icon- folder.gif   Conference Reports for NATAP  
  16th CROI
Conference on Retroviruses and Opportunistic Infections Montreal, Canada
February 8-11, 2009
Back grey_arrow_rt.gif
Resistance Report
  CROI 2009 Feb 8-12 Montreal
Mark A. Wainberg
McGill University AIDS Centre, Montreal
The XVIth Conference on Retroviruses and Opportunistic Infections contained a wealth of new information on the topic of HIV drug resistance. The coverage of subjects presented on Resistance at the Conference will be broken down into sections.
1. Resistance to Protease Inhibitors
Oral Session 18 contained a number of presentations relating to fundamental mechanisms of drug resistance. First, in Abstract 65LB, Celia Schiffer and colleagues documented that protease inhibitors (PIs) could potentially be designed that would interact with the active sites of the HIV-1 protease (PR) in such fashion as to preclude the development of drug resistance. This work was initiated on the basis of crystal structure analysis of HIV-1 PR and its ability to bind to specific protease inhibitors. The results indicated that a number of inhibitors could bind to PR notwithstanding the presence of a number of classical mutations that are associated with resistance to many other members of the PI family of drugs. Results were confirmed on the basis of phenotyping analysis to demonstrate that many of these inhibitors possessed high level potency against a panel of resistant viruses. These results validate approaches based on efforts to specifically design novel PIs that are able to fit within the substrate envelope of PR and give hope that more robust PIs will be developed in the future that will less susceptible to problems of resistance.
Abstract 126 by F. Clavel and colleagues was on the topic of the HIV-1 Gag gene and its interactions with viral protease and the role of Gag in regard to HIV drug resistance to protease inhibitors (PIs). As shown by other groups, it is now evident that cleavage site mutations within Gag can account for a substantial proportion of resistance to protease inhibitors (PIs). However, it has been difficult until now to develop meaningful assays that are able to quantitate the numbers of such mutations within Gag or to define with precision the extent to which such mutations are likely to occur within select patient populations. Dr. Clavel has pointed to the advent of pyrosequencing as an important method that will help to resolve many of the issues that relate to the role of cleavage site and other mutations within Gag in this regard. Dr. Clavel also observed that mutations within Gag can often compensate for loss of viral replication capacity that may be due to mutations within PR and that are responsible for high levels of drug resistance to the PI family of compounds. This is a point well made and both retrospective and prospective analyses of clinical samples that have shown resistance to protease inhibitors will shed further light on this subject. Dr. Clavel further noted that both protease mutations and mutations within Gag that confer resistance to PIs may affect the antiviral activity of maturation inhibitors such as beviramat, although this topic requires additional study. In all likelihood, the use of pyrosequencing to determine the presence of minority viral species both at baseline and in species that emerged as a consequence of development of resistance to protease inhibitors will shed further light on this topic.
Abstract 655 by R. Dierynck et al was a follow up investigation from the ARTEMIS trial in which drug naïve patients were dosed with either Darunavir/r (800/100mg qd) vs LPV/r (800/200mg daily). This analysis has now documented that rates of mutational accumulations in patients who received DRV/r were lower than among individuals who received LPV/r at 96 weeks of follow up. These findings are consistent with those reported earlier by the ARTEMIS study team in regard to numbers of virological failures being higher in those individuals who received LPV/r than DRV/r.
2. Resistance to NRTIs/NtRTIs
In Abstract 66LB, E. Lansdon et al, working at Gilead Sciences, have performed further work on a novel nucleotide analogue termed GS-9148-disphosphate. This molecule has a unique resistance profile and can continue to be active against viruses possessing a number of classical NRTI resistance mutations including M184V, K65R, L74V and several thymidine analogue mutations (TAMs). Co-crystallization of this molecule together with covalently tethered double-stranded DNA documented that it bound in a highly specific way to RT, such that conformational changes ensued that resulted in closing of the fingers' domain. The authors believe that this constitutes firm evidence of a unique mechanism of action of GS-9148 and helps to explain why only one mutation associated with drug resistance, i.e. the multidrug resistance Q151M substitution, was unique in being able to confer a moderate level of resistance against this compound. Further studies are underway in regard to the clinical development of GS-9148 and/or derivative molecules that will hopefully prove to be non-toxic and fully efficacious in regard to blockage of HIV replication.
