American Journal of Hematology 40:295-298 (1992)
Interleukin-5 in Eosinophilic Gastroenteritis Takayuki Takahashi, Kishiko Nakamura, Satoru Nishikawa, Reiko Tsuyuoka, Akira Suzuki, Masao Murakami, Masahiro Amenomori, Yoshiaki Okuno, and Hiroo lmura Second Department of Internal Medicine, Faculty of Medicine (T.T..S.N.,R.T.,A.S.,M.M.,M.A..Y.O.,H.I.) and College of Medical Technology (K.N.), Kyoto University, Kyoto, Japan
During a 4-year period a 28-year-old female had 4 episodes of eosinophilia of over 10,000/ pl ;these episodes were associated with nausea, vomiting, diarrhea, and abdominal pain. On one occasion, she had ascites and pleural effusion which contained numerous mature eosinophils. On each occasion, these attacks disappeared within several weeks without any specific treatment. A diagnosis of eosinophilic gastroenteritis was made. A plasma sample obtained during the eosinophilia generated in vitro eosinophilic colonies when added to granulocyte/macrophage-progenitor (CFU-GM) cultures without exogeneous growth factors. Colony formation was inhibited by anti-interleukin-5 (IL-5) antibody but not by antibodies toward IL-3, granulocyte colony-stimulating factor (G-CSF) or GM-CSF. A high plasma interleukin-5 (IL-5) level was noted when measured by enzyme-linked immunosorbent assay, while IL-3, G-CSF, and GM-CSF were undetectable. During remission the plasma gave negative results both for colony formation and IL-5 level. These results indicate that the eosinophilia of this disease is mediated by IL-5. c 1992 Wiley-Liss, Inc.
Key words: eosinophilic colonies, anti-IL-5 antibody, IL-5
INTRODUCTION
Eosinophilic gastroenteritis is characterized by intermittent abdominal distress, such as nausea, vomiting. diarrhea. and abdominal pain, which is accompanied by increased numbers of mature eosinophils in the peripheral blood [ 1-31. Ascites or pleural effusion is frequently observed, and many mature eosinophils are present in these fluids (21. In most patients, the disease resolves spontaneously; however, it recurs repeatedly. Although the pathogenesis of this disease is unknown, Klein et al. [2] speculated that food allergy is the cause of this disease. Recent studies have shown that three cytokines, granulocyte/macrophage colony-stimulating factor (GMCSF) 141,interleukin-3 (IL-3) [ 5 ] , and lL-5 [6], are involved in eosinophilopoiesis. Of these 3 cytokines, 1L-5 acts on the most mature precursors of eosinophils [7], and induces only eosinophilic colonies in vitro [6], in contrast to the other two, which generate colonies of multiple cell lineages [4.5]. Furthermore, IL-5 has been demonstrated to be the main factor involved in reactive eosinophilia such as Kimura's disease, reactive hypereosinophilic syndrome, and eosinophilic angioedema [S-lo]. In the present study, we found eosinophilic (Eo)-CSF activity i n
0 1992 Wiley-Liss, Inc.
the plasma from a patient with eosinophilic gastroenteritis and identified this activity as IL-5.
CASE REPORT A 28-year-old female was admitted to our hospital because of nausea, vomiting, diarrhea. and abdominal pain in July 1989. In the 2-year period before admission, she had had two similar episodes associated with eosinophilia. which resolved spontaneously in a few weeks. She had not taken any medication prior to the onset of abdominal distress in each occasion. She had had a skin rash due to food allergy to mackerel twice when she was a teenager. On physical examination, ascites was noted. Superficial lymph node enlargement and hepatosplenomegaly were not noted. Laboratory examination revealed a white cell count of 24,0OO/plwith 74% mature eosinophils, 16% neutrophils, 1% monocytes, and 9% lymphocytes.
