Drug-resistant mutants of HIV-1 in patients exhibiting increasing CD4 cell count despite virological failure of highly active antiretroviral therapy

The role of HIV-1 resistance-conferring mutations in predicting discordant viro-immunological responses to highly active antiretroviral therapy was investigated. Reverse transcriptase (RT) M184V, and proteases L24I and V82A were significantly associated with an increased CD4 cell count despite virological failure, whereas RT Y181C reduced the probability of immunorecovery. Lower HIV-1-RNA levels in discordant patients and higher values in those carrying RT Y181C suggested a role of viral replicative capacity. Selected mutations should be considered in evaluating the benefit of continuing therapy in discordant patients. Highly active antiretroviral therapy (HAART) has significantly reduced morbidity and mortality rates in HIV-positive patients [1], even though a significant number of patients eventually experience a virological failure to these drug treatment combinations, which can be partly explained by the development of drug resistance [2,3]. However, despite virological failure, in a considerable proportion of patients the CD4 cell count remained elevated above the pretherapy baseline level [4–6]. Several mechanisms may account for such discrepant responses to treatment, including the prevention of T cell apoptosis by protease inhibitors [7], an alteration in T cell turnover kinetics, and changes in viral fitness [4,6]. Moreover, it has been reported that patients who exhibit increasing CD4 cell counts, even in the absence of significant virological response to HAART, displayed a similar incidence and pattern of resistance mutations to protease inhibitors as patients who exhibit both immunological and virological failure to therapy [8]. In order to investigate the role of resistance-conferring genotypic mutations selected during HAART in the reverse transcriptase (RT) and protease gene as predictive factors of the dissociation of immunological and virological responses to HAART, we studied a group of HIV-infected patients with virological failure consecutively observed at two HIV reference centres in Rome, in whom a genotypic resistance test was performed. Virological failure was defined as a reduction of the HIV-1-RNA level of less than 1 log10 after 2 months of treatment or persistent detectable (> 80 copies/ml) HIV-1 RNA after 6 months, or any rebound (confirmed with at least two consecutive tests) after reaching undetectable viraemia. For investigational purposes a standardized definition of discordant immunological response was used, consisting of a greater than 100 CD4 cells/μl increase from starting HAART to the genotyping test, despite persistent virological failure. In order to exclude bias caused by recent virological rebound, a more stringent definition considering only patients with more than 6 months of persistent virological failure was applied. HIV-1-RNA samples were prospectively analysed by direct sequencing of RT and protease using two commercial assays: the HIV-1 Perkin-Elmer genotyping kit (Perkin-Elmer, Foster City, CA, USA) and the TrueGene HIV-1 assay (Visible Genetics Inc., Toronto, Ontario, Canada). For each patient detailed data on antiretroviral history were collected. Statistical analysis was performed by independent t-test for continuous and Fisher's exact test for discrete variables. A forward stepwise logistic regression model was employed to estimate univariate and multivariate odds ratios (OR) for categorical and continuous variables. Baseline characteristics of 354 enrolled patients showed: a median age of 38 years [interquarterile range (IQR) 34–42]; male sex 245 (69%); intravenous drug users 92 (26%), men who have sex with men 87 (25%), heterosexuals 121 (34%); and previous AIDS 143 (40%). The median CD4 cell count and HIV-1-RNA level at genotyping were 275 cells/μl (IQR 138–394) and 4.63 log10 copies/ml (IQR 4.11–5.07), respectively. Patients were on HAART for a median of 26 months (IQR 17–33). A mean increase of 129 CD4 cells/μl [95% confidence interval (CI) 109–149] and a mean decrease of 0.09 log10 copies/ml of HIV-1 RNA (95% CI 0.03–0.23) was observed from starting HAART to genotyping. A CD4 cell count increase of over 100 cells/μl from baseline to genotyping despite virological failure was observed in 159 (45%) patients. This proportion of patients with discordant immunovirological response increased to 51% (84/166) among those with over 6 months of persistent viraemia from the definition of virological failure. Table 1 shows univariate analysis of factors related to an immunological recovery despite virological failure. A prolonged time on HAART, a history of an ever-undetectable HIV-1-RNA level, the presence of RT M184V and proteases L24I and V82A were all significantly associated with an increased probability of immunological recovery of more than 100 CD4 cells/μl from baseline despite virological failure. By contrast, an immunological recovery despite virological failure was significantly decreased by