 |
|
|
| |
"Deep" Sequencing Finds X4 Virus in Screening Samples of Maraviroc Failures
|
| |
| |
International HIV and Hepatitis Drug Resistance Workshop, June 8-12, 2010, Dubrovnik, Croatia
Mark Mascolini
"Deep" 454 sequencing of the HIV-1 V3 loop often detected minority populations of virus that can use the CXCR4 coreceptor in patients determined to have CCR5-using HIV by the original Trofile phenotypic assay, and in whom maraviroc-based treatment later failed in the MOTIVATE trial of treatment-experienced patients [1]. Researchers from the University of British Columbia determined that early maraviroc failure was not associated with "obvious" mutations in V3.
This analysis involved 184 MOTIVATE patients judged to have R5 virus at screening by the original Trofile assay but who had virologic failure during the trial. Investigators in Richard Harrigan's lab used the geno2pheno algorithm to interpret V3 sequence data from both standard sequencing and from 454 "deep" sequencing, using cutoffs of 5.75 for standard sequencing and 3.5 for "deep" sequencing. They also performed standard sequencing at virologic failure. Luke Swenson and colleagues used ClustalX neighbor-joining trees and Figtree to compare V3 sequences of failure samples with pretreatment "deep" sequencing data.
Standard V3 sequencing at virologic failure made the same coreceptor call as screening Trofile in 168 of 184 cases for a concordance of 91%. Both methods called virus non-R5 in 91 samples and R5 in 77. Of the 16 discordant samples, Trofile read 14 as non-R5 and standard genotyping called 2 non-R5. Non-R5 virus emerged in a median of 4 weeks, whereas virologic rebounds with R5 virus occurred after a median of 8 weeks.
The median geno2pheno false-positive rate in people with non-R5 variants was 1.1 at failure, indicating "very non-R5" virus, compared with a median screening false-positive rate of 23.9. Swenson and colleagues suggested that the 65 people who had R5 virus by both standard sequencing and Trofile at all weeks probably had virologic failure because of resistance to maraviroc or a drug in the regimen taken with maraviroc.
People in whom standard genotyping detected non-R5 virus at virologic failure did worse clinically than people with R5 virus at failure. Those with non-R5 virus at failure had a smaller median drop in viral load from study entry to week 8: 0.4 log versus 2.0 log. And people with non-R5 virus at failure gained fewer CD4 cells from study entry to week 24--a median of 34 CD4s versus 90 cells in those with R5 virus at failure.
"Deep" 454 sequencing determined that 51 of 184 patients (28%) whose maraviroc regimen failed had non-R5 virus comprising at least 2% of the viral population at screening. In 63 of 93 people (68%) with non-R5 sequences detected by "deep" sequencing in screening samples, screening sequences were closely related to failure sequences. That result suggests that outgrowth of pre-existing non-R5 virus explains the virologic failure. "Deep" sequencing also found that people whose regimen failed with non-R5 virus had an average of 15% non-R5 virus at screening, whereas people whose regimen failed with R5 virus had an average of 1% non-R5 virus at screening.
Swenson and collaborators concluded that "most non-R5 virologic failures show outgrowth of pre-existing minority variants that could be detected earlier with 'deep' sequencing." An unanswered question is how well "deep" sequencing would perform compared with the enhanced Trofile assay.
Reference
1. Swenson LC, Chui CKS, Brumme CJ, et al. Genotypic analysis of V3 in patients experiencing virological failure on maraviroc-containing regimens and correlation with pre-treatment 'deep' sequencing results. HIV and Hepatitis Drug Resistance Workshop. June 8-12, 2010. Dubrovnik, Croatia. Abstract 9.
|
| |
|
|
|
|
|
