Main

It is now recognized that, unlike most other prophylactic vaccines for human viral pathogens, an effective vaccine against HIV-1 will probably need to completely block the establishment of a productive infection within a very short time frame (1–3 days of transmission). Such protection has, in fact, been achieved by administering polyclonal and monoclonal anti-HIV-1-neutralizing antibodies (NAbs) to humanized mice or macaques before challenge with SIV/HIV chimaeric viruses (SHIVs)1,2,3,4,5,6,7,8.

During the past 7 years, monoclonal antibodies (MAbs) have been isolated from selected HIV-1 infected individuals, who generate anti-viral NAbs (bNAbs) with broad and potent activity against isolates of diverse genetic and geographical origin13. Several of these bNAbs have been used to suppress ongoing viral infections in humanized mice, macaques, and humans14,15,16,17,18. Pre-exposure immunoprophylaxis with bNAbs has also been evaluated in macaque models. In most of these experiments, a single dose of antibody, typically infused 24–48 h before a single high-dose virus challenge, was sufficient to block infection by a virus challenge, capable of establishing an infection in all untreated animals4,19,20,21. Humans, however, are usually exposed to much lower doses of virus on several occasions before becoming infected with HIV-1 (ref. 22).

It is worth noting that before the development of an effective hepatitis A virus vaccine, pre-exposure immunoprophylaxis with hepatitis A immune globulin was common practice for travellers to endemic regions of the world; protective effects lasted 3–5 months23. Prophylactic administration of antibodies against other microbial pathogens has also been used to prevent disease24. On the basis of this idea, we explored the possibility that a single administration of a potent neutralizing anti-HIV MAb, in the setting of repeated low-dose SHIV challenges, might protect for extended periods of time, thereby providing a proof of concept for periodic administration of MAb as an alternative to HIV-1 vaccination.

We initially selected three MAbs for the repeated low-dose SHIV challenge experiment on the basis of their previously described activity in blocking virus acquisition in a cohort of 60 macaques after a single high-dose SHIV challenge21. Two of these antibodies (VRC01 (ref. 12) and 3BNC117 (ref. 11)) target the gp120 CD4bs and one (10-1074 (ref. 10)) is dependent on the presence of HIV-1 gp120 N332 glycan, located immediately downstream of the V3 loop. The challenge virus selected for the present study was SHIVAD8-EO (ref. 25), an R5-tropic molecular-cloned derivative of the clade B SHIVAD8 (ref. 26), which possesses multiple properties typical of pathogenic HIV-1 isolates27.

When tested against large HIV-1 pseudovirus panels including multiple clades, 3BNC117 and VRC01 neutralize more than 80% of the viral isolates and 10-1074 neutralizes between 60% and 70%. Against sensitive viruses, 10-1074 is the most potent, followed by 3BNC117 and VRC01 (ref. 28). Consistent with this trend, the 50% inhibitory concentration (IC50) values for VRC01, 3BNC117, and 10-1074 against SHIVAD8-EO were 0.67, 0.06, and 0.08 μg ml−1, respectively, and the 80% (IC80) values were 2.04, 0.19, and 0.18 μg ml−1, respectively (Extended Data Fig. 1a). Neutralization sensitivities were also measured using the SHIV challenge stock in a single round of infection assay in TZM-bl cells, using replication-competent SHIVAD8-EO. The IC50 and IC80 values for VRC01, 3BNC117, and 10-1074 in this assay system were 2.06, 0.12, and 0.05, and 7.14, 0.32, and 0.14 μg ml−1, respectively (Extended Data Fig. 1b).

In an initial experiment designed to simulate low-dose mucosal transmission in humans, a cohort of nine monkeys was challenged weekly by the intrarectal route with ten 50% tissue culture infectious doses (TCID50) of SHIVAD8-EO, in the absence of antibody treatment. As shown in Fig. 1a, plasma viraemia became detectable after two to six challenges, with a median of 3.0 weekly virus exposures needed to infect all nine animals. On the basis of these results, the inoculum size administered to each monkey per challenge was estimated to be 0.27 50% animal infectious doses (AID50).

Figure 1: HIV MAbs delay virus acquisition after repeated low-dose intrarectal SHIVAD8-EO challenges.
figure 1

a, Plasma viral loads in macaques receiving no MAb (controls, n = 9). IR, intrarectal. b, Representation of the regimen used to assess the protective efficacy of MAbs. Macaques were intravenously (i.v.) administered the indicated MAbs at a dose of 20 mg kg−1and challenged 1 week later and every week thereafter. c–f, Plasma viral loads in macaques administered VRC01, 3BNC117, 10-1074, and VRC01-LS bNAbs, respectively.

