Highly active antiretroviral therapy failure and protease and reverse transcriptase human immunodeficiency virus type 1 gene mutations.
T. Scarabaggio,Gioacchino Angarano
DOI: https://doi.org/10.1086/314909
1999-08-01
Abstract:To the Editor—To assess the role that genotypic mutations within the protease (PR) domain play in determining treatment failure, Young et al. [1] determined the sequence of PR and reverse transcriptase (RT) genes of human immunodeficiency virus type 1 (HIV-1) isolates from 35 patients in whom highly active antiretroviral therapy (HAART) was unsuccessful. Even though the authors found an unexpectedly low number of PR mutations (range, 0–6) and observed that failure may occur with only a few mutations, they emphasized the importance of patient drug exposure. Unquestionably, compliance and pharmacologic factors are important in determining the efficacy of any antiretroviral regimen, but they should not be considered the sole cause of treatment failure. For this study, we selected 8 severely immunodeficient (median CD4 cell count, 107) patients whose treatment had failed (as defined by a rise in plasma virus load or a decrease of <1 log10). The patients had received a combination regimen including a PR inhibitor, and their therapy was directly supervised by a responsible family member, leaving no doubt of treatment adherence. Most patients had received multiple nucleoside RT inhibitors for 124 months, either as monotherapy or as combination therapy, and at least two PR inhibitors (median duration, 11.5 months; range, 6–18). The sequences of plasmaderived PRand RT-coding regions were determined with a 373 sequencer (Perkin-Elmer Cetus, Norwalk, CT) and compared by use of Factura software (BCM Search Launcher; Baylor College of Medicine, Houston) with the reference HXB2–HIV1. At the time of plasma collection, all but 2 patients were receiving combination therapy with stavudine, lamivudine, and indinavir (or nelfinavir [1 patient]); the mutations detected in the PR and RT genes are listed in table 1. Mutations were categorized as either “primary” (with discernible effect on the drug susceptibility of the virus) or “secondary” [2, 3]. The primary mutations associated with PR inhibitor resistance were V82A (63% of patients), L90M (50% of patients), M46I/L (25% of patients), and G48V (12.5% of patients). Secondary mutations mainly included L63P (100% of patients) and L10I and I54V (63% each of patients). Except for 2 patients in whom L63P was detected as a single mutation, the remaining 6 patients presented a variable pattern with up to 7 mutations; however, in agreement with published reports [4, 5], we estimated that only a modest (0to 30-fold) susceptibility decrease resulted in most patients and that only 2 patients (nos. 7 and 8) had a higher reduction of PR inhibitor susceptibility. Multiple RT mutations associated with zidovudine resistance were detected in all patients, only 1 of whom was receiving zidovudine; additional RT mutations compatible with previous RT inhibitor therapy were T69D (25% of patients) and L74V (12.5% of patients). With the current lamivudine-based regimen, 86% of patients also had mutation M184V; and mutations K101E, K103N, and V108I were detected in a patient who was receiving efavirenz. Overall, in our experience, analysis of PR and RT genes gave results closely resembling those of Young et al. [1]. However, we do not feel that a large number of mutations are essential for establishment of PR inhibitor resistance, and because we are certain of our patients’ adherence, some comments might be appropriate. First, it appears that it is the association or interaction of mutations, not the single mutation in itself, that determines the decrease in drug susceptibility. Moreover, particular mutations, such as L90M, are more efficient than others, such as V82A, in determining drug resistance, and most data regarding drug resistance derive from monotherapy studies, so little is known concerning the development of phenotypic resistance to multidrug regimens. In other words, it is the quality and the association rather than the quantity of site substitutions that are crucial for treatment failures. In addition, as stated by Young et al. [1], mutations at the PR cleavage sites, which are currently not under investigation, would also augment the possibility of treatment failure. Second, the importance that RT inhibitor resistance has for HAART failure must be stressed. In our experience, most patients currently on PR inhibitor therapy have been previously administered multiple RT inhibitors, and the effect of these previous treatments on successive options remains uncertain. Furthermore, as observed in our series, after prolonged zidovudine administration, a large number of patients will become candidates for a stavudine-based regimen, but stavudine phosphorylation is normally reduced in these circumstances [6], thus leading to development of some cellular resistance, which differs according to the individual patient. Last, the extensive use of earlier RT inhibitors, such as zidovudine and didanosine, obliges clinicians to adopt newer drugs, such as lamivudine, which selects for M184V. Although the role of this mutation is not certain in the context of a multidrug regimen, it is known that M184V as a single mutation confers high-level phenotypic resistance to lamivudine [7]. Therefore, the reasons for treatment failure are extremely complex, and each patient has an accumulation of individual factors that distinguishes him or her from other patients. The effect that single mutations have on phenotypic resistance in the context of a combination regimen should be carefully investigated.