High prevalence of CXCR4 in HAART-naïve HIV-1 CRF01_AE infected patients

Abstract

BACKGROUND: HIV-1 entry pathway can be blocked by a recently developed CCR5 antagonist, maraviroc, and it only has inhibitory effects on R5-tropic viruses. Population-based studies revealed that R5-tropic viruses usually pre-dominated in treatment-naïve patients and in the stage of primary infection. X4-tropic viruses are often associated with late disease progression. The prevalence of X4-tropic viruses in treatment-naïve patients in non-B subtypes remains unclear. In this study, we compared a cohort of treatment-naïve patients infected by both subtype B and CRF01_AE on their co-receptor usage predicted by genotypic algorithms, in order to determine the prevalence of tropism in CRF01_AE infected patients. MATERIALS & METHODS: A total of 105 subtype B and 86 subtype CRF01_AE antiretroviral-naïve patients were recruited at the Government Integrated Treatment Centre between June 2009 and June 2011 in Hong Kong. Peripheral blood was collected from all patients after their consent. Viral RNA was first extracted from plasma, followed with amplification and sequencing by a previously published protocol (McGovern et al 2010). HIV-1 co-receptor usage was predicted by two sets of algorithms: (i) Geno2Pheno [co-receptor] 1.2, false positive rate was set at 10%; (ii) Web PSSM, using subtype B x4r5 matrix. Chi-square test was used to calculate the significance of tropism distribution in two subtypes. RESULTS: Recommendations from the European Consensus Group on clinical management of HIV-1 tropism testing (10% FPR)’ was used in Geno2Pheno analysis. 26 (24.8%) of subtype B samples and 47 (54.7%) of subtype CRF01_AE samples were predicted to use CXCR4 co-receptor. Chi-square test showed that the distribution of R5-tropic and X4-tropic viruses was significantly different between two subtypes (P < 0.0001). WebPSSM used amino acid sequence for co-receptor prediction. All possible amino acid sequences were generated when there were any mixed base-pair occurred in the V3 sequences. Only 15 (14.3%) subtype B and 38 (44.2%) subtype CRF01_AE had X4-tropic variants detected. Statistical analysis suggested CRF01_AE had more X4-tropic variants in viral population than subtype B (P < 0.0001). CONCLUSIONS: Our data demonstrated that the prevalence of X4-tropic virus was significantly higher in CRF01_AE treatment-naïve patients than subtype B, despite of any genotypic algorithms. X4-tropic viruses were expected to be more virulence and accelerated disease progression during late stage of infection. It may suggest CRF01_AE viruses in Hong Kong are more virulent, better in viral fitness or even have a faster disease progression rate than subtype B. More research and clinical monitoring are required to confirm this phenomenon. With the possible high prevalence of CXCR4 in CRF01_AE treatment-naïve patients, our findings indicate the use of maraviroc will be more cost-effective for first-line therapy instead of salvage therapy.