Receives for publication August 29, 1991: accepted January 30, 1Y92. Address reprint requests to Takayuki Takahashi, 1M.D.. The Second Department of Internal Medicine. Faculty of Medicine, Kyoto University, 54 Shogoin Kawahara-cho. Sakyo-ku, Kyoto 606, Japan.
296
Case Report: Takahashi et al.
The hemoglobin level and platelet count were within normal limits. A bone marrow aspirate showed normocellular marrow with 3 1 % eosinophils; however, neither maturation arrest nor dysplastic features were seen. In addition, a hemopoietic progenitor assay of the marrow cells, using the multilineage method [ 1 I], disclosed normal hemopoietic colony formation in number, size. and maturation. Examinations revealed no parasite infection. Computed tomography of the abdomen was non-specific. Endoscopic examination of upper gastrointestinal tract and barium enema were performed 2 weeks after admission; however, no abnormal findings were observed and no tissue biopsy was made. On chest X-ray examination, pleural effusion of the left side was noted. Cytological examination of the pleural fluid showed numerous mature eosinophils. A diagnosis of eosinophilic gastroenteritis was made. The treatment was only supportive (intravenous fluids), and the abdominal distress, eosinophilia, ascites, and pleural effusion disappeared completely 3 weeks after admission. In July 199 I,she had a 4 t h attack of the disease, which also resolved spontaneously. MATERIALS AND METHODS Plasma and Pleural Fluid
Heparinized blood and pleural fluid at various clinical states were obtained from the patient after informed, written consent was obtained. Also pleural fluids were obtained from patients with lung cancer who did not have any eosinophilia or leukocytosis in the process of diagnosis. After centrifugation at 800g for 10 min, plasma and pleural fluid were filtered through a 0.45 - pni niembrane filter (Corning, NY) and kept at -20°C until examination. Assay of Granulocyte/Macrophage Progenitor (CFU-GM)
Bone marrow cells were obtained with informed, written consent from patients with possible early lymphoma as part of the process of clinical staging. The CFU-GM assay used has been described previously [12]. Briefly, non-T, non-adherent bone marrow mononuclear cells ( 1 X 10’ cells/ml) were mixed with Iscove’s modified Dulbecco’s MEM (IMDM) (Gibco, Grand Island, NY), 1 0 30% patient’s plasma, S% FCS, and 0.9% methylcellulose. One-milliliter aliquots were cultured in 35 X 12-mm plastic dishes (Lux, Naperville, IL) for 14 days at 37°C in 5% CO,. Cultures were performed in triplicate. Brownish colonies (eosinophilic colony, Eo colony) of more than 50 eosinophils were counted using an inverted microscope (Olympus Co., Tokyo). Aggregates of 10-49 eosinophils were scored as Eo clusters. Granulocyte (G) or macrophage (M) colonies containing more than 50 cells were scored simultaneously. In some experiments, to confirm the cellular component of Eo colonies, colo-
nies were picked up with a fine-drawn Pasteur pipette, spread out over a slide glass, and stained with Bieblich scarlet. In another set of experiments, prior to making the culture cocktail, patient’s plasma was mixed with antimurine IL-5 monoclonal antibody (MoAb), NC17 (Olympus) [ 131, at a protein concentration of 50 pg/ml, and incubated at room temperature for 30 min. NC17 neutralizes human IL-5 [8,9]. The plasma was also treated with anti-G-CSF MoAb (25 pg/ml; Chugai Pharm. Co., Tokyo) or anti-GM-CSF MoAb (5 Fg/ml; Genzyme, Boston) in the same way. In some experiments, antibody treatment of the plasma was performed using anti-human IL-3 MoAb (10 kg/ml; Genzyme). As a positive control for Eo colony formation, recombinant (r) IL-5 (10 ngiml; Suntory Co., Osaka, Japan) was added to the CFU-GM assay with 30% plasma from a healthy volunteer (normal plasma). Enzyme-Linked lmmunosorbent Assay (ELISA)
