Workshop in Québec City, Canada, Dirk Jochmans (Tibotec) described a new type of reverse transcriptase inhibitor -- a nucleotide-competing reverse transcriptase inhibitor (NcRTI).2 These compounds are not nucleoside analogs (and therefore do not require triphosphorylation). However, they bind close to the same site in the enzyme -- the active site -- as do nucleoside reverse transcriptase inhibitors (NRTIs) and are therefore nucleoside-competing. At this year's Resistance Workshop, Jochmans described the activity of the prototype NcRTI compound versus 1,700 clinical HIV-1 isolates [Abstract 16].3 The researchers also selected for NcRTI mutations by culturing HIV-1 in the presence of increasing concentrations of this drug. Only M184V (three-fold decreased susceptibility) and Y115F (10-fold reduced susceptibility when present in combination with M184V) reduced NcRTI susceptibility. Viruses with multiple thymidine analog mutations (TAMs) and with the T69 insertion or with Q151M mechanisms of multi-drug resistance retained susceptibility to the NcRTI. Viruses with NNRTI-resistance mutations also retained susceptibility. Of note, the K65R mutation, which reduces susceptibility to all NRTIs other than zidovudine (ZDV), increased NcRTI susceptibility in a way similar to that observed with ZDV. These in vitro data suggest that such a compound may have clinical synergy with other NRTIs, such as ZDV (as it is not affected by TAMs) and tenofovir (TDF) (as it is highly active against viruses with K65R), and that it could be a useful addition to the current reverse transcriptase inhibitor armamentarium.
Rafael Landovitz (Schering-Plough) provided drug susceptibility data from the recently completed phase 2 trial of the CCR5 inhibitor vicroviroc, in which vicroviroc was compared with efavirenz (EFV) in combination with Combivir [Abstract 18].4 Landovitz reported the results of susceptibility testing of baseline viruses as well as viruses tested after the 14-day lead-in of vicroviroc (during which time it was given as monotherapy). The authors found that baseline susceptibility of viruses from patients who had persistent virus suppression while receiving vicroviroc and Combivir did not differ from the baseline susceptibility of viruses obtained from patients who eventually developed virologic failure with vicroviroc. The authors also found no evidence to suggest that the two-week monotherapy lead-in with vicroviroc selected for virus strains with reduced susceptibility. This study suggests that baseline resistance to vicroviroc does not seem to be the explanation for the poorer performance of vicroviroc relative to EFV when combined with Combivir in this early phase 2 study.
Although of little clinical relevance in the near term, researchers from Merck & Co. presented structural data on the interaction between the Fab of a monoclonal antibody D5, which recognizes and binds to the heptad repeat 1 (HR1) region of gp41, thereby blocking the entry of HIV into cells by a mechanism similar to that of the fusion inhibitor enfuvirtide (ENF) [Abstract 14].5 The long-term implications of this line of work may become extremely important should Merck & Co. be able to develop an immunogen that induces endogenous, broadly neutralizing antibodies that inhibit HIV-1 by this mechanism. Such an immunogen could lead to a vaccine that might prevent initial HIV infection in a significant proportion of vaccinated individuals.
Marie-Pierre Bethune (Tibotec-Virco) presented the first analysis of the phenotypic and genotypic determinants of resistance to TMC-114 based on a comprehensive analysis of pooled data from the POWER 1, 2, and 3 studies [Abstract 73].6 In these studies, 458 patients received the currently recommended dose of 600 mg TMC-114 plus a 100 mg sub-therapeutic dose of ritonavir (RTV). Baseline phenotype (fold-change reduction in TMC-114 susceptibility as determined using Virco's Antivirogram) was correlated with the virologic response to therapy. At week 24, 50%, 25%, and 13% of patients with TMC-114 fold-change levels of <10, 10 to 40, and >40 had reached an RNA level <50 copies/mL.