Abstract 67 by K. Das et al provided further information in regard to a drug combination that has long been thought to represent hope in regard to HIV therapeutics. Notably, it is well known that the K65R mutation and TAMs are for the most part mutually exclusive and are rarely seen in combination in clinical isolates. Is it therefore rational to combine zidovudine (ZDV) together with Tenofovir for the treatment of HIV disease? The current study has used an X-ray crystolagraphic approach to study the structural basis for the potential efficacy of this combination. The results document that K65R and TAM mutations such as K70R and T215Y will be unlikely to co-occur in regard to preservation of enzymatic function. Of course, it remains to be seen whether the strategy of using ZDV together with TDF will be embraced by clinical investigators, given a wide range of other therapeutic options and the fact that concern continues to be expressed in regard to potential toxicity over long periods of time of ZDV, even at relatively low doses.
Abstract 68, presented by J. Radzio et al, was on the topic of the N348I mutation that is located within the connection domain of HIV-1 reverse transcriptase (RT). Previous studies have documented that patients receiving a combination of ZDV and Nevirapine (NVP) are more likely to develop the N348I mutation than are patients not treated with these drugs. In addition, the development of N348I is frequently accompanied by the appearance of both TAMs as well as the Y181C mutation associated with resistance against non-nucleoside reverse transcriptase inhibitors (NNRTIs). These investigators therefore created molecular clones in which the Y181C mutation was added to viruses that contained either the K70R mutation or a combination of M41L/T215Y. They showed that the addition of Y181C to the aforementioned molecular clones resulted in renewed sensitivity to ZDV. However, viruses containing Y181C, TAMs, as well as N348I displayed high level resistance against ZDV. Hence, N348I seems to be able to down-regulate the naturally occurring antagonism between TAMs and Y181C and this can be demonstrated in both tissue cultures as well as in biochemical assays. The results further suggest that N348I may be an important connection domain mutation that plays a role in regard to establishment of double resistance against both ZDV as well as the NNRTI family of compounds.
Abstracts 623 by V. Von Wil et al and 624 by M. Biondi were both on the topic of the N348I mutation within the connection domain of HIV reverse transcriptase. There is now accumulating evidence that this substitution might compensate for deficits associated with the M184V mutation in RT that confers high level resistance against both 3TC and FTC. Most importantly, these abstracts have documented the potential importance of mutations in the connection domain and the potential future need to sequence for relevant mutations within this region of RT. As an example, Abstract 624 documents that N348I may exhibit an influence on RNAase H activity and that this substitution may compromise the benefits of ZDV/NVP therapy.
Abstract 628 by C. Invernizzi et al was on the topic of the selection of the K65R mutation by subtype C viruses. This group constructed molecular clones that were derived from subtype B viruses and introduced into them subtype C sequences from the 64 and 65 amino acid coding regions. The result was that these novel viruses behaved like subtype C viruses in regard to their ability to rapidly select the K65R substitution when drug pressure applied was applied with any of ABC, TDF, ATC, D4T, TDI or combinations thereof in the presence or not of 3TC or FTC. In contrast, wild type subtype C viruses also were able to quickly select for the K65R substitution under these conditions but subtype B viruses were very slow to select for K65R if this occurred at all. Thus, these findings help to further underscore the apparent predilection of subtype C viruses to select for the K65R mutation, as has now been shown in clinical studies from three different countries in Sub Saharan Africa in which subtype C viruses are predominant.
