Resistance to amprenavir before and after treatment with lopinavir/ritonavir in highly protease inhibitor-experienced HIV patients

Genotypes in nine highly protease inhibitor (PI)-experienced patients were studied before and after lopinavir/ritonavir (LPV/r) treatment. Resistance to amprenavir was the rule both before and after LPV/r treatment. Treatment with LPV/r can select for the 50V mutation. In this setting, significant differences in the inference of the amprenavir phenotype from genotype were observed when using different algorithms. Amprenavir is an HIV PI that may have reduced cross-resistance within its class [1], retaining antiviral activity in HIV isolates from patients failing PI-including regimens [2]. Furthermore, the co-administration of ritonavir doses of as little as 100 mg twice a day can increase amprenavir minimum concentration (Cmin) levels to above those that inhibit wild-type virus strains and may exceed the IC50 of many PI-resistant isolates [3,4]. Although resistance or an impaired response to amprenavir can be expected in the presence of 84V plus at least two other PI mutations [2,5], it has been claimed that amprenavir is potentially effective in highly PI-experienced patients. Lopinavir co-formulated with a ritonavir boosting dose (LPV/r) has been found to be effective in the same clinical setting [6]. LPV/r treatment of PI-experienced patients may select the 84V and other PI mutations associated with amprenavir resistance, including the 50V mutation [7]. Furthermore, cross-resistance between these two drugs has been reported [8] and may be underestimated [9]. This raises the concern that a failure to respond to an LPV/r-including regimen may hamper the subsequent efficacy of amprenavir-based therapies. The aim of this retrospective study was to investigate resistance to amprenavir before and after a failed response to LPV/r-containing regimens in highly PI-experienced patients. Among 94 patients enrolled in the Lopinavir/Ritonavir Expanded Access Programme with two available resistance tests (the first performed before LPV/r treatment and the second performed in the case of virological failure: HIV RNA > 400 copies/ml after at least 3 months of treatment), we selected the amprenavir-naive patients who were receiving a PI-containing regimen at the time of the baseline resistance test, and subsequently received LPV/r without any other PI. Genotyping was performed by extracting HIV RNA using the QIAmp Viral RNA kit (Qiagen GmbH, Hilden, Germany). Expand reverse transcriptase (Hoffman-LaRoche, Mannheim, Germany) was then used for RNA retro-transcription to complementary DNA, which was then amplified by means of two nested reactions using the Expand high fidelity polymerase chain reaction (PCR) system kit (Hoffman-LaRoche) with oligonucleotide primers provided by Virco (Mechelen, Belgium). The amplified fragments were purified using the QIAquick kit (Qiagen) and sequenced using Megabace1000 (Pharmacia-Amersham Freiburg, Germany); the sequences were analysed by means of Sequencer (Genecodes Corp., MI, USA) and the HIVdb HIV reverse transcriptase and protease sequence database (http://hivdb.stanford.edu/). The following algorithms provided by HIValg 3.0.1 (Stanford University database: http://hivdb.stanford.edu/) were used to interpret the genotype results: the Agence Nationale de Recherches sur le SIDA (ANSR), the beta test version of the drug-resistance interpretation program (`HIVdb’ algorithm, Stanford), the Rega Institute algorithm, GuideLines Rules 4.0, Visible Genetics VGI (Toronto, Ontario, Canada), RetroGram (release 1.6i: http://retrogram.com) and VirtualPhenotype (Virco). VirtualPhenotype resistance to amprenavir was defined as a more than twofold change in drug susceptibility in comparison with the wild type. Nine of the 94 subjects (eight men; median age 38 years; range 33–47) fulfilled the inclusion criteria. Table 1 shows their characteristics at baseline and after a median of 24 weeks of treatment with a LPV/ r-including regimen (range 14–44 weeks). They had all received antiretroviral therapy for a median of 5.8 years (range 4.5–7.9), including the use of at least three PI for a median of 3.3 years (range 1.4–5.8).Table 1: Patient characteristics.The median baseline HIV-RNA level was 4.67 log10 copies/ml (range 3.93–5.65), and the median CD4 T-lymphocyte count was 241 cells/μl (range 77–422). After a median of 24 weeks of LPV/r treatment, the level of HIV-RNA decreased to 4.23 log10 copies/ml (range 2.97–4.95) and the CD4 cell count increased to 342 cells/μl (range 186–447). Baseline mutations in positions 10, 32, 46, 47, 50, 54, 84 and 90 were detected in eight, none, seven, none, none, five, five and five patients, respectively. After LPV/r treatment, six patients developed new substitutions: one in position 10, one in position 50 and four in position 54. The median number of