Combination therapy with anti-HIV-1 antibodies maintains viral suppression

Abstract

Individuals infected with HIV-1 require lifelong antiretroviral therapy, because interruption of treatment leads to rapid rebound viraemia. Here we report on a phase 1b clinical trial in which a combination of 3BNC117 and 10-1074, two potent monoclonal anti-HIV-1 broadly neutralizing antibodies that target independent sites on the HIV-1 envelope spike, was administered during analytical treatment interruption. Participants received three infusions of 30 mg kg-1 of each antibody at 0, 3 and 6 weeks. Infusions of the two antibodies were generally well-tolerated. The nine enrolled individuals with antibody-sensitive latent viral reservoirs maintained suppression for between 15 and more than 30 weeks (median of 21 weeks), and none developed viruses that were resistant to both antibodies. We conclude that the combination of the anti-HIV-1 monoclonal antibodies 3BNC117 and 10-1074 can maintain long-term suppression in the absence of antiretroviral therapy in individuals with antibody-sensitive viral reservoirs.

Figures

Extended Data Figure 1.. Study participant selection and demographics.

a, Flow diagram indicating the selection of study participants. b, Table showing individual participant demographics and baseline clinical characteristics. * EVG – elvitegravir, cobi - cobicistat, TDF - tenofovir disoproxil fumarate, FTC - emtricitabine, RPV - rilpivirine, EFV - efavirenz, TAF - tenofovir alafenamide fumarate, D, RTV – ritonavir, ABC – abacavir, 3TC - lamivudine, DTG – dolutegravir. ** NNRTI-based regimens were switched four weeks before ART interruption due to longer half-lives of NNRTIs. *** Pre-screening of bulk outgrowth virus obtained from PBMC culture by TZM-bl assay. # All participants harbored clade B viruses. Viral load < 20 D: Plasma HIV-1 RNA detected but not quantifiable by clinical assay. Dx, diagnosis; Scr, screening; Wk −2, week −2; d0, day 0. Grey shaded rows indicate participants who were found to have detectable viremia (HIV-1 VL > 20 copies/ml) at week −2 or day 0. These participants were not included in the efficacy analyses given lack of viral suppression at baseline.

Extended Data Figure 2.. Demographics, CD4+ T cells during study period in participants and pharamacokinetics of 3BNC117 and 10–1074.

a, Table showing baseline participant demographics. *NNRTI - Non-nucleoside reverse transcriptase inhibitor. b, Absolute CD4+ T cell counts and percentage of CD4+ T cells among CD3+ T cells at screening (n=15), day 0 (n=15), at the time of viral rebound (n=13) and at the end of the study are shown (n=15) (see also Supplementary Table 2). The last available time point after resuppression was used as end of the study time point for the participants that reinitiated ART. Red lines indicate mean and error bars indicate standard deviation. P values were obtained using a two-tailed paired t-test comparing CD4+ T cell counts between day 0 and the time of viral rebound. c,d, 3BNC117 (red) and 10–1074 (blue) levels in serum (n=15) as determined by TZM-bl assay (c) and ELISA (d). Curves indicate mean serum antibody concentrations and error bars represent standard deviation. Red and blue triangles indicate 3BNC117 and 10–1074 infusions, respectively. In the TZM-bl assay, lower limits of detection were 0.46 µg/ml and 0.01 µg/ml for 3BNC117 and 10–1074, respectively (c). In the ELISA, lower limits of detection were 0.78 µg/ml and 0.41 µg/ml, respectively (d). In cases where participants only received 2 infusions due to early viral rebound (9245, 9249 and 9253), only antibody concentrations up to the second infusion were included. Half-life of each bNAb is indicated in days. e,f, Half-lives of both antibodies as measured by TZM-bl assay (e) and ELISA (f). Each dot represents a single participant. The half-lives of both antibodies from the 15 participants enrolled in the study are represented. Black lines indicate the mean value and standard deviation (n=15). P values were obtained using a two-tailed unpaired t-test comparing the two antibodies.

Extended Data Figure 3.. Phylogenetic tree of viruses from all enrolled participants.

Maximum likelihood phylogenetic trees of full-length env sequences containing all the sequences obtained from Q2VOA cultures and rebound viruses from SGA or rebound outgrowth of the 15 participants enrolled in the study. Participants are indicated by individual colors.

