A Protein Epitope Targeted by the Antibody Response to Kawasaki Disease

Abstract

Background: Kawasaki disease (KD) is the leading cause of childhood acquired heart disease in developed nations and can result in coronary artery aneurysms and death. Clinical and epidemiologic features implicate an infectious cause but specific antigenic targets of the disease are unknown. Peripheral blood plasmablasts are normally highly clonally diverse but the antibodies they encode are approximately 70% antigen-specific 1-2 weeks after infection.

Methods: We isolated single peripheral blood plasmablasts from children with KD 1-3 weeks after onset and prepared 60 monoclonal antibodies (mAbs). We used the mAbs to identify their target antigens and assessed serologic response among KD patients and controls to specific antigen.

Results: Thirty-two mAbs from 9 of 11 patients recognize antigen within intracytoplasmic inclusion bodies in ciliated bronchial epithelial cells of fatal cases. Five of these mAbs, from 3 patients with coronary aneurysms, recognize a specific peptide, which blocks binding to inclusion bodies. Sera from 5/8 KD patients day ≥ 8 after illness onset, compared with 0/17 infant controls (P < .01), recognized the KD peptide antigen.

Conclusions: These results identify a protein epitope targeted by the antibody response to KD and provide a means to elucidate the pathogenesis of this important worldwide pediatric problem.

Keywords: Kawasaki disease; monoclonal antibodies; plasmablast.

Figure 1.

Experimental strategy and characteristics of PB isolated from peripheral blood of 11 children 1–3 weeks after KD fever onset. A, Study overview. B–D, Analysis of single cells from 11 children with KD. B, Most PB were VH3, VH4, or VH1. C, Number of PB by isotype; IgA and IgG PB were most commonly identified. D, Of 1156 PB sequenced, 42 sets of oligoclonal PB were identified for antibody production and 15 somatically mutated IgA plasmablasts were also selected for production. Of 60 monoclonal antibodies prepared, 10 strongly bound to intracytoplasmic inclusion bodies in ciliated bronchial epithelial cells of children who died of KD using immunohistochemistry and 22 were found to bind weakly. Strongly binding antibodies were tested on an animal virus peptide array. Abbreviations: IgA, immunoglobulin A; IgG, immunoglobulin G; KD, Kawasaki disease; PB, plasmablasts; RT-PCR, reverse transcription and polymerase chain reaction; VH, variable heavy; VL, variable light.

Figure 2.

A and D, KD monoclonal antibodies recognize ICI in KD ciliated bronchial epithelium by immunohistochemistry (arrows). A, ICI are identified by monoclonal antibody KD1-2G11 in a 4-month-old male infant (KD patient A) who died of acute KD in the United States at 3 weeks after fever onset. B, No staining is observed using KD1-2G11 in a 3-month-old US female infant who died of influenza. D, ICI are identified in KD patient A using antibody KD4-2H4. A–D, ×20. Abbreviations: ICI, intracytoplasmic inclusion bodies; KD, Kawasaki disease; mAb, monoclonal antibody; pt, patient.

Figure 3.

Identifying the peptide motif recognized by Kawasaki disease (KD) monoclonal antibody KD4-2H4 using animal virus peptide array and substitution array. A, Animal virus peptide array demonstrates that KD4-2H4 recognizes multiple related peptides with averaged median foreground fluorescence intensities above a cutoff value of 200; the epitope identity number from the Immune Epitope Database (www.iedb.org) is listed for each reacting peptide. B, The motif of KD4-2H4 binding identified as statistically significant by MEME bioinformatics analysis. C, Substitution matrix array of peptide AIIPDREALYQEFDEME using KD4-2H4. Preferred amino acids for binding are those on the left of the substitution matrix while nonpreferred amino acids are those on the right.

Figure 4.

Monoclonal antibodies from additional KD patients recognize KD peptide. A, KD monoclonal antibodies bind to KD peptide by ELISA. The OD reading of a scrambled version of the peptide was subtracted as a negative control. Antibodies KD4-2H4 and KD6-2B2 bind most strongly to the peptide, followed by KD8-1D4, KD8-1B10, and KD6-1A10. Samples were assayed in triplicate at each dilution of 10, 1, and 0.1 µg/mL. Dots represent individual assay results and horizontal lines represent mean of 3 assays. B, Substitution matrix array of peptide AVIPDREALYQDIDEME (derived from KD peptide) using KD6-2B2 demonstrates that KD6-2B2 shares an epitope with KD4-2H4. Preferred amino acids for binding are those on the left of the substitution matrix while nonpreferred amino acids are those on the right. Abbreviations: KD, Kawasaki disease; OD, optical density.

Figure 5.

KD peptide blocks binding of KD monoclonal antibodies to KD ICI in ciliated bronchial epithelium of KD patient A. A, ICI are identified in KD patient A using antibody KD4-2H4 preincubated with scrambled peptide (arrows). B, ICI staining is blocked when antibody KD4-2H4 is preincubated with KD peptide. C, ICI are identified in KD patient A using antibody KD6-2B2 preincubated with scrambled peptide (arrows). D, ICI staining is blocked when antibody KD6-2B2 is preincubated with KD peptide. A–D, ×20. Abbreviations: ICI, intracytoplasmic inclusion bodies; KD, Kawasaki disease.

Figure 6.

Western blot analysis of sera for reactivity to KD peptide fusion protein. For each blot, lane 1 contains GST alone, lane 2 contains GST-KD peptide fusion protein (GST-3X), and lane 3 contains human IgG as a positive control. Blots were stripped and polyclonal anti-GST antibody was applied to ensure that the GST fusion proteins were present; each serum sample was tested 2–5 times and a representative blot is shown. Molecular weight of IgG HC is 50 kD, GST-3X is approximately 35 kD, and GST alone is approximately 26 kD (arrows). Abbreviations: FC, febrile control; GST, glutathione sulfur transferase; IC, infant control; IgG HC, immunoglobulin G heavy chain; IVIG, intravenous immunoglobulin; KD, Kawasaki disease; M, All Blue Standard (Biorad).