Background: In this study we have used HIV-1 isolates from individuals receiving a protease inhibitor (PI) based regimen, to analyze the impact of basal genetic background on viral fitness evolution.

Methods:  16 paired plasma samples and HIV-1 isolates, corresponding to weeks 0, 12, 24, and 48 post-treatment, were obtained from 4 randomly selected AZT-experienced, PI naïve patients enrolled in the ACTG 315 study (i.e., zidovudine + lamivudine + ritonavir). Viral fitness was measured using competition experiments, against a set of 3 wild-type HIV-1 controls, followed by HTA to estimate the final ratio of the 2 viruses in the mixture.  Relative viral fitness values were then expressed as a percentage of the replicative capacity of a wild-type subtype B primary U.S. isolate.  Viral RNA was extracted from the plasma samples and genomic regions encoding PR and RT (pol gene) were RT-PCR amplified and sequenced.

Results: Pre-therapy genotypic analyses of PR and RT revealed 2 patients harboring wild-type viruses, whereas 2 had AZT resistant strains. 1 of the subjects who harbored AZT mutations also had polymorphisms in the PR gene at codons associated with PI resistance (i.e., 36, 77, and 93).  After 48 weeks, only the patients with baseline AZT mutations had developed resistant virus to all 3 antiretroviral drugs (PR, 82; RT, 41-219, 184).  Compensatory mutations in PR (10, 36, 54, 77, and 93) were observed only in the virus that contained PR polymorphisms at baseline. Ex vivo viral fitness evolution paralleled these genotypic changes.  At week 48, high viral replicative capacity was observed in both wild-type viruses (71.8% and 96.3%), while lower viral fitness was estimated for the drug resistant viruses (34.9% and 61.5%).  Interestingly, in those patients failing therapy, 2 distinct evolution patterns were observed when viral fitness at baseline and post-therapy were compared.  An increase in viral fitness (+50% of the initial fitness value) was observed in the drug resistant isolate carrying natural compensatory mutations in the PR gene.  In contrast, a reduction in replicative capacity (-28%) was calculated for the virus harboring only primary PR mutations.  In addition, despite a similar viral load rebound, an inverse relationship was observed between viral fitness and CD4 counts (61.5%, 267 cells/mL vs 34.9%, 443 cells/mL) in these 2 subjects.         

Conclusions: Selection of drug resistant viruses with impaired fitness is often followed by generation and selection of additional compensatory mutations, which enhance replication capacity.  However, in this study, we observed that the evolution of viral fitness depends on the initial HIV-1 genetic background and it correlates well with the virologic response to antiretroviral therapy.  Thus, natural polymorphisms associated with resistance to PI- or RT-inhibitors, may lead to a more rapid recovery of viral replication capacity and further HIV pathogenesis.