Extended Data Figure 4.. Viral rebound, amino acid variants at 10–1074 contact sites and sensitivities of latent and rebound viruses in the participants with detectable viremia >20 copies/ml 2 weeks prior to or at the start of ATI.

a, Plasma HIV-1 RNA levels (black; left y-axis) and bNAb serum concentrations (3BNC117, red; 10–1074, blue; right y-axis). Red and blue triangles indicate 3BNC117 and 10–1074 infusions, respectively. Serum antibody concentrations were determined by TZM-bl assay. Grey shaded areas indicate time on ART. Lower limit of detection of HIV-1 RNA was 20 copies/ml. b, Kaplan-Meier plots summarizing time to viral rebound. Y-axis indicates percentage of participants that maintain viral suppression. X-axis indicates weeks after start of ATI. Participants receiving the combination of 3BNC117 + 10–1074 are indicated by the blue line (n=4). Dotted red line indicates a cohort of individuals receiving 3BNC117 alone during ATI (n=13) and dotted black line indicates a cohort of participants who underwent ATI without any intervention (n=52). c, Color charts show Env contact sites of 10–1074 at the G(D/N)IR motif (positions 324–327, according to HXB2 numbering) and the glycan at the potential N-linked glycosylation site at position 332 (NxS/T motif at positions 332–334). LR indicates latent reservoir viruses isolated by Q2VOA (week −2) and RB indicates rebound viruses isolated by SGA (plasma) or viral outgrowth (PBMCs). 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 indicate the lack of variation between latent reservoir virus and rebound virus at the indicated position. Full-color rectangles represent amino acid residues with changes in distribution between reservoir and rebound viruses. d, Dot plots indicating IC80 (µg/ml) of 3BNC117 (left) and 10–1074 (right) against latent and rebound viruses determined by TZM-bl neutralization assay. Q2VOA-derived latent viruses from week −2 are shown as black circles. For outgrowth culture-derived rebound viruses, the highest IC80 determined is shown as red circle. For 9250 and 9253, no viruses could be obtained from rebound outgrowth cultures and pseudoviruses were made from env sequences of the latent reservoir (Q2VOA) and rebound viruses (plasma SGA). Note that 9249 and 9253 had pre-existing resistant viruses in the reservoir (IC50 > 2 μg/ml). 9248 and 9250 had pre-existing viruses that failed to reach an IC100 when tested up to 50 µg/ml for 3BNC117 (Extended Data Fig. 5). Rebound viruses of all 4 participants showed IC80 or IC100 of > 50 μg/ml for both 3BN117 and 10–1074.

Extended Data Figure 5.. Phylogenetic env trees and TZM-bl neutralization curves for individuals with viral blips.

a, Circulating reservoir and viral rebound in study participants with detectable viremia at week −2 or day 0. Maximum likelihood phylogenetic trees of full-length env sequences of viruses isolated from week −2 Q2VOA cultures, rebound plasma SGA and rebound outgrowth from the 4 participants with viral blips. Open black rectangles indicate Q2VOA-derived viruses from week −2. Viruses obtained at the time of rebound are indicated by red rectangles (plasma SGA) and red stars (rebound PBMC outgrowth cultures), respectively. Asterisks indicate nodes with significant bootstrap values (bootstrap support ≥ 70%). Clones are denoted by colored lines. Boxes indicate IC80s (µg/ml) of 3BNC117 and 10–1074 against individual clones, with asterisks indicating IC100 values of > 50 μg/ml. b, Latent reservoir virus TZM-bl neutralization curves for two participants that had a viral load > 20 copies/ml at day 0 (9248 and 9250). Curves show neutralization titers by 3BNC117 (blue), 10–1074 (red) and other bNAbs, when available, for week −2 Q2VOA-derived viruses present in the circulating reservoir. Three representative viruses from 9248 (top) and 9250 (bottom) are shown, respectively. Note that, although these viruses had low 3BNC117 and 10–1074 IC50 or IC80 titers, IC100 (black dotted line) is reached only at a high concentration or not reached at all. The neutralization titer was measured by TZM-bl neutralization assay using a 5-parameter curve fit method.

Extended Data Figure 6.