The IL-5 concentration in the plasma was measured using an ELISA kit [ 141 kindly provided by Suntory Co. G-CSF concentration was also measured by ELISA [ 151. GM-CSF and IL-3 concentrations were measured using ELISA kits purchased from Genzyme and R&D Systems (Minneapolis, MN), respectively. Statistical Analysis
The statistical significance of the values obtained was evaluated using Student’s t-test. RESULTS Eo-CSF Activity in Plasma and Its Identification
Table I shows the Eo-CSF activity of the patient’s plasma obtained at the time of admission. The plasma generated 13.3 Eo clusters when added to the CFU-GM culture at a volume of 10%. The number of Eo clusters was increased and 10 Eo colonies were generated when the plasma volume was increased to 30%; however, G. or M colonies were rarely seen. Eo colony or cluster formation was inhibited by treatment of the plasma with antiIL-5 MoAb, while anti-G-CSF or anti-GM-CSF MoAbs did not affect Eo colony or cluster formation. Normal plasma (30%) and 10 ng/ml of rIL-5 generated a greater number of Eo colonies or clusters without G , or M colony formation. Eo colony and cluster formation was also inhibited by anti-IL-5 MoAb. Pleural effusion showed lower Eo-CSF activity than plasma; 30% pleural effusion generated nine Eo clusters but not colonies. During disease-free periods, the patient’s plasma did not induce any type of colonies or Eo clusters. While, a plasma sample obtained during the 4th attack of the disease generated 8.7 1.2 (mean 2 SE) Eo colonies per 1 X 10’ bone marrow cells without any
*
Case Report: IL-5 in Eosinophilic Gastroenteritis
297
TABLE 1. Eo-CSF Activity in the Patient’s Plasma and Identification by Antibody Neutralirationt
Number of colonies/l X 10’ BM cells*’ ~~~~
Culture conditions
~
Eo colony*2
10% P-PI.*~,S% FCS 30% P-pl., 5% FCS 30% P-pl., 5% FCS, (Y IL-5 MoAb*’ (50 pg/ml) 30% P-pl., 5% FCS, (Y G-CSF MoAb (25 crgiml) 30% P-PI., 5% FCS, (Y GM-CSF MoAb (5 pgiml) 30% N-PI.*~,5% FCS, rhIL-S*9 30% N-pl., 5% FCS, rhIL-5, (Y IL-SMoAb (50 pglml)
Eo cluster*3 G colony*4 M colony*’
0 10.0 k 1.0 0
13.3 f 2.0 34.0 f 4.6 2.0 f 1.2
0.3 f 0.3 0.7 +- 0.3 0.3 f 0.3
0.7 f 0.3 0.3f 0.3
11.0 f 2.1
31.0 f 4.4
0
0
8.3 k 1.8
29.3 f 4.0
0.3 f 0.3
0
35.7 f 4.0 0
51.7 f 7.6 4.0 k 2.3
1.0 f 0.6 0.7 f 0.3
0.3 f 0.3 0.7 k 0.3
0
?.Data represent themean fSE (n = 3).Nos. in parentheses indicate the protein concentration of the antibody during preincubation (30 min, room temperature) with plasma. * I , non-T, non-adherent bone marrow mononuclear cells; *2, eosinophilic colony (more than 50 cells); *3, eosinophilic cluster (10-49 cells); *4, granulocytic colonies (more than SO cells); *S, macrophage colony (more than SO cells); *6, plasma from the patient during the period of eosinophilia (3rd attack); *7, anti IL-5 monoclonal antibody; *8, plasma from a healthy donor (normal plasma); *9, recombinant human IL-5 ( I 0 ngiml). TABLE II. Concentration 01 IL-5 in Plasma and Pleural Fluid?
