icon-folder.gif   Conference Reports for NATAP  
  Targeting HIV Entry: 4th International Workshop
Rio Grande, Puerto Rico
December 8-9, 2008
Back grey_arrow_rt.gif
Almost 40% of Viruses in Treatment-Experienced German Group Use CXCR4
  Targeting HIV Entry: 4th International Workshop,
December 8-9, 2008, Rio Grande, Puerto Rico
Mark Mascolini
Testing coreceptor use with the enhanced Trofile assay, German investigators determined that nearly 40% of viruses isolated from treatment-experienced people could use the CXCR4 coreceptor, a result meaning CCR5 antagonists will not help much in controlling these patients' viral replication. As in earlier studies, a gene-based algorithm used to predict coreceptor use had high specificity but relatively low sensitivity in yielding the same results as Trofile. Contrary to other studies, however, the German survey found that genotyping predicted Trofile-determined coreceptor use better in people with a non-B HIV-1 subtype than in people infected with subtype B virus.
The researchers used both the original Trofile phenotypic assay and the enhanced assay to figure coreceptor use in 738 viral isolates, 642 (87%) from people with subtype B virus and 92 (13%) from people with other HIV-1 subtypes. (Subtype could not be determined in 4 isolates.) Because of the small number of individual non-B subtypes, the investigators combined all non-B viruses for this analysis. The multicenter team also sequenced the V3 loop of HIV gp120 and applied the geno2pheno algorithm to predict coreceptor use based on genotype.
The investigators used the original Trofile to characterize coreceptor use of 619 isolates (83.9%) and the enhanced Trofile to characterize 119 isolates (16.1%). The original Trofile rated 414 isolates (66.9%) R5 and 205 (33.1%) dual-mixed (DM), meaning those viruses could use CXCR4. The enhanced Trofile called 72 isolates (60.5%) R5 and 47 (39.5%) DM.
Adjusting the false-positive rate of the geno2pheno predicting tool results in different sensitivities and specificities in determining the same coreceptor use as Trofile. With the false-positive rate set at 10%, geno2pheno had a sensitivity of only 63.4% compared with the standard assay and 59.6% compared with the enhanced assay. With the same false-positive cutoff of 10%, specificity with geno2pheno was much higher--89.6% compared with original Trofile results and 91.7% compared with enhanced Trofile results.
The German group had virologic outcomes for 43 people who started a salvage regimen including the CCR5 antagonist maraviroc. For 15 of these people, the researchers had both geno2pheno and Trofile coreceptor calls, and for 26 they had only geno2pheno calls. Fourteen of 15 people rated R5 by Trofile reached a viral load below 50 copies, while 32 of 37 rated R5 by geno2pheno reached a sub-50 load. One of 3 people rated X4 by Trofile and 3 of 4 rated X4 by geno2pheno reached a viral load below 50 copies. Although the latter findings suggest some people whose HIV got rated X4 actually had R5 virus, interpreting that result is difficult until the researchers analyze the number of active antiretrovirals taken with maraviroc. A CCR5 antagonist can stymie R5 viruses in a mixed viral population while other drugs in the regimen shut down the X4 viruses.
With the geno2pheno false-positive rate set at 10%, the algorithm predicted Trofile-determined coreceptor use in subtype B isolates with a sensitivity of only 60.3% but a specificity of 90.3%. In contrast, with non-B subtype isolates, geno2pheno sensitivity jumped to 85.7% and specificity stayed good at 87.3%. These results run counter to earlier studies that found better geno2pheno coreceptor predictions with subtype B virus.
1. Obermeier M, Sichtig N, Braun P, et al. Genotypic and phenotypic analysis of coreceptor usage in a large cohort of therapy-experienced patients in Germany. 4th International Workshop. December 8-9, 2008. Rio Grande, Puerto Rico. Abstract 9.