November 19, 1999
Several presentations offered data on the HIV genotypes observed in patients failing antiretroviral therapy. I will report on several abstracts that focused on resistance to the nucleoside reverse transcriptase inhibitors.
Until recently there has been little in vivo data reporting the mutations associated with resistance to abacavir. Understanding abacavir resistance is important in order to learn what options exist in the nucleoside reverse transcriptase inhibitor class for patients who fail on abacavir. Two presentations from the Glaxo Wellcome virology group described mutations in abacavir recipients who had previous experience with nucleoside reverse transcriptase inhibitors, and in a group of patients who were antiretroviral naïve before experiencing virologic failure on a two drug combination with abacavir and one of several protease inhibitors.
One report by Ross et al., described the experience of 87 patients who stopped their mono- or dual nucleoside reverse transcriptase inhibitor therapy, switched to the combination of zidovudine, lamivudine, and abacavir., and subsequently developed virologic failure. 56 patients maintained a viral load < 400 copies/mL after 48 weeks of treatment. Prior to switching to ZDV/3TC/ABC, 30 (54%) of these patients had a 184V mutation associated with 3TC resistance and 29 (52%) had a mutation associated with a thymidine analog (ZDV or d4T). Fifteen patients experienced virologic failure, ten of whom had genotypes available at baseline and at the time of treatment failure for comparison. 80% of these patients with treatment failure had a 184V mutation at baseline and 70% had at least one mutation associated with resistance to a thymidine analog. The baseline viral load was higher in those patients with experienced treatment failure compared to those who had a durable response, 21,476 copies/mL versus 853 copies/mL. Decreases in viral load following the switch to the triple regimen were similar in patients with and without virus with the codon 184V mutation.
One of the 15 patients experiencing virologic failure had wild type virus at the time of viral load rebound. Four patients rebounded with only a codon 184V mutation, though additional mutations associated with resistance to ZDV accumulated over time in two patients as they continued on therapy. The remainder had an assortment of ZDV-associated mutations together with the 184V mutation. The most commonly encountered mutations were M41L (7 patients), D67N (6), T69D/N (2), K70R (4), and T215Y/F (8). One patient developed a L74V mutation after 48 weeks of therapy that is associated with ddI resistance; this mutation can also be selected by abacavir. Phenotyping of virus from eight patients showed reduced susceptibility to abacavir in 6 patients, to 3TC in 7, and to ZDV in 4.
What can we conclude from all of this? First, abacavir is active in patients with virus that has the 184V mutation associated with resistance to 3TC. Patients who fail on a ZDV/3TC/ABC combination will likely have virus resistant to all three drugs, and therefore, should probably have a completely different combination used in the next regimen. As yet, we still do not know what options exist in the nucleoside reverse transcriptase inhibitor class for patients who fail on this triple combination (they may not have many or any). Lastly, I think this data provides a strong case for the use of abacavir to intensify antiretroviral therapy in patients with low viral loads that have not reached undetectable levels, or perhaps in patients with are experiencing early treatment failure with low viral loads.
Another report by Ross with a different group of colleagues described the genotypic profile of virus from 17 of the 18 patients (out of 74 patients originally treated) failing treatment with a two-drug combination of abacavir plus indinavir, nelfinavir, saquinavir, ritonavir, or amprenavir. (Abstract 362) No mutations associated with resistance to abacavir was present at baseline. Five patients had mutations associated with abacavir resistance at the time of virologic failure -- either M184V or K65R. Six patients had mutations in the protease cleavage region, and ten had polymorphisms in the protease gene that are commonly seen in patients failing on a protease inhibitor.
This study demonstrates that resistance to abacavir does not develop rapidly, as is the case with 3TC or any of the non-nucleoside reverse transcriptase inhibitors. In addition, since the 184V mutation can be selected by abacavir (the mutation associated with 3TC resistance), one has to wonder if abacavir should ever be given in a combination without 3TC.