Abstract 646 by R. Paradis et al also documented that a mutation in the reverse transcriptase connection domain, notably A376S also seems to be associated with diminished responsiveness to nevirapine in previously NNRTI treated subjects. Similar findings were reported in Abstract 643 by A. McCormick et al from the DART study among patients who had received a treatment regimen of 3TC/ZDV/ABC. Similar observations on the relationship between drug resistance and viral fitness were reported in Abstract 651 by M. Juzon et al who documented that each of mutations at positions Q148/N155 and Y143 impact on viral fitness while contributing to HIV drug resistance.
Abstract 656 was by C. Hoffman et al reporting on the development of resistance to antiretroviral drugs in South African patients receiving antiretroviral therapy. Not surprisingly, these individuals received treatment with ZDV/3TC/EFV and had a preponderance of M184V mutations in the aftermath of treatment failure. NNRTI resistance was also reported in this group as were the presence of TAMs. Patients with HIV viremia had a preponderance of NNRTI mutations. Clearly, patients who fail a regimen of ZDV/3TC/EFV with multiple mutations as noted above will be limited in regard to a number of future second or third line therapeutic options.
3. Resistance to Integrase Strand Transfer Inhibitors (INSTIs)
Abstract 69 by S. Fransen and colleagues was on the topic of mutations within the integrase (IN) gene that help to define resistance against Raltagravir (Ral). These investigators studied individuals who had failed therapy with Ral and have created a series of site-directed mutants containing mutations at position 143, 148, 155, either alone or in combination with other secondary IN-related resistance mutations. Thus, these investigators created combinations of mutations that may appear early in the multiple pathways that are associated with resistance to integrase inhibitors alongside mutations of a more secondary nature that have been shown to both augment viral replication fitness, when coupled with primary mutations, as well as increase overall levels of drug resistance to integrase strand transfer inhibitors (INSTIs). In general, the results of the studies now show that the mutations at positions 143 and 148 may have less of an effect on viral replication and cause higher levels of drug resistance against Ral than do mutations at position 155. These findings may help to explain earlier clinical observations that suggest that the 155 mutation may be more likely to appear first in clinical samples but may, over time, be supplanted by mutations at positions 143 and 148, accompanied by the appearance of numerous secondary mutations associated with resistance against this family of compounds. In this context, it will not be surprising if this story ultimately appears to be a similar one to the initial appearance of the M184I mutation prior to M184V in the context of resistance against either 3TC or FTC. The basis for this progression is likely to be based on considerations of viral fitness, with viruses that impact to a lesser degree on viral replicative capacity e.g. those containing N155H, being supplanted over time by viruses containing mutations at positions 143 and 148, with the latter becoming predominant in the context of selection by Ral and other INSTIs.
A topic that is germane, to our overall understanding of drug resistance is that of mutations within the IN gene that cause diminished responsiveness to INSTIs. In abstract 125, Dr. M. Miller and colleagues presented data from clinical studies that had been carried out with Raltagravir (Ral) that had revealed that three major pathways existed in regard to appearance of first mutations that are responsible for resistance against this compound.
These primary mutations first emerged at positions Y143C/H/R, Q148H/K/R, and N155H. As stated, increasing evidence indicates that Y143 and Q148 pathways are preferred over time and that switches are likely to occur from N155H to Q148H/K/R, largely because of reasons of viral fitness, as explained above in regard to abstract 69.
Abstract 125 also dealt with the role of polymorphisms and concluded that these are unlikely to influence responsiveness to either Ral or to Elvitagravir (Elv) as is also suggested in other abstracts at the conference. In a retrospective analysis of the BENCHMARK studies, polymorphisms that were identified at baseline were not associated with differential responsiveness to therapy. In particular, one substitution, T97A, seemed to have been observed at baseline in equal proportions in both responders and non-responders to Ral. Other abstracts at the conference showed that polymorphisms are present at very low levels in most samples as determined by pyrosequencing. Furthermore, there does not seem to be any correlation between the baseline presence of polymorphisms in the integrase gene and responsiveness to therapy. Similar results were obtained in a retrospective analysis of study samples from the Merck 004 phase 2 study. It should be mentioned, however, that the field of integrase sequencing and our understanding of both polymorphisms and major mutations associated with resistance to integrase strand transfer inhibitors (INSTIs) is still in its infancy and further research on this topic is needed.