Concentration of cytokines (pg/ml) IL-5
IL-3
GM-CSF
G-CSF
1,800+ 104 1,560f 131
Interleukin-5 in Eosinophilic Gastroenteritis Takayuki Takahashi, Kishiko Nakamura, Satoru Nishikawa, Reiko Tsuyuoka, Akira Suzuki, Masao Murakami, Masahiro Amenomori, Yoshiaki Okuno, and Hiroo lmura Second Department of Internal Medicine, Faculty of Medicine (T.T..S.N.,R.T.,A.S.,M.M.,M.A..Y.O.,H.I.) and College of Medical Technology (K.N.), Kyoto University, Kyoto, Japan
During a 4-year period a 28-year-old female had 4 episodes of eosinophilia of over 10,000/ pl ;these episodes were associated with nausea, vomiting, diarrhea, and abdominal pain. On one occasion, she had ascites and pleural effusion which contained numerous mature eosinophils. On each occasion, these attacks disappeared within several weeks without any specific treatment. A diagnosis of eosinophilic gastroenteritis was made. A plasma sample obtained during the eosinophilia generated in vitro eosinophilic colonies when added to granulocyte/macrophage-progenitor (CFU-GM) cultures without exogeneous growth factors. Colony formation was inhibited by anti-interleukin-5 (IL-5) antibody but not by antibodies toward IL-3, granulocyte colony-stimulating factor (G-CSF) or GM-CSF. A high plasma interleukin-5 (IL-5) level was noted when measured by enzyme-linked immunosorbent assay, while IL-3, G-CSF, and GM-CSF were undetectable. During remission the plasma gave negative results both for colony formation and IL-5 level. These results indicate that the eosinophilia of this disease is mediated by IL-5. c 1992 Wiley-Liss, Inc.
Key words: eosinophilic colonies, anti-IL-5 antibody, IL-5
INTRODUCTION
Eosinophilic gastroenteritis is characterized by intermittent abdominal distress, such as nausea, vomiting. diarrhea. and abdominal pain, which is accompanied by increased numbers of mature eosinophils in the peripheral blood [ 1-31. Ascites or pleural effusion is frequently observed, and many mature eosinophils are present in these fluids (21. In most patients, the disease resolves spontaneously; however, it recurs repeatedly. Although the pathogenesis of this disease is unknown, Klein et al. [2] speculated that food allergy is the cause of this disease. Recent studies have shown that three cytokines, granulocyte/macrophage colony-stimulating factor (GMCSF) 141,interleukin-3 (IL-3) [ 5 ] , and lL-5 [6], are involved in eosinophilopoiesis. Of these 3 cytokines, 1L-5 acts on the most mature precursors of eosinophils [7], and induces only eosinophilic colonies in vitro [6], in contrast to the other two, which generate colonies of multiple cell lineages [4.5]. Furthermore, IL-5 has been demonstrated to be the main factor involved in reactive eosinophilia such as Kimura's disease, reactive hypereosinophilic syndrome, and eosinophilic angioedema [S-lo]. In the present study, we found eosinophilic (Eo)-CSF activity i n
0 1992 Wiley-Liss, Inc.
the plasma from a patient with eosinophilic gastroenteritis and identified this activity as IL-5.
CASE REPORT A 28-year-old female was admitted to our hospital because of nausea, vomiting, diarrhea. and abdominal pain in July 1989. In the 2-year period before admission, she had had two similar episodes associated with eosinophilia. which resolved spontaneously in a few weeks. She had not taken any medication prior to the onset of abdominal distress in each occasion. She had had a skin rash due to food allergy to mackerel twice when she was a teenager. On physical examination, ascites was noted. Superficial lymph node enlargement and hepatosplenomegaly were not noted. Laboratory examination revealed a white cell count of 24,0OO/plwith 74% mature eosinophils, 16% neutrophils, 1% monocytes, and 9% lymphocytes.
