Report from the 4th International Workshop on HIV Drug Resistance and Treatment Strategies.
Unlike 2 years ago, almost none of the presentations at this year's workshop were presented at the International AIDS Conference (Durban, South Africa). This was unfortunate, and those authors missed out on an important meeting that, as a consequence of their absence, was not as deep in basic science as it should have been. The focus of most presentations at this workshop was related to the application of commercial resistance assays and to the mechanisms of HIV drug resistance. There were also sessions on new drugs and on immune reconstitution, but these sessions were not substantial.
Thymidine analog mutations
Zidovudine (Retrovir) resistance has been well described, with identification of 6 mutations associated with diminished response to zidovudine. New reports suggest that zidovudine resistance mutations may be important for resistance to other nucleoside reverse transcriptase inhibitors (NRTIs), and a novel mechanism for resistance has been described. A number of reports, including 2 presented at this meeting, have shown that stavudine (Zerit) can also select for mutations typically associated with zidovudine resistance, even though there is not a corresponding shift in phenotypic resistance. Lennerstrand and colleagues from Virco, UK, confirmed the presence of ATP(*)-dependent primer unblocking with zidovudine resistance, shown in a published report from Meyer last year. This is a form of pyrophosphorolysis in which the drug residue can be cleaved more efficiently by mutant reverse transcriptase, allowing the extension of reverse transcriprase to continue. Reverse transcriptase bearing the multi-nucleoside drug resistance (MNDR) insertion after position 69, especially together with zidovudine resistance mutations, appears to allow unblocking of reverse transcriptase arrest caused by NRTIs through an ATP-dependent mechanism. In contrast, the V75T mutation (associated with stavudine resistance) and the Q151M MNDR mutation appear to be involved in decreased binding of stavudine independent of ATP. Thus, it is probably appropriate to refer to zidovudine resistance mutations as thymidine analog mutations, although the absolute level of resistance to these mutations by stavudine is much lower.
Meyer extended his previously published studies and showed that the unblocking mechanism was highest in reverse transcriptase bearing both M41L and T215Y, the 2 mutations most important for zidovudine resistance. Of note, introduction of foscarnet resistance mutations W88G, E89K, Q161L and H208Y in either wild-type or zidovudine-resistant reverse transcriptase decreased ATP-dependent unblocking of reverse transcriptase, suggesting a mechanism for zidovudine resistance reversal with foscarnet mutations. Unblocking may also occur with guanosine analogs, such as abacavir (Ziagen), perhaps explaining the diminished response to abacavir in patients with pre-existing zidovudine resistance.
New prospective trials of HIV drug resistance
Two new prospective trials examining the utility of HIV drug resistance testing in clinical management were presented in their entirety for the first time, and both of these new studies employed phenotypic resistance testing (PRT). VIRA3001, an open-label, randomized, 16-week study compared therapy switches guided by the Virco PRT with those guided by standard of care (SOC). Study subjects (N=268) had failed their first protease inhibitor (PI) containing regimen and had viral loads greater than 2000 copies/mL. Median baseline viral load was 4 log copies/mL and baseline CD4 T cell count was 340 cells/[mm.sup.3]. Physicians were asked to indicate a new regimen twice: first before the PRT results, and again after PRT results or after a 4-week interval. Over the month of enrollment and provision of PRT results, 44% of subjects in the SOC arm had a different choice for a new regimen, compared with 79% of subjects in the PRT arm. Patients in the PRT arm received significantly more active drugs and fewer non-nucleoside reverse transcriptase inhibitors (NNRTIs). Based on the primary endpoint of the study (the proportion of patients achieving viral load reduction less than 400 copies/mL), there was a trend toward improved outcome in the PRT arm that was not quite significant using an intent-to-treat analysis where missing samples equal failures (p=0.099). However, this was significant using an as-treated analysis (p=0.033). In a secondary analysis of the change in viral load over 16 weeks, a significant difference was seen when limited to observed values (on-treatment analysis) with a 1.23 log copies/mL decrease for PRT compared with 0.87 log copies/mL for SOC (p=0.004). Despite the general support for resistance testing in this new prospective study, the results must be considered equivocal since the difference between treatment groups was not significant as determined by the primary study analysis.
Another new study was presented from a French cohort, the NARVAL/ARNS 088, involving 541 subjects randomized to group P (phenotypic results by an in-house recombinant virus assay), group G (genotypic results by TruGene, Visible Genetics, Inc.) and group SOC. At the primary 12-week study endpoint, the proportion of patients with viral loads greater than 200 copies/mL was 33%, 41% and 34% in groups P, G and SOC, respectively, with no difference in those having viral load decreases greater than or equal to 1 log copies/mL. Preliminary but incomplete 24-week results showed no significant difference between the P and SOC groups, with group G doing better than SOC. This study disputes what has become dogma as espoused in numerous expert commentaries, namely that genotypic testing is most useful early when genotypic changes are simple, but that phenotypic testing is needed with treatment experienced patients since they have complex genotypes. In this study, performed in treatment experienced patients at very advanced disease stages, genotypic testing was in fact marginally better than phenotypic testing and significantly better than SOC. Part of the problem arises from the way the study was conducted, with rigid recommendations and the use of an in-house phenotypic assay. Resistance to stavudine or abacavir was identified more than twice as often by genotype as by phenotype, and stavudine and didanosine (Videx) were prescribed much more often in groups P and G, compared with SOC. The relatively low response rates (only about one third of patients overall at 12 weeks) suggest that resistance testing may not be that helpful when there are limited treatment options available.