A series of 11 protease mutations were associated both with reduced phenotypic susceptibility in the Virco database and with reduced virologic response: V11I, V32I, L33F, I47V, I50V, I54L or M, G73S, L76V, I84V, and L89V in the POWER studies. When three or more of these mutations were present, there was a diminished response to TMC-114 in terms of achieving a plasma HIV-1 reduction to <50 copies/mL: zero to two mutations >40%, versus three or more mutations TMC-114 has a very similar structure to amprenavir (APV). Therefore, it is not surprising that many of the 11 TMC-114 mutations are well-known APV-related mutations, particularly V32I, I47V, I50V, I54M/L, and I84V. L33F and L76V are also APV-related mutations, although these associations are not as widely known. G73S is associated with resistance to all protease inhibitors (PIs). V11I and L89V are mutations that are highly correlated with other drug-resistance mutations and may not be of much predictive power by themselves.
HIV-1 reverse transcriptase has 1,680 nucleotides encoding 560 amino acids. It is a multifunctional enzyme that has a polymerase region with reverse transcriptase and DNA polymerase activity, and an RNAseH domain which catalyzes the removal of RNA nucleotides from the DNA-RNA dimer created following reverse transcription. The 5' (or N-terminal) part of the molecule (approximately the first 350 amino acids) is involved in polymerization. The 3' part of the molecule (approximately amino acids 440 to 560) is responsible for the RNAseH activity. Amino acids 350 to 440 encode the connection subdomain.
However, standard HIV-1 genotypic resistance testing for reverse transcriptase inhibitor resistance usually involves sequencing only the first 750 of the 1,680 nucleotides (encoding the first 250 amino acids) of the reverse transcriptase gene. It has previously been reported that several mutations beyond position 250 may influence NRTI or NNRTI susceptibility. The mutation L318F causes delavirdine (DLV) resistance and low-level nevirapine (NVP) resistance. It usually occurs in combination with other drug-resistance mutations. L283I, a common polymorphism, may reduce NVP susceptibility about two-fold. G333E, another common polymorphism, may reduce ZDV susceptibility in combination with other ZDV-resistance mutations.
Jessica Brehm (University of Pittsburgh) and colleagues [Abstract 125]7 described two novel mutations in the connection and RNAseH subdomains -- A371V and Q509L -- that are selected in vitro in the presence of ZDV. These mutations appear to occur only in the presence of other TAMs, and in this setting they further reduce ZDV susceptibility. They also slightly reduce susceptibility to lamivudine (3TC) (15-fold versus seven-fold), abacavir (ABC) (three-fold versus 2.4-fold), and TDF (2.7-fold versus 1.5-fold) in the presence of certain patterns of TAMs. Krista Frankenberry (US National Cancer Institute) and colleagues [Abstract 126] described other mutations in the connection region -- G335C, N348I, and A360I -- that also appear to reduce ZDV susceptibility.
Soumi Gupta and colleagues (Monogram Biosciences) [Abstract 127]8 compared the results of their standard assay, which includes reverse transcriptase residues 1 to 305, with an assay that includes the entire patient-derived pol gene (1 to 560) on 52 clinical virus isolates. Susceptibility testing revealed that the results were similar for the majority of isolates. However, seven of 52 results and one of 52 EFV results showed a more than five-fold difference. Mutations at three positions in the extended pol fragment (348, 369, and 399) appeared to be responsible for these changes, usually in combination with other drug-resistance mutations. Several residues beyond position 250 may influence the level of resistance to different NRTIs and NNRTIs. However, the changes in susceptibility are usually subtle and usually occur only in the presence of other reverse transcriptase inhibitor resistance mutations. Therefore, there is little evidence for sequencing beyond position 250 during routine genotypic resistance testing for clinical purposes.
It has been known for more than five years that changes in protease cleavage sites -- particularly those that are in the Gag region -- modulate the effect of the mutations in protease that cause PI resistance. Protease cleavage site mutations have nearly always been shown to be compensatory, contributing to virus resistance and increased fitness only in the presence of primary protease mutations. In addition to the protease cleavage sites in Gag, other Gag mutations have been reported during the past several years to influence PI drug susceptibility.
