Viraemia suppressed in HIV-1-infected humans by broadly neutralizing antibody 3BNC117

Abstract

HIV-1 immunotherapy with a combination of first generation monoclonal antibodies was largely ineffective in pre-clinical and clinical settings and was therefore abandoned. However, recently developed single-cell-based antibody cloning methods have uncovered a new generation of far more potent broadly neutralizing antibodies to HIV-1 (refs 4, 5). These antibodies can prevent infection and suppress viraemia in humanized mice and nonhuman primates, but their potential for human HIV-1 immunotherapy has not been evaluated. Here we report the results of a first-in-man dose escalation phase 1 clinical trial of 3BNC117, a potent human CD4 binding site antibody, in uninfected and HIV-1-infected individuals. 3BNC117 infusion was well tolerated and demonstrated favourable pharmacokinetics. A single 30 mg kg(-1) infusion of 3BNC117 reduced the viral load in HIV-1-infected individuals by 0.8-2.5 log10 and viraemia remained significantly reduced for 28 days. Emergence of resistant viral strains was variable, with some individuals remaining sensitive to 3BNC117 for a period of 28 days. We conclude that, as a single agent, 3BNC117 is safe and effective in reducing HIV-1 viraemia, and that immunotherapy should be explored as a new modality for HIV-1 prevention, therapy and cure.

Extended Data Figure 1

HIV-1 neutralizing activity of 3BNC117. a. Summary of 3BNC117 neutralizing in vitro activity based on 237 HIV-1 isolates comprising 6 different clades. Data were retrieved from the ‘AntibodyDatabase’ by Anthony West (West et al., PNAS, 2013). b. Illustration of the fraction (i.e. % coverage; y-axis) of HIV-1 isolates that are neutralized at a given IC50 (μg/ml; x-axis) using the same data set.

Extended Data Figure 2

CD4+ and CD8+ T cell counts before and after 3BNC117 infusion. a. Absolute numbers (cells/μl) of CD4+ and CD8+ T cell counts of all enrolled HIV-1-infected participants at screen, on the day of 3BNC117 infusion (day 0), and at day 28 after infusion. b. Percentage of CD4+ and CD8+ T cells for the same subjects and time points. Mean and standard deviation are indicated in red and black, respectively. No significant differences between pre-infusion and post infusion levels were detected by using One-way ANOVA.

Extended Data Figure 3

3BNC117 serum concentration and activity in single subjects. Serum levels of 3BNC117 in all uninfected (a) and HIV-1-infected (b) individuals that received 1, 3, 10, or 30 mg/kg 3BNC117 at day 0. Antibody levels were measured by a sandwich ELISA using an anti-3BNC117 specific antibody (green) or by measuring the 3BNC117 serum activity in a TZM.bl neutralization assay (blue).

Extended Data Figure 4

3BNC117 sensitivity and changes in viremia in 2 ART-treated subjects 3BNC117 sensitivity and changes in viremia of subjects 2B2 and 2C2. Both subjects were on ART when enrolled in the study and received a single dose (2B2, 3 mg/kg; 2C2, 10 mg/kg) of 3BNC117 at day 0. The left y-axis shows log₁₀ change in viremia from baseline, and right y-axis shows antibody level measured in ELISA. Blue line reflects change in VL and dotted gray line antibody level. Numbers indicate IC50s for 3BNC117 of autologous viral isolates measured by TZM.bl assay, color-coded as indicated on the right.

Extended Data Figure 5

Correlating viral decay with 3BNC117 sensitivity and starting viral load. a. Maximum decline in viral load in ART-untreated HIV-1-infected participants with baseline 3BNC117-sensitive viruses (IC50 < 1 μg/ml) versus pretreatment (day 0) viral load (Pearson coefficient r = 0.72 p = 0.03; Spearman coefficient rho = 0.78, p = 0.02). b. Maximum drop in viral load in HIV-1-infected and 3BNC117-sensitive individuals receiving a 10 or 30 mg/kg dose of 3BNC117 (y-axis) versus baseline autologous virus sensitivity to 3BNC117 (x-axis; Pearson coefficient r = 0.69 p = 0.03; Spearman coefficient rho = 0.41, p = 0.23).

Figure 1. Pharmacokinetics of 3BNC117 in healthy and HIV-1-infected individuals

a. Diagrammatic representation of the study. Time of 3BNC117 infusion indicated by the red arrow, and sampling for 3BNC117 serum levels, HIV-1 viral load, CD4⁺/CD8⁺ counts and env sequencing indicated below. b. Antibody decay measured in TZM.bl assays (solid lines) and ELISA (dotted lines). Mean values and SEM for uninfected individuals (3 per group) are shown in blue and for HIV-1-infected individuals (2–5 per group) in red. Light gray indicates lower level of accuracy by the ELISA assay and dark gray by the TZM.bl assay. Open circles indicate levels lower than the accuracy threshold.

Figure 2. HIV-1 viral load measurements

3BNC117 dose indicated in red. Plots (left-hand column) show absolute VLs in HIV-1 RNA copies/ml (y-axis) vs. time in days after infusion (x-axis; left panel). Log10 changes (right-hand column) in VL from day 0. Red line illustrates the average (LS-mean, by mixed-effect linear model). Individual subjects are indicated on the right side. Subjects 2E1-5 were pre-screened for 3BNC117-sensitivity. At 30 mg/kg dose level, the change in viremia was significant (p= 0.004, <0.001, <0.001, <0.001, 0.011 at days 4, 7, 14, 21, and 28, respectively) when compared to all available pretreatment values (Extended Data Table 2b).

Figure 3. 3BNC117 sensitivity, changes in viremia, and 3BNC117 levels

3BNC117 dose is indicated on the top of the graphs. The left y-axis shows log10 change in viremia from baseline, and right y-axis shows antibody level measured in ELISA. Blue line reflects change in VL and dotted gray line antibody level. Numbers indicate IC₅₀s for 3BNC117 of autologous viral isolates measured by TZM.bl assay, color-coded as indicated on the top right.

Figure 4. HIV-1 envelope sequence analysis after 3BNC117 infusion

a. HIV-1 envelopes were cloned from plasma samples. Logogram showing Env gp120 regions (amino acid positions; 270–285, 360–371, and 455 to 471–485, according to HXBc2 numbering) indicating sequence changes from day 0 (blue bar) to day 28 (red bar). The frequency of each amino acid is indicated by its height. Red residues represent mutations that were only found after treatment, black residues represent amino acids that changed in frequency after treatment, empty boxes represent gaps, and + symbols represent 3BNC117 contact sites on gp120. b. Phylogenetic trees show gp120 evolution from day 0 to 28 after treatment for 2C1, 2C5, 2D1, 2E1 and 2E2. Blue and red bars represent sequences obtained on days 0 and 28 (Supplementary Fig. 1). Illustrated values represent the IC₅₀ of 3BNC117 in μg/ml against the cloned HIV-1 pseudoviruses from the analyzed sequences (Extended Data Table 3). The geometric means of the pseudoviruses’ IC₅₀s on days 0 and 28 respectively are for 2C1: 0.06 and 0.14 μg/ml; 2C5: 0.02 and 7.09 μg/ml; 2D1: 0.15 and 0.52; 2E1: 0.09 and 0.23 μg/ml; 2E2: 0.01 and 0.03 μg/ml.