Interpretation of phenotypic resistance assays
A major area of controversy involves interpretation of tests, both genotypic and phenotypic. Genotypic resistance test interpretation requires software to deduce resistance patterns based on genotypic mutations. This will always be imprecise since the impact of "minor" mutations on phenotypic manifestations of known genotypic patterns will probably never be fully sorted out. There is general consensus on the consequence of certain mutational patterns, but there still is great debate over the meaning of zidovudine resistance mutations as well as the ranking and importance of the primary PI mutations.
Phenotypic resistance assays, although more straightforward, may pose even more problems for interpretation. Here the [IC.sub.50] is presented as a change in the concentration of drug needed to inhibit 50% of [patient] viruses, compared with that needed to inhibit wild type [naturally occurring] virus. The cut-off for resistance is uniformly applied at 2.5 fold for ViroLogic and 4 fold for Virco. There is ample evidence that different cutoff points need to be established for each drug. At one extreme is stavudine resistance, in which greater than 2-fold increases in [IC.sub.50] are seldom seen unless mutations include the MNDR position 69 insertion. A 2-fold increase in [IC.sub.50] may likely be significant for stavudine resistance, but this small change cannot be reliably measured. At the other extreme is the definition of resistance for PI combinations, in which the drug levels of PIs rise so high that an [IC.sub.50] cut-off much higher than 2.5 or 4 fold may be needed. New data on correlates of resistance to lopinavir (Kaletra, formerly ABT-378/r) in isolates from patients with prior PI failure was presented by Kempf and colleagues from Abbott Laboratories. Responses to lopinavir in 112 isolates from patients who had failed other PIs were correlated with numbers of mutations. Despite the presence of 6 or 7 mutations, which typically yielded a 14-fold rise in [IC.sub.50], most isolates could be suppressed, suggesting that the resistance cut-off for this PI combination should probably be set at approximately 20 fold.
A number of other presentations suggested that phenotypic resistance testing may be misleading when compared with clinical outcomes. Studies of individuals recently infected with HIV, but who have not been exposed to NNRTI treatment, are commonly found to have decreased susceptibility to NNRTIs, as high as 30%. This phenotypic resistance usually results in [IC.sub.50] increases of 4 to 10 fold and the key NNRTI mutations K103N and Y181C are usually not seen. Harrigan and colleagues from the British Columbia Centre for Excellence in HIV/AIDS performed a retrospective analysis on stored plasma samples from 272 patients treated with NNRTIs. While 4- to 10-fold decreases in baseline NNRTI susceptibility were relatively common, this was not predictive of therapy failure since 92% of subjects with modestly decreased susceptibility who were prescribed an NNRTI still achieved viral load reduction to less than 400 copies/mL. Here phenotypic resistance studies would tell you to avoid a class of drugs that in fact should be effective.
Another phenomenon that has generated a great deal of interest is the opposite case, where some virus strains exhibit increased susceptibility to NNRTIs compared with wild-type virus, often termed hypersusceptibility. Using the ViroLogic assay, researchers found that the [IC.sub.50] actually fell in some cases with increasing drug concentration. In the analysis of virus strains that exhibited this phenomenon, Haubrich and colleagues in San Diego noted that this was most commonly seen in patients experienced with NRTI therapy, particularly if mutations to zidovudine and to lamivudine (Epivir) were present. This generated a great deal of interest, and may have some implications regarding interactions of mutations that occur while on therapy. While of interest to many, this could simply be an in vitro artifact that really does not have much relevance in test interpretation. After all, susceptibility is susceptibility.
In a plenary session, David Back from the University of Liverpool presented an overview of cellular resistance to HIV drugs, including interactions between inducible cellular gene products, such as P glycoprotein and a multidrug resistance protein (MRP), and PIs. P glycoprotein is an endogenous, energy-dependent cell transporter that has been extensively studied in cancer chemotherapy. Induction and high-level expression of P glycoprotein result in cells refractory [resistant] to chemotherapeutic agents through an efflux transport system. Although it seems to many that this is a new phenomenon associated with HIV treatment, a whole family of these efflux transporter systems have been identified and investigated for cancer chemotherapy. Different members of the family seem to have different specificities for PIs, but all of them appear to result in some efflux of all known agents. These may also reduce cellular uptake of PIs, again with variable rates for different agents. Other recent studies have shown that MRPs may also play a role in NRTI resistance by decreasing the amount of intracellular drug to levels below that necessary to inhibit HIV replication, allowing growth and evolution of drug-resistant HIV. One of these, MDR4, markedly increases the [IC.sub.50] of NRTIs, indicating that class resistance for reverse transcriptase inhibitors may also be induced. This should not be a surprise, since HIV drugs (similar to antineoplastic drugs) are toxic to cells, and one would expect that inducible mechanisms to protect the cells would be invoked. There will be a great deal of work needed to develop model systems to examine these transporters for HIV drugs, but most of the previous work has been performed in cell lines rather than more relevant primary cells. There are no available assays to detect cellular resistance, but this must be considered in individuals who are not responding to drugs despite good medication adherence, and who do not have detectable virologic resistance.
(*) Adenosine triphosphate (ATP) is a molecule that supplies energy to cells for process like muscle contraction and sugar metabolism.
|Printer friendly Cite/link Email Feedback|
|Publication:||Research Initiative/Treatment Action!|
|Date:||Sep 1, 2000|
|Previous Article:||Three HIV-treating physicians express their professional opinions regarding a case scenario.|
|Next Article:||A look at HIV immune research.|