Antibody-antigen kinetics constrain intracellular humoral immunity

Abstract

During infection with non-enveloped viruses, antibodies stimulate immunity from inside cells by activating the cytosolic Fc receptor TRIM21. This intracellular humoral response relies on opsonized viral particles reaching the cytosol intact but the antigenic and kinetic constraints involved are unknown. We have solved the structure of a potent TRIM21-dependent neutralizing antibody in complex with human adenovirus 5 hexon and show how these properties influence immune activity. Structure-guided mutagenesis was used to generate antibodies with 20,000-fold variation in affinity, on-rates that differ by ~50-fold and off-rates by >175-fold. Characterization of these variants during infection revealed that TRIM21-dependent neutralization and NFκB activation was largely unaffected by on-rate kinetics. In contrast, TRIM21 antiviral activity was exquisitely dependent upon off-rate, with sub-μM affinity antibodies nevertheless unable to stimulate signaling because of fast dissociation kinetics. These results define the antibody properties required to elicit an efficient intracellular immune response during viral infection.

Figure 1. 9C12 Fab: Adv5 hexon complexed structure.

(a) Model of the adenovirus virion based on PDB 1P30 as a gray surface. Surface representations of the hexon trimer from the complexed structure (monomers shown in shades of green). Secondary structures of the bound 9C12 Fabs are in light gray. (b) Superposition of free (gray; PDB 1P30) and complexed hexon (green) monomers with labeled domains. The view is from the interior of the molecule, perpendicular to the three-fold molecular axis. (c) Side by side images of the interface between 9C12 Fab (gray-blue) and hexon, either showing superposed free structure (gray) or complexed (green). The hexon loops only present in the complexed structure are labeled. Note that two hexon monomers are involved at the interface (light and dark green). (d) Surface representation of 9C12 Fab V_L (blue) and V_H (gray) bound to HVR2 and HVR8 from two different hexon monomers. Residues from HVR loops are shown as sticks.

Figure 2. 9C12 CDR interactions with hexon.

(a) Surface representation of hexon (monomers in shades of green). Secondary structure of V_L (blue) and V_H (gray) with CDR loops marked in yellow. The 9C12 CDRs interact with HVRs from two hexon monomers. (b,c) Interacting residues from hexon (top, cyan & green) with residues from 9C12 (bottom, V_H in yellow, V_L in blue/gray). Dashed lines indicate putative hydrogen bonds. Two different views are shown corresponding to residues mutated in Figs 3(b) and 4(c).

Figure 3. Structure-guided 9C12 mutants prevent hexon binding, Adv5 neutralization and NFκB activation.

(a) SPR titrations to determine kinetics and affinity of mouse and human WT 9C12 and selected h9C12 CDR-engineered mutants that abrogate hexon binding. (b) Ability of h9C12 variants to prevent Adv5 infection in 293 T WT cells or 293 T TRIM21 KO cells. (c) NFκB induction by h9C12 variants in 293 T WT and TRIM21 KO cells upon adenovirus infection. Cellular experiments were performed in triplicate.

Figure 4. Modulation of h9C12 on and off-rate kinetics have disproportionate affects on antiviral activity.

(a) SPR titrations of selected h9C12 variants. (b) Neutralization of Adv5 infection by h9C12 variants in 293 T WT cells or 293 T TRIM21 KO cells. (c) Induction of NFκB in 293 T WT and TRIM21 KO cells by h9C12 variants upon adenovirus infection. Cellular experiments were performed in triplicate.

Figure 5. TRIM21-dependent neutralization and signaling is independent from on-rate but proportional to off-rate.

Neutralization and signaling efficacy, expressed as percentage of WT activity, and the EC50 of selected h9C12 variants plotted against their off rates (a–c), on rates (d–f) and affinities (g–i) as determined by SPR. The Pearson correlation (Pearson’s correlation coefficient r) and the statistical significance (two-tailed P value) between these kinetic measures are shown.