Excellent research on resistance to INSTIIs was also presented in the poster session.
In Poster Abstract 615, A. Delelis et al documented that the Q148H mutation in integrase was impaired in regard to integration of viral DNA into host cell. They have now documented that the G140S secondary mutation is capable of rescuing this defect if these two mutations are located together on the same viral genome. It appears as though G140S is capable of increasing HIV replicative fitness in the presence of Q148H while also contributing toward higher levels of drug resistance. These results help to explain why the double combination of G140S/Q148H is common among treatment failures that have employed Ral as an antiviral drug. Abstract 618 by C. Charpentier et al was on the topic of use of allele-specific PCR assays to detect specific mutations associated with resistance to Ral. This group developed assays to detect each of the Q148R, Q148H, and N155H substitutions. Among 32 individuals followed, they noted that Q148R variants were sometimes present even prior to initiation of therapy with Ral, albeit at low levels (<1%). The authors note that the presence of minority species were not necessarily associated with treatment failure. It is intriguing that they did not detect the N155H mutation by this methodology in view of the fact that the later mutation seems to be the first resistance associated substitution to emerge under conditions of Raltagravir treatment failure. In Abstract 618, I. Malet et al performed clonal analysis of viral variants that displayed resistance to Ral. These investigators have documented that the Q148R/H and N155H mutations, when present simultaneously in patients, had occurred on different viral DNA strands. This documents that there are likely to be two independent pathways that lead to development of resistance to Ral and that neither the Q148 mutations nor N155H are likely to co-exist on the same viral backbone. This is an important observation that establishes the likelihood that resistance to Ral may emerge independently through a variety of routes. It should be recognized that use of future treatment regimens may dictate that resistance to Ral and other inhibitors of IN may emerge through pathways are yet to be discovered.
Similar observations were obtained by G. Anis and colleagues in poster 619 on the basis of clonal analysis. These investigators documented that mutations at amino acid positions 143 and 155 could be demonstrated within the same patients but not on the same viral genomic backbone. In Abstract 621, H. Hatano et al concluded that mutations at positions Y143 may occur in the absence of other mutations and that this substitution is associated with high levels of drug resistance. In contrast, the N155H substitution was prone to disappear over time and was associated with lower levels of resistance than substitutions at positions 143 or 148. Finally, Abstract 622 by M. Wirden and colleagues documented that Raltagravir may be removed from drug regimens under conditions of treatment failure without any subsequent increase in viral load. This suggests that Raltagravir is unlikely to exert a residual antiviral effect, based on fitness or other considerations as a consequence of the presence of resistance-confering mutations within the IN gene. Furthermore, a removal of Ral from the failing regimen should ensure that no further Ral-related resistance mutations become selected. Accordingly, this strategy is probably a good one to be followed under conditions of Ral treatment failure.
4. Resistance to entry inhibitors
Session 32 at CROI included issues of drug resistance and treatment responsiveness as well as lectures on the transmission of HIV drug resistance and development of assays that might predict responsiveness to CCR5 antagonists. These topics will be dealt with by other reporters who were present at the conference.
