HIV therapy by a combination of broadly neutralizing antibodies in humanized mice

Abstract

Human antibodies to human immunodeficiency virus-1 (HIV-1) can neutralize a broad range of viral isolates in vitro and protect non-human primates against infection. Previous work showed that antibodies exert selective pressure on the virus but escape variants emerge within a short period of time. However, these experiments were performed before the recent discovery of more potent anti-HIV-1 antibodies and their improvement by structure-based design. Here we re-examine passive antibody transfer as a therapeutic modality in HIV-1-infected humanized mice. Although HIV-1 can escape from antibody monotherapy, combinations of broadly neutralizing antibodies can effectively control HIV-1 infection and suppress viral load to levels below detection. Moreover, in contrast to antiretroviral therapy, the longer half-life of antibodies led to control of viraemia for an average of 60 days after cessation of therapy. Thus, combinations of potent monoclonal antibodies can effectively control HIV-1 replication in humanized mice, and should be re-examined as a therapeutic modality in HIV-1-infected individuals.

Figure 1. Monotherapy using broadly neutralizing antibodies in HIV-1_YU2-infected hu-mice

a, Left panel shows viral loads (RNA copies/ml, y-axis) measured over time (days, x-axis) in untreated HIV-1_YU2-infected hu-mice (control group). Each line represents a single mouse and symbols reflect viral load measurements. Symbol characters correspond to individual mice as indicated (right). Hu-mice were infected with HIV-1_YU2 (i.p.) between day −22 and −16 (orange square) and baseline viral loads were measured between day −4 and −2 (Supplementary Table 1). Dotted line represents limit of detection for viral load determination (800 copies/ml). Right panel shows changes in log₁₀ (RNA copies/ml) from baseline (grey line) at day 0 with green line representing the average in viral load changes. b, Illustration of HIV-1 viral loads as in (a) but with single antibody treatment (shaded in grey) starting at day 0. Red line represents the average in viral load changes superimposed with averages of the control group (green line, a). c, Changes in log₁₀ (RNA copies/ml) 6-7 days and 27-32 days after treatment initiation. Columns and error bars represent mean and standard deviation (SD), respectively. Significant statistical differences among groups were determined by performing a Kruskal-Wallis-test with Dunn-multiple comparison post-hoc test using GraphPad Prism version 5.0b for Mac OS X, GraphPad Software, San Diego California, USA. Significant differences among groups were detected at day 6-7 but not at the later time point (27-32 days). Asterisks (*, p≤0.05; **, p≤0.01) indicate a significant difference compared to the control group. d, Sequence analysis of HIV-1 gp120 after viral rebound while on therapy with single bNAbs (Supplementary Fig. 7). Pie charts show the dsitribution of amino acid changes at the antibodies’ respective target sites. A selection of amino acid substitutions was tested in vitro and confirmed antibody escape (Supplementary Fig. 8 and Table 2b). Mutations are relative to HIV-1_YU2 and numbered according to HXBc2. Total number of analyzed mice/gp120 sequences is indicated in the center of the charts (Supplementary Fig. 7, Supplementary Table 2b).

Figure 2. HIV therapy by a combination of three (tri-mix) or five (penta-mix) broadly neutralizing antibodies in HIV-1_YU2-infected hu-mice

a, Left panel shows viral loads (RNA copies/ml, y-axis) over time (days, x-axis) in HIV-1_YU2-infected hu-mice treated with a combination of 3BC176, PG16 and 45-46^G54W (tri-mix; grey shading). Each line represents a single mouse and symbols indicate viral load measurements (Supplementary Table 1). Infection and viral load determination was performed as in Fig. 1. Right panel shows changes in log₁₀ (RNA copies/ml) from baseline at day 0. Red and green lines represent the average values in viral load change of tri-mix and control group (Fig. 1a), respectively. b, Individual gp120 envelope sequences cloned from single mice (y-axis) during tri-mix therapy after viral rebound. gp120 sequences are represented by horizontal gray bars with silent mutations indicated in green and amino acid replacements in red. Black asterisks represent mutations generating a stop codon and bold gray bars deletions. All mutations are relative to HIV-1_YU2 and numbered according to HXBc2. Mutations at sites highlighted in blue can confer resistance to PG16- or 45-46^G54W-mediated neutralization in vitro (Supplementary Fig. 8, Supplementary Table 3a). c, Pie charts as in Fig. 1d illustrating distribution of amino acid changes in gp120 (b) at PG16 (left) or 45-46^G54W (right) respective target sites (Supplementary Table 3a). d, As in a, for HIV-1_YU2-infected hu-mice treated with a combination of 3BC176, PG16, 45-46^G54W, PGT128 and 10-1074 (penta-mix; grey shading).

Figure 3. Viral rebound in HIV-1_YU2-infected hu-mice after cessation of antibody therapy

Viral load in RNA copies/ml (blue, left y-axis) and antibody concentration reactive to YU2 gp120 in μg/ml (orange, right y-axis) over time (x-axis) after the last antibody injection (day 0). Green dotted line indicates the viral load average of the control group (Fig. 1a). a, Viral load and YU2 gp120–reactive antibody concentration after stopping tri-mix therapy in mice that effectively controlled viremia below limit of detection. b, gp120 sequences illustrated as in Fig. 2b for viruses obtained after viral rebound from mice previously treated with tri-mix therapy. Vertical blue bars highlight sites in which mutations are able to confer resistance (Supplementary Fig. 2b, 3a). c, Viral load and YU2 gp120–reactive antibody concentration after stopping penta-mix therapy (Supplementary Fig. 3b). d, gp120 sequences for viruses obtained after rebound from mice previously treated with penta-mix therapy. e, Viral load of four mice re-treated with penta-mix therapy after viral rebound (c). Treatment period is highlighted in grey. Mouse numbers correspond to the mice in c.