In-depth characterization of viral isolates from plasma and cells compared with plasma circulating quasispecies in early HIV-1 infection

Abstract

Background: The use of in vitro models to unravel the phenotypic characteristics of circulating viral variants is key to understanding HIV-1 pathogenesis but limited by the availability of primary viral isolates from biological samples. However, overall in vivo genetic variability of HIV-1 within a subject may not be reflected in the viable viral population obtained after isolation. Although several studies have tried to determine whether viral populations expanded in vitro are representative of in vivo findings, the answer remains unclear due to the reduced number of clonal sequences analyzed or samples compared. In order to overcome previous experimental limitations, here we applied Deep Pyrosequencing (DPS) technology in combination with phenotypic experiments to analyze and compare with unprecedented detail the composition of viral isolates and in vivo quasispecies.

Methodology/principal findings: We amplified by DPS HIV-1 genomic regions covering gag, protease, integrase and env-V3 to characterize paired isolates from plasma and peripheral blood mononuclear cells and compare them with total plasma viral RNA in four recently HIV-1 infected subjects. Our study demonstrated the presence of unique haplotypes scattered between sample types with conservation of major variants. In addition, no differences in intra- and inter-population encoded protein variability were found between the different types of isolates or when these were compared to plasma viral RNA within subjects. Additionally, in vitro experiments demonstrated phenotypic similarities in terms of replicative capacity and co-receptor usage between viral isolates and plasma viral RNA.

Conclusion: This study is the first in-depth comparison and characterization of viral isolates from different sources and plasma circulating quasispecies using DPS in recently HIV-1 infected subjects. Our data supports the use of primary isolates regardless of their plasma or cellular origin to define genetic variability and biological traits of circulating HIV-1 quasispecies.

Figure 1. Comparison of the efficiencies of HIV-1 isolation methods from PBMCs or plasma samples.

To determine the efficiency of HIV-1 isolation from PBMCs and plasma, we compared virus recovery from 56 plasma samples and 38 PBMCs samples with viral load ranging from 10 to >10⁶ copies/ml. (A) Efficiency of HIV-1 recovery from PBMCs in percentages per viral load range. (B) Efficiency of HIV-1 recovery from plasma samples in percentages per viral load range. Bars represent mean values. Numbers next to the bars indicate (number of positive samples/total number of samples tested).

Figure 2. Schematic representation HIV-1 sample types analyzed for comparative purposes per study subject.

Total blood was separated into plasma and PBMCs for the following: 1. Total viral RNA extraction (RNA) used for DPS and virus tropism. 2. Plasma virus isolation (VP) used for DPS, virus tropism and Replicative Capacity (RC): VP isolates were obtained by mixing plasma extracted anti-CD44 HIV-1 particles with a pool of CD8⁺-depleted PBMCs from three seronegative-donors (D1, D2, and D3) and culture during 2 to 3 weeks for virus in vitro expansion. 3. Cell virus isolation (VC) used for DPS, virus tropism and RC. VC isolates were obtained by co-culture of HIV+ cells with a pool of total PBMCs from three seronegative-donors (D1, D2, and D3) and culture during 3 to 4 weeks for virus in vitro expansion. Colored red, orange and dark blue circles represent cells from CD8+ depleted seronegative donors D1, D2 and D3 respectively. Colored red, green and purple circles represent cells from seronegative donors D1, D2 and D3 respectively. Light blue circles represent HIV+ cells. Red stars indicate virus production.

Figure 3. Phylogenetic trees for HIV-1 gag, protease, integrase and env-V3 sequences extracted from VC, VP, and RNA using DPS.

Symbols represent unique haplotypes extracted from DPS for subjects P20 ○, P21 ▵, P22 ⋄, and P23 □, according to sample type: VC (light symbols), VP (dark symbols), and RNA (empty symbols). Numbers next to symbols indicate haplotype frequencies obtained from the total reads; only values above 20% are indicated in the figure. Node numbers indicate bootstrap values over 75%. (A) gag maximum-likelihood phylogenetic tree based on the TrN model. (B) Integrase maximum-likelihood phylogenetic tree based on the HKY model. (C) Protease maximum-likelihood phylogenetic tree based on the HKY+ G (α = 0.565) model. (D) env-V3 maximum-likelihood phylogenetic tree based on the TrN model.

Figure 4. Replicative capacity of VC and VP HIV-1 isolates in primary cells.

PBMCs stimulated from seronegative donors were infected in triplicate with each viral variant, and virus growth was monitored by p24 production over one week. Slopes were compared for VC and VP pairs, and p<0.001 was considered significant. Squares represent mean values and bars represent the standard error of the mean. Light squares correspond to VC and dark squares correspond to VP. (A) P20, (B) P21, (C) P22, and (D) P23.