Several experts have speculated on whether or not changes outside the protease may explain the fact that many patients who develop virologic failure while receiving a boosted PI have viruses that do not contain PI-resistance mutations. Studies to date, however, have suggested that this is unlikely to be the explanation, as nearly all studies have shown that recombinant viruses lacking protease mutations associated with resistance remain drug-susceptible, despite the fact that the PhenoSense assay contains the 3' part of the Gag, which is the part of the Gag for which most PI-related changes have been described.
Mounir Ait-Khaled (GlaxoSmithKline) and colleagues [Abstract 32]9 sequenced Gag cleavage sites from 32 patients in whom an antiretroviral regimen containing fosamprenavir (FPV)/RTV failed without PI-resistance mutations. Although several polymorphisms were identified in Gag, there was no evidence of selection as the samples obtained at baseline and at virological failure had nearly identical Gag sequences. Two years ago, Monique Nijhuis (University Medical Center, Utrecht) and colleagues [XIII International HIV Drug Resistance Workshop, Abstract 133]10 described a Gag cleavage site change that by itself appeared to reduce susceptibility to an investigational PI by two- to six-fold. Working with the same inhibitor (RO033-4649), this group performed additional selection experiments and identified additional non-cleavage site Gag mutations, which in in vitro experiments appeared to also reduce RO033-4649 susceptibility.
Most studies to date suggest that changes at Gag sites both within and outside of cleavage sites are compensatory changes that do not cause resistance to currently approved PIs. The provocative data by Nijhuis and colleagues, whether they apply only to RO033-4649 or to other PIs as well, suggest that much is still unknown about the initial processes by which HIV-1 protease dimerizes and then cleaves itself out of the Gag-pol polypeptide before acting upon the remaining cleavage sites in Gag and pol. Learning how this occurs might provide more insight into how changes in the Gag can decrease susceptibility to a PI without the presence of mutations in the protease itself.
Michael Norton (Abbott Laboratories) presented data on the development of drug resistance in the MONARK trial, which included 83 patients receiving lopinavir (LPV)/RTV monotherapy and 53 patients receiving LPV/RTV plus Combivir for 96 weeks [Abstract 74].11 Study participants were required to have CD4 counts >100 cells/µL and plasma HIV-1 RNA levels <100,000 copies/mL prior to study entry. The main findings of the study were that virologic failure, defined as plasma HIV-1 RNA levels >500 copies/mL, occurred in 25% (21 patients) of the monotherapy group and 6% (three patients) of the combination therapy group. However, PI-resistance mutations were found in only two patients of the monotherapy group (M46I; L10F + V82A) and in no patients in the combination therapy group.
This study suggests that virological failure with resistance is rare, even during prolonged LPV/RTV monotherapy. Virologic failure occurs at a higher rate among persons receiving LPV/RTV compared with those receiving LPV/RTV plus Combivir. However, the absence of drug resistance in all but two of the 21 patients receiving LPV/RTV suggests that nonadherence to LPV/RTV may be responsible for the higher rate of virologic rebound in the absence of other active drugs to suppress virus replication.
Michael Kozal (Yale University) presented data on the clinical implications of drug resistance from a large recently completed Community Programs for Clinical Research on AIDS (CPCRA) study [Abstract 79].12 The FIRST study followed 1,397 patients for five years following their randomization to an NNRTI-containing regimen or a triple-class regimen. The primary results of the study will be presented at the XVI International AIDS Conference later this year in Toronto. At this year's Resistance Workshop, Kozal reported that although the rate of first virological failure was highest in persons starting a PI-containing regimen, the rate of virologic failure with resistance was only marginally higher in those beginning therapy with a PI-containing regimen relative to an NNRTI-containing regimen.
