Kawasaki disease (KD) is an acute systemic vasculitisoccuring in infants and children younger than 4 years ofage.1 The target vessels are small and medium-sized arteries,especially coronary arteries. The etiology of KD is stillunknown and the pathophysiology of systemic vasculardamage is also unresolved. Although high-dose intravenousimmunoglobulin therapy (IVIG) is effective for reducing theincidence of coronary arterial lesions (CAL),2,3 the mechanismfor the effectiveness of IVIG is also still unclear.4–6 The earlystage of KD is marked by increased levels of immatureneutrophils, such as band neutrophils and metamyelocytes.Several reports have indicated that the decrease in neutro-phils after IVIG was very important for reducing theincidence of CAL.7 However, it was shown by histologicexamination that a few neutrophils had infiltrated in themedia of coronary arteries in an autopsy case who died after9 days of illness.8 This discrepancy is the reason why therole of neutrophils in vascular damage in KD is not clear.

Neutrophil-activation cytokines, such as interleukin (IL)-8and granulocyte colony stimulating factor (G-CSF), areproduced by monocytes and vascular endothelial cellsstimulated with lipopolysaccharide (LPS), tumor necrosisfactor (TNF)-α and IL-1.9–12 However, the pathophysiologicroles of IL-8 and G-CSF in infectious and inflammatorydiseases are still poorly understood. There have been noreports in which these two cytokines were simultaneouslyexamined in the clinical course of KD.13–15 In the presentstudy, we examined serial changes in neutrophil counts andserum levels of IL-8 and G-CSF in an attempt to elucidatethe role of neutrophils in vascular damage in KD.

Methods

From October 1994 to June 1998, a total of 52 patients withKD were enrolled in the present study.16 The ages of thesubjects ranged from 2 months to 6 years (mean age 2 yearsand 1 month). Forty-eight cases were treated with both IVIG(a daily dose of 200–400 mg/kg for 5 consecutive days) andacetyl salicylic acid (ASA; 30–50 mg/kg per day), whilefour cases were treated with ASA alone (30–50 mg/kg perday). Informed consent was given by parents for all caseswho received IVIG. In all KD patients, two-dimensional

Pediatrics International (2001) 43, 115–119

Original Article

Serum levels of neutrophil activation cytokines in Kawasaki disease

HIROYUKI SUZUKI, EISAKU NODA, MASAKAZU MIYAWAKI, TAKASHI TAKEUCHI,SHIGERU UEMURA AND MICHIO KOIKE

Department of Pediatrics, Wakayama Medical University, Wakayama, Japan

Abstract BBaacckkggrroouunndd: The aim of the present study was to investigate whether neutrophils are early effector cells forvascular endothelial damage in the acute phase of Kawasaki disease (KD) by examining serial changes inneutrophil counts and serum levels of neutrophil activation cytokines, such as granulocyte colony stimulatingfactor (G-CSF) and interleukin (IL)-8.MMeetthhooddss: From October 1994 to June 1998, a total of 52 patients with KD were included in the study.Thirty-three patients had some infectious diseases, while 20 healthy children served as control subjects.Serial changes in neutrophils were counted by the optimal Wright–Giemsa staining method and serum levelsof IL-8 and G-CSF in patients with KD were measured by an enzyme-linked immunosorbent assay system.RReessuullttss: Serum G-CSF levels both before and after intravenous immunoglobulin therapy (IVIG; P<0.05)and neutrophil counts after IVIG (P<0.005) were higher in KD patients with coronary arterial lesions(CAL) than those without CAL. However, serum IL-8 levels before and after IVIG showed no significantdifferences in these two groups.CCoonncclluussiioonnss: These data suggest that neutrophils may be important as early effector cells for vascularendothelial damage and that G-CSF may play a more important role than IL-8 in KD.

Key words granulocyte colony stimulating factor, interleukin-8, Kawasaki disease, neutrophil.

