Elsevier

Human Immunology

Volume 59, Issue 1, January 1998, Pages 29-38
Human Immunology

Kawasaki Disease and the T-Cell Antigen Receptor

https://doi.org/10.1016/S0198-8859(97)00233-4Get rights and content

Abstract

ABSTRACT: We investigated the evidence for an infectious etiology of Kawasaki disease (KD), an acute vasculitis of unknown etiology, by assessing the effects of KD on the T cell antigen receptor variable beta region families (Vβ). Using 3-color flow cytometry, we studied KD patients pre- and post-intravenous gamma globulin (IVIG) therapy and at >40 days post therapy, additionally comparing them to matched pediatric control patients (PCC) and their own healthy parents (one parent/KD child). Of all the Vβ families examined, only Vβ2 exhibited statistically significant differences, between the pre- and postIVIG samples and preIVIG and parent samples. No associations were found between Vβ2 findings and T cell memory, activation, or adhesion markers. For 2 KD patients, 4 parents, and 1 PCC participant, >15% of resting CD8+ lymphocytes and >15% of blastic CD8+ lymphocytes expressed a single Vβ family, which varied by individual, without similar expansions in the CD4+ cell populations. One of the participants with this abnormality was the only one with significant cardiac abnormalities. For all participants with the Vβ abnormality, other T-cell abnormalities were extensive and involved both CD4+ and CD8+ cells. We suggest that Vβ2 changes do occur in KD, as previously reported. However, these may not be involved in disease pathogenesis. Other Vβ changes also occur. Those occurring in parents may reflect asymptomatic reinfection with an infectious agent causing KD. Further, some KD patients may have restricted cytotoxic T-cell responses to that as yet unidentified agent; this restricted response may be associated with more severe cardiac involvement. Published by Elsevier Science Inc.

Introduction

Kawasaki Disease (KD) is an acute and transient vasculitis occurring in infants and children. It is an extremely common disorder, with the reported incidence in the United States being greatest between February and May [1]. Until the advent of intravenous gammaglobulin (IVIG) therapy, KD was associated with a 15%–30% incidence of inflammatory coronary arteritis 2, 3, 4and a mortality rate of up to 5%, due to cardiac complications [5]. With IVIG therapy in the acute phase of disease, cardiac symptomatology is inhibited; fever and toxic symptomatology usually resolve in hours 6, 7, 8, 9. The mechanism of IVIG’s action is unknown. However, because its effect is clinically significant, IVIG is now the treatment of choice for any case of suspected KD.

The etiology of KD is unknown despite intense investigation by numerous researchers. KD is generally assumed to be infectious in origin because of epidemiologic characteristics shared with other infectious diseases, i.e., a distinct age predilection (6 months to 5 years) 2, 10, 11, 12; epidemic occurrences and seasonality 1, 10, 11; and distinct clinical symptomatology [10]. This symptomatology includes fever, an erythematous rash, edematous hands and feet, subsequent digital desquamation, bilateral nonexudative conjunctivitis, inflammation of mucous membranes, and cervical adenopathy.

Because these symptoms are similar to those occurring with known superantigen-related infections, particularly staphylococcal and streptococcal, it has been proposed that KD involves superantigen (SAG) activity. It has recently been suggested that KD is associated with selective expansion of T lymphocytes expressing T-cell antigen receptor (TCR) variable beta region families (Vβ) 2 and 8 13, 14; selective Vβ expansion is consistent with SAG activity. The organisms suggested as causing KD have included streptococcus, retroviruses, parvovirus, herpesvirus, and most recently, a toxic shock syndrome toxin-secreting Staphylococcus aureus [15]. Research reports have supported 16, 17or contradicted 18, 19, 20, 21the possibility of SAG activity and/or involvement of various suggested organisms 22, 23, 24, 25, 26, 27, 28. One recent study presented evidence for TCR complementarity-determining region 3 (CDR3) size profiles suggestive of nominal antigen activity in KD, rather than SAG activity [21]. Most of these studies have included few participants and some, no, or poorly matched comparison groups. Some key studies 13, 14of the TCR repertoire employed polymerase chain reaction techniques, which are unreliable for this purpose 29, 30, 31, and/or monoclonal reagents to only a few Vβ families 13, 14, 16, 17, 18.

To address the issues of pathogenesis, the role of SAGs, and the involvement of the TCR in KD, we studied patients during acute KD illness (within 10 days of fever onset), at 24–72 h post receipt of IVIG, and during convalescence (>40 days following fever onset). In addition, we studied one genetic parent residing with the child and a genetically unrelated, ethnically- and age-matched child without KD.