Patients with PI resistance at first virologic failure did not have an increased risk of progression of disease compared to those without virologic failure. However, patients with NNRTI resistance were 4.4 times more likely to have progression of disease than those without virologic failure. This prospective study confirms a trend observed in several recent retrospective studies: although initial ART with an NNRTI-containing regimen has a lower risk of virologic failure than initial ART with a PI-containing regimen, the consequences of virologic failure for a patient on an NNRTI-containing regimen are usually more severe than the consequences of virologic failure of a PI-containing regimen, because NNRTI failure is more often associated with drug resistance and the loss of the NNRTI class as a future treatment option.
The desire to limit the costs and toxicity associated with ART through the use of intentional or structured treatment interruptions (STIs) is not unreasonable. However, the practical implementation has been associated with several negative consequences for patients. First, the SMART study was terminated prematurely last year when an interim analysis revealed that patients receiving intermittent CD4-guided therapy had greater HIV-related morbidity than patients who had received continuous therapy. Second, multiple small studies of STIs have shown that the development of new NNRTI resistance occurs more commonly in patients randomized to STIs than in control patients receiving continuous therapy.
The AIDS Clinical Trials Group (ACTG) 5170 study examined the clinical and virological outcome of treatment discontinuations in patients who had initiated ART when their CD4 counts were >350 cells/µL and viral load levels were <55,000 copies/mL, and whose viral loads on therapy were <400 copies/mL. The primary results of this study, which were reported by Daniel Skiest (Baystate Medical Center, Springfield, Massachusetts) and colleagues at the 13th Conference on Retroviruses and Opportunistic Infections (CROI) in February 2006,1 suggested that treatment discontinuation in this population is safe, provided that therapy is resumed if CD4 counts decrease to <350 cells/mL. Discontinuation of the NNRTI was staggered by two days.
Brad Hare (University of California, San Francisco) reported the results of an in-depth substudy of HIV-1 drug resistance in viruses from the 54 patients in ACTG 5170 undergoing interruption of an NNRTI-containing regimen after plasma HIV-1 RNA levels reached >5,000 copies/mL [Abstract 34].14 In this analysis, 11 of 54 subjects discontinuing ART had detectable mutations by either standard genotyping or by real-time polymerase chain reaction (PCR). Mutations were more likely to be found in persons with plasma HIV-1 RNA levels between 50 and 400 copies/mL, and appeared to increase the risk of subsequent virologic failure in five patients when treatment was restarted at a later time with an NNRTI-containing regimen.
This study is consistent with previous reports of STIs that suggest an increased risk of selecting for drug resistance during each cycle of treatment interruption/treatment re-institution. The researchers suggest that if treatment of an NNRTI-resistant virus is necessary, the lag between NNRTI discontinuation and the discontinuation of the remaining drugs in the regimen should be longer than two days. Although the number of patients in the study is low, there is a suggestion that those with plasma HIV-1 RNA levels between 40 and 500 copies/mL are at high risk of having detectable NNRTI-resistance mutations often only detected by more sensitive assays.
HIV-1 drug resistance can be acquired (developing in a patient receiving ART) or transmitted (occurring because a virus with drug-resistance mutations was transmitted to a treatment-naive patient). Although both acquired and transmitted resistance are public health concerns, transmitted resistance has the potential to more rapidly reverse the effectiveness of first-line therapy at the population level. Patients with transmitted resistance begin ART with a lower genetic barrier to resistance, a higher risk of virologic failure, and a higher risk of developing resistance even to those drugs in their regimen that were originally fully active.
Annemarie Wensing (University Medical Center, Utrecht) and colleagues [Abstract 98]15 reported that from 2002 to 2003, 96 of 1,050 previously untreated patients from 17 European countries in the SPREAD study had viruses with a drug resistance mutation. Seventy-one percent harbored only one drug-resistance mutation; whereas 29% harbored two or more drug resistance mutations. Fifty-two percent of resistant viruses contained at least one revertant mutation at reverse transcriptase position 215 or a mixture of wild-type and resistant virus. The 215 revertant mutations are transitional or sentinel mutations because, in contrast to the double nucleotide mutants 215Y and 215F, the revertant mutations (such as CDESNIV) do not reduce drug susceptibility, but rather indicate that an individual was probably infected with a virus containing T215F or T215Y and that the revertant evolved from this virus by back mutation.