Correspondence: Hiroyuki Suzuki, Department of Pediatrics,Wakayama Medical University, 811-1 Kimiidera, Wakayama 641-0012, Japan. Email: hsuzuki@mail.wakayama-med.ac.jp

Received 14 February 2000; revised 29 September 2000;accepted 2 October 2000.

echocardiography was performed at least twice a weekduring the first month after the onset of KD. We observed 12cases with CAL, two cases had transient dilatation andanother 10 had coronary arterial aneurysms (not giantaneurysm). Twenty-four patients (age range 2 months to 11years; mean age 3 years and 5 months) had some infectiousdisease, while 20 healthy children (age range 2 months to4 years; mean age 2 years and 10 months) served as controlsubjects. Neutrophils were counted by the optimal Wright–Giemsa staining method. Serum samples from peripheralblood were obtained at least three times; that is, before IVIG (3–7 days), after IVIG (8–12 days) and during theconvalescent stage of the illness (20–25 days). Control serafrom patients with infectious diseases were obtained duringthe febrile periods (disease day 1–3). Sera were separated bycold centrifugation and stored at –40°C until analysis. Serumlevels of IL-8 and G-CSF were measured by an enzyme-linked immunosorbent assay system (IL-8: TFB, Tokyo,Japan; G-CSF: Amersham Pharmacia Biotech UK,Buckinghamshire, UK). Statistical analysis was performedby unpaired Student’s t-test (neutrophils) and the Mann–Whitney U-test (G-CSF and IL-8).

Results

Neutrophils

Serial changes in neutrophil counts in peripheral venousblood of all cases with KD are shown in Fig. 1a. In patientswithout CAL, the increased neutrophil counts before IVIGpromptly decreased after IVIG. In contrast, patients withCAL maintained high neutrophil counts and, in someinstances, neutrophil counts increased further after IVIG.Thus, although there was no significant difference in neutro-phil counts before IVIG between patients with and withoutCAL, neutrophil counts after IVIG were higher in patientswith CAL than in those without CAL (P<0.005; Table 1).

Granulocyte colony stimulating factor

Serum G-CSF levels in 12 of 15 healthy children were lessthan 50 pg/mL (Fig. 2a). In KD patients, before IVIG, serumG-CSF levels were higher than 50 pg/mL except for twocases. After IVIG, serum G-CSF levels decreased promptly,except for two of nine cases with CAL (Fig. 1b). Serum G-CSF was significantly higher in the acute phase (beforeIVIG; mean (~SD) 589.4~655.1 pg/mL) than in the con-valescent phase (32.7~87.4 pg/mL; P<0.01) and in healthycontrols (33.2~20.5 pg/mL; P<0.01). We compared serumG-CSF levels in patients with and without CAL before andafter IVIG. Serum G-CSF levels before and after IVIG were

116 H Suzuki et al.

Fig. 1 (a) In patients without coronary arterial lesions (CAL;�---�), neutrophil counts before intravenous immunoglobulin(IVIG) promptly decreased after IVIG. In contrast, patients withCAL (�—�) maintained a high neutrophil count and, in someinstances, the neutrophil count increased further after IVIG. (b) Before IVIG, serum granulocyte colony stimulating factor(G-CSF) levels were all higher than 50 pg/mL except for twocases. After IVIG, serum G-CSF levels decreased promptlyexcept for two of nine cases with CAL. (�---�), patients withCAL; (�—�), patients without CAL. (c) Serum interleukin(IL)-8 levels in seven of 24 patients with Kawasaki disease(KD) before IVIG were higher than 20 pg/mL. Three of theseven cases were patients with CAL. After IVIG, serum IL-8levels in all patients with KD decreased promptly. (�---�),patients with CAL; (�—�), patients without CAL.

higher in patients with CAL than those without CAL(P<0.05; Table 1). No significant correlation was observedbetween serum G-CSF levels and neutrophil counts in anystage of KD. Although serum G-CSF levels in asepticmenigitis were as low as those of healthy children, those ofbacterial meningitis due to Hemophilus influenzae, measlesand hand–foot–mouth disease increased moderately (Fig. 2a).

Interleukin-8

Serum IL-8 levels in 18 of 20 healthy children were lessthan 20 pg/mL. Serum IL-8 levels in seven of 24 patientswith KD before IVIG were higher than 20 pg/mL. Three ofseven cases showed CAL. After IVIG, serum IL-8 levels inKD patients decreased promptly (Fig. 1c). Serum IL-8 levelswere significant higher in the acute phase (before IVIG22.8~24.7 pg/mL) than during the convalescent phase(8.9~6.5 pg/mL; P<0.05) and in healthy controls(10.3~11.2 pg/mL; P<0.05). We compared serum IL-8levels in patients with and without CAL before and afterIVIG and there were no significant differences either beforeor after IVIG in these two groups (Table 1). In addition, nosignificant correlation was observed between serum IL-8levels and neutrophil counts in any stage of KD. Serum IL-8levels were high in four of five cases with bacterialmeningitis due to H. influenzae. Although serum IL-8 levelsin four cases with measles were slightly increased, those ofseveral viral infectious diseases, such as hand–foot–mouthdisease and aseptic meningitis were as low as those ofhealthy children. (Fig. 2b).