Color charts show 3BNC117 contact sites in Env according to HXB2 numbering. Diagram shows the 13 participants that experienced viral rebound before week 30. LR indicates latent reservoir viruses isolated by Q2VOA (on weeks −2 and 12 when available). RB indicates rebound viruses isolated by SGA (plasma) and viral outgrowth (PBMCs). 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 indicate the lack of variation and full-color rectangles represent amino acid residues with changes in the distribution between the reservoir and rebound.

Extended Data Figure 7.. Comparison of the circulating latent reservoir and rebound viruses.

Maximum likelihood phylogenetic trees of full-length env sequences of viruses isolated from Q2VOA, rebound plasma SGA, and rebound PBMC outgrowth cultures from participants 9241, 9244, 9246 and 9247, that rebounded before week 30. Open and closed black rectangles indicate Q2VOA-derived viruses from week −2 and week 12, respectively. Viruses obtained at the time of rebound are indicated by red rectangles (plasma SGA) and red stars (rebound PBMC outgrowth cultures). Asterisks indicate nodes with significant bootstrap values (bootstrap support ≥ 70%). Clones are denoted by colored lines mirroring the colors of slices in Figure 5. Boxes indicate IC80s (μg/ml) of 3BNC117 and 10–1074 against representative viruses throughout the phylogenetic tree and clones, when possible (Supplementary Table 4). Asterisks in boxes indicate IC100 values of > 50 μg/ml.

Extended Data Figure 8.. Recombination events in rebound viruses.

a, Maximum likelihood phylogenetic trees of full-length env sequences of viruses isolated from Q2VOA cultures and rebound SGA in the 4 participants where rebound viruses showed recombination events. Open and closed black rectangles indicates Q2VOA-derived viruses from week −2 and week 12, respectively. Rebound plasma SGA- or outgrowth-derived viruses are indicated by closed red rectangles. Green stars represent parent sequences that underwent recombination to produce the child sequences (red stars). b, Circos plots indicating the relationship between the parent sequences and the recombinants. Open and closed black rectangles indicate Q2VOA-derived sequences from week −2 and week 12, respectively. Rebound virus sequences are indicated by red rectangles. The thickness of the black outer bars represents the number of sequences obtained from that particular clone.

Extended Data Figure 9.. Phylogenetic trees of participants 9245, 9251, 9254 and 9255.

Maximum likelihood phylogenetic trees of full-length env sequences of viruses isolated from Q2VOA cultures and rebound plasma SGA and rebound outgrowth from the 2 participants with pre-existing resistance to one of the 2 antibodies (9245 and 9251) and the 2 sensitive participants (9254 and 9255) who maintained viral suppression for > 30 weeks (end of the study). Open and closed black rectangles indicate Q2VOA-derived viruses from week −2 and week 12, respectively. Rebound plasma SGA viruses are indicated by closed red rectangles. Asterisks indicate nodes with significant bootstrap values (bootstrap support ≥ 70%). Clones are denoted by colored lines beside the phylogenetic tree. Numbers correspond to 3BNC117 and 10–1074 IC80 neutralization titers.

Extended Data Figure 10.. Clonal distribution of the circulating latent reservoir and IUPM changes.

a, Pie charts depicting the distribution of Q2VOA-derived env sequences obtained at weeks −2 (W-2) and week 12 (W12). Number in the inner circle indicates the total number of env sequences analyzed. White represents sequences isolated only once across both time points and colored slices represent identical sequences that appear more than once (clones). The size of each pie slice is proportional to the size of the clone. Red arrows denote clones that significantly change in size (P ≤ 0.05 (two-sided Fisher’s exact test)) between the two time points. b, Table indicating the summary of clonal env sequences and IUPM in the 9 individuals with an antibody-sensitive reservoir. c, IUPM versus time of viral rebound in the antibody-sensitive individuals (n=7) that rebounded within the study observation period (30 weeks). P values were obtained using a two-tailed Pearson correlation test comparing the two variables.