Lisa Ross (GlaxoSmithKline) and colleagues [Abstract 107]16 described the prevalence of genotypic resistance in 1,795 previously untreated patients from 33 states in the United States enrolling in GlaxoSmithKline clinical trials between 2000 and 2004. The overall frequency of resistance was 10%, and included 6% NNRTI resistance, 4% NRTI resistance, and 3% PI resistance. Overall resistance doubled, and NNRTI resistance increased from 2% to 7% between 2000 and 2004.
To monitor and treat transmitted virus, genotypic testing to assist physicians in choosing initial therapy is now recommended by the US Department of Health and Human Services (DHHS) prior to the initiation of ART.17 This relatively new recommendation is based on the increasing number of reports in the United States and Europe that HIV-1 drug resistance can be detected genotypically in about 10% to 20% of newly infected individuals and in 5% to 15% of newly diagnosed individuals. Jeffrey Johnson (US Centers for Disease Control and Prevention [CDC]) and colleagues [Abstract 69]18 applied a highly sensitive real-time PCR assay to detect the common mutations K103N, Y181C, and M184V in plasma at frequencies as low as 0.5% to 138 previously untreated persons enrolled in a study of ABC/3TC/EFV from 2001 to 2002. The 138 patients included 69 patients who had developed virologic failure with this combination and 69 control patients who were successfully treated during the course of the study. Mutations at positions K103N, Y181C, and M184V were found in two, one, and one person by standard genotyping and in six, two, and four persons by real-time PCR.
Overall, 10 patients had at least one of these mutations. The subsequent response to therapy suggested that these low-level mutants were clinically relevant, in that two of 10 with a mutation never achieved virologic suppression, four of 10 failed within two months, two of 10 failed by month four, and two of 10 by month six. Three of four patients with available genotypes had the same mutations as the low-frequency variants present at baseline. This is a provocative study in that it not only shows that mutations may be missed at baseline but that the minor variants may be clinically relevant. It is not clear to what extent the six patients in whom K103N, Y181C, or M184V were detected only by real-time PCR had other mutations on their genotypes that may have pointed their physicians to the likelihood of primary HIV-1 resistance.
Eoin Coakley (Monogram Biosciences) described clinical cut-offs for tipranavir (TPV), LPV, saquinavir (SQV), and APV derived from correlating baseline phenotype data with virologic response (change in RNA at week four) in more than 500 patients in the RESIST trial [Abstract 71].19 The lower clinical cut-off was defined as the fold change where HIV RNA response first begins to decline relative to those with fully susceptible virus (fold-change <1). The upper clinical cut-off was defined as the fold-change above which the attributable RNA decline from baseline was <0.3 logs. The challenge, of course, was for the model to take into account the potency of the non-PI part of the regimen, and this was done using a continuous phenotypic susceptibility score for each drug in the regimen. Among the 253 patients who received TPV, a decrease in virologic response relative to susceptible virus occurred at a fold change of 2 or higher. Nearly complete loss of activity was observed at a fold change of 8 or higher. Comparable cut-offs for SQV/RTV, which was used in 106 patients, were 2.3 and 12; for APV/RTV 4 and 11, and for LPV/RTV 9 and 55. The fact that 53% of the baseline isolates were fully susceptible to TPV whereas only 18%, 17%, and 26% were fully susceptible to LPV, SQV, and APV underscores the unique resistance profile of TPV.
In a presentation highlighting the growing problem of resistance in developing world countries, Florence Doualla-Bell (McGill AIDS Center, Montreal) and colleagues [Abstract 46]20 described the genetic mechanisms of resistance in 23 Botswana patients with HIV subtype C virus developing virologic failure while receiving one or more consecutive dual NRTI/NNRTI regimens. Of 11 patients who had received ART containing the combination of stavudine (d4T) and didanosine (ddI), seven developed the reverse transcriptase mutation K65R. In three patients, that mutation occurred in combination with Q151M, which is selected more often by ddI than by any other drug, and which is associated with high-level resistance to all drugs except 3TC, emtricitabine (FTC), and TDF. In four patients, K65R occurred alone or with M184V.