Receives for publication August 29, 1991: accepted January 30, 1Y92. Address reprint requests to Takayuki Takahashi, 1M.D.. The Second Department of Internal Medicine. Faculty of Medicine, Kyoto University, 54 Shogoin Kawahara-cho. Sakyo-ku, Kyoto 606, Japan.
296
Case Report: Takahashi et al.
The hemoglobin level and platelet count were within normal limits. A bone marrow aspirate showed normocellular marrow with 3 1 % eosinophils; however, neither maturation arrest nor dysplastic features were seen. In addition, a hemopoietic progenitor assay of the marrow cells, using the multilineage method [ 1 I], disclosed normal hemopoietic colony formation in number, size. and maturation. Examinations revealed no parasite infection. Computed tomography of the abdomen was non-specific. Endoscopic examination of upper gastrointestinal tract and barium enema were performed 2 weeks after admission; however, no abnormal findings were observed and no tissue biopsy was made. On chest X-ray examination, pleural effusion of the left side was noted. Cytological examination of the pleural fluid showed numerous mature eosinophils. A diagnosis of eosinophilic gastroenteritis was made. The treatment was only supportive (intravenous fluids), and the abdominal distress, eosinophilia, ascites, and pleural effusion disappeared completely 3 weeks after admission. In July 199 I,she had a 4 t h attack of the disease, which also resolved spontaneously. MATERIALS AND METHODS Plasma and Pleural Fluid
Heparinized blood and pleural fluid at various clinical states were obtained from the patient after informed, written consent was obtained. Also pleural fluids were obtained from patients with lung cancer who did not have any eosinophilia or leukocytosis in the process of diagnosis. After centrifugation at 800g for 10 min, plasma and pleural fluid were filtered through a 0.45 - pni niembrane filter (Corning, NY) and kept at -20°C until examination. Assay of Granulocyte/Macrophage Progenitor (CFU-GM)
Bone marrow cells were obtained with informed, written consent from patients with possible early lymphoma as part of the process of clinical staging. The CFU-GM assay used has been described previously [12]. Briefly, non-T, non-adherent bone marrow mononuclear cells ( 1 X 10’ cells/ml) were mixed with Iscove’s modified Dulbecco’s MEM (IMDM) (Gibco, Grand Island, NY), 1 0 30% patient’s plasma, S% FCS, and 0.9% methylcellulose. One-milliliter aliquots were cultured in 35 X 12-mm plastic dishes (Lux, Naperville, IL) for 14 days at 37°C in 5% CO,. Cultures were performed in triplicate. Brownish colonies (eosinophilic colony, Eo colony) of more than 50 eosinophils were counted using an inverted microscope (Olympus Co., Tokyo). Aggregates of 10-49 eosinophils were scored as Eo clusters. Granulocyte (G) or macrophage (M) colonies containing more than 50 cells were scored simultaneously. In some experiments, to confirm the cellular component of Eo colonies, colo-
nies were picked up with a fine-drawn Pasteur pipette, spread out over a slide glass, and stained with Bieblich scarlet. In another set of experiments, prior to making the culture cocktail, patient’s plasma was mixed with antimurine IL-5 monoclonal antibody (MoAb), NC17 (Olympus) [ 131, at a protein concentration of 50 pg/ml, and incubated at room temperature for 30 min. NC17 neutralizes human IL-5 [8,9]. The plasma was also treated with anti-G-CSF MoAb (25 pg/ml; Chugai Pharm. Co., Tokyo) or anti-GM-CSF MoAb (5 Fg/ml; Genzyme, Boston) in the same way. In some experiments, antibody treatment of the plasma was performed using anti-human IL-3 MoAb (10 kg/ml; Genzyme). As a positive control for Eo colony formation, recombinant (r) IL-5 (10 ngiml; Suntory Co., Osaka, Japan) was added to the CFU-GM assay with 30% plasma from a healthy volunteer (normal plasma). Enzyme-Linked lmmunosorbent Assay (ELISA)
