Monoclonal antibodies to meningococcal factor H binding protein with overlapping epitopes and discordant functional activity

Abstract

Background: Meningococcal factor H binding protein (fHbp) is a promising vaccine candidate. Anti-fHbp antibodies can bind to meningococci and elicit complement-mediated bactericidal activity directly. The antibodies also can block binding of the human complement down-regulator, factor H (fH). Without bound fH, the organism would be expected to have increased susceptibility to bacteriolysis. Here we describe bactericidal activity of two anti-fHbp mAbs with overlapping epitopes in relation to their different effects on fH binding and bactericidal activity.

Methods and principal findings: Both mAbs recognized prevalent fHbp sequence variants in variant group 1. Using yeast display and site-specific mutagenesis, binding of one of the mAbs (JAR 1, IgG3) to fHbp was eliminated by a single amino acid substitution, R204A, and was decreased by K143A but not by R204H or D142A. The JAR 1 epitope overlapped that of previously described mAb (mAb502, IgG2a) whose binding to fHbp was eliminated by R204A or R204H substitutions, and was decreased by D142A but not by K143A. Although JAR 1 and mAb502 appeared to have overlapping epitopes, only JAR 1 inhibited binding of fH to fHbp and had human complement-mediated bactericidal activity. mAb502 enhanced fH binding and lacked human complement-mediated bactericidal activity. To control for confounding effects of different mouse IgG subclasses on complement activation, we created chimeric mAbs in which the mouse mAb502 or JAR 1 paratopes were paired with human IgG1 constant regions. While both chimeric mAbs showed similar binding to fHbp, only JAR 1, which inhibited fH binding, had human complement-mediated bactericidal activity.

Conclusions: The lack of human complement-mediated bactericidal activity by anti-fHbp mAb502 appeared to result from an inability to inhibit binding of fH. These results underscore the importance of inhibition of fH binding for anti-fHbp mAb bactericidal activity.

Figure 1. Binding of anti-fHbp mAb JAR 1 to recombinant fHbp as measured by ELISA.

Panel A. Binding of JAR 1 to fHbp ID 1 (variant group 1); ID 77 (variant group 2); or ID 28 (variant group 3). Panel B. Binding of a control anti-fHbp, JAR 13, which is specific for fHbp in variant groups 2 and 3 . Sequence variants tested and respective symbols are those shown in Panel A. Panel C. Binding of JAR 1 to variants of fHbp in variant group 1. ID 6, 14, 15, 55 and 74 are negative. ID 1, 2, 4, 9 and 13 are positive. Panel D. Binding of a positive control anti-fHbp mAb, JAR 65, to recombinant fHbp variants shown in Panel C. Wells of the microtiter plates were coated with 2 µg/ml of each recombinant protein.

Figure 2. Network analysis of the relatedness of the fHbp amino acid sequence variants tested for JAR 1 binding.

The network analysis was generated by the program, Splits Tree . All sequence variants were in variant group 1. Each variant is designated by its peptide ID number as described in the public fHbp sequence variant database (http://pubmlst.org/neisseria/fHbp/) followed by Roman numeral I or IV, which designates the fHbp modular group , . Sequence variants represented by filled black circles were positive for binding to JAR 1; sequence variants represented by open circles were negative for binding to JAR 1 (See Figure 1).

Figure 3. Anti-fHbp mAbs JAR 1 and mAb502 recognize overlapping epitopes on the C-terminal domain of fHb.

Panel A. Inhibition of binding of a fixed concentration (1 µg/ml) of anti-fHbp mAbs to recombinant fHbp as measured by ELISA. Red triangles: inhibition of binding of mAb502 (IgG2a) by JAR 1 (IgG3). Green circles: Inhibition of binding of JAR 1 by mAb502. Error bars represent ranges in independent assays. Panel B. Binding of JAR 1 to yeast clones displaying randomly mutated fHbp on their surface. A representative clone is positive for binding to JAR 1 (red solid line) and a control anti-fHbp mAb, JAR 5 (blue dashed line). Panels C, D, E. Binding of anti-fHbp mAbs to mutants of recombinant fHbp as measured by ELISA. Crosses, wild-type (WT) fHbp ID 1; Blue squares, D142A mutant; White triangles, K143A mutant; Orange circles, R204A mutant; Pink diamonds with dashed black line, R204H mutant. Panel C. Binding of JAR 5 (positive control, data are superimposed). Panel D. Binding of JAR 1. Panel E. Binding of mAb502. Panel F. Binding of JAR 1 (red triangles with solid red line) or mAb502 (green circles with dashed line) to fHbp ID 2, which has H204 instead of R204 in ID 1. Wells were coated with 2 µg/ml of wild-type or mutant fHbp.