PowerPoint slide

Full size image

The regimen used to assess the protective efficacy of the three anti-HIV-1 MAbs against a repeated low-dose rectal challenge of SHIVAD8-EO is shown in Fig. 1b. Individual MAbs (20 mg kg−1) were administered a single time intravenously to three cohorts of six animals. Starting 1 week later, each group was challenged weekly by the intrarectal route with ten TCID50 of SHIVAD8-EO. Samples of blood, collected at regular intervals, were monitored for levels of viral RNA, concentrations of MAb, and anti-SHIV-neutralizing titres. The number of virus challenges required to establish a SHIVAD8-EO infection, indicated by measurable viraemia (>100 viral RNA copies per millilitre of plasma), in the recipients of the anti-HIV-1 MAbs was compared with that needed for virus acquisition in the control group.

In all cases, the administration of MAbs delayed virus acquisition. Animals receiving VRC01 required 4–12 challenges; 3BNC117 required 7–20 challenges; and 10-1074, 6–23 challenges (Fig. 1c–e). The differences in the number of challenges required for infection, and thus the median times to virus acquisition compared with control monkeys, were 8 weeks for VRC01, 13 weeks for 3BNC117, and 12.5 weeks for 10-1074.

The pharmacokinetic profile of VRC01 was altered by introducing two amino-acid mutations (M428L and N434S, referred to as ‘LS’) into its Fc domain, which increased its half-life in both plasma and tissues9. The neutralization activity of this VRC01 derivative, designated VRC01-LS, was first tested in the TZM-bl assay and, as expected, it exhibited IC50 and IC80 values similar to VRC01 (Extended Data Fig. 1a, b). When administered to six macaques in the previously described repeated low-dose SHIVAD8-EO challenge system, the VRC01-LS-treated animals required 9–18 challenges (median = 14.5) for all of the monkeys to become infected (Fig. 1f). Thus the modified VRC01-LS Fc domain conferred an estimated 1.8-fold increase in the number of challenges, resulting in successful acquisition compared with the parental VRC01 MAb.

The protective effects of the four anti-HIV-1 MAbs are described by Kaplan–Meier analysis in which the percentage of macaques remaining uninfected is plotted against the number of SHIVAD8EO challenges (Fig. 2a). Significantly increased numbers of challenges were required to establish infections in the recipients of the VRC01, 3BNC117, 10-1074, and VRC01-LS MAbs than in the control animals (P = 0.007, 0.002, 0.002, and 0.002, respectively), using the Wilcoxon rank-sum test (Fig. 2b). A comparison of the individual pairs of Kaplan–Meier curves revealed that 10-1074, 3BNC117, and VRC01-LS were not significantly different from each other in blocking infection.

Figure 2: Kaplan–Meier analysis and magnitude of protection by HIV MAbs in repeated low-dose challenge.
figure 2

a, Kaplan–Meier survival curves for recipients of the four bNAbs and the cohort of control animals. The percentage of macaques remaining uninfected is plotted against the number of ten TCID50 SHIVAD8-EO intrarectal challenges required to establish infections. b, Statistical differences are represented as P values (Wilcoxon rank-sum test) by comparing the number of challenges resulting in infection between control animals and an individual MAb recipient group or between different MAb-treated groups.

PowerPoint slide

Full size image

Ultrasensitive nested quantitative reverse-transcription PCR (qRT–PCR) and qPCR assays for plasma viral RNA and cell-associated viral RNA and DNA29 were performed on plasma and peripheral blood mononuclear cell samples, collected from recipients of the different neutralizing MAbs, before SHIVAD8-EO breakthrough infections, as assessed by plasma viraemia measured in our standard assay. In all cases, the levels of viral RNA and DNA measured with the ultrasensitive assays were below detectable limits (Extended Data Table 1).

The plasma concentrations of the infused MAbs were measured longitudinally in individual animals beginning 1 week after infusion. (Fig. 3 and Extended Data Tables 2 and 3). The median plasma concentrations at the times of virus breakthrough for the 10-1074 and 3BNC117 recipient cohorts were 0.169 and 0.330 μg ml−1, respectively (Fig. 3e). These values are comparable to the IC80 values determined in vitro, using the TZM-bl assay with replication competent SHIVAD8-EO (Extended Data Fig. 1b). The median plasma concentrations at the times of virus acquisition for VRC01 and VRC01-LS were 10- to 20-fold higher (1.825 and 6.446 μg ml−1) and were also in the same range as the IC80 values determined in vitro (Extended Data Fig. 1b). It is worth noting that three of the six recipients of the 10-1074 MAb experienced rapid decay of plasma antibody, which fell to background levels between weeks 4 and 6 after administration (Fig. 3c). A similar pattern occurred for three of the 3BNC117 MAb and VRC01-LS recipients, although the decline of antibody in plasma was delayed in these two groups animals (Fig. 3b). This rapid clearance of plasma MAbs in the subgroups of the 10-1074 and 3BNC117 recipients tracked with the emergence of anti-antibody responses to the infused anti-HIV-1 human MAbs (Extended Data Fig. 2). For monkeys infused with the 10-1074 MAb, the median number of challenges for successful infection in the three-animal subgroup not experiencing the rapid anti-antibody induced decay was 17.0 weeks compared with 12.5 for the entire 10-1074 recipient cohort.

Figure 3: Plasma concentrations of the infused MAbs in macaques correlate with long-term protection from SHIV infection.
figure 3

a–d, Plasma antibody concentrations in macaques administered VRC01, 3BNC117, 10-1074, and VRC01-LS decay over time. e, Median plasma concentrations at the times of virus breakthrough in bNAb recipients were 0.169 (10-1074), 0.330 (3BNC117), 1.825 (VRC01), and 6.446 (VRC01-LS), respectively. Boxes represent the twenty-fifth and seventy-fifth percentiles, and the heavier line represents the median value for each group. The top and bottom horizontal bars outside the boxes represent the maximum and minimum of the data, respectively.

PowerPoint slide

Full size image

Probit analysis was also used to estimate the probability of infection as a function of the imputed plasma MAb concentration at the time of each challenge. The probability of infection per infection for the control monkeys was 0.27, estimated by pooling all of the SHIVAD8-EO challenges to this group of animals; this is indicated by the single open circle along the ordinate of Extended Data Fig. 3. Not unexpectedly, the curves relating antibody concentration and virus acquisition for VRC01 and VRC01-LS were superimposed on one another even though VRC01-LS had a longer half-life in vivo. In this same analysis, the curves for 10-1074 and 3BNC117, which conferred lower probabilities for infection at each plasma MAb concentration, reflected their greater neutralization potency against the challenge virus, relative to the VRC01 antibodies. At a 10-1074 MAb plasma concentration of 1 μg ml−1, the model predicts a probability of infection, for a single challenge, of 0.044, approximately sixfold less than that estimated for animals receiving no antibodies.

The plasma neutralization titre was also determined for each of the MAb recipients at multiple times after infusion (Fig. 4a). The median plasma-neutralizing titres for the four groups of macaques at the time of SHIVAD8-EO acquisition were low: <1:20 (below the level of detection) for 10-1074 and 3BNC117 recipients; 1:27 for the VRC01 group; and 1:51 for the VRC01-LS cohort (Fig. 4b). As noted earlier for plasma MAb concentrations, the levels of detectable neutralizing activity in members of each cohort inversely correlated with the emergence of anti-antibodies (compare Fig. 4a and Extended Data Fig. 2).

Figure 4: The decline of neutralizing antibody titres in plasma over time in macaques corresponds to the time of virus acquisition.
figure 4

a, Plasma IC50 titres of the indicated MAbs were determined longitudinally using the TZM-bl cell assay. IC50 values are colour coded: 21–99 as green; 100–999 as yellow; and ≥1,000 as red. b, Plasma-neutralizing titres at the time of virus acquisition for the four groups of MAb recipients. Boxes represent the twenty-fifth and seventy-fifth percentiles, and the heavier line represents the median value for each group. The top and bottom horizontal bars outside of the boxes represent the maximum and minimum of the data, respectively.

PowerPoint slide

Full size image

In conclusion a single administration of potent anti-HIV-1-neutralizing MAbs to naive macaques was protective against repeated low-dose SHIV infection for several months. The duration of protection was directly related to antibody potency and half-life. When considered in the context of a potential exposure to HIV-1 in regions of the world where the HIV-1 is endemic, the barrier to infection when antibody concentrations remain above protective levels in infused individuals could have a profound impact on virus transmission. As noted earlier, anti-antibodies directed against some of the administered MAbs emerged quite rapidly in some macaques, and diminished their prophylactic efficacy. However, this is not likely to occur in humans as reported in a recent study of VRC01 (ref. 30). On the basis of the results obtained with VRC01 and VRC01-LS, it is also anticipated that the creation and use of 3BN117 and/or 10-1074 derivatives with the LS mutation should exhibit increased durability in vivo, resulting in protection of up to 6 months against SHIVAD8-EO-infected macaques. The administration of a multivalent cocktail of these anti-viral bNAbs could augment their efficacy by increasing overall breadth and their capacity to block the transmission of resistant HIV-1 strains.

