AUTHOR=Judicate George P. , Barabona Godfrey , Kamori Doreen , Mahiti Macdonald , Tan Toong Seng , Ozono Seiya , Mgunya Amina Shaban , Kuwata Takeo , Matsushita Shuzo , Sunguya Bruno , Lyamuya Eligius , Tokunaga Kenzo , Ueno Takamasa TITLE=Phenotypic and Genotypic Co-receptor Tropism Testing in HIV-1 Epidemic Region of Tanzania Where Multiple Non-B Subtypes Co-circulate JOURNAL=Frontiers in Microbiology VOLUME=12 YEAR=2021 URL=https://www.frontiersin.org/journals/microbiology/articles/10.3389/fmicb.2021.703041 DOI=10.3389/fmicb.2021.703041 ISSN=1664-302X ABSTRACT=

HIV human immunodeficiency virus type I (HIV-1) entry inhibitor potency is dependent on viral co-receptor tropisms and thereby tropism determination is clinically important. However, phenotypic tropisms of HIV-1 non-B subtypes have been poorly investigated and the genotypic prediction algorithms remain insufficiently validated. To clarify this issue, we recruited 52 treatment-naïve, HIV-1-infected patients in Tanzania, where multiple HIV-1 non-B subtypes co-circulate. Sequence analysis of 93 infectious envelope clones isolated from their plasma viral RNA revealed the co-circulation of subtypes A1, C, D, and inter-subtype recombinant forms (isRFs). Phenotypic tropism assays revealed that lentivirus reporters pseudotyped with 75 (80.6%) and 5 (5.4%) envelope clones could establish infection toward U87.CD4 cells expressing CCR5 (R5) and CXCR4 (X4), respectively; whereas the remaining 13 (14%) clones could infect both cells. Genotypic analyses by widely used algorithms including V3 net charge, Geno2pheno, WebPSSM, and PhenoSeq showed that almost all phenotypic X4-tropic clones and only 15 of 75 phenotypic R5-tropic clones were concordantly predicted. However, the remaining 60 phenotypic R5-tropic clones were discordantly predicted by at least one algorithm. In particular, 2 phenotypic R5-tropic clones were discordantly predicted by all algorithms tested. Taken together, the results demonstrate the limitation of currently available genotypic algorithms for predicting co-receptor inference among co-circulating multiple non-B subtypes and emerging isRFs. Also, the phenotypic tropism dataset presented here could be valuable for retraining of the widely used genotypic prediction algorithms to enhance their performance.