Abstract 638 by A. Tsibris dealt with the use of vicriviroc (VCV) therapy over 28 weeks and the emergence of high level resistance against VCV. Resistance was associated with mutations within the V3 loop and, over time, following withdrawal of VCV, these study investigators observed that the viral quasispecies again became dominated by VCV-sensitive viruses. This study also identified a dominant novel mutated V3 loop form after 48 weeks of therapy, that persisted even following withdrawal of VCV from treatment. These findings suggest that novel V3 loops that may be involved in resistance to VCV probably emerge from baseline viral variants. The data also underscore that reversion to more fit variants will quickly occur if drug pressure is removed. Abstract 639 by B. Jubb et al was on the topic of long term use of MVC and the occurrence of low level viral replication despite the use of this compound. However, true resistance to MVC was seen in only one third of virologic failures whose viruses maintained tropism for CCR5. These findings suggest that true CCR5-tropic resistance to MVC is very uncommon even among patients who achieve virologic rebound and/or who never attain true virologic suppression. Finally, K. Yoshimura et al in Abstract 640 documented the presence of a number of mutations within the V4 and C4 envelope regions of a circulating recombinant form of HIV-1. Interestingly, the emergence of these mutations seem to be associated with resistance against CCR5 inhibitors other that MVC as well as to MVC itself. This study raises the likelihood that cross-resistance among CCR5 inhibitors will be likely to occur, regardless of the precise mechanism that may give rise to initial resistance against such compounds as MVC.
5. Resistance to Bevirimat
Several abstracts dealt with the topic of resistance to this novel first-of-class maturation inhibitor. In one of them (Salzwedel et al, Abstract 635), it was demonstrated that subtype-specific polymorphisms at or near the CA-SP1 junction may determine responsiveness to this compound. Should it turn out that not all HIV subtypes can respond equally well to bevirimat, this might complicate its future use in therapy. A second abstract by NA Margot et al (Abstract 637) suggested that many patients who have not yet received bevirimat may be unable to respond to it in ideal fashion due to polymorphisms in SP1 and the capsid junction.
6. Studies in Developing Countries
Abstract 656 was by C. Hoffman et al reporting on the development of resistance to antiretroviral drugs in South African patients receiving antiretroviral therapy. Not surprisingly, these individuals received treatment with ZDV/3TC/EFV and had a preponderance of M184V mutations in the aftermath of treatment failure. NNRTI resistance was also reported in this group as were the presence of TAMs. Patients with HIV viremia had a preponderance of NNRTI mutations. Clearly, patients who fail a regimen of ZDV/3TC/EFV with multiple mutations as noted above will be limited in regard to a number of future second or third line therapeutic options.
In Abstract 657, C. Wallis et al followed patients who had initiated therapy with d4T-containing regimens. Most of the individuals had also received 3TC and either EFV or NVP. Of interest, the K65R substitution was detected in 5% of individuals while the Q151M complex was detected in 4%. The K65R mutation was sometime seen in concert with Q151M mutations and the most common TAM was D67N. Of note, as well, the V106M substitution that is commonly seen in subtype C patients was more frequent in EFV containing regimens than it was in patients who received NVP. Abstract 658 by R. Murphy et al was also on the topic of ARV usage in South Africa and documented the use in second line treatment of LPV/r plus 2 NRTIs against which previous resistance had not been demonstrated. In general, good treatment outcomes were obtained, suggesting that LPV/r is an appropriate second line treatment regimen in such settings.
In addition to the foregoing, a number of abstracts contained important information on the presence of mutations associated with drug resistance at baseline. These findings are mostly confirmatory of results that have already been reported from other geographic settings but are important because they establish that the problem of transmitted resistance and baseline mutations associated with drug resistance is likely to be universal. In this context, it is important to support the WHO program for monitoring of transmitted drug resistance and the presence of drug resistance in developing country settings. Abstracts 688 and 689 by S. Bertagnolio and D. Bennett, respectively, deal with this issue and the establishment of quality assurance programs toward this purpose
7. The Use of Ultrasensitive Procedures including Pyrosequencing for Analysis of HIV Drug Resistance
Session 42 was a themed oral discussion of abstracts that were also presented as posters in session 119 on the topic of novel technologies to better understand HIV drug resistance. In particular, pyrosequencing, also referred to as g454h sequencing, provides ultra-sensitive analyses that can detect the presence of minority viral populations including specific mutations that may be present in the total viral population at frequencies <1%. In contrast, population-based sequencing which is the classic technique employed by most diagnostic labs will only detect species that are present at a level of 20% or more of the total viral population. An important question to now ask is whether the advent of such techniques as pyrosequencing will render more traditional procedures obsolete.
