Circulating CD4⁺CD8⁺ T lymphocytes in patients with Kawasaki disease

Abstract

We serially monitored cell surface antigen expression on mononuclear cells in peripheral blood isolated from patients with Kawasaki disease (KD), and found, for the first time, that a markedly increased number of CD4⁺CD8⁺ T lymphocytes was present in some of the patients (11 of the 24 cases). The cases of five of these 11 patients were complicated with coronary artery lesion (CAL); the 13 patients with normal numbers of CD4⁺CD8⁺ T lymphocytes did not have CAL. The patients' age, sex and grade of systemic inflammation evaluated by peripheral leucocyte count and serum C-reactive protein levels were not correlated to the number of CD4⁺CD8⁺ T lymphocytes. Other cell surface antigen characteristics of the CD4⁺CD8⁺ T lymphocytes included CD3⁺, CD45RA⁺, CD45RO⁺, CD16⁻, and HLA-DR⁺. These results indicate that the surface antigen characteristics of the KD peripheral blood examined were the same as those of Epstein–Barr virus infection without CD45RA⁺. These findings provide useful information for the analysis of the pathogenesis of KD.

INTRODUCTION

Kawasaki disease (KD) [1] is a syndrome with fever that occurs mainly in childhood, with unknown aetiology. Approximately 10% of patients with KD develop arteritis of the coronary vessels over the course of the disease process [2]. A staphylococcal toxin has recently received special attention as a causative substance of KD [3]. It has also been reported that an abnormal increase of various circulating cytokines was observed in KD, along with an increase in the number of CD4⁺ T cells and activated monocytes in peripheral blood at the acute phase of disease development [4–8].

Recent advances in flow cytometry have made it possible to identify cell surface markers by double- and triple-staining methods. The identification of cell surface markers on T cells revealed CD4⁺CD8⁻ and CD4⁻CD8⁺ T cells in the peripheral blood, in addition to a small number of CD4⁺CD8⁺ and CD4⁻CD8⁻ T cells [9, 10]. An abnormal increase in the number of CD4⁺CD8⁺ T cells in the peripheral blood is known to occur in cord blood [11, 12] and in patients with certain diseases including myasthenia gravis [13, 14], idiopathic thrombocytopenic purpura [15], multiple sclerosis [16], Behçet's syndrome [17] and infectious mononucleosis [18].

In this study, we identified (for the first time) an abnormal increase in the number of CD4⁺CD8⁺ T cells in the peripheral blood among some KD patients at the second to third week from the onset of the disease.

PATIENTS AND METHODS

KD patients and controls

Twenty-four subjects (4 months to 6 years old with an average age of 2 years 7 months) were selected from among the KD patients who had been admitted to Dokkyo University Hospital between November 1994 and December 1996. The blood samples, laboratory test data during the course of the disease and follow-up records of all patients were available for this study. All subjects received aspirin (30 mg/kg per day) and high-dose gamma-globulin (400 mg/kg × 5 days). Ninety children from whom blood samples were available for this study, from 3 months to 5 years old (average 1 year 11 months old) were selected as controls from the out-patient population during the same period. Subjects with haematological diseases and with known disorders accompanied by increased numbers of CD4⁺CD8⁺ T cells, including infectious mononucleosis, were excluded from this study. One patient (female, 7 years old) with Guillain–Barré syndrome who received high-dose gamma-globulin treatment (400 mg/kg × 5 days) was selected as a disease control, and immunostaining of the peripheral blood from this patient was conducted.

Blood samples from all patients were obtained after informed consent was given by the parents.

Monoclonal antibodies

FITC-conjugated anti-CD4 mouse MoAb (Leu-3a; Becton Dickinson, San Jose, CA) and PE-conjugated anti-CD8 mouse MoAb (Leu-2b; Becton Dickinson) were used for double-colour staining. Three-colour staining was performed by the following combinations. For detection of CD3, CD16, HLA-DR, CD45RA or CD45RO antigen within CD4⁺CD8⁺ cells, PE-cyanine 5-conjugated anti-CD4 mouse MoAb (Pharmingen, San Diego, CA), PE-conjugated anti-CD8 antibody and FITC-conjugated anti-CD3 (anti-Leu-4; Becton Dickinson), CD16 (anti-Leu-11c; Becton Dickinson), HLA-DR (Becton Dickinson), CD45RA (HI100; Pharmingen) or CD45RO antibody (UCHL 1; Pharmingen) were used, respectively. Mouse IgG1 and IgG2a antibodies (Becton Dickinson) were used as controls.

Cell surface marker analysis

Heparin-added peripheral blood (50 μl) and 10 μl of the respective staining MoAbs were mixed and incubated for 20 min at 4°C, followed by haemolysis of the erythrocytes. After washing, a flow cytometric analysis was conducted to investigate the surface markers by setting a gate specific for lymphocytes on a FACScan (Becton Dickinson) flow cytometer. In the triple-staining, the gate was first set for lymphocytes, and then for CD4⁺CD8⁺ cells to investigate the surface markers within these cell populations.