This study adds data relevant to two questions. The first relates to the activity of d4T in treating viruses with K65R. Phenotypic data show that virus isolates with K65R have slightly reduced d4T susceptibility, and that K65R has emerged during in vitro passage with d4T. Conversely, some clinical studies, including one presented at this year's Resistance Workshop,21 suggest that in multivariate models d4T retains in vivo activity in treating viruses with K65R. The current study supports the idea that K65R does interfere with d4T activity, as K65R rarely if ever emerges in patients receiving ZDV plus ddI.
The second question relates to the genetic mechanisms of resistance in different subtypes. Previous studies have shown that each of the subtypes appears to respond equally well to ART. In addition, the mutations responsible for resistance in most subtypes appear to be similar. However, subtle differences exist between the mechanisms by which different subtypes develop resistance to different drugs. Although the authors claim that K65R may be a more common escape mutation in patients receiving subtype C viruses, many attendees believed that additional confirmatory reports are needed. Moreover, the combination of ddI/d4T is not recommended and is fortunately rarely being used, even in resource-limited countries.
Anne-Geneviève Marcelin (Hôpital Pitie-Salpetriere, Paris) and colleagues [Abstract 77]22 described the resistance outcomes in a group of 109 patients in Mali infected with viruses belonging to the CRF_02 subtype, who were treated with a fixed-dose combination of d4T, 3TC, and NVP. After a median time of seven months, 22 patients had a virus load >200 copies/mL. Eleven had no drug-resistance mutations and eleven had M184V (10 patients) and/or an NNRTI-resistance mutation (10 patients). The mutations observed were similar to those observed in subtype B viruses, suggesting that in this population of CRF_02 viruses, the response to ART and the mechanisms of resistance are similar to those observed in subtype B viruses.
Walid Heneine (US Centers for Disease Control and Prevention) and colleagues from the CDC and Emory University [Abstract 94]23 challenged 27 rhesus macaques with 14 weekly rectal exposures to a low dose of simian/human immunodeficiency virus (SHIV) (400,000 virions but a tissue culture infectious dose of just 10 units). Tenofovir and FTC (six patients) or FTC alone (six patients) were administered subcutaneously to 12 macaques. Fifteen control macaques were not given preventive therapy. Fourteen of the 15 control macaques became infected after a median of two challenges. In the FTC group, four of the six macaques became infected after a median of 11 challenges. In the combination therapy group, no animal became infected throughout the complete challenge period of 14 months. Although not directly related to drug resistance, this study is one of the most convincing demonstrations of the potential efficacy of a drug combination for preventing infection. The reasons for the success of TDF plus FTC despite the failure of FTC alone also remain to be determined.
The entire first half of the first day of the XV International HIV Drug Resistance Workshop was devoted to a hepatitis B (HBV) and hepatitis C virus (HCV) drug resistance workshop, highlighting the importance of hepatitis coinfection in HIV-infected patients. The introduction of HBV and HCV drug resistance presentations at this workshop is a logical consequence of three trends in antiviral therapy:
Anti-HBV therapy, especially small molecules that specifically inhibit the virus (and therefore lead to specific mutational changes responsible for drug resistance), is further developed than specific anti-HCV therapy. Three small molecules that inhibit HBV polymerase -- 3TC, adefovir, and entecavir -- have been licensed for use in treating HBV. Two other commonly used anti-HIV drugs -- TDF and FTC -- have potent anti-HBV activity and are also occasionally used off-label for treating HBV. A sixth compound, telbivudine, is in advanced clinical development.
