Complex Interplay between Epitope Specificity and Isotype Dictates the Biological Activity of Anti-human CD40 Antibodies

Abstract

Anti-CD40 monoclonal antibodies (mAbs) that promote or inhibit receptor function hold promise as therapeutics for cancer and autoimmunity. Rules governing their diverse range of functions, however, are lacking. Here we determined characteristics of nine hCD40 mAbs engaging epitopes throughout the CD40 extracellular region expressed as varying isotypes. All mAb formats were strong agonists when hyper-crosslinked; however, only those binding the membrane-distal cysteine-rich domain 1 (CRD1) retained agonistic activity with physiological Fc gamma receptor crosslinking or as human immunoglobulin G2 isotype; agonistic activity decreased as epitopes drew closer to the membrane. In addition, all CRD2-4 binding mAbs blocked CD40 ligand interaction and were potent antagonists. Thus, the membrane distal CRD1 provides a region of choice for selecting CD40 agonists while CRD2-4 provides antagonistic epitopes.

Keywords: CD40; Fc receptors; TNFR; agonist; antagonist; crystal structure; epitope; immunotherapy; isotype; monoclonal antibody.

Graphical abstract

Figure 1

Mouse IgG1 Constant Regions Cannot Confer Agonistic Activity to all Anti-h CD40 mAbs (A) Purified hCD40Tg (WT) or hCD40Tg/Fcgr2b^−/− (KO) splenic mouse B cells were incubated with increasing concentrations of Lob 7/4-m1, SGN40-m1, or CP870,893-m1 for 4 days, and proliferation measured by [³H]thymidine incorporation. Means ± SEM, n = 3, data representative of at least three experiments. (B) Purified splenic WT or KO B cells in the absence or presence of irradiated hFcγRIIB-overexpressing CHO-K1 cells (XL) were treated as in (A) with 1 μg/mL each mAb (m1). Means ± SEM, n = 3, data representative of at least three experiments. (C) 1 × 10⁵ OTI cells were adoptively transferred to hCD40Tg mice 1 day before treatment with 30 μg of the indicated mAbs (m1). Phycoerythrin (PE)-labeled SIINFEKL+ cells, expressed as a percentage of total CD8⁺ cells. Values for individual mice are shown; error bars represent means ± SEM. Data representative of at least two experiments. Comparison of all CRD1-targeting mAbs individually with Lob 7/6-m1, ^∗∗p < 0.01 using one-way ANOVA followed by Tukey's multiple comparisons test.

Figure 2

Crystal Structure of ChiLob 7/4 Fab' in Complex with CD40 Extracellular Domain (A) The crystal structure of the CD40:ChiLob 7/4 Fab' complex shows the Fab' in gray (dark gray, heavy chain; light gray, light chain) and CD40 colored by CRD (blue, CRD1-A1; red, CRD1-B2; green, CRD2-A1; magenta, CRD2-B2; orange, CRD3-A1; cyan, CRD3-B2 [partial]). (B) Interface, with structural elements removed for clarity. Shown are the CDR loops of ChiLob 7/4 in red for CDR1, green for CDR2, and blue for CDR3. Light color shades represent the light chain, dark shades the heavy chain. CD40 colored as in (A) (translucent). See also Table S1.

Figure 3

Epitope Mapping of Anti-hCD40 mAbs (A) C-terminally His-tagged CD40 proteins consisting of 4, 3, or 2 CRD domains (4 = full-length; 3 = CRD1 deleted; 2 = CRD1 and CRD2 deleted) were analyzed by western blotting with the mAbs indicated above each panel used for detection. A composite image from multiple blots with different antibodies is shown. (B) hCD0Tg mouse B cells were incubated with FITC-labeled Lob 7/4 or Lob 7/6, as indicated, in the absence or presence of a 10-fold excess of competitor (comp) mAbs indicated above each plot. Cell labeling was assessed by flow cytometry; filled histogram, unlabeled cells; light gray, FITC-labeled mAbs alone; black line, FITC-labeled mAbs + competitor. (C) Surface plasmon resonance analysis of mAb competition for CD40 binding. The mAbs indicated above each plot were flowed over immobilized hCD40 ECD for 800 s to allow saturation of hCD40 binding sites, followed by the addition of individual competitor mAb for 350 s. Non-competitive mAbs (i.e., mAbs without overlapping epitope) give a response greater than 500 response units (RU). (D) Untransfected 293 cells (none), or cells transfected with full-length (WT) human CD40 or CD40 in which N-terminal amino acids 23–37 had been deleted, were incubated with the indicated anti-hCD40 mAbs (h2 isotype). Bound mAbs were detected with anti-human Fc-FITC. The percentage of cells in the boxed region, denoting positive staining, is shown for each plot. Results from one of two experiments shown. (E). Schematic of the approximate locations of epitopes for nine human CD40 mAbs analyzed. See also Figure S1.