The IL-5 concentration in the plasma was measured using an ELISA kit [ 141 kindly provided by Suntory Co. G-CSF concentration was also measured by ELISA [ 151. GM-CSF and IL-3 concentrations were measured using ELISA kits purchased from Genzyme and R&D Systems (Minneapolis, MN), respectively. Statistical Analysis
The statistical significance of the values obtained was evaluated using Student’s t-test. RESULTS Eo-CSF Activity in Plasma and Its Identification
Table I shows the Eo-CSF activity of the patient’s plasma obtained at the time of admission. The plasma generated 13.3 Eo clusters when added to the CFU-GM culture at a volume of 10%. The number of Eo clusters was increased and 10 Eo colonies were generated when the plasma volume was increased to 30%; however, G. or M colonies were rarely seen. Eo colony or cluster formation was inhibited by treatment of the plasma with antiIL-5 MoAb, while anti-G-CSF or anti-GM-CSF MoAbs did not affect Eo colony or cluster formation. Normal plasma (30%) and 10 ng/ml of rIL-5 generated a greater number of Eo colonies or clusters without G , or M colony formation. Eo colony and cluster formation was also inhibited by anti-IL-5 MoAb. Pleural effusion showed lower Eo-CSF activity than plasma; 30% pleural effusion generated nine Eo clusters but not colonies. During disease-free periods, the patient’s plasma did not induce any type of colonies or Eo clusters. While, a plasma sample obtained during the 4th attack of the disease generated 8.7 1.2 (mean 2 SE) Eo colonies per 1 X 10’ bone marrow cells without any
*
Case Report: IL-5 in Eosinophilic Gastroenteritis
297
TABLE 1. Eo-CSF Activity in the Patient’s Plasma and Identification by Antibody Neutralirationt
Number of colonies/l X 10’ BM cells*’ ~~~~
Culture conditions
~
Eo colony*2
10% P-PI.*~,S% FCS 30% P-pl., 5% FCS 30% P-pl., 5% FCS, (Y IL-5 MoAb*’ (50 pg/ml) 30% P-pl., 5% FCS, (Y G-CSF MoAb (25 crgiml) 30% P-PI., 5% FCS, (Y GM-CSF MoAb (5 pgiml) 30% N-PI.*~,5% FCS, rhIL-S*9 30% N-pl., 5% FCS, rhIL-5, (Y IL-SMoAb (50 pglml)
Eo cluster*3 G colony*4 M colony*’
0 10.0 k 1.0 0
13.3 f 2.0 34.0 f 4.6 2.0 f 1.2
0.3 f 0.3 0.7 +- 0.3 0.3 f 0.3
0.7 f 0.3 0.3f 0.3
11.0 f 2.1
31.0 f 4.4
0
0
8.3 k 1.8
29.3 f 4.0
0.3 f 0.3
0
35.7 f 4.0 0
51.7 f 7.6 4.0 k 2.3
1.0 f 0.6 0.7 f 0.3
0.3 f 0.3 0.7 k 0.3
0
?.Data represent themean fSE (n = 3).Nos. in parentheses indicate the protein concentration of the antibody during preincubation (30 min, room temperature) with plasma. * I , non-T, non-adherent bone marrow mononuclear cells; *2, eosinophilic colony (more than 50 cells); *3, eosinophilic cluster (10-49 cells); *4, granulocytic colonies (more than SO cells); *S, macrophage colony (more than SO cells); *6, plasma from the patient during the period of eosinophilia (3rd attack); *7, anti IL-5 monoclonal antibody; *8, plasma from a healthy donor (normal plasma); *9, recombinant human IL-5 ( I 0 ngiml). TABLE II. Concentration 01 IL-5 in Plasma and Pleural Fluid?
Concentration of cytokines (pg/ml) IL-5
IL-3
GM-CSF
G-CSF
1,800+ 104 1,560f 131