a high HIV-1-RNA level at genotype, a high CD4 cell count at baseline, nevirapine exposure among non-nucleoside reverse transcriptase inhibitors (NNRTI) and the emergence of RT Y181C. At multivariate analysis only the HIV-1-RNA level at genotype, CD4 cell count at baseline, and RT Y181C were associated with a reduced probability of sustained immunological recovery, whereas overall time on HAART, an ever-undetectable HIV-1-RNA level and protease L24I significantly predicted persistently elevated CD4 cell counts. When only the 166 patients with more than 6 months from last virological failure were considered, HIV-1 RNA at genotype, CD4 cell count at baseline and RT Y181C were predictive of failing, whereas overall time on HAART was predictive of achieving an immunological recovery despite virological failure (Table 1). Patients with immunological recovery had lower mean values of HIV-1 RNA at genotype (4.52 versus 4.82;P < 0.01). Patients with RT Y181C mutants had higher mean values of log10 HIV-1 RNA (4.86 versus 4.59;P < 0.01), lower absolute values of CD4 cell counts at genotyping (231 versus 313 cells/μl;P < 0.01) and a lower change of CD4 cell count between baseline and genotyping (+68 versus +142 cells/μl;P < 0.01).Table 1: Factors associated with probability of over 100 CD4 cells/μl increase from baseline despite virological failure. Our data suggest that Y181C may negatively affect sustained immunological recovery despite virological failure (OR 0.40; 95% CI 0.20–0.80). A reduction in viral fitness has been postulated to explain the preservation of CD4 cells in patients with persistent viral replication. Decreased replicative capacity, as observed in viruses containing protease inhibitor or nucleoside reverse transcriptase inhibitor-associated mutations would result in reduced CD4 cell death, increased CD4 cell production, and finally the achievement of a new steady state [6]. Our finding of lower HIV-1-RNA levels in patients with CD4 cell recovery confirmed the role of viral fitness. According to these observations most HIV-1 strains in patients with a discordant viro-immunological response harboured primary and secondary mutations in the protease gene [8]. NNRTI resistance is largely restricted to codons from 98 to 110 and from 179 and 190, and those occurring more frequently in clinical isolates during early virological failure were Y181C, G190A, G106A and K103N [9]. At present, little is known about the effects of NNRTI resistance mutations on the catalytic activity of HIV-1 RT, even though the rapid rebound observed for strains containing RT Y181C or K103N suggests that these mutations have a limited impact on viral fitness. In-vitro models show only a modest reduction in the replication kinetics of strains containing Y181C compared with wild-type virus [10], and it is possible that the distance of NNRTI mutations from the active site of enzyme does not allow a reduction of enzymatic activity and consequently viral replicative capacity. This modest impact on the viral fitness of selected NNRTI mutations, such as K103N and Y181C, is in agreement with the higher HIV-1-RNA level in patients carrying Y181C mutants and with the negative predictive value of Y181C on sustained CD4 cell recovery observed in our study. On the contrary, our data confirm the role of V82A in the protease and of M184V in the RT as mutations conferring a reduced replicative capacity to the virus [8,11]. The role of accessory mutation protease L24I has to be further elucidated in larger cohorts (only 20 out of 354 failing patients carried such a mutation), yet its position in the tertiary structure of HIV-1 protease (just beside the catalytic core formed by amino acids 25–27) strongly suggests that such a mutation may be relevant in affecting virus fitness. These observations may have several clinical implications. It has been observed that among patients with persistent viraemia, despite the presence of reduced drug susceptibility, antiretroviral therapy is associated with sustained immunological benefit, with the maintenance of a viral population with reduced replicative capacity [12]. It is conceivable that these viro-immunological discrepancies could be at least partly mediated by the different effects of mutations on viral fitness. On one hand, this opens the possibility of considering, in situations of remarkable viro-immunological dissociation and in the absence of valid therapeutic alternatives, the maintenance of drugs selecting for mutations strongly affecting virus fitness. On the other hand, the emergence of a single point mutation able to lead rapidly to broad cross-resistance to the entire NNRTI drug class [13], in the absence of sustained immunological response (as is the case with Y181C), should suggest a rapid discontinuation of NNRTI after a confirmed virological failure. Andrea Antinoria Giuseppina Liuzzia Antonella Cingolanib Ada Bertolia Simona Di Giambenedettob Maria P. Trottaa Maria G. Rizzob Enrico Girardia Andrea De Lucab Carlo F. Pernoa

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.