Discussion

In the present study, we have demonstrated that serum G-CSFlevels before and after IVIG were significantly higher in KD

Neutrophil-activation cytokines in KD 117

Table 1 Neutrophil counts and granulocyte colony stimulatingfactor and interleukin-8 levels before and after intravenousimmunoglobulin therapy in patients with and without coronaryarterial lesions

CAL (+) CAL (–) P(n=12) (n=36)

Neutrophil count (/µL)Before IVIG 9469~4238 10 550~4168 NSAfter IVIG 15 252~8942 5536~3464 <0.005*

Granulocyte colony stimulating factor (pg/mL)Before IVIG 1249~891 342~372 <0.05†

After IVIG 207~256 34.2~78.3 <0.05†

Interleukin-8 (pg/mL)Before IVIG 25.2~22.1 22.0~26.3 NSAfter IVIG 11.6~7.6 9.1~6.8 NS

Data are the mean~SD. *Student’s t-test; †Mann–Whitney U-test.

IVIG, intravenous immunoglobulin. ‘After IVIG’ meansimmediately after 5 consecutive days of IVIG.

Fig. 2 Serum (a) granulocyte colony stimulating factor (G-CSF) and (b) interleukin (IL)-8 levels in patients with variousdiseases. (a) Serum G-CSF levels in 12 of 15 healthy childrenwere less than 50 pg/mL. In Kawasaki disease (KD), serum G-CSF levels before intravenous immunoglobulin (IVIG) werehigher than 50 pg/mL except for two cases. Serum G-CSFlevels in seven cases with coronary arterial lesions (CAL) weremaximum before IVIG; in contrast, those in two cases withCAL increased to maximum levels after IVIG. Although serumG-CSF levels in aseptic menigitis were as low as those ofhealthy children, those of bacterial meningitis due to Hemo-philus influenzae, measles and hand–foot–mouth diseaseincreased moderately. (b) Serum IL-8 levels in 18 of 20 healthychildren were less than 20 pg/mL. In KD, serum IL-8 levelsbefore IVIG were higher than 20 pg/mL in only seven patientswith KD. Serum IL-8 levels in some cases of bacterialmeningitis due to H. influenzae and measles increased, but thosein patients with several different viral infectious diseases, suchas hand–foot–mouth disease and aseptic meningitis, were as lowas those of healthy children. (�), patients without CAL; (�),patients with CAL.

patients with CAL than those without CAL. In addition, wehave also demonstrated that neutrophil counts immediatelyafter IVIG were significantly higher in KD patients with CALthan those without CAL. However, serum IL-8 levels beforeand after IVIG showed no significant difference.

Both G-CSF and IL-8 are major neutrophil activationcytokines. The monocyte/macrophage lineage is one of themost prominent sources of G-CSF and IL-8, but thesefactors can also be produced by vascular endothelial cellsand fibroblasts stimulated with LPS, TNF-α and IL-1.9–12 Inacute KD, previous studies have reported the activation ofmonocytes/macrophages with the secretion of cytokines,such as neopterin,6 monocyte chemoattractant protein-1(MCP-1),15 TNF-α17 and macrophage colony stimulatingfactor (M-CSF).18 These immune conditions may upregulatethe production of G-CSF and IL-8 in acute KD. G-CSFplays an important role in the regulation of granulopoiesisbut, more importantly, it modulates the function and activityof developing and matured neutrophils, such as chemotaxis,phagocytosis and cell surface receptor expression.11,12

Although a number of studies have shed light on thephysiologic roles of G-CSF, the pathophysiologic roles of G-CSF in infectious and inflammatory diseases are stillpoorly understood. There are several reports in which serumlevels of G-CSF have been measured in trauma and variousinfectious diseases in adults.19–21 Our study has revealed thatserum G-CSF levels in healthy children were less than50 pg/mL, but those in patients with bacterial meningitis andother viral infectious diseases, such as measles and hand–foot–mouth disease, increased moderately. These findingssuggest that serum G-CSF levels increase in not only bacterialbut also viral infectious diseases. In KD, serum G-CSFlevels after IVIG in patients with CAL are still high, butthey decrease to one-sixth of the levels before IVIG. Incontrast, neutrophil counts increase 1.5-fold after IVIGwhen CAL begins to appear. Although neutrophil countsmay be controlled by many other factors in addition to G-CSF, these findings suggest that neutrophils are earlyeffector cells for vascular endothelial damage22 and that G-CSF plays an important role not only in the regulation ofgranulopoiesis but also in the modulation of the function andactivity of neutrophils in KD.14,23,24

