A Genotypic Test for HIV-1 Tropism Combining Sanger Sequencing with Ultradeep Sequencing Predicts Virologic Response in Treatment-Experienced Patients

• Published in: PloS one Vol. 7; no. 9; p. e46334
• Main Authors: Kagan, Ron M., Johnson, Erik P., Siaw, Martin, Biswas, Pinaki, Chapman, Douglass S., Su, Zhaohui, Platt, Jamie L., Pesano, Rick L.
• Format: Journal Article
• Language: English
• Published: United States Public Library of Science 27.09.2012; Public Library of Science (PLoS)
• Subjects: Acquired immune deficiency syndrome; Adult; Aged; AIDS; Algorithms; Anti-HIV Agents - metabolism; Anti-HIV Agents - pharmacology; Antiretroviral drugs; Biological Assay; Biology; Biomarkers, Pharmacological; CCR5 protein; CCR5 Receptor Antagonists; Clinical trials; CXCR4 protein; Cyclohexanes - metabolism; Cyclohexanes - pharmacology; Diagnosis; Female; Genetic aspects; Genotyping Techniques; Glycoproteins; High-Throughput Nucleotide Sequencing; HIV; HIV infections; HIV Infections - drug therapy; HIV Infections - metabolism; HIV Infections - virology; HIV tests; HIV-1 - drug effects; HIV-1 - physiology; Human immunodeficiency virus; Humans; Infectious diseases; Male; Medical research; Medicine; Middle Aged; Patients; Predictive Value of Tests; Receptors, CCR5 - metabolism; Receptors, CXCR4 - metabolism; Retrospective Studies; Sensitivity; Sensitivity enhancement; Triazoles - metabolism; Triazoles - pharmacology; Tropism; Viral Tropism; Viruses; United States; New York; California; United States > US; San Juan Capistrano California
• ISSN: 1932-6203; 1932-6203
• DOI: 10.1371/journal.pone.0046334

A tropism test is required prior to initiation of CCR5 antagonist therapy in HIV-1 infected individuals, as these agents are not effective in patients harboring CXCR4 (X4) coreceptor-using viral variants. We developed a clinical laboratory-based genotypic tropism test for detection of CCR5-using (R5) or X4 variants that utilizes triplicate population sequencing (TPS) followed by ultradeep sequencing (UDS) for samples classified as R5. Tropism was inferred using the bioinformatic algorithms geno2pheno([coreceptor]) and PSSM(x4r5). Virologic response as a function of tropism readout was retrospectively assessed using blinded samples from treatment-experienced subjects who received maraviroc (N = 327) in the MOTIVATE and A4001029 clinical trials. MOTIVATE patients were classified as R5 and A4001029 patients were classified as non-R5 by the original Trofile test. Virologic response was compared between the R5 and non-R5 groups determined by TPS, UDS alone, the reflex strategy and the Trofile Enhanced Sensitivity (TF-ES) test. UDS had greater sensitivity than TPS to detect minority non-R5 variants. The median log(10) viral load change at week 8 was -2.4 for R5 subjects, regardless of the method used for classification; for subjects with non-R5 virus, median changes were -1.2 for TF-ES or the Reflex Test and -1.0 for UDS. The differences between R5 and non-R5 groups were highly significant in all 3 cases (p<0.0001). At week 8, the positive predictive value was 66% for TF-ES and 65% for both the Reflex test and UDS. Negative predictive values were 59% for TF-ES, 58% for the Reflex Test and 61% for UDS. In conclusion, genotypic tropism testing using UDS alone or a reflex strategy separated maraviroc responders and non-responders as well as a sensitive phenotypic test, and both assays showed improved performance compared to TPS alone. Genotypic tropism tests may provide an alternative to phenotypic testing with similar discriminating ability.