Complex patterns of protease inhibitor resistance among antiretroviral treatment-experienced HIV-2 patients from Senegal: implications for second-line therapy.

Protease inhibitor (PI)-based antiretroviral therapy (ART) can effectively suppress HIV-2 plasma load and increase CD4 counts; however, not all PIs are equally active against HIV-2, and few data exist to support second-line therapy decisions. To identify therapeutic options for HIV-2 patients failing ART, we evaluated the frequency of PI resistance-associated amino acid changes in HIV-2 sequences from a cohort of 43 Senegalese individuals receiving unboosted indinavir (n = 18 subjects)-, lopinavir/ritonavir (n = 4)-, or indinavir and then lopinavir/ritonavir (n = 21)-containing ART. Common protease substitutions included V10I, V47A, I54M, V71I, I82F, I84V, L90M, and L99F, and most patients harbored viruses containing multiple changes. Based on genotypic data, we constructed a panel of 15 site-directed mutants of HIV-2ROD9 containing single- or multiple-treatment-associated amino acid changes in the protease-encoding region of pol. We then quantified the susceptibilities of the mutants to the HIV-2 "active" PIs saquinavir, lopinavir, and darunavir using a single-cycle assay. Relative to wild-type HIV-2, the V47A mutant was resistant to lopinavir (6.3-fold increase in the mean 50% effective concentration [EC50]), the I54M variant was resistant to darunavir and lopinavir (6.2- and 2.7-fold increases, respectively), and the L90M mutant was resistant to saquinavir (3.6-fold increase). In addition, the triple mutant that included I54M plus I84V plus L90M was resistant to all three PIs (31-, 10-, and 3.8-fold increases in the mean EC50 for darunavir, saquinavir, and lopinavir, respectively). Taken together, our data demonstrate that PI-treated HIV-2 patients frequently harbor viruses that exhibit complex patterns of PI cross-resistance. These findings suggest that sequential PI-based regimens for HIV-2 treatment may be ineffective.