Several studies have already reported IL-8 in associationwith various inflammatory diseases, including rheumatoidarthritis,25 meningitis,26–28 psoriasis29 and reperfusion injury.30

In the present study, serum IL-8 levels were high in patientswith meningitis due to H. influenzae26 and measles, but notas high as in those patients with some other viral infection.In KD, there was no significant difference in serum IL-8levels between patients with and without CAL. Lin et al.13

have reported that there is a significant difference in serumIL-8 levels between these two groups. A possible explan-zation for this disparity is that the patients in the study of

Lin et al.13 were not treated with IVIG but only with ASAalone. In addition, Lin et al.13 measured serum IL-8 levels byradioimmunoassay and the measured values were shown asseveral thousand pg/mL; in contrast, we measured theselevels by ELISA and the measured values were shown asless than 100 pg/mL. It is also possible that these bigdifferences in serum IL-8 levels may create a discrepancy inthe relationship between serum IL-8 levels and CAL.

Recent studies31,32 have shown that ulinastatin, a potentinhibitor of neutrophil elastase, may be effective in reducingthe incidence of CAL in KD. These reports may be inagreement with our results. In conclusion, our resultssuggest that neutrophils may be important as early effectorcells for vascular endothelial damage and that G-CSF mayplay a more important role than IL-8 in KD.

References

1 Kawasaki T. Acute febrile mucocutaneous syndrome withlymphoid involvement with specific desquamation of fingersand toes in children: Clinical observation of 50 cases. Jpn. J.Allergol. 1967; 16: 178–222.

2 Furusho K, Kamiya T, Nakano H et al. High dose intravenousgammaglobulin for Kawasaki disease. Lancet 1984; ii:1055–8.

3 Newburger JW, Takahashi M, Burns JC et al. The treatment of Kawasaki syndrome with intravenous gammaglobulin. N. Engl. J. Med. 1986; 315: 341–7.

4 Suzuki H, Iizuka T, Uemura S, Koike M, Maeda J. Thepathophysiological significance of serum tumor necrosis factorin Kawasaki disease. In: Kawasaki T (ed.). Proceedings of the Third International Kawasaki Disease Symposium, 29 November–2 December, 1998, Tokyo, Japan. Japan HeartFoundation, Tokyo, 1988; 117–19.

5 Suzuki H, Uemura S, Tone S et al. The effects of immuno-globulin and interferon-γ on the production of tumor necrosisfactor-α and interleukin-β by peripheral blood monocytes inthe acute phase of Kawasaki disease. Eur. J. Pediatr. 1996;155: 291–6.

6 Iizuka T, Minatogawa Y, Suzuki H et al. Urinary neopterin asa predictive marker of coronary artery abnormalities inKawasaki syndrome. Clin. Chem. 1993; 39: 600–4.

7 Furusho K. Changes of leukocyte counts in children withKawasaki disease under different treatments. Prog. Med. 1988;8: 86–90 (in Japanese).

8 Masuda H, Naoe S, Tanaka N. A pathological study ofcoronary artery in Kawasaki disease (MCLS) with specialreference to morphogenesis of aneurysm. Angiology 1981; 21:899–912 (in Japanese).

9 Baggiolini M, Dewald B, Moser B. Interleukin-8 and relatedchemotactic cytokines: CXC and CC chemokines. Adv.Immunol. 1994; 55: 97–179.

10 Hoch RC, Schraufstatter IU, Cochrane CG. In vivo, in vitro,and molecular aspects of interleukin-8 and the interleukin-8receptors. J. Lab. Clin. Med. 1996; 128: 134–45.

11 Dale DC, Liles WC, Summer WR, Nelson S. Granulocytecolony-stimulating factor: Role and relationships in infectiousdisease. J. Infect. Dis. 1995; 172: 1061–75.

118 H Suzuki et al.

12 Demetri GD, Griffin JD. Granulocyte colony-stimulatingfactor its receptor. Blood 1991; 78: 2791–808.

13 Lin CY, Lin CC, Hwang B, Chiang B. Serial changes of seruminterleukin- 6, interleukin-8, and tumor necrosis factor alphaamong patients with Kawasaki disease. J. Pediatr. 1992; 121:924–6.

14 Inoue Y, Kato M, Kobayashi T, Shinohara M, Sone K,Morikawa A. Increased circulating granulocyte colony-stimulating factor in acute Kawasaki disease. Pediatr. Int.1999; 41: 330–3.

