Background:  In a phase I/II evaluation of the CXCR4 inhibitor AMD3100, HIV RNA was significantly reduced (0.9 log₁₀) in the single study subject that harbored CXCR4-tropic virus, but not in the remaining subjects that harbored either dual or mixed (DM)-tropic or R5-tropic virus_. In this study, we further evaluated the targeted antiviral activity of AMD3100 in vivo by characterizing the tropism composition of the baseline and on-treatment DM virus populations.

Methods:  Subjects from the AMD3100 phase I/II study that were determined to have DM virus at baseline were included in this study (n = 15). Co-receptor tropism of virus population and individual envelope clones was evaluated using a single replication cycle envelope pseudovirus assay (Trofile) and patient-derived gp160 env clones were sequenced.

Results:  By comparing the infectivity in U87/CD4/CXCR4 cells of paired baseline and day 11 virus populations, 2 groups were identified. The suppressor group (n = 11) displayed a reduction in CXCR4 infectivity at day 11 relative to baseline; most subjects in this group experienced >5-fold reductions in CXCR4 infectivity and 8 subjects became R5-tropic. In contrast, no changes in CXCR4 infectivity of paired baseline and day 11 samples were observed in the non-suppressor group (n = 4). No significant viral load reductions were observed between the baseline and day-11 samples in any subjects from either group. Furthermore viruses from the suppressor group showed a lower level of luciferase activity on CXCR4 cells (10² to 10⁴ relative light units [rlu] ) than the non-suppressor group (10⁵ to 10⁶ rlu) at baseline. Clonal analysis of baseline populations indicated that the non-suppressor group was enriched for dual-tropic variants while the suppressor group contained a higher proportion of R5-tropic variants. The mean percentage of CXCR4-using clones in the non-suppressor group was 95% compared to only 27% in the suppressor group.

Conclusions:  This study indicates that AMD3100 has the ability to inhibit both X4-tropic and certain dual-tropic variants in vivo. Dual-tropic viruses exhibit considerable variation in their efficiency of CXCR4 and CCR5 co-receptor use. The efficiency of CXCR4 usage and the tropism composition of the baseline virus population appears to influence the suppression of dual-tropic variants by AMD3100. Further characterization of dual-tropic variants is needed to fully appreciate the susceptibility of dual tropic viruses to CXCR4 and CCR5 targeted therapies.