Figure 1.. Delayed viral rebound with 3BNC117 and 10–1074 combination therapy during ATI.

a, Study design. Red and blue triangles represent 3BNC117 and 10–1074 infusions, respectively. b, Plasma HIV-1 RNA levels (black; left y-axis) and bNAb serum concentrations (3BNC117, red; 10–1074, blue; right y-axis) in the 9 bNAb-sensitive participants (left) and the 2 participants with pre-existing resistance against one of the antibodies (right). Red and blue triangles indicate 3BNC117 and 10–1074 infusions, respectively. Serum antibody concentrations were determined by TZM-bl assay. Grey shaded areas indicate time on ART. Lower limit of detection of HIV-1 RNA was 20 copies/ml. c, Kaplan-Meier plots summarizing time to viral rebound for the participants with HIV-1 RNA < 20 copies/ml 2 weeks before and at the start of ATI (n=11, left), for the participants sensitive to both antibodies (n=9, center), and for the participants that showed pre-existing resistance to one of the antibodies (n=2, right). Y-axis indicates percentage of participants that maintain viral suppression. X-axis indicates weeks after start of ATI. Participants receiving the combination of 3BNC117 + 10–1074 are indicated by the blue line. Dotted red lines indicate a cohort of individuals receiving 3BNC117 alone during ATI (n=13) and dotted black lines indicate a cohort of participants who underwent ATI without intervention (n=52).

Figure 2.. Amino acid variants at 10–1074 contact sites and bNAb sensitivity of reactivated latent and rebound viruses.

a, Color charts show Env contact sites of 10–1074 at the G(D/N)IR motif (positions 324–327, according to HXB2 numbering) and the glycan at the potential N-linked glycosylation site at position 332 (NxS/T motif at positions 332–334). Diagram shows the 7 bNAb-sensitive participants that rebounded before week 30 (left) and the 2 individuals with pre-existing resistance to one of the 2 antibodies (right). LR indicates latent reservoir viruses isolated by Q2VOA and RB indicates rebound viruses isolated by SGA (plasma) or viral outgrowth (PBMCs). 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 indicate the lack of variation between latent reservoir virus and rebound virus at the indicated position. Full-color rectangles represent amino acid residues with changes in distribution between reservoir and rebound viruses. b,c, Dot plots indicating IC80 (µg/ml) of 3BNC117 (b, left panel) and 10–1074 (c, right panel) against latent and rebound viruses determined by TZM-bl neutralization assay. Q2VOA-derived latent viruses from week −2 and week 12 are shown as black and grey circles, respectively. For outgrowth culture-derived rebound viruses, the highest IC80 determined is shown as red circle. For 9246, 9252, 9245 and 9251 viruses could not be obtained from rebound outgrowth cultures and pseudoviruses were made from env sequences from Q2VOA and plasma SGA.

Figure 3.. Comparison of the circulating latent reservoir and rebound viruses.

Maximum likelihood phylogenetic trees of full-length env sequences of viruses isolated from Q2VOA, rebound plasma SGA, and rebound PBMC outgrowth cultures from 3 out of 7 participants (9242, 9243 and 9252) that rebounded before week 30 (9241, 9244, 9247 and 9246 are depicted in Extended Data Fig. 7). Open and closed black rectangles indicate Q2VOA-derived viruses from week −2 and week 12, respectively. Viruses obtained at the time of rebound are indicated by red rectangles (plasma SGA) and red stars (rebound PBMC outgrowth cultures). Asterisks indicate nodes with significant bootstrap values (bootstrap support ≥ 70%). Clones are denoted by colored lines mirroring the colors of slices in Extended Data Fig. 10a. Boxes indicate IC80s (µg/ml) of 3BNC117 and 10–1074 against representative viruses throughout the phylogenetic tree and clones, when possible (Supplementary Table 4). Asterisks in boxes indicate IC100 values of > 50 μg/ml.

Figure 4.. Distribution of the circulating latent reservoir and rebound viruses.

a, Venn diagrams showing sequence identity between env sequences obtained from Q2VOA at week −2 (blue) and week 12 (grey), and plasma SGA or rebound PBMC outgrowth culture at the time of viral rebound (red). Area of overlap is proportional to the number of identical sequences. Number of sequences obtained is indicated. b, Infectious units per million (IUPM) CD4+ T cells at weeks −2 and 12 as determined by Q2VOA. Participants with IUPMs higher and lower than 0.1 are shown on the top and bottom, respectively. Participant 9254 is not shown due to lack of sample availability. The 2 time points were not statistically different (P = 0.078 (paired t-test)).