0031-3998/92/3202-0160$03.00/0 PEDIATRIC RESEARCH Vol. 32, No.2, 1992 Copyright © 1992 International Pediatric Research Foundation, Inc. Printed in U.S.A. Activation of Human Eosinophils In Vitro by Respiratory Syncytial Virus' JAN L. L. KIMPEN, ROBERTO GAROFALO,2 ROBERT C. WELLIVER, AND PEARAY L. OGRA2 School ofMedicine, State University ofNew York at Buffalo, Departments ofPediatrics and Microbiology, and Division ofInfectious Diseases, The Children's Hospital, Buffalo, New York 14222 ABSTRACT. To determine the nature of the interaction RSV is the leading cause of pneumonia and bronchiolitis between viruses and eosinophils, normodense eosinophils during infancy (1). It has been speculated that RSV infection in were separated from the blood of healthy volunteers and infancy plays a role in the development of hyperreactive airway incubated in vitro with respiratory syncytial virus (RSV). disease in later life (2). The pathogenesis ofthe inflammation of After incubation for 2 h with the virus, