Enhanced neonatal Fc receptor function improves protection against primate SHIV infection

Abstract

To protect against human immunodeficiency virus (HIV-1) infection, broadly neutralizing antibodies (bnAbs) must be active at the portals of viral entry in the gastrointestinal or cervicovaginal tracts. The localization and persistence of antibodies at these sites is influenced by the neonatal Fc receptor (FcRn), whose role in protecting against infection in vivo has not been defined. Here, we show that a bnAb with enhanced FcRn binding has increased gut mucosal tissue localization, which improves protection against lentiviral infection in non-human primates. A bnAb directed to the CD4-binding site of the HIV-1 envelope (Env) protein (denoted VRC01) was modified by site-directed mutagenesis to increase its binding affinity for FcRn. This enhanced FcRn-binding mutant bnAb, denoted VRC01-LS, displayed increased transcytosis across human FcRn-expressing cellular monolayers in vitro while retaining FcγRIIIa binding and function, including antibody-dependent cell-mediated cytotoxicity (ADCC) activity, at levels similar to VRC01 (the wild type). VRC01-LS had a threefold longer serum half-life than VRC01 in non-human primates and persisted in the rectal mucosa even when it was no longer detectable in the serum. Notably, VRC01-LS mediated protection superior to that afforded by VRC01 against intrarectal infection with simian-human immunodeficiency virus (SHIV). These findings suggest that modification of FcRn binding provides a mechanism not only to increase serum half-life but also to enhance mucosal localization that confers immune protection. Mutations that enhance FcRn function could therefore increase the potency and durability of passive immunization strategies to prevent HIV-1 infection.

Conflict of interest statement

The authors declare no competing financial interests. Readers are welcome to comment on the online version of the paper.

Figures

Extended Data Figure 1. Enhanced FcRn-binding mutants of VRC01 are released from human FcRn at pH 7.4

VRC01 or the indicated FcRn-binding mutants were injected in PBS (pH 6.0) at a concentration of 100 nM over human-FcRn-immobilized Biacore CM5 sensor chips (~500 response units (RU)) and were dissociated using PBS at pH 6.0 followed by PBS at pH 7.4.

Extended Data Figure 2. Mutants with enhanced human FcRn binding persist at higher serum concentrations than VRC01 in the human-FcRn(276)-transgenic mouse

The indicated VRC01-derived monoclonal antibodies (2 mg kg−1) were injected intravenously into 6–8-week-old human-FcRn(276)-transgenic mice (n = 4, male and female mice evenly distributed). Serum concentrations were assessed by indirect ELISA against RSC3 over time. The percentage of the monoclonal antibodies remaining in the serum is shown compared with the percentage on day 1 (set at 100%).

Extended Data Figure 3. Binding of VRC01, VRC01-LS and other FcRn-binding mutants to FcγRIIa and FcγRIIb

Binding to the indicated human Fc receptors was evaluated by ELISA as described in Methods.

Extended Data Figure 4. Serum pharmacokinetics, rectal tissue accumulation and mucosal secretion of VRC01-derived monoclonal antibodies in cynomolgus macaques

VRC01 or VRC01-LS (10 mg kg−1) were injected intravenously into cynomolgus macaques. a, The serum levels (n = 4) were analysed over time. b, c, The amounts of monoclonal antibody per mg of total tissue protein in rectal biopsy samples (b) and per mg of total IgG in rectal secretions (c) from each monoclonal-antibody-injected cynomolgus macaque (n = 2 per group) were quantitated and are shown over time. The values for each macaque and the mean values for the groups are shown as dotted lines and heavier solid lines, respectively (b). Pharmacokinetic parameters were calculated with a two-compartment model (VRC01 versus VRC01-LS, half-life, 9.0 versus 30.3 days; clearance, 15.7 versus 3.7 ml day−1 kg−1; area under the curve (AUC), 896 versus 2,812 day × μg ml−1).

Extended Data Figure 5. Viral load measurements over time after passive antibody transfer and SHIV challenge of non-human primates

The indicated monoclonal antibodies (0.3 mg kg−1) were administered intravenously to rhesus macaques (n = 12). Macaques were challenged with SHIV BaLP4 5 days later, and the plasma viral loads were measured over time. Normal human IgG was given to the control group. Five of 12 VRC01-LS-injected macaques and 10 of 12 VRC01-injected macaques were infected (P = 0.0447, one-tailed Fisher’s exact test).

Extended Data Figure 6. Accumulation and localization of VRC01 in vaginal tissue

a, VRC01 or VRC01-LS (10 mg kg−1) were injected intravenously into female cynomolgus macaques (n = 2). The monoclonal antibody concentration per mg of total tissue protein in vaginal biopsy samples was quantitated at the indicated times. The values for each macaque and the mean values for the groups are shown as dotted and solid lines, respectively. b, VRC01 (20 mg kg−1) was injected intravenously into rhesus macaques (n = 2), and vaginal biopsy samples were taken and processed. Immunohistochemical staining was performed before and after the antibody dosing. The arrows indicate VRC01 (RSC3 staining, red) in basal and parabasal epithelial cells. The sections are representative of the sections assessed for two macaques. Scale bar, 50 μm.