Participants were enrolled from hospitals throughout the Atlanta area in 1994–1996. Index participants were (a) 6 months to 8 years of age; (b) febrile for >5 to 10 days, with a documented temperature of >38.5°C; (c) without other discernible cause of illness; and (d) with at least 4 of the following findings, upon examination by a cardiologist: an erythematous rash, nonexudative bilateral conjunctivitis, a reddened inflamed tongue, fissured lips, cervical adenopathy, and arthralgia or arthritis. Only one participant had significant cardiac findings; he received 2 doses of IVIG over a 24-h period. Echocardiographic findings had not resolved completely by the time the convalescent sample was obtained. For all 30 participants with acute KD, we enrolled one genetic parent: 26 at the time of the patients’s preIVIG sample was drawn; two, 1 day after the preIVIG sample; one, 4 days after the preIVIG sample; and one, 1 month after the preIVIG sample. Genetic siblings were not enrolled, to avoid clinically nonindicated venipuncture of a healthy child. For 20 KD participants, we enrolled an unrelated child being catheterized for a cardiac disorder not associated with KD, streptococcal infection, or immune deficiency. The pediatric “cardiac controls” (PCC) were matched to index cases by gender (19 male/11 female), ethnicity (15 white/15 black), and age ±3 years (mean 3.7 years, median 3.3 years, range 0.8–8.8 years), since these may be related to activation marker and T-cell subset profiles. This protocol was reviewed and approved by the Emory University and Centers for Disease Control and Prevention (CDC) human subjects review committees. Informed consent was obtained from the adult participants and parents of the pediatric participants.

Fluorescein isothiocyanate-conjugated (FITC) murine monoclonal antibodies against the human TCR were purchased from three sources, with reactivity specified to date as follows: Immunotech1 (Westbrook, ME; Vβ2, Vβ3, Vβ5S2 Vβ8S1/8S2, Vβ13S6, Vβ17S1, Vβ21S3, Vβ22S1); T Cell Diagnostics (Woburn MA; Vα2, Vβ3S1, Vβ5S1 Vβ5S2/5S3, Vβ5S3, Vβ6S7, Vβ8 [referred to herein as “8A”], Vβ8 [referred to herein as “8B”], Vβ12S1, Vβ13S1/13S3); and Becton Dickinson Immunochemistry (BD) (San Jose, CA) BD (TCRα,β and TCR δ chain). Phycoerythrin-conjugated (PE) monoclonal anti-CD8 and Peridininin chlorophyll protein (PerCP) monoclonal anti-CD4 and anti-CD8 were obtained from BD, as were fluorochrome-conjugated monoclonal antibodies: CD3, CD19, CD16, CD56, CD45RA, CD45RO, HLA-DR, CD25, CD38, and CD71. Conjugated murine monoclonal antibody to CD29 was purchased from Coulter Corporation (Miami, FL) and to CD62L, from PharMingen (San Diego, CA). Natural killer cells (NK) are defined herein as those CD3− and CD16 or CD56+. Cells positive for CD3 and for CD16 or CD56 will be referred to as CD3+NK.

Peripheral venous blood lymphocyte (PBL) samples were obtained on all participants. Three-color cytofluorometry was done using a whole-blood technique, a FACScan or FACSort (BD) flow cytometer, and Lysis II software. Negative control aliquots were stained with FITC mouse IgG1, PE-mouse IgG2a, and PerCP-mouse IgG1. From each sample, 20,000–30,000 ungated events were collected. For all parameters, results are reported for CD4+ and CD8+ cells in a lymphocyte scatter gate. In addition, for Vβ’s and CD71, up to 3000 events were collected from each sample using the following gating: (a) forward and side scattering of TCRαβ+ cells indicative of large, less dense (blastic) cells [as opposed to small, dense (resting) lymphocytes]; and (b) positivity for CD4 or CD8, but not both [32]. Vβ analyses are reported for the following four populations: resting CD4+ cells, resting CD8+ cells, blastic CD4+ cells, and blastic CD8+ cells. CD45RA, CD45RO, HLA-DR, CD25, CD38, and CD71 were assessed in resting CD4+ and resting CD8+ cells (6 antigens × 4 populations = 12 non-Vβ variables evaluated); CD71 was also assessed in blastic CD4+ and CD8+ cells (2 additional variables, for a total of 14 evaluated, non-Vβ variables). Detailed results for these variables, not in association with Vβ findings, are provided elsewhere [33].

