Glycan dependent phenotype differences of HIV-1 generated from macrophage versus CD4+ T helper cell populations

Abstract

Human immunodeficiency virus type 1 (HIV-1) is able to infect a variety of cell types with differences in entry efficiency and replication kinetics determined by the host cell type or the viral phenotype. The phenotype of the virus produced from these various cell types, including infectivity, co-receptor usage and neutralisation sensitivity, may also be affected by the characteristics of the producing cell. This can be due to incorporation of variant cell-specific molecules or differences in post-translational modifications of the gp41/120 envelope. In this study we produced genetically identical virus strains from macrophages, CD4-enriched lymphocytes as well as Th1 and Th2 CD4⁺ cell lines and compared each different virus stock for their infectivity in various cell types and sensitivity to neutralisation. In order to study the effect of the producer host cell on the virus phenotype, virus stocks were normalised on infectivity and were sequenced to confirm env gene homogeneity. Virus production by Th1 or Th2 cells did not compromise infectivity of the variant cell types tested. We observed no difference in sensitivity to co-receptor blocking agents upon viral passage through Th1 and Th2 CD4⁺ cell lineages nor did this affect DC-SIGN-mediated viral capture as measured in a transfer assay to CD4⁺ lymphocytes. Virus produced by macrophages was comparably sensitive to CC-chemokine inhibition as was virus generated from the array of CD4⁺ lymphocytes. We identified that virus produced from macrophages was fourteen times more resistant to 2G12 neutralisation than virus produced from CD4⁺ lymphocytes. Macrophage-produced dual-tropic (R5/X4) virus was six times more efficiently transmitted to CD4⁺ cells than lymphocyte-derived HIV-1 (p<0.0001) after DCSIGN capture. These results provide further insights to what extent the host cell influences viral phenotype and thereby various aspects of HIV-1 pathogenesis but suggest that viruses generated from Th1 versus Th2 cells are consistent in phenotype.

Keywords: HIV-1; T helper; glycan; macrophages; phenotypes.

Figure 1

TZM-bl cell infections. TZM-bl cell infections using 1,000 TCID₅₀ virus clones were measured by luciferase activity, depicted in log scale on the y-axis. (A) Th1-produced (white) and Th2-produced (red) virus (n=5). (B) Macrophage-produced (mφ) (white) and lymphocyte-produced (CD4) (blue) virus (n=5). Virus clones were produced in five replicates for each producing cell type and were used to infect TZM-bl cells in triplicate. For each virus isolate replicate, the median value of a triplicate infection is shown as a single bar on the graph with error bars representing the range. ***, P<0.001; ns, not significant.

Figure 2

Infection of T helper cells with Th1- and Th2-produced HIV-1. SF162 and H671-B10 virus clones were produced in four or five replicates from Th1 and Th2 cells and used to infect either Th1 or Th2 cells in duplicate. Each line represents a virus clone replicate that established a productive infection, with Th1-produced virus presented in black and Th2-produced virus in red. The percentage of isolates resulting in productive infection is displayed in brackets. Infections were performed with three different TCID₅₀ values; 500, 100 and 20, of SF162 (A, CCR5 using) and H671-B10 (B, dual-tropic). CA-p24 production is depicted on the y-axis in logarithmic scale over the course of infection.

Figure 3

Determination of co-receptor affinity. Affinity for both the CCR5 and CXCR4 co-receptor was determined by HIV-1 infection of CD4-enriched lymphocytes in addition of 3-fold dilutions of RANTES and AMD3100, respectively, up to fully blocking concentrations. Logarithmic values are presented on the x-axis, while the y-axis depicts the percentage of inhibition. In a separate graph, we compared IC₅₀ values between both viral stocks using the Wilcoxon signed rank test. Virus clones were produced in each cell type in 3, 4 or 5 replicates and used to infect each cell type in inhibition assays in quadruplicate. Each virus clone replicate is plotted as a single line and the IC₅₀ value derived from these inhibition curves are plotted as a single point for each virus isolate replicate. (A) RANTES inhibition of Th1- (black) and Th2-produced (red) virus stocks (n=3) of NSI-18. (B, C) AMD3100 inhibition of Th1- (black) and Th2-produced (red) viral stocks of H671-B10 (dual-tropic) (n=4) and LAI (CXCR4 using) (n=5). (D) RANTES inhibition of macrophage (mφ)- (black) and lymphocyte-produced (blue) viral stocks of NSI-18 (CCR5 using) (n=5). *, P<0.05; ns, not significant.

Figure 4

DC-SIGN-mediated transmission to CD4-enriched lymphocytes. NSI-18, H671-B10 and (A–C) Transmission of Th1- (white) versus Th2-produced (red) NSI-18 (CCR5 using) (n=4), H671-B10 (dual-tropic) (n=4) and LAI (CXCR4 using) (n=5). Three to five clones were produced from each cell type and infection experiments were performed in triplicate. The bars represent median values of HIV-infected lymphocytes for each clone. A separate graph depicts the values of all clones from each cell type and we used the Wilcoxon signed rank test to determine statistical significance on transmission of Th1- and Th2-produced variants. (D–F) Transmission of macrophage (mφ)- (white) versus lymphocyte-derived (blue) NSI-18 (CCR5 using) (n=5), SF162 (CCR5 using) (n=5) and H671-B10 (dual-tropic) (n=4). Transmissions with NSI-18 and H671-B10 were repeated once. The bars represent median values of HIV-infected lymphocytes for each clone. A separate graph depicts the values of all clones from each cell type and we used the Wilcoxon signed rank test to determine statistical significance on transmission of mφ - and lymphocyte-produced variants. (G) Sensitivity of macrophage (mφ)- (white) and lymphocyte-derived (blue) SF162 HIV-1 to the carbohydrate dependent 2G12 antibody neutralisation was determined by infecting CD4-enriched lymphocytes with virus, which was neutralized with 3-fold increasing concentrations of antibody. Inhibition curves were constructed based on CA-p24 values from the peak of viral replication. The experiment was conducted twice with one representative profile shown. *, P<0.05; ****, P<0.0001; ns, not significant.