Figure 4. Inhibition of binding of fH in relation to anti-fHbp mAb bactericidal activity.

Panel A. Inhibition of fH binding to recombinant fHbp as measured by ELISA. A fixed concentration of human fH (2 µg/ml) was incubated with different concentrations of anti-fHbp mAb JAR 1 (red triangles with solid red line) or mAb502 (green circles with dashed green line). Panels B. Inhibition of fH binding to live bacteria of wild-type group B strain H44/76 as measured by flow cytometry. A fixed concentration of fH (∼90 µg/ml) was incubated with 2 µg/ml of JAR 1, or 2 or 50 µg/ml of mAb502. The source of human fH was 20% IgG-depleted human serum. Dark gray filled area, bacteria without fH as a negative control; Light gray filled area, fH without the addition of a mAb; fH and JAR 1 (red line); fH and mAb502, (green dashed line). Data for 2 or 50 µg/ml of mAb502 were similar; only 50 µg/ml histogram is shown. Panel C. Inhibition of fH binding to live bacteria of a mutant of group A strain Senegal 1/99 with over-expressed fHbp ID 5 as measured by flow cytometry. Symbols same as in Panel B. Panel D. Survival of N. meningitidis group B H44/76 strains after 60 min incubation with anti-fHbp mAbs and 20% human complement. JAR 1 (red triangles) and the anti-PorA P1.7 mAb control were bactericidal but not mAb502 (green circles). Panel E. Same as Panel D except infant rabbit complement was used instead of human complement. All three mAbs had activity. For each panel, the respective results were replicated in two independent experiments. Error bars represent ranges in values.

Figure 5. Activity of chimeric human IgG1 mouse anti-fHbp mAbs.

Panel A. Binding of chimeric mAbs to fHbp by ELISA using anti-human kappa light chain secondary antibody (See Methods). Red triangles with solid red line, chimeric JAR 1; Green circles with dashed line, chimeric mAb502. Panel B. Inhibition of fH binding to recombinant fHbp as measured by ELISA. Human fH (2 µg/ml) was incubated with different concentrations of chimeric JAR 1 or chimeric mAb502. Symbols as in Panel A. Panel C. Inhibition of binding of fH to live bacteria of group B strain H44/76 as measured by flow cytometry. A fixed concentration of human fH (∼90 µg/ml) was incubated with 2 µg/ml of chimeric JAR 1, or 2 or 50 µg/ml of chimeric mAb502. The source of fH was 20% IgG-depleted human serum. Dark gray filled area, bacteria without fH as a negative control; Light gray filled area, fH without the addition of a mAb; fH and JAR 1 (solid red line); fH and mAb502 (green dashed line). Data for 2 or 50 µg/ml of chimeric mAb502 were similar; only the 50 µg/ml histogram is shown. Panel D. Survival of wild-type group B strain H44/76 after 60 min incubation with chimeric anti-fHbp mAbs and 20% human complement. Symbols for chimeric anti-fHbp mAbs are same as in Panel A; Blue diamonds with dashed blue line, positive control anti-PorA P1.7 mAb. Data are from two experiments. Error bars, range in values. Panel E. Same as Panel D except that infant rabbit complement was used instead of human complement.

Figure 6. Structural model of fH-fHbp complex.

fH fragment is shown in gray. The amino-terminal domain of fHbp is shown in light blue, while the carboxy-terminal domain is shown in light green. Panel A. Alanine substitutions for arginine at residue 204 (R204A) or for lysine at position 143 (K143A) decreased or eliminated JAR 1 binding. His substitution for Arg at position 204 (R204H) did not affect JAR 1 binding (See Figure 5). Panel B. Residues shown previously by NMR to be part of the mAb502 binding site . In addition, we found that alanine or histidine substitutions for arginine at residue 204 (R204A or R204H) eliminated binding of mAb502, and an alanine substitution for aspartic acid at residue 142 (D142A) decreased binding (See Figure 3). The figure was constructed with PyMol (http://www.pymol.org).