Safety and antiviral activity of combination HIV-1 broadly neutralizing antibodies in viremic individuals

Abstract

Monotherapy of HIV-1 infection with single antiretroviral agents is ineffective because error-prone HIV-1 replication leads to the production of drug-resistant viral variants^1,2. Combinations of drugs can establish long-term control, however, antiretroviral therapy (ART) requires daily dosing, can cause side effects and does not eradicate the infection^3,4. Although anti-HIV-1 antibodies constitute a potential alternative to ART^5,6, treatment of viremic individuals with a single antibody also results in emergence of resistant viral variants^7-9. Moreover, combinations of first-generation anti-HIV-1 broadly neutralizing antibodies (bNAbs) had little measurable effect on the infection^10-12. Here we report on a phase 1b clinical trial ( NCT02825797 ) in which two potent bNAbs, 3BNC117¹³ and 10-1074¹⁴, were administered in combination to seven HIV-1 viremic individuals. Infusions of 30 mg kg^-1 of each of the antibodies were well-tolerated. In the four individuals with dual antibody-sensitive viruses, immunotherapy resulted in an average reduction in HIV-1 viral load of 2.05 log₁₀ copies per ml that remained significantly reduced for three months following the first of up to three infusions. In addition, none of these individuals developed resistance to both antibodies. Larger studies will be necessary to confirm the efficacy of antibody combinations in reducing HIV-1 viremia and limiting the emergence of resistant viral variants.

Figure 1.. Study design and pharmacokinetics of 3BNC117 and 10–1074 in HIV-1-infected individuals.

(a) Schematic representation of the study design. (b) Baseline demographics of study participants. (c) Serum concentrations (μg/ml) of 3BNC117 (red) and 10–1074 (blue) in viremic individuals after a single infusion (upper panels) and three infusions given every two weeks (lower panels) of 3BNC117 and 10–1074 (30 mg/kg of each antibody). bNAb concentrations were determined by TZM-bl assay (left) and ELISA (right). Lines indicate arithmetic mean concentration and standard deviation. Dotted grey lines indicate lower limits of quantitation (TZM-bl, 0.46 μg/ml and 0.1 μg/ml for 3BNC117 and 10–1074, respectively; ELISA, 0.78 μg/ml and 0.41 μg/ml for 3BNC117 and 10–1074, respectively). Grey circles indicate antibody levels below the limit of quantitation. Numbers show average half-life.

Figure 2.. Viral load following 3BNC17/10–1074 infusions in HIV-1-infected participants.

(a-c) Changes in viremia and bNAb serum concentrations in HIV-1-infected participants showing (a) late rebound, (b) early rebound or (c) no response after 3BNC117 and 10–1074 combination therapy. Upper graphs show HIV-1 RNA in copies/ml (black, left y-axis), and 3BNC117 (red) and 10–1074 (blue) serum levels (right y-axis, as determined by TZM-bl). X-axis shows weeks after the first antibody infusion. Dashed line indicates the lower limit of detection of HIV-1 RNA (20 copies/ml). Arrows indicate antibody infusions. Lower graphs show log₁₀ changes of HIV-1 RNA copies compared to day 0. Green shading depicts viral suppression compared to day 0. (d-f) Simultaneous confidence band estimation to determine time of significant suppression (red dotted lines) of HIV-1 viremia in (d) all viremic participants (n=7, a-c), (e) individuals harboring 3BNC117- and 10–1074-sensitive viruses (n=4, a), and (f) participants carrying viruses with partial or full bNAb resistance (n=3, b-c). Each dot represents a viral load measurement. Solid and dashed lines represent the regression fit and simultaneous confidence bands at 95% certainty level, respectively, and were computed using the Gaussian family for the local likelihood function using R package locfit (version 1.5–9.1).

Figure 3.. Phylogenetic sequence analysis of viremic participants.

(a) Maximum likelihood phylogenetic tree of all SGA-derived env gene sequences (n=382) obtained from plasma of viremic study participants (n=7). (b) Maximum likelihood phylogenetic trees of env sequences (n=356) obtained from plasma of single participants before antibody therapy (light grey) and at viral rebound (blue). Dark grey indicates sequences amplified at week 6 after the antibody infusion (participant 91C34). Env sequences that were used to produce pseudoviruses for neutralization testing are indicated. Black asterisks indicate nodes with significant bootstrap values (bootstrap support ≥ 70%). (c) Circos plots indicating the relationship between parent sequences and recombinants in single participants (n=6). SGA sequences are depicted by light grey (day 0), dark grey (week 6) and blue (rebound) rectangles. Grey lines indicate recombination events between different viruses. Thickness of the black outer bars represents the number of sequences obtained from that particular clone.

Figure 4.. HIV-1 escape analysis of individuals receiving 3BNC117/10–1074 therapy.

(a) Viral sensitivities (IC₈₀, μg/ml) of pseudoviruses constructed from SGA-derived env sequences obtained on day 0 (grey) and at the time of rebound (blue). Columns reflect geometric mean IC₈₀s of viruses tested against 3BNC117 (left panel) and 10–1074 (right panel). Each dot represents one viral isolate. Fully resistant viruses (IC₈₀ > 50 μg/ml) are depicted by red circles. Green shading indicates the range of IC₈₀s. (b) Frequency of amino acids in and around known 3BNC117 and 10–1074 contact residues in Env (3BNC117, aa 274–283, 364–374, and 455 to 471; 10–1074, aa 324–327 and 332–334). Amino acids are numbered according to HXB2. + indicate 3BNC117 contact sites. D0 indicates viruses isolated from plasma by SGA before antibody infusions (day 0) and R indicates rebound viruses isolated by SGA. Each amino acid is represented by a color and the frequency of each amino acid is indicated by the height of the rectangle. Shaded rectangles represent instances in which amino acids that were found in rebound viruses were also found in day 0 viruses at the indicated position. Full-color rectangles represent instances in which an amino acid was found in rebound sequences but not in day 0 sequences.