Cellular markers were analyzed using Lysis II software. For all antigens except the Vβ’s; all elements of an analytic cluster were analyzed at one time. All samples for a given cluster were collected on the same machine. PreIVIG, postIVIG, and parents’ readings were done using identical settings, at the same time or within 72 h of one another. Convalescent and PCC samples were assessed at later times and not necessarily with the earlier settings. The same Vβ reagent panel was used for every sample in a given cluster, but the Vβ panel was enlarged as new reagents became available.

Values for the index case at presentation were compared with each of the following: those of the parent; the PCC; and his/her own findings at 24–72 h postIVIG therapy. For 20 patients, we obtained an additional blood sample at >40 days following acute illness (i.e., during convalescence, and compared these findings with those preIVIG). Because most of these parameters are not normally distributed, a wilcoxon signed rank test was used, rather than parametric-based analyses that assume normal distributions (e.g., means, standard deviations, etc.). The wilcoxon signed rank test provides a nonparametric matched comparison of the median value for the preIVIG sample to the median value of each of the other samples in that analytic cluster. A result will be referred to as significant if the two-sided p-value was <.01, rather than <.05, as a correction for multiple comparisons.

Section snippets

Results

Of all the Vβ families examined, only Vβ2 exhibited statistically significant differences between the preIVIG samples and multiple other comparison groups (Fig. 1). For both resting CD4+ and resting CD8+ cells, a lower proportion of the postIVIG samples (p = .0039 for CD4+ cells and p = .0015 for CD8+ cells) and parent samples (p = .0013 for CD4+ cells and p = .0110 for CD8+ cells) expressed Vβ2. None of these parents had symptoms of KD. For 17 participant clusters, both the child’s postIVIG

Discussion

In this controlled study, we addressed the effects of KD on the TCR, to determine if there were evidence for or against SAG activity, cytotoxic T cell restriction, and an infectious cause of the disease. One laboratory has suggested that KD is associated with selective expansion of T lymphocytes expressing Vβ’s 2 and 8 13, 14. Our data, as well as those in one other study [17], support an effect of either KD or IVIG therapy on Vβ2, but not Vβ8. However, we found no association between pre- to

Acknowledgements

We are grateful for the cooperation of the participants in this study, and the assistance of numerous fellow, resident, and faculty physicians who identified potential participants. We are especially grateful to Ms. Lynda Gregg, who helped in the clinical aspects of the study, and Sharon Collins, Ph.D., who obtained blood specimens from some participants.

References (44)

  • Y Nakamura et al.

    Cardiac sequelae of Kawasaki disease in JapanStatistical analysis

    Pediatrics

    (1991)
  • R Dhillon et al.

    Management of Kawasaki disease in the British Isles

    Arch Dis Child

    (1993)
  • JW Newburger et al.

    The treatment of Kawasaki syndrome with intravenous gamma globulin

    N Engl J Med

    (1986)
  • JW Newburger et al.

    A single intravenous infusion of gamma globulin as compared with four infusions in the treatment of acute Kawasaki syndrome

    N Engl J Med

    (1991)
  • H yanagawa et al.

    Nationwide surveillance of Kawasaki disease in Japan, 1984 to 1993

    Pediatr Infect Dis J

    (1995)
  • Nakamura Y, Yanagawa H, and Kawasaki T. Temporal and geographical clustering of Kawasaki disease in Japan. In: Shulman...
  • EJ Tizard et al.

    Clinical aspects of 100 patients with Kawasaki disease

    Arch Dis Child

    (1991)
  • J Abe et al.

    Selective expansion of T cells expressing T-cell receptor variable regions Vβ2 and Vβ8 in Kawasaki disease

    Prov Natl Acad Sci USA

    (1992)
  • J Abe et al.

    Characterization of T cell repertoire changes in acute Kawasaki disease

    J Exp Med

    (1993)
  • N Curtis et al.

    Evidence for a superantigen mediated process in Kawasaki disease

    Arch Dis Child

    (1995)
  • Y Yamashiro et al.

    Selective increase of Vβ2+ T cells in the small intestinal mucosa in Kawasaki disease

    Pediatr Res

    (1996)
  • BA Peitra et al.

    TCR Vβ family repertoire and T cell activation markers in Kawasaki disease

    J Immunol

    (1994)
  • Cited by (8)

    • Kawasaki disease

      2004, Archives of Disease in Childhood: Education and Practice Edition
    • The pathogenesis of Kawasaki disease and superantigens

      1999, Japanese Journal of Infectious Diseases
    • The role of superantigens in vasculitis

      1999, Current Opinion in Rheumatology
    View all citing articles on Scopus
    View full text