Poveda E, Seclén E, González Mdel M et al. J Antimicrob Chemother 2009;63:1006–10.
In this timely report, Poveda and colleagues describe their efforts to optimize an existing web-based algorithm (position-specific scoring matrix [PSSM]) designed to predict HIV-1 coreceptor usage based on the V3 region of the HIV envelope protein gp120 (the primary determinant of viral tropism). Previous bioinformatic studies have derived the PSSM, a bioinformatic method estimating the propensity of a V3-amino acid sequence to use the CXCR4 receptor, based on the presence of single amino acid substitutions at positions 11, 24, and 25, as well as the net charge of these amino acids.
CCR5 chemokine receptor antagonists such as maraviroc and vicriviroc selectively inhibit HIV-1 strains that exhibit growth preference as determined by use of the CCR5 coreceptor. Treatment failures on these regimens are linked to the presence, prior to treatment, and subsequent emergence of CXCR4- using (X4) mixed and dual-tropic virus. As recommended by licensing authorities and expert guidelines, clinicians use the Trofile® assay, a reliable but costly and highly specialized phenotype assay to assess viral tropism (this is performed by shipping samples to the manufacturers of the assay, Monogram Bioscience [South San Francisco, CA, USA]). Poveda and colleagues identified one group of samples on which to base their improved algorithm design (202 samples archived from the Spanish maraviroc expanded-access program), and then validated their results using a second group (148 samples from a second study assessing the prevalence of CCR5-tropic virus in Spain). From each study, they split blood samples into two and performed two tests, the Trofile assay using one-half of the sample, and a much simpler process of amplification and genetic sequencing of the Env V3 region using the second part of the sample.