Antiretroviral Therapy

Resistance

The potential for varying antiretroviral resistance patterns in different body compartments was addressed. No differences in RT and protease sequences were found between 42 paired viral samples obtained from plasma and lymph node of 28 patients. (Abstract I-122.) Drug resistance was infrequently detected in virus isolated from the genital tract of women when compared to paired virus isolated from their plasma. (Abstract I-124.)

A consensus is emerging that current methodologies for detecting in vitro resistance of HIV-1 to antiretroviral agents may be useful in predicting which drugs will be ineffective, but that the methodologies are insufficiently sensitive to minority viral populations to be reliably predictive of drug success.

There is, in addition, a continuing battle among varying technologies for determining genotypic and phenotypic resistance. Three methods of genotyping were compared: gene chip hybridization-based sequencing (Affymetrix), direct solid phase sequencing (Applied Biosystems), and point mutation probe assay (Murex). A panel of nine RT plasmids and 28 clinical isolates were examined. While gene sequencing detected the most mutations, the other two methods also performed well. (Abstract I-114.)

Polymorphic changes in the HIV-1 gene pool not associated with drug resistance consistently developed prior to the appearance of mutations associated with drug resistance. This suggests that their detection may serve as an early warning system. (Abstract I-76.)

An evaluation of treatment-experienced patients given ritonavir and saquinavir in combination found, by multivariate analysis, that both baseline genotype (by VircoGENTM sequence analysis) and phenotype (by Antivirogram TM) antiretroviral susceptibility tests were highly predictive of response to this combination. Patients with baseline phenotypically susceptible virus were 12-fold (95% CI 1.2-111; P < 0.05) more likely to respond (HIV RNA < 500 copies/mL) than those with non-susceptible patterns. Virologic response was four-fold (95% CI 1.0-13; P < 0.05) more likely in those with genotypic susceptibility. Nonetheless, only approximately one-half of patients with in vitro evidence of drug susceptibility responded to the salvage regimen with a reduction in viral load below the limits of detection. (Abstract I-78.) This is consistent with the emerging consensus noted above, that genotypic and phenotypic tests are useful in predicting treatment failure, but not success.

One potential mechanism of resistance not detectable by current assays is reduced phosphorylation of nucleoside analogs. The thymidine kinase content of peripheral blood mononuclear cells is decreased in patients with high viral load and/or low CD4 count, theoretically leading to decreased phosphorylation of some NRTIs and resultant reduced antiretroviral activity. (Abstract A-70.)

Multinucleoside analog resistance (MNR) has previously been associated with the accumulation of multiple mutations, as well as with the codon 151 multidrug resistance cluster. MNR has also been found in association with insertions in the RT gene at or near codon 69, most commonly 69S (-S-S). While this alteration alone does not cause significant resistance, its presence together with ZDV resistance mutations is associated with cross-resistance to d4T, ddI, and ddC. (Sunday, Abstract LB-4.)

The ZDV-associated resistance mutations, M41L and M41L + K70R, are associated with reduced viral fitness. (Abstract I-111.)

The complete protease genes of HIV-1 isolates from 58 PI naïve patients were sequenced prior to the initiation of therapy, which included a PI to determine if the presence of polymorphisms predict failure of response to treatment. No mutations known to be associated with PI resistance were detected, but differences from the subtype B consensus sequence were detected at 29 individual codons, with a median number of differences per isolate of four. There was no correlation between the number of amino acid substitutions and response to therapy, indicating that these natural polymorphisms are not relevant to PI treatment outcome. (Abstract I-112.)

Failure of ritonavir therapy is initially associated with increased drug resistance and reduced viral replicative capacity (fitness). At this stage, the viral protease demonstrates an increased protease Ki to ritonavir, consistent with resistance to this drug; the reduced fitness is associated with diminished catalytic efficiency of the protease. With continued drug administration (and, thus, continued selective pressure), however, a quadruple mutant (36I-54V-71V-82T) that has increased replicative efficiency may subsequently arise. Evolution of this increased viral fitness is associated with an increased catalytic efficiency of the protease compared to wild type viral protease, while the ritonavir Ki remains stable. The unfortunate consequence is a drug-resistant virus with increased replicative capacity. (Abstract I-118.)