The prevalence and determinants of the K65R mutation in HIV-1 reverse transcriptase in tenofovir-naive patients

The K65R mutation in HIV-1 reverse transcriptase is associated with reduced susceptibility to abacavir and tenofovir. We established its prevalence within a large clinical database, and investigated correlations with other resistance-associated mutations and antiretroviral history. The presence of K65R is associated with previous abacavir use. Although rare, it is preferentially selected within non-thymidine analogue-containing regimens, compared with concurrent zidovudine or stavudine use, which is associated with thymidine analogue mutations. Both genetic routes may compromise abacavir and tenofovir activity. The long-term efficacy of antiretroviral therapy is constrained by the emergence of drug-resistant virus, with mutations that may confer widespread cross-resistance. This is probably the case for the nucleoside analogues. Tenofovir, a newly licenced nucleotide analogue, is active against some HIV-1 mutant isolates that are resistant to zidovudine, didanosine, abacavir and lamivudine. It is also active against multinucleoside drug-resistant variants carrying the Q151M mutation within reverse transcriptase (RT). A recent study had shown reduced susceptibility to tenofovir in virus isolates carrying three or more thymidine analogue mutations (TAM), including the M41L and L201W mutations. In-vitro and in-vivo studies [1] have suggested that isolates carrying the K65R mutation in RT have reduced susceptibility to tenofovir. We therefore assessed the prevalence of the K65R mutation in antiretroviral-experienced, tenofovir-naive patients and assessed its drug treatment and virological correlates. All genotypic resistance tests (VIRCO) performed in our department from 1995 to October 2001 were included in the study. The full antiretroviral drug history was available and matched to each resistance test for analysis. Data were analysed using SAS statistical software. Qualitative data were analysed using χ2 test statistics. When the numbers of frequencies in cells were low, these were adjusted using Yates’ correction. Data are presented as proportions with 95% confidence intervals (CI). When the distribution of quantitative data were hypergeometric between groups, comparisons were made using Mann–Whitney U test statistics. All P values presented are two tailed. Of 1757 HIV drug resistance tests on 999 patients, 17 patients had the K65R mutation (1.70%). The median number of previous antiretroviral combinations in the 17 patients was three (range one to 11). Nine cases were abacavir experienced and eight were abacavir naive. This represents 6.2% (95% CI 3.0–11.5%) of all patients ever exposed to abacavir carrying virus with the K65R mutation versus 1.0% (95% CI 0.4–2.0%) of patients not exposed to abacavir carrying this mutation. There were no other significant differences in previous drug combinations, didanosine use, thymidine analogue use or mutations present in the abacavir and non-abacavir groups. Resistance data were available within the database in 186 individuals currently on abacavir. Nine patients had virus with the K65R mutation (group 1) and 177 patients had virus without the K65R mutation (group 2). A comparison of these two groups identified no difference in duration of time on abacavir, total duration of time on nucleoside reverse transcriptase inhibitors (NRTI) or the median number of nucleoside mutations (Table 1).Table 1: Correlations and associated mutations in abacavir-experienced individuals.Genotypic resistance testing of stored serum samples from all patients is in group 1 before abacavir therapy demonstrated only one case in which the K65R mutation was present before abacavir use. The NRTI-associated mutations that were analysed are shown in Table 1. An increase in the prevalence of A62V, V75I and Y115F mutations in HIV-1 RT was observed in group 1 (K65R present) compared with group 2, which was statistically significant. There was also a higher prevalence of L74V, F116Y and Q151M mutations in group 1, but the difference did not reach statistical significance. Further sub-analysis controlling for the presence of the Q151M mutation in groups 1 and 2 showed no difference in the prevalence of the associated mutations (A62V, V75I, F77L and Y115F); an expected observation in the light of the association between this cluster. There was an increased incidence of concurrent thymidine analogue (zidovudine, stavudine) use in group 2. This was associated with an increased likelihood of two or more TAM being present in this group (mean 3.32), although there was no difference in individual TAM between the groups. By contrast, no difference in the presence of M41L and L210W viruses or the combination of K70R, T215F/Y and K219Q mutations was observed between each group. In tenofovir-naive patients the K65R mutation was present in 1.70% of individuals having a resistance test. There was a significant increase in incidence of this mutation associated with previous abacavir experience (6.2%) but not with didanosine. Abacavir-experienced patients without the K65R mutation were more likely to be currently on a thymidine analogue and have associated TAM. The presence of three or more TAM, including M41L and L210W, has been shown to reduce susceptibility to tenofovir. We were not able to demonstrate an increased frequency of these TAM in our cohort. Abacavir-experienced patients with the K65R mutation were less likely to be on a thymidine analogue and were more likely to carry the mutations associated with the Q151M multinucleoside drug-resistant complex. There may be two pathways of viral evolution that can be selected when on abacavir: either via the K65R mutation if not currently on thymidine analogues or via the TAM if using thymidine analogues. Either pathway can lead to reduced susceptibility to tenofovir. As tenofovir becomes more widely used as part of antiretroviral therapy in HIV disease, physicians must be aware of this pre-existing resistance especially in abacavir-experienced individuals.