Many reports have described genotypic changes of virus in patients failing on combinations that include ZDV and 3TC. There are far fewer reports of the resistance profiles of patients with viral load rebounds on combinations that don't include 3TC. Trip Gulick reported on the genotypic and phenotypic changes of virus from patients experiencing virologic failure in the START I and II trials, two trials that compared different dual nucleoside regimens in triple drug combinations with indinavir (ZDV + 3TC versus d4T + 3TC in START I, and ZDV + 3TC versus d4T + ddI in START II). Only three of 61 patients had baseline mutations (one RT 184M - 3TC mutation; one RT 41L - ZDV mutation; and one PR 10I - indinavir mutation.
Twelve of 34 patients with virologic failure on ZDV/3TC/IDV had wild type genotype and phenotype. Twenty-one patients developed a RT M184V/M mutation that was most often associated with a > 100-fold loss of phenotypic sensitivity to 3TC. Two of these patients also had single mutations to indinavir (PR V82A in one case and PR L10V in another), though isolates remained phenotypically sensitive to indinavir. One patient had a PR L10L/I mutation that remained phenotypically sensitive to indinavir.
Thirteen patients had virologic failure on d4T/3TC/IDV. Five rebounded with virus that a wild type genotype and phenotype. One patient had a RT M41L mutation at baseline and had no new mutations at the time of failure. Seven patients had the RT M184V/M mutation and phenotypic resistance to 3TC. One of these patients also had the PR L10V mutation.
Fourteen patients had virologic failure on d4T/ddI/IDV. Interestingly, eight of these patients had no genotypic changes and had phenotypically wild type virus. Two patients had baseline PR L10I/V mutations, developed no new mutations at the time of rebound, and had phenotypically wild type virus. One patient had RT V75V/M mutant virus that was phenotypically wild type. Two patients had PR V82A mutations associated with resistance to indinavir, though the virus remained phenotypically sensitive to indinavir. Lastly, one patient had virus with a RT D67D/N mutation and PR M46M/L, PR I54I/V, and V82A mutations that was 6-fold less sensitive to indinavir. Thus, all patients had virus that was phenotypically sensitive to d4T and ddI at the time of virologic rebound.
This study confirms that patients on a 3TC- and protease inhibitor-inclusive combination will likely have only 3TC genotypic and phenotypic resistance when first experiencing virologic failure. It also suggests that it may be hard for the virus to develop resistance to d4T and ddI when those two drugs are used in combination. The high proportion of patients who rebounded with wild type virus in all treatment arms is difficult to explain. These patients may have been poorly adherent to medications and were, in effect, on a suboptimally active combination. It is possible that a minority population of resistant virus was present but not detected. Alternatively, there may be pharmacokinetic or cellular mechanisms that result in virologic failure in some cases. Whether a greater number of treatment options remain for patients failing on a d4T/ddI combination than d4T/3TC or ZDV/3TC combinations is a question that will only be resolved with the strategic antiretroviral trials that are underway.
Several groups have described nucleotide insertions at codon 69 in the reverse transcriptase encoding region of the HIV genome that is associated with the development of virus that is broadly resistant to all of the nucleoside reverse transcriptase inhibitors. At this meeting, Ross and colleagues identified four patients with three-nucleotide deletions at RT codons 67, 69, and 70 . (Abstract 365) These were always in association with other, more commonly described mutations in RT. All of these patients had extensive treatment histories including at least three nucleoside reverse transcriptase inhibitors. Virus with the codon 69 was phenotypically resistant to ZDV, 3TC, and abacavir, but sensitive to d4T, ddC, and ddI. Virus with the codon 70 deletion was resistant to all of the nucleotide reverse transcriptase inhibitors. Of the two patients with codon 67 deletions, one viral isolate was phenotypically resistant to only zidovudine, the other was sensitive only to d4T. To understand the role the codon 67 deletion mutation has on virus independent of other mutations, a recombinant virus with the codon 67 deletion was constructed. This virus retained phenotypic sensitivity to all of the reverse transcriptase inhibitors. Therefore, unlike the codon 69 and 70 deletions, the codon 67 deletion is not associated with phenotypically altered virus.
The clinical significance of this report is uncertain. We do not know how commonly these codon deletions occur. If present, they may be missed by the currently available genotyping assays, illustrating another potential limitation of these tests. This report does underscore the fact that new paths for the virus to develop resistance to the currently available drugs continue to be described.