Therapeutic antibody activation of the glucocorticoid-induced TNF receptor by a clustering mechanism

Abstract

GITR is a TNF receptor, and its activation promotes immune responses and drives antitumor activity. The receptor is activated by the GITR ligand (GITRL), which is believed to cluster receptors into a high-order array. Immunotherapeutic agonist antibodies also activate the receptor, but their mechanisms are not well characterized. We solved the structure of full-length mouse GITR bound to Fabs from the antibody DTA-1. The receptor is a dimer, and each subunit binds one Fab in an orientation suggesting that the antibody clusters receptors. Binding experiments with purified proteins show that DTA-1 IgG and GITRL both drive extensive clustering of GITR. Functional data reveal that DTA-1 and the anti-human GITR antibody TRX518 activate GITR in their IgG forms but not as Fabs. Thus, the divalent character of the IgG agonists confers an ability to mimic GITRL and cluster and activate GITR. These findings will inform the clinical development of this class of antibodies for immuno-oncology.

Fig. 1.. Structure of mGITR in complex with DTA-1 Fabs.

(A) Cryo-EM density map of mGITR in complex with DTA-1 Fabs. The receptor subunits, Fab heavy chain (HC) and light chain (LC), and micelle are colored as indicated in the legend. (B) Molecular model of the mGITR dimer with two Fabs and colored as in (A). The transmembrane helices are illustrated to show the orientation of mGITR with respect to the cell membrane. (C) Molecular model of mGITR dimer highlighting the CRD1 (pink), CRD2 (maroon), and CRD3 (purple) regions. The receptor is shown parallel to the membrane (left) and as viewed from the extracellular space (right). (D) The two mGITR subunits are shown after separating them and then placing each subunit in the same orientation for comparison. The flexible apical loop of each subunit is highlighted in yellow. The subunits are shown parallel to the membrane (top) and as viewed from the extracellular space (bottom).

Fig. 2.. The mGITR dimer interface is locked by an intersubunit disulfide bond.

(A) mGITR dimer interface with CRD1 (pink), CRD2 (maroon), and CRD3 (purple) and apical loops of CRD1 (yellow). Spheres on the apical loops mark Cys⁵⁵. (B) Illustration of the mGITR protein with relevant domains. The full-length protein is shown at the top, and the mGITR_ΔCTD construct with C55A and C164A mutations is shown at the bottom. (C) Western blots of full-length mGITR and the four CTD deletion constructs. Nonreduced (−DTT) and reduced (+DTT) conditions are on the left and right, respectively. Each lane is numbered according to the sample it contains, and sample descriptions are provided in the legend. MW, molecular weight. (D) FSEC traces showing EGFP fluorescence of lysate from HEK293 cells expressing full-length mGITR_egfp (dashed line), mGITR_ΔCTD-egfp (solid black line), mGITR_ΔCTD-egfp-C55A (red line), and mGITR_ΔCTD-egfp-C164A (blue line). AU, arbitrary units.

Fig. 3.. Human and mouse agonistic antibodies bind distinct residues on one epitope.

(A) A single mGITR subunit (gray) bound to a DTA-1 Fab with light chain (light blue) and heavy chain (dark blue). (B) Extracellular view of an mGITR subunit (left) and mGITR with residues proximal to the DTA-1 Fab and swapped to hGITR residues (right). (C) Sequence alignment between mGITR and hGITR in the region where DTA-1 binds mGITR. The residues on mGITR that are swapped to hGITR residues are indicated in the mGITR-swap sequence. WT, wild type. (D to F) Flow cytometry of cells expressing wild-type mGITR and treated with no primary antibody (negative control) (D), cells expressing wild-type mGITR and treated with DTA-1 IgG (E), and cells expressing the mGITR-swap mutant and treated with DTA-1 IgG (F). A fluorescein isothiocyanate (FITC)–labeled secondary antibody was used in all three experiments. The percentages correspond to the fraction of cells reporting an FITC fluorescent signal.

Fig. 4.. DTA-1 and mGITRL bind overlapping epitopes on mGITR.

(A and B) Surface rendering of the mGITR/DTA-1 Fab molecular model colored as indicated in the legend. (C and D) Proposed molecular arrangement of the mGITR dimer in complex with two copies of the mGITRL dimer. This model was built using the mGITR dimer structure (present study) as a template and aligning two copies of the mGITR monomer/mGITRL monomer cocrystal structure to the dimer. The effect was to create a model of the mGITR dimer with one mGITRL monomer bound to each mGITR subunit. Last, two copies of the mGITRL dimer were aligned to each mGITRL monomer to generate the final model presented in (C) and (D). The model is colored as indicated in the legend. (E) Illustration of discrete complexes formed by mGITR dimers with DTA-1 Fabs. (F) Illustration of clusters formed by mGITR dimers with DTA-1 IgG. (G) Illustration of clusters formed by mGITR dimers with mGITRL.

Fig. 5.. DTA-1 mimics the receptor clustering of mGITRL.

(A to F) FSEC traces of mGITR fused to EGFP (mGITR_egfp). The molar ratio used throughout is 1:15:4:2 for mGITR_egfp:mGITRL:Fab:IgG. Each trace shows mGITR_egfp alone (green) and the resulting trace from mGITR_egfp mixed with another protein or proteins (black). The traces are mGITR_egfp alone (A), mGITR_egfp with mGITRL (B), mGITR_egfp with DTA-1 Fab (C), mGITR_egfp with DTA-1 IgG (D), mGITR_egfp incubated (5 min, 4°C) with mGITRL before adding DTA-1 Fab (E), and mGITR_egfp incubated with DTA-1 Fab before adding mGITRL (F). The inset in (C) shows isolation of the mGITR_egfp-Fab peak and analysis by SDS-PAGE with bands for mGITR_egfp, Fab heavy chain, and light chain. Cartoon illustrations show mGITR (magenta), mGITRL (yellow), DTA-1 IgG (cyan) (large symbol), and DTA-1 Fab (cyan) (small symbol).

Fig. 6.. Full-length IgG is required for GITR agonism.

(A to C) Flow cytometry experiments with mGITR and no primary antibody (negative control) (A), mGITR with DTA-1 Fab (B), and mGITR with DTA-1 IgG (C). (D to F) Flow cytometry experiments with hGITR and no primary antibody (negative control) (D), hGITR with TRX518 Fab (E), and hGITR with TRX518 IgG (F). (G) Illustration of luciferase assay used to measure GITR activation by DTA-1 and TRX518 Fab and IgG. (H) Luciferase assay measuring mGITR activation in cells that are untreated or treated with DTA-1 Fab or DTA-1 IgG. (I) Luciferase assay measuring hGITR activation in cells that are untreated or treated with TRX518 Fab or TRX518 IgG. One-way analysis of variance (ANOVA) followed by a Tukey’s multiple comparison test was used to determine statistical significance between groups (****P < 0.0001).