HIV-1 biological phenotype and predicted coreceptor usage based on V3 loop sequence in paired PBMC and plasma samples

Paired PBMCs and plasma samples from 34 HIV-infected patients were studied to verify the relationship between coreceptor use based on genotyping of V3 region of HIV-1 envelope gp120 and biological phenotype with virus isolation and subsequent correlation to clinical characteristics. The "11/25" rule, geno2pheno and PSSM were compared. All SI patients were HIV-1 subtype B (p=0.04) and had a lower CD4 count than NSI patients (p=0.01), while no differences were observed in mean HIV-RNA (log) (p=0.6). SI phenotype was not associated with AIDS-defining events (p=0.1) or with concurrent antiretroviral therapy (p=0.4). With geno2pheno, which shows the highest sensibility (83%), an X4 or X4/R5 genotype in PBMC DNA was also associated to B-subtype and lower CD4 count (p=0.01) compared to R5 isolates. Based on plasma RNA sequences, the predicted coreceptor usage agreed with PBMC DNA in 79% of cases with the "11/25" rule, 82% with geno2pheno, and 82% with PSSM. A X4 virus in plasma (but not in PBMCs) was significantly associated with HAART in all three methods (p=0.01 for "11/25" rule, p=0.01 for geno2pheno and p=0.03 for PSSM). Due to viral mixtures and/or difficulties in genotype interpretation, current V3 sequence-based methods cannot accurately predict HIV-1 coreceptor use.