Abstract 679 by J. Archer et al was on the topic of the use of pyrosequencing to detect minority species of CXCR4-using viruses. The results demonstrated that sequences considered to represent CXCR4 were detected at very low frequencies in almost all subjects at times prior to treatment with Maraviroc (MVC). In contrast, higher frequencies of such sequences emerged at later time points after selection of viruses that displayed dual or dual/mixed (D/M) tropism in regard to use of both CCR5 and CXCR4 co-receptors.
Abstract 680 by L. Swenson et al was on the topic of attempting to quantitate tropism by pyrosequencing. This abstract documented a prevalence of CXCR4-utilizing variants in clinical samples that had been unable to respond to MVC therapy. Furthermore, it was not possible to detect these X4-related sequences by standard genotyping. These authors have shown that a low prevalence of X4-using viruses, seems to be associated with improved responsiveness to MVC.
Abstract 681 by A. Joyce et al was on the topic of minority K65R variants in patients who had interrupted therapy with the combination of TDF/ZDV/3TC as a sub-study of the DART virology group. Although this group documented that it is possible through use of pyrosequencing to detect minority resistant variants in as many as 30% of individuals who stopped therapy, K65R minority variants were not detected in any of 18 individuals who stopped and/or restarted ZDV/3TC/TDF. Furthermore, lower than anticipated levels of the M184V mutations were detected in this analysis. However, it should be noted that patients who receive both ZDV and TDF may be unlikely to develop the K65R mutation due to the antagonism of TAMs. Furthermore, the patients in this study did not possess subtype C variants of HIV-1. This is an important consideration in view reports from several centers that subtype C variants may be especially prone to development of the K65R substitution.
Abstract 682 by Ceccherini-Silverstein et al was on the use of pyrosequencing to better define mutations associated with resistance to Ral. This group followed 74 multi experienced patients and showed that minority mutations associated with resistance to Ral could be demonstrated more easily by pyrosequencing than through use of population-based sequencing studies. For example, at baseline, patients who ultimately went on to fail treatment with Ral seemed to possess minor mutational species associated with resistance to this drug. However, the presence of such minor species at baseline was not necessarily associated with treatment failure in all individuals.
In Abstract 683, L. Swenson et al also reported on the use of pyrosequencing and found that this method did not in general detect drug resistance mutations at baseline. Similar finding were reported in Abstract 683 by T. Lee et al working with plasma samples from 22 treatment experienced individuals who had undergone treatment failure. Abstract 685 by J. Liu et al demonstrated that mutations associated with resistance to Ral were very uncommon in treatment-naïve patients on the basis of work performed with a parallel allele specific sequencing procedure.
It should be recognized that the use of pyrosequencing and other ultrasensitive procedures for detection of drug resistance mutations may be very costly. The equipment to carry out pyrosequencing is itself expensive, but an even greater problem may be the costs of reagents associated with the performance of these assays. This point was made by several speakers during the question period. In general, it is not yet clear that the use of this ultrasensitive yet expensive procedure may be more cost-effective than use of more standard genotyping to assess the presence of drug resistance and the provision of appropriate routine diagnostic information to clinicians. In contrast, it is becoming increasingly apparent that pyrosequencing and other ultrasensitive procedures for monitoring drug resistance offer significant advantages in regard to a multitude of specific research questions that relate to monitoring of minority viral populations. As this technique becomes more refined as well as cheaper overtime, which seems likely to happen over the next several years, it is likely to see greater use. Furthermore, the interpretation of results pertaining to pyrosequencing is likely to become more straightforward.