Neutralization Takes Precedence Over IgG or IgA Isotype-related Functions in Mucosal HIV-1 Antibody-mediated Protection

Abstract

HIV-1 infection occurs primarily through mucosal transmission. Application of biologically relevant mucosal models can advance understanding of the functional properties of antibodies that mediate HIV protection, thereby guiding antibody-based vaccine development. Here, we employed a human ex vivo vaginal HIV-1 infection model and a rhesus macaque in vivo intrarectal SHIV challenge model to probe the protective capacity of monoclonal broadly-neutralizing (bnAb) and non-neutralizing Abs (nnAbs) that were functionally modified by isotype switching. For human vaginal explants, we developed a replication-competent, secreted NanoLuc reporter virus system and showed that CD4 binding site bnAbs b12 IgG1 and CH31 IgG1 and IgA2 isoforms potently blocked HIV-1_JR-CSF and HIV-1_Bal26 infection. However, IgG1 and IgA nnAbs, either alone or together, did not inhibit infection despite the presence of FcR-expressing effector cells in the tissue. In macaques, the CH31 IgG1 and IgA2 isoforms infused before high-dose SHIV challenge were completely to partially protective, respectively, while nnAbs (CH54 IgG1 and CH38 mIgA2) were non-protective. Importantly, in both mucosal models IgG1 isotype bnAbs were more protective than the IgA2 isotypes, attributable in part to greater neutralization activity of the IgG1 variants. These findings underscore the importance of potent bnAb induction as a primary goal of HIV-1 vaccine development.

Keywords: Antibodies; HIV-1; IgA; IgG; Mucosal immunology; Neutralizing antibodies; Non-human primate rectal challenge model; Vaginal explants.

Fig. 1

In human vaginal tissue, macrophages and monocytes are the main FcR-expressing effector cells and FcγRII is the most widely expressed FcR on vaginal leukocytes. Major leukocyte populations were identified in single cell suspensions isolated from whole vaginal tissue and their FcR expression profiles were determined by flow cytometry. (a) The frequency of each leukocyte population (e.g., B cells) identified in vaginal cell suspensions relative to total live leukocytes (CD45 +). (b) The proportion of vaginal leukocytes that express FcγRI, FcγRII, FcγRII or FcαRI. The contribution from each leukocyte population is show as a percentage of total live leukocytes.

Fig. 2

Broadly-neutralizing but not the non-neutralizing mAbs block snLuc.HIV-1_JR-CSF infection in human vaginal explants. (a) Summary of ex vivo HIV-1 inhibition results from kinetic snLuc infection assay conducted with tissue from 5 donors. Each square represents the result from an individual donor tissue for the listed mAb or small molecule. bnAbs are outlined in blue. (b) Representative data from vaginal explant infectivity model. PHA-activated vaginal tissue explants (3 per well) were exposed to snLuc-.HIV-1_JR-CSF (1 × 10⁶ IU/well) that was pre-opsonized with the indicated mAbs at 20 μg/mL. The next day, the inoculum was collected (Day 1 samples in plots above), the tissues were washed several times and individual mAbs were added back at 20 μg/mL to the culture media. Explants were maintained in culture for up to 21 days, with 80% media collection and replacement every 1–3 days. Decay in snLuc activity (RLU) after day 2 indicates inhibition of HIV-1 infection. By contrast, an increase or low level maintenance in snLuc activity over time indicates productive infection. Delayed infection kinetics were defined as productive infection that is delayed by at least two time points compared to the matched virus only conditions. The CH65 IgG and IgA mAbs were included as isotype controls. Indinivir (IDV, protease inhibitor), Zidovudine (AZT, reverse transcription inhibitor), and b12 IgG were included as positive controls for virus inhibition. The antiretrovirals (ARV) were maintained up to day 7 post-exposure.

Fig. 3

Passive infusion of CH54 IgG or CH38 mIgA2 does not protect rhesus macaques against high dose mucosal SHIV challenge. Groups of Indian-origin rhesus macaques received two IV infusions of either CH54 IgG (a), CH38 mIgA2 (b) or isotype matched anti-flu mAb CH65 at 50 mg/kg. Six hours post-1st infusion monkeys were challenged with SHIV_Bal. Plasma viral RNA levels of individual monkey are shown.

Fig. 4

HIV-1-specific, non-neutralizing IgG mAbs in combination do not inhibit ex vivo HIV-1 infection of human vaginal explants. Representative data from two separate experiments in two donor tissues showing infection with snLuc.HIV-1_JR-CSF and (top) snLuc.HIV-1_Bal26 (bottom). Three IgG mAbs directed against the C1 region of g120 (CH29, CH54 and CH57), one against the gp41 immunodominant (7b2 IgG) and one against a V1/V2 non-glycan conformational epitope (CH58 IgG) were combined at 4 μg/mL each, for a total of 20 μg/mL anti-HIV-1 mAb cocktail. These experiments were conducted as described in Fig. 2 with the following modifications: Culture media was supplemented with 10% heat-inactivated human serum AB in lieu of FBS, and an env-deleted snLuc reporter construct (delta env) was included as a negative control for snLuc.HIV infection.

Fig. 5

Protection by CH31 mAbs against SHIV_SF162P3 intrarectal challenge. Plasma viral RNA levels following high dose rectal challenges with SHIV_SF162P3 in monoclonal antibody-treated rhesus monkeys. (a) Rhesus monkeys were treated by rectal instillation with four different isotypes of CH31 monoclonal antibody including CH31 IgG, CH31 mIgA2, CH31 dIgA2 and CH31 sIgA2 followed by high dose intra-rectal viral challenge. Pavilizumab or CH65 IgG or CH65 mIgA2 were used as control. (b) Monkeys were passively infused with either CH31 IgG or CH31 mIgA2 prior to high dose rectal challenge. Pavilizumab or CH65 IgG or CH65 mIgA2 were used as control.

Fig. 6

Ex vivo protection of human vaginal explants from HIV-1 infection by CH31 bnAbs reflects the relative potencies of each isotype variant in classic in vitro neutralization assays. (a) Summary of ex vivo snLuc.HIV-1_Bal26 infection of vaginal explants in the presence of serially diluted isotype variants of 7b2 and CH31. Each line represents the readout from a single infection well containing 3 explants; symbols correspond to donors; and the 3 color gradients correspond to Ab concentrations that are considered high, medium and low. For 7b2 and CH58 mAbs these concentrations were 20, 4 and 0.8 μg/mL, respectively. For CH31 mAbs for which additional concentrations were tested, the individual infection curves were categorized into 3 color intensity groupings coinciding with the dose-response inhibition profiles of each isotype variant. Infection curves for (1) all mAb concentrations that consistently blocked infection to undetectable levels are depicted in dark blue; (2) the lowest concentration range that effectively blocked infection in most donor tissues are shown in medium blue; (3) and lower mAb concentrations that showed limited or no impact on HIV-1 infection are shown light blue. (b) The relationship between ex vivo (LED80) protection and in vitro neutralization potency IC50 (left) and IC80 (right) for CH31 mAbs. Neutralization IC50 and IC80 values were taken from (Table S2) except for the IC50 value for CH31 IgG which was taken from CATNAP (median of all reported values), www.hiv.lanl.gov. LED80 values were taken from Fig. S5. One-tailed Spearman's correlation analysis was done in GraphPad Prism.