Many anti-HCV small molecule inhibitors of the NS3 serine protease and the NS5B RNA-dependent RNA polymerase with both in vitro and in vivo activity have been developed and are in various stages of clinical development. These inhibitors belong to three mechanistic classes: protease inhibitors, nucleoside polymerase inhibitors, and allosteric nonnucleoside polymerase inhibitors.
Much of the optimism for the future of HCV therapy has been focused on HCV protease inhibitors, because several of these, particularly Vertex 950, have been further along in clinical development and because phase 1 clinical data have shown marked decreases in plasma virus levels with their use. Hepatitis C virus levels have decreased more than 4 logs in a two-week study of VX-950 and 2 to 3 logs in a two-day viral kinetic study of BILN-2061.
In contrast, most of the polymerase inhibitors have produced less marked 1- to 2-log reductions in plasma HCV levels. There are two different types of investigational small molecule polymerase inhibitors. The first are nucleoside analogs, which cause chain termination; the second are nonnucleoside polymerase inhibitors, which bind to different pockets in HCV polymerase. In contrast to HIV NNRTIs, nonnucleoside HCV polymerase inhibitors bind to at least three different sites in HCV polymerase. These inhibitors have also been characterized by the region to which they bind in the enzyme. As most polymerase enzymes bear some resemblance to a right hand, with regions corresponding to a thumb, palm, and fingers, these regions have also been used to more specifically describe the activity of HCV polymerase inhibitors.
This year's Resistance Workshop featured three oral presentations on the activity and mechanisms of drug resistance associated with anti-HCV polymerase inhibitors. Investigators from Merck & Co. presented an important proof of concept study, which showed that two chimpanzees receiving the novel nucleoside polymerase inhibitor MK-0608 had 2- to 3-log reductions in virus levels after a single intravenous dose, and up to a 5-log reduction after seven days of therapy [Abstract 5].24 Whether or not this compound will be found to be useful for treating human HCV infection will not be known for a few years. Several days following the discontinuation of MK-0608, investigators detected a mixture of wild-type and mutant virus at a position in the enzyme associated with HCV resistance. Nonetheless, this study suggests that we may soon be seeing more potent polymerase inhibitors with activity rivaling that of the HCV protease inhibitors.
Isabel Najera (Roche Pharmaceuticals) and Akhter Molla (Abbott Laboratories) [Abstracts 3 and 4, respectively]25,26 each described data on the in vitro activity and in vitro selection of mutations associated with resistance to nonnucleoside polymerase activity. Najera described the activity of two nonnucleoside polymerase inhibitors -- one that binds to the polymerase thumb region and the other to the palm region. During in vitro passage with each drug alone, mutations developed in the region of the molecule that was targeted by the inhibitor, as well as secondary mutations for the palm site inhibitor. During in vitro passage with both inhibitors, both thumb and palm site mutations emerged allowing the virus to develop resistance to both drugs. Molla described a new highly active polymerase inhibitor to which virus resistance developed rapidly during in vitro passage but which was highly synergistic when used in combination with interferon (IFN).
Taken together, these and previous studies of new small molecule inhibitors of HCV suggest that although some inhibitors will be highly potent, most will have a low genetic barrier to resistance. If these drugs are used as monotherapy agents, the rapid development of resistance appears to be inevitable. This resistance may be more rapid than that observed for HIV-1, as it is estimated that HCV is replicating at a level that is 2 to 3 logs higher than HIV-1 (i.e., the production of about 1 trillion virions per day), and because it appears that even within tissue culture there may be more heterogeneous genotypes than observed with HIV-1.