Evaluation of coronary artery lesions

Echocardiography was performed every 2–3 days from the time of hospitalization. The echocardiographic criteria used were those of the Kawasaki disease Research Committee [19], namely coronary arteries with diameters of 3 mm or more were defined as coronary artery lesions (CAL).

Measurement of peripheral leucocyte counts and determination of C-reaction protein

The number of peripheral leucocytes was analysed by a Coulter automated blood counter, Model STKS (Miami, FL) and the C-reactive protein (CRP) was analysed by a Hitachi Autoanalyzer, Model 7450 (Tokyo, Japan).

Statistical analysis

Statistical analyses were conducted using the Mann–Whitney U-test and Fisher's exact probability test, and P < 0.05 was considered significant.

RESULTS

Circulating CD4⁺CD8⁺ cells in controls

In all 90 controls patients, the number of CD4⁺CD8⁺ cells was 2.15 ± 0.92% (mean ± s.d.) of the total T cells. Based on this result, the normal value for the number of CD4⁺CD8⁺ cells in this study was set at not more than 4.9% (within +3 s.d.) (Fig. 1).

Fig. 1.

Maximum ratio of CD4⁺CD8⁺ cells in the Kawasaki disease patients and control children.

Circulating CD4⁺CD8⁺ cells in KD

Eleven of the 24 KD patients had > 4.9% CD4⁺CD8⁺ cells during the course of the disease (Fig. 1).

All of the CD4⁺ cells among the CD4⁺CD8⁺ cells were CD4⁺CD8⁺. In contrast, the CD8⁺ cells were composed of both CD4⁺CD8⁺ and CD4⁻CD8⁺ cells. The fluorescent intensity against anti-CD8 antibody of the CD4⁺CD8⁺ cells was weaker than that of the CD4⁻CD8⁺ cells (Fig. 2).

Fig. 2.

Representative patterns of CD4 (abscissa) and CD8 (ordinate) determinant expression by (a) normal peripheral blood lymphocytes, (b) lymphocytes from patient 1 with a large number of CD4⁺CD8⁺ cells on day 112 of illness, (c) lymphocytes from patient 2 with a large number of CD4⁺CD8⁺ cells on day 26 of illness, (d) lymphocytes from patient 3 with a large number of CD4⁺CD8⁺ cells on day 20 of illness.

In the sequential analysis of the number of CD4⁺CD8⁺ cells in the peripheral blood from the KD patients, three of 11 KD patients with > 4.9% CD4⁺CD8⁺ cells showed high values over a 1-year period. Of the three patients with a prolonged increase in number of CD4⁺CD8⁺ cells, one had CAL. The other eight of these 11 patients showed a transient increase, peaking between weeks 2 and 3 after the onset of disease. Also, the patients whose CD4⁺CD8⁺ population was not more than 4.9% at the highest point showed a transient increase, peaking between weeks 2 and 3 after the onset of disease (Fig. 3).

Fig. 3.

Serial examinations of the ratio of CD4⁺CD8⁺ cells in the 24 Kawasaki disease patients.

The cases of five of the 24 KD patients were complicated with CAL. The intergroup comparison regarding the incidence of CAL between the group with high CD4⁺CD8⁺ cell numbers (n = 11) and the group with normal value showed that the rate in the group with high CD4⁺CD8⁺ cells was significantly higher than that in the group with normal values. No significant differences between the two groups were observed in age, gender distribution, peripheral leucocyte counts before treatment, or serum CRP values (Table 1).

Table 1.

Clinical features of the cases with high numbers of CD4⁺CD8⁺ T lymphocytes and with normal values of CD4⁺CD8⁺ T lymphocytes

Phenotypical analysis of CD4⁺CD8⁺ cells

The triple-staining showed that all of the CD4⁺CD8⁺ cells were CD3⁺ T cells. The results of the additional cell surface marker tests were as follows: positive for activating antigens (HLA-DR⁺), negative for natural killer (NK) cell antigen (CD16⁻), and positive for naive cell marker (CD45RA⁺) and memory cell marker (CD45RO⁺) (Fig. 4).

Fig. 4.

Representative results of the phenotypical analysis of CD4⁺CD8⁺ cells from the Kawasaki disease patients by triple-colour flow cytometry. An electronic gate was set on CD4⁺CD8⁺ cells in order to analyse the expression of the other surface antigens. Patients 1 and 2 are the same as those described in Fig. 2. Abscissa, Fluorescence intensity (log scale); ordinate, relative cell number. *Percentage of positive cells in CD4⁺CD8⁺ cells.