PI mutations was seven (range 4–8) at baseline, and seven (range 7–9) after LPV/r treatment. Table 2 shows the genotype interpretations obtained using the different algorithms before and after LPV/r treatment. According to the ANSR algorithm, four patients harboured amprenavir-sensitive variants at baseline, only one of whom remained susceptible to amprenavir after LPV/r treatment: on the basis of this algorithm, amprenavir/ritonavir could have been proposed to nine patients at baseline and seven after LPV/r treatment. The interpretations provided by the HIVdb, Rega Institute and VGI algorithms indicated that amprenavir was not recommended in any of the patients at baseline or after LPV/r treatment. According to the RetroGram algorithm, amprenavir (with or without ritonavir) could have been proposed to three patients at baseline and two after LPV/r treatment. VirtualPhenotype was evaluable in six patients at baseline and seven after LPV/r treatment. The median-fold change in amprenavir susceptibility was 3.7 (range 2–19.5) at baseline and 7.9 (range 2.3–19.5) after treatment with LPV/r. None of the patients showed a less than twofold change in amprenavir susceptibility at baseline or after LPV/r treatment.Table 2: Inference of susceptibility to amprenavir on the basis of different algorithm interpretations of genotype.Our results suggest that HIV susceptibility to amprenavir is reduced in most patients experiencing multiple virological failures. The great majority of our highly drug-experienced patients had mutations in positions 10, 46, 54, 84 and 90, which have been associated with amprenavir resistance, both before and after treatment with LPV/r [2]. However, a further viral evolution was observed during a median 24 weeks of LPV/r treatment, which was unable to suppress viral replication completely. It is worth noting that the number of individuals harbouring virus variants bearing mutations at codon 54 increased from five at baseline to nine after LPV/r treatment. Mutations in this codon are often selected when patients fail to respond to amprenavir-including regimens [10], and are associated with a resistance to amprenavir [1,2]. Furthermore, LPV/r treatment led to the selection of the 50V mutation in one of our patients. Both the 50V and the 54L/M/V substitution are associated with amprenavir resistance [1,2], thus suggesting that the number of highly drug-experienced patients who can benefit from therapy with an amprenavir-including regimen may further decrease after treatment with LPV/r. However, pharmacokinetic enhancement with ritonavir can increase amprenavir Cmin levels in such a way as to overcome low-level resistance [3,4], and it can be speculated that some patients might have benefited from a boosted amprenavir-including regimen at baseline and, similarly, a treatment including boosted amprenavir could still be offered to some of these patients even in the case of a failure to LPV/r. We found a small increase both in number of PI mutations and in the fold change in amprenavir susceptibility, as assessed by VirtualPhenotype, after LPV/r therapy. It must be stressed that we found considerable differences in the interpretation of the same genotype across algorithms. A recent comparative study [11] found less discordance, but did not report the proportion of sequences obtained from highly experienced patients; furthermore, the authors detected a higher degree of discordance in the presence of complex mutational patterns, and found that amprenavir was responsible for 47% of the PI discordances between algorithms. These findings underline the conflicting or disputable nature of the published data concerning HIV resistance to amprenavir in the presence of multiple PI mutations, which greatly complicate the task of clinicians responsible for managing patients who are potential candidates for an amprenavir-including salvage regimen. We have previously reported a high degree of discordance between PI genotype and phenotype after more than one failure to a PI-including regimen [12], and these results were subsequently confirmed in a larger study [13]. Taken together, these data suggest that the combined use of genotype and phenotype may be useful in guiding the choice of PI in patients who have experienced more than one failure on a PI-including regimen [11,13]. In conclusion, in our cohort of highly PI-experienced patients, amprenavir resistance was the rule rather than the exception both before and after LPV/r treatment. However, we cannot exclude the possibility that some of these patients might benefit from treatment with amprenavir provided that it is boosted with ritonavir. Treatment with LPV/r may select the 50V mutation. There are considerable between-algorithm differences in terms of the inference of amprenavir phenotypes from genotypes in this setting.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.