15 Terai M, Jibiki T, Harada A et al. Dramatic decrease ofcirculating levels of monocyte chemoattractant protein-1 inKawasaki disease after gamma globulin treatment. J. Leukoc.Biol. 1999; 65: 566–72.

16 Research Committee on Kawasaki Disease. Report of Sub-committee on Standadization of Diagnostic Criteria andReporting of Coronary Artery Lesion in Kawasaki Disease.Ministry of Health and Welfare, Tokyo, 1984.

17 Matsubara T, Furukawa S, Yabuta K. Serum levels of tumornecrosis factor, interleukin 2 receptor, and interferon-gammain Kawasaki disease involved coronary-artery lesions. Clin.Immunol. Immunopathol. 1990; 56: 29–36.

18 Igarashi H, Hatake K, Tomizuka H, Yamada M, Gunji Y,Momoi MY. High serum levels of M-CSF and G-CSF inKawasaki disease. Br. J. Haematol. 1999; 105: 613–15.

19 Tanaka H, Ishikawa K, Nishino M, Shimazu T, Yoshioka T.Changes in granulocyte colony-stimulating factor concentra-tion in patients with trauma and sepsis. J. Trauma 1996; 40:718–26.

20 Kawakami M, Tsutsumi H, Kumakawa T et al. Levels ofserum granulocyte colony-stimulating fator in patients withinfections. Blood 1990; 76: 1962–4.

21 Omori F, Okamura S, Shimoda K, Otsuka T, Harada M, NihoY. Levels of human serum granulocyte colony-stimulatingfactor and granulocyte–macrophage colony-stimulating factorunder pathological conditions. Biotherapy 1992; 4: 147–53.

22 Hirao J, Yamashita T. Circulating soluble L-selectin levels inKawasaki disease with coronary artery lesions. Pediatr. Int.1997; 39: 290–2.

23 Niwa Y, Somia K. Enhanced neutrophilic functions inmucocutaneous lymph node syndrome (MCLS): With specialreference to the possible role of increased oxygen intermediategeneration in the pathogenesis of coronary thromboarteritis. J. Pediatr. 1984; 104: 56–60.

24 Okada M, Satoh T, Hayashi T. Effects of intravenous γ-globulin on neutrophil function in Kawasaki disease. ActaPaediatr. Jpn. 1991; 33: 785–90.

25 Seitz M, Dewald B, Ceska M, Gerber N, Baggiolini M.Interleukin-8 in inflammatory rheumatic diseases: Synovialfluid levels, relation to rheumatoid factors, production bymononuclear cells, and effects of gold sodium thiomalate andmethotrexate. Rheumatol. Int. 1992; 12: 159–64.

26 Noda A, Suzuki H, Miyawaki M et al. The interleukin-8concentrations in serum and cerebrospinal fluid from patientswith Haemophilus influenzae meningitis and aseptic meningitis.Infect. Immun. Child. 1997; 9: 249–53 (in Japanese).

27 Ishigro Y, Suzuki Y, Inaba A et al. The production of IL-8 incerebrospinal fluid in aseptic meningitis of children. Clin. Exp.Immunol. 1997; 109: 426–30.

28 Yokoyama T, Oda M, Akazai A, Kitamura T, Seino Y.Cerebrospinal fluid levels of interleukin-8 in asepticmeningits: Relationship of IL-8 to neutrophil migration. Infect.Immun. Child. 1996; 8: 209–12 (in Japanese).

29 Sticherling M, Bornscheuer E, Schroder J-M, Christophers E.Localization of neutrophil-activating peptide-1/interleukin-8immunoreactivity in normal and psoriatic skin. J. Invest.Dermatol. 1991; 96: 26–30.

30 Frering B, Philip I, Dehoux M, Rolland C, Langlois JM,Desmonts JM. Circulating cytokines in patients undergoingnormothermic cardiopulomonary bypass. J. Thorac. Cardiovasc.Surg. 1994; 108: 636–41.

31 Okada M, Nakai S, Kobayashi Y, Satoh T. Effects ofintravenous gamma globulin and ulinastatin on patients withKawasaki disease and predicted giant coronary arteryaneurysm. J. Jpn. Pediatr. 1997; 101: 1165–70 (in Japanese).

32 Nakano M, Yamada S, Hayashi Y, Morishita M, Iwaki T,Toyota M. New therapy for Kawasaki disease: The efficacy ofulinastatin. J. Jpn. Pediatr. 1996; 49: 435–44 (in Japanese).

Neutrophil-activation cytokines in KD 119