Advances in research technology eventually have a positive effect on the clinical treatment of patients. Ralf Bartenschlager (University of Heidelberg, Germany) reviewed the recent rapid progress made during the past several years in the propagation of HCV in cell culture, and how this has facilitated both drug screening in cell culture and the identification of drug-resistance mutations in in vitro passage experiments [Abstract P1].27 In vitro passage experiments to identify mutations responsible for drug resistance have a major role in the development of new antiviral drugs. The development of mutations during in vitro passage with a new compound provides the following essential information:
Stephen Locarnini (Victorian Infectious Diseases Reference Laboratory and University of Melbourne, Australia) reviewed the genetic mechanisms and clinical consequences of HBV resistance [Abstract P2].28 Hepatitis B is a DNA virus with an RNA intermediate. Its polymerase enzyme catalyzes a reverse transcription step and has an error rate much higher than that of other DNA viruses and is similar to that of HIV-1 reverse transcriptase. The amount of heterogeneity observed in HBV, however, is lower than that observed for HIV because a large proportion of the HBV genome consists of overlapping reading frames; therefore pol evolution is constrained not just because pol gene mutations can decrease polymerase function, but also because these mutations can decrease envelope function.
Current HBV drugs consist of L-nucleosides including 3TC, FTC, and telbivudine; acyclic phosphonates including adefovir and TDF; and cyclopentene compounds including entecavir. More than 15 drug-resistance mutations have been described. Some of these mutations cause resistance to one drug, whereas others cause resistance to multiple drugs. Some drug-resistance mutations are primary and appear to directly reduce the affinity of HBV polymerase to an inhibitor. Others are compensatory and appear to either compensate for the decreased fitness or increase the level of resistance associated with another drug-resistance mutation.
Lamivudine resistance has an annual incidence of about 20%, with cumulative rates approaching 80% by year four. Although similar data are not available for FTC and telbivudine, the mechanisms of resistance to these drugs are expected to be equally high, as the genetic mechanisms of resistance are similar. Adefovir resistance is slower, with cumulative rates of up to 10% over a three-year period. Entecavir resistance in persons receiving it as their first anti-HBV drug has not yet been reported despite two years of follow-up in some patients. The frequency of TDF resistance is also low but few data are available.
It is well known that M204V/I is the major mutation associated with 3TC resistance. This mutation is analogous to the M184V mutation in HIV-1 reverse transcriptase, as both mutations are in the active site YMDD motif. Locarnini, however, described eight patterns of mutation observed in patients developing virologic failure during 3TC therapy. All but two patterns have M204V/I with or without additional drug resistance mutations including L80V/I, I169T, V173L, L180M, A181T, T184S, and Q215S. Whereas M204V/I causes high-level cross-resistance to other L-nucleosides and entecavir, several other 3TC-resistance mutations also confer clinically significant cross-resistance to adefovir and entecavir
N236T and A181V/T confer resistance to adefovir, although virological failure can occur without these changes. Cross-resistance between adefovir and the L-nucleosides is minimal, as only one mutation, A181T, appears to cause resistance to both drugs. There are two major mechanisms of resistance to entecavir, both of which include the 3TC-resistance mutation M204V (M204V + L180M + M250V +/- I169T and M204V + L180M + 202I + T184G).
Locarnini reviewed the large amount of data showing that drug-resistant virus is not benign and argued that an increased consideration of drug resistance is required to improve the current approaches to treating HBV infection. For historical reasons, 3TC has been the first drug used, but because of its high rate of failure with drug resistance and often cross-resistance, other drugs, and possibly drug combinations, are now preferable options. Locarnini also made the case for genotypic resistance testing in persons with virologic failure, to confirm the presence of drug resistance and to determine its mechanism, and thus the pattern of cross-resistance, before changing therapy.
Finally, Locarnini described several examples of how certain uncommon pol drug-resistance mutations cause changes in two key regions of envelope: a region in which mutations lead to envelope changes that interfere with vaccine-induced neutralizing antibodies, and a region recognized by antibodies relied upon for serologic testing.
Robert W. Shafer is Associate Professor of Medicine at the Stanford University School of Medicine.
Editor's Note: This article was compiled from and expands upon AIDScan's daily coverage of the XV International HIV Drug Resistance Workshop, which was made possible through an educational grant from Bristol-Myers Squibb to the International Association of Physicians in AIDS Care.
This article was provided by International Association of Physicians in AIDS Care. It is a part of the publication IAPAC Monthly.