Background:  Preliminary virologic analyses have identified 2 phenotypic pathways associated with virologic failure in the phase 3 studies of maraviroc (MVC) (MOTIVATE 1 and 2). Either CXCR4-using virus was detected on MVC treatment (tropism) or virus was selected that was able to infect cells using MVC-bound CCR5 (resistance). In each case, changes in V3 loop sequence could be associated with the phenotypic pathway selected. The aims of this genotypic analysis were firstly to characterize the V3 sequences of the CXCR4-using virus selected in some patients on MVC and secondly to determine the V3 sequences for all patients with phenotypic evidence of MVC resistance at week 24.

Methods: Paired baseline and on-treatment samples from 35 patients were studied by clonal analysis. Env (gp160) sequencing and tropism confirmation were performed on 12 env clones from each time point. Clones from patients failing with R5 virus were analyzed for evidence of phenotypic resistance. Dual-tropic clones were further categorized as "dual-R" or "dual-X" based on V3 loop similarity to R5 or X4 clones isolated from the same patient. The V3 loops were analyzed using 8 freely available genotypic algorithms.

Results:  Clonal analyses confirmed results for the original population samples. Amino acid changes were identified in the V3 loops of all patients with evidence of phenotypic resistance. The amino acid changes differed between patients, reflecting the heterogeneity of the gp160 sequence and multiple genetic pathways to resistance. Dual-tropic clones were identified in samples from 18 out of 20 patients:  12 patients who received MVC had "dual-X" virus on-treatment, in 7 of whom the virus was identified at baseline; in 5 MVC-treated patients and 2 placebo-treated patients, the "dual-X" virus emerged during treatment. A total of 6 patients had "dual-R" clones at baseline; 4 received MVC. The "dual-R" virus was not identified during treatment in 3 of the MVC patients but persisted in both placebo patients. An analysis of the infectivity values from the Trofile assay showed a different profile between the two types of dual-tropic virus.

Conclusions: Mutations in the V3 loop appear to play a key role in conferring the MVC CCR5-tropic resistant phenotype. This limited dataset suggests that some dual-tropic clones are responsive to MVC in vivo. However, in the patients studied, these clones always co-existed with non-responsive "dual-X" clones.