Extended Data Figure 7. Similar antibody responses to VRC01 and VRC01-LS in non-human primates

VRC01 or VRC01-LS (10 mg kg−1) was injected intravenously into Indian rhesus macaques (n = 4). a, Anti-VRC01 antibody responses (left) or anti-VRC01-LS antibody responses (right) were evaluated by ELISA. b, Sera from animals that were administered VRC01 or VRC01-LS were absorbed with 20 μg ml−1 VRC01, and binding to VRC01-LS was then evaluated by ELISA.

Figure 1. In vitro evaluation of VRC01 and its FcRn-binding mutants

a, The FcRn-binding mutations used in this study are shown. The IHH mutant does not bind to human FcRn, and the other mutations enhance binding to human FcRn. b, Binding to HIV-1 Env (RSC3, a resurfaced, stabilized core of HIV-1 gp120) was examined by ELISA. c, Binding to human FcRn was assessed at pH 6.0 and pH 7.4 by ELISA. d, Transcytosis was assessed by adding 1.0 μg ml−1 of each monoclonal antibody (mAb) to the basolateral side of a monolayer of MDCK cells that express human FcRn, which were grown on a transwell filter plate, and incubating for 2 h. The concentration of each antibody on the apical side was measured by quantitative RSC3 ELISA, and the fold increase in antibody concentration over VRC01 (WT) is shown. *, P < 0.001 versus VRC01, two-tailed Student’s t-test. e, Binding to human FcγRIIIa was examined by ELISA. f, The ability of each monoclonal antibody to induce ADCC was analysed using human peripheral blood mononuclear cells as effector cells and HIV-1-infected CEM-NKR cells as targets and staining with the fluorescent dyes PKH26 and carboxyfluorescein succinimidyl ester (CFSE). The percentage killing was calculated from the percentage of CFSE− cells within the PKH26hi population. Assays were performed in duplicate (b, c, e, f) or quadruplicate (d), and the data points represent the mean values (b, c, e, f) or the mean ± s.e.m. (d). The data are representative of two independent experiments (b–f). A450, absorbance at 450 nm.

Figure 2. Pharmacokinetic study in Indian rhesus macaques

a, The indicated VRC01 variant monoclonal antibodies were injected intravenously at 10 mg per kg body weight into rhesus macaques on day 0 (VRC01 and VRC01-LS, n = 4; VRC01-IHH, n = 2), and the serum levels over time are shown. The pharmacokinetic parameters were calculated with a two-compartment model (VRC01 versus VRC01-LS, half-life, 4.7 versus 11.8 days; clearance, 16.5 versus 6.5 ml day−1 kg−1; area under the curve (AUC), 632 versus 1,645 day × μg ml−1). b, The amount of monoclonal antibody per mg of total tissue protein in rectal biopsy samples from each monoclonal-antibody-injected rhesus macaque (n = 4 per group) was quantitated and is shown over time. VRC01-LS persisted at higher levels than did VRC01 (P < 0.001, two-tailed t-test). The data points are the mean ± s.e.m. c, Immunohistochemical staining of the rectal biopsy samples (n = 6 at 10 or 20 mg kg−1 dose) after VRC01 administration shows co-localization of VRC01, as measured by indirect binding to an Env probe, RSC3 (VRC01, red; top) and FcRn, stained with an anti-human FcRn polyclonal antibody (FcRn, red; bottom). The sections shown are representative of those observed in the six macaques. The adjacent sections were stained with RSC3 and the anti-human FcRn polyclonal antibody in test samples. Buffer without RSC3 and control rabbit IgG were used as negative controls for VRC01 and FcRn staining, respectively. The arrows indicate VRC01 or FcRn staining in the cytoplasm of epithelial cells. The images were captured at ×20 original magnification, except for the right column, which shows magnified views of the centre column (×40 original magnification).

Figure 3. VRC01-LS affords greater protection against intrarectal SHIV BaLP4 challenge than VRC01

The indicated monoclonal antibodies were administered intravenously to rhesus macaques at a dose of 0.3 mg kg−1 on day 0. a, Monoclonal antibody levels per mg of total tissue protein in rectal biopsy samples (n = 4 per group) are shown over time. The data points are mean ± s.e.m. b, The macaques (n = 12 per group) were challenged with SHIV BaLP4 on day 5 after monoclonal antibody administration, and the percentage of uninfected macaques was plotted over time. Normal human IgG was given to the control group. Five of 12 VRC01-LS-treated macaques and 10 of 12 VRC01-treated macaques were infected (P = 0.0447, one-tailed Fisher’s exact test).