Methods

Animal experiments

Thirty-three male and female rhesus macaques (Macaca mulatta) of Indian genetic origin from 2 to 4 years of age were housed and cared for in accordance with Guide for Care and Use of Laboratory Animals Report number NIH 82-53 (Department of Health and Human Services, Bethesda, Maryland, 1985) in a biosafety level 2 National Institute of Allergy and Infectious Diseases (NIAID) facility. All animal procedures and experiments were performed according to protocols approved by the Institutional Animal Care and Use Committee of NIAID, NIH. Animals were not randomized and the data collected were not blinded. Phlebotomies, euthanasia, and sample collection were performed as previously described31. All of the macaques used in this study were negative for the major histocompatibility complex (MHC) class I Mamu-A*01, Mamu-B*08, and Mamu-B*17 alleles. No animals were excluded from the analysis.

Antibodies

The VRC01, 3BNC117, 10-1074, and VRC01-LS anti-HIV-1 monoclonal NAbs were isolated and produced as described elsewhere9,10,11,12. MAb 10-1074 was produced by transient transfection of IgH and IgL expression plasmids into the human embryonic kidney cells whereas VRC01, VRC01-LS, and 3BNC117 were produced from Chinese hamster ovary cells. All of the MAbs were IgG1. All of the monoclonal antibodies were purified by chromatography and sterile filtration and were endotoxin free. A single dose (20 mg kg−1) of each MAb was administered intravenously to individual animals in four cohorts of monkeys.

Virus challenge

The origin and preparation of the tissue-culture-derived SHIVAD8-EO stock have been previously described25. One week after MAb infusion, animals were challenged intrarectally with ten TCID50 of SHIVAD8-EO, and every week thereafter, until a virus infection was established. A paediatric nasal speculum was used to gently open the rectum and a 1 ml suspension of virus was slowly infused into rectal cavity using a plastic tuberculin syringe. An intrarectal challenge SHIVAD8-EO inoculum size of ten TCID50 was chosen for repeated low-dose experiments on the basis of previous results indicating that (1) 1,000 TCID50 of SHIVAD8-EO administered by the intrarectal route resulted in the establishment of infections of 30 of 30 rhesus monkeys and (2) an intrarectal virus titration suggested that 1,000 TCID50 of SHIVAD8-EO was equivalent to approximately ten AID50 (ref. 21).

Quantification of viral nucleic acids

Viral RNA levels in plasma were determined by qRT–PCR (ABI Prism 7900HT sequence detection system; Applied Biosystems) as previously described31. Ultrasensitive measurement of plasma SIV gag RNA was performed as described, and cell-associated levels of SIV RNA and DNA were determined by a nested, hybrid real-time/digital PCR assay, essentially as reported previously29.

Antibody concentrations in plasma

Plasma antibody levels were quantified by ELISA using purified MAbs as a standard and anti-antibody responses in plasma were also evaluated as reported earlier9. These assays were performed twice.

Neutralization assays

The titres of each MAb against SHIVAD8-EO was assessed by two types of in vitro neutralization assay: (1) TZM-bl entry assay with pseudotype challenge virus25,27 and (2) a single-round TZM-bl infectivity assay with replication competent challenge virus32. Antibody concentrations required to inhibit infection by 50% or 80% are reported as IC50 or IC80, respectively. TZM-bl cells were obtained through the NIH AIDS Reagent Program, Division of AIDS, NIAID, NIH, from J. C. Kappes, X. Wu and Tranzyme33.These cells were not authenticated for this study and not tested for mycoplasma contamination. The neutralization activity present in plasma samples collected from rhesus macaques was assessed by TZM-bl entry assay with pseudotype challenge virus. The IC50 titre was calculated as the plasma dilution causing 50% reduction in RLUs compared with virus controls. The neutralization assays were repeated twice.

Statistical analyses

No statistical methods were used to predetermine sample size. The experiments were not randomized. The investigators were not blinded to allocation during experiments and outcome assessment.

A Wilcoxon rank-sum test was used to compare number of challenges until infection between each MAb group and control; these comparisons were considered primary and were compared with a Bonferroni-adjusted α of 0.05/4 = 0.0125 to determine significance. Comparisons between antibodies were considered secondary and not adjusted for multiple comparisons. Finally, probit models were used to model the probability of infection at each challenge as a function of concurrent antibody concentration. Since these values were not always measured at the precise time of challenge, antibody concentrations were modelled separately for each animal over time, and these models were used to impute the concentration at the exact time of each challenge for the probit model.