Figure 4

Alanine Scanning Epitope Mapping of Anti-hCD40 mAbs CHO-k1 cells stably expressing mutant hCD40 proteins were incubated with the indicated m1 mAb and binding detected by flow cytometry with anti-mouse Fc-FITC. (A) Example plots for the F54-T55 mutant to illustrate positive (tick) and negative (cross) binding results. (B) Summary of data for additional hCD40 mutants (at least two experiments per combination). (C) Deduced mAb epitopes are illustrated on the hCD40 crystal structure resolved in Figure 2, and are color coded with those for ChiLob 7/4, SGN40, and Lob 7/2 overlapping. See also Figure S2.

Figure 5

Influence of Human IgG Isotype on Epitope-Dependent Agonism (A) Purified hCD40Tg mouse splenic B cells were incubated with the indicated mAb isotypes at various concentrations for 4 days, and proliferation was measured as in Figure 1A. Means ± SEM, n = 3, data representative of at least three experiments. (B) Purified human B cells were incubated with the indicated mAbs at 1 μg/mL in the absence (−) or presence (+) of crosslinking cells overexpressing hFγRIIB, as in Figure 1B. Histograms show expression of CD23 (black lines) compared with untreated cells (gray histograms) as determined by flow cytometry. (C) Purified hCD40Tg WT or KO mouse splenic B cells were treated as in (A) with 1 μg/mL of each mAb of m1, h1, or h2 isotype, and proliferation determined by [³H]thymidine incorporation. Means ± SEM, n = 5, data representative of at least three experiments. (D) 1 × 10⁵ OT1 cells were adoptively transferred to hCD40Tg mice 1 day before treatment with 100 μg h1 or h2 of the indicated mAbs and 100 μg OVA. OT1 cells in peripheral blood were quantified on day 5 using PE-labeled SIINFEKL tetramer, expressed as percentage of total CD8⁺ cells. Data points represent individual animals from at least two independent experiments per mAb. Horizontal bars indicate mean values. ^∗∗p < 0.001, ^∗∗∗∗p < 0.0001, h1 versus h2 of same mAbs using the Student's unpaired t test. See also Figure S3.

Figure 6

Influence of Hinge Conformation on SGN40 and CP870,893 Agonistic Activity (A) CE-SDS profiles of the mAbs chemically skewed toward their h2A and h2B conformations (h2′A′ and h2′B′, respectively) or “locked” in h2A (through point mutation C233S) and h2B (through point mutation C127S). (B) Ability of the different h2 mAbs to stimulate proliferation of hCD40Tg mouse B cells was analyzed as described in Figure 1A. Data are presented as means ± range of duplicate samples. (C) The ability of the mAb to stimulate OT1 cell expansion in vivo was analyzed as described for Figure 1C. Results for individual animals are plotted. Horizontal bars indicate mean values. ^∗p < 0.05 for h2′A′ versus h2′B′ using the Student's unpaired t test.

Figure 7

Antagonist Anti-hCD40 mAbs Engage CRD2-4 Domains (A) Purified hCD40Tg mouse splenic B cells were incubated with or without the indicated mAb, all of m1 isotype, followed by hexameric mCD40L-h1Fc. Bound CD40L was then detected with anti-human-FITC and flow cytometry. (B) Structure superposition of CD40:Lob 7/4 Fab' complex (colored as in Figure 2A) with the CD40:CD40L complex (PDB: 3QD6). CD40L is shown as a trimer binding to the opposite side of CD40 than Lob 7/4 and to a different CRD (CD40L shown as white surface representation). (C) Purified hCD40Tg mouse splenic B cells were incubated with CD40L alone (10 μg/mL) or together with the indicated mAbs (m1 at 1 μg/mL) for 20 hr. Activation was assessed by homotypic adhesion (top) or flow cytometry to assess upregulation of activation markers. Results from one of at least three experiments for each mAb are shown. (D) Purified hCD40Tg mouse splenic B cells were incubated with the indicated mAbs (m1 at 1 μg/mL) under control conditions (left panel) or in the presence of CD40L (10 μg/mL, right panel). Proliferation was assessed by [³H]thymidine incorporation as in Figure 1A and plotted as means ± SEM. (E) As in (D), except that h2 mAbs were used. Means ± SEM of triplicate samples representing results from two or three experiments per mAb. ^∗∗p < 0.01, ^∗∗∗p < 0.001, ^∗∗∗∗p < 0.0001 using the unpaired Student's t test, left hand plots for indicated groups, right hand plots versus CD40L alone.