Effects on the number of circulating CD4⁺CD8⁺ cells by high dose gamma-globulin treatment

In the patient with Guillain–Barré syndrome who had received a high dose of gamma-globulin, the number of CD4⁺CD8⁺ cells in all T cells was low (1.2–3.5%) (data not shown).

DISCUSSION

CD4⁺CD8⁺ T cells are present in small amounts (2–3%) in normal individuals [9], and their level occasionally reaches 10% [18, 20, 21]. Their number is also known to be high in cord blood cells [11, 12]. The number of CD4⁺CD8⁺ T cells is relatively high in patients with myasthenia gravis [13, 14], idiopathic thrombocytopenic purpura [15], multiple sclerosis [16], Behçet's syndrome [17] and infectious mononucleosis [18]. The present study demonstrated that the number of CD4⁺CD8⁺ T cells in the peripheral blood of some KD patients is also high. It is possible that the administration of a large dose of gamma-globulin to KD patients may facilitate an increase in the numbers of CD4⁺CD8⁺ T cells in the peripheral blood. However, the administration of gamma-globulin may not affect the expression of surface antigens on T cells, since the number of CD4⁺CD8⁺ T cells was not increased in the present patient with Guillain–Barré syndrome who received a high dose of gamma-globulin.

We observed an association between the complication of CAL and the abnormal increase of T cell surface markers, since abnormal increases of CD4⁺CD8⁺ T cells were observed only in the group with this complication. However, the ages and gender of the patients did not affect the abnormal increase of CD4⁺CD8⁺ T cells. Peripheral leucocyte count and serum CRP values (which are used as indicators of the degree of inflammation) before treatment did not correlate with the abnormal increase of CD4⁺CD8⁺ T cells.

The antigenic characteristics of CD4⁺CD8⁺ T cells in healthy individuals are positive for NK cell antigens and negative for activation antigens, and both CD45RA and CD45RO antigens are coexpressed on the surface of these T cells, suggesting that CD4⁺CD8⁺ T cells are in a stage between naive and memory cells [21]. It has been reported that CD4⁺CD8⁺ T cells from patients with infectious mononucleosis caused by Epstein–Barr virus (EBV) infection were negative for NK cell antigens and positive for activation antigens and memory-type T cells, since they were positive for CD45RO [18]. The CD4⁺CD8⁺ T cells in the present KD patients showed activation antigens (+) and NK cell surface antigens (−), results which were similar to those in the patients with EBV infection. However, these T cells were between naive and memory T cells, since both CD45RA and CD45RO were coexpressed on these T cells; a similar result was observed in healthy individuals.

The correlation between virus infection and the appearance of CD4⁺CD8⁺ T cells was also observed in patients with disorders other than EBV infection. Furukawa et al. reported that CD4⁺CD8⁺ T cells were detected in the peripheral blood of patients with congenital immunodeficiency diseases complicated with human T cell lymphotropic virus type I infection [22]. Hayashi et al. demonstrated that a CD4⁺CD8⁺ T cell line could be established when normal cord leucocytes were co-cultured with a human T lymphotropic virus type II-producing simian leucocyte cell line (Si-IIA) [23]. Lusso et al. reported that the expression of CD4 antigen was induced when CD8⁺ T cells were infected with human herpes virus 6 [24]. On the basis of these results, it is possible that the abnormal increase of CD4⁺CD8⁺ T cells in the peripheral blood leucocytes of KD patients correlates with a virus infection.

Regarding the mechanisms that lead mature T cells to become CD4⁺CD8⁺ T cells, several studies suggest the contribution of IL-4 to this process. This concept is supported by the following observations: the expression of CD8⁺ antigen was induced in CD4⁺ T cell clones stimulated by IL-4 [25], and the increase of IL-4 production in CD4⁺CD8⁺ T cell lines, established by co-culturing normal cord leucocytes and Si-IIA, was observed [23]. Regarding the correlation between KD and IL-4, we have found that the plasma IL-4 concentration reaches a maximum in the acute KD phase, and returns to normal in the convalescent phase [6]. Taken together, these findings suggest that the abnormal increase of CD4⁺CD8⁺ T cell numbers in the peripheral blood leucocytes of KD patients might be correlated to the increases of IL-4 production.

In conclusion, we observed (for the first time) a new immune response, i.e. there was an abnormal increase of CD4⁺CD8⁺ T cells in the peripheral leucocytes in some KD patients during the course of the disease. We found that among the KD patients with an abnormal increase of CD4⁺CD8⁺ T cells, the rate of CAL complication was high. The antigenic characteristics of the CD4⁺CD8⁺ T cells are activation antigen (+) and NK cell antigen (−), results similar to those observed in patients with EBV infection. However, both CD45RA and CD45RO are co-expressed on these T cells in KD patients, a finding which is similar to that observed in healthy individuals but different from that in patients with EBV infection. These results yield critical information to investigate the pathogenesis of KD.