Feasibility analysis on genotypic prediction of coreceptor usage of HIV-1 CRF01_AE strains in China

Abstract

Objective To analyze feasibility of using genotypic prediction for coreceptor usage of HIV-1 CRF01_AE strains circulating in China. Methods Twenty-two CRF01_AE strains were isolated from HIV-1 infected patients in Beijing, Guangxi and Sichuan. The genotypic methods (11/25 rule, WebPSSM and geno2pheno) were used to predict coreceptor usage of the strains. The prediction were compared with the results of phenotypic testing based on Ghost cell line. Results Of the 22 CRF01_AE strains, 4 strains used CCR5/CXCR4 coreceptor and 18 strains used the CCR5 coreceptor, as detected by the GHOST cell line test. The consistency rate between genotypic prediction by 11/25 rule and phenotypic test was 86.4%. For geno2pheno, the consistency rates between genotypic predictions by G2p-clinical, NGS-sanger and G2p-colon models and phenotypic tests were 95.5%, 86.4% and 68.2%, respectively. For WebPSSM, the consistency rates between genotypic predictions by PSSM-sinsi_B, PSSM-x4r5, and PSSM-sinsi_C models and phenotypic tests were 86.4%, 72.7% and 45.5%, respectively. Conclusions For the patients infected with HIV-1 CRF01_AE strains in China, coreceptor usage should be predicted before treatment of coreceptor antagonists. Geno2pheno is a more accurate way for genotypic prediction of HIV coreceptor usage. Key words: HIV-1; CRF01_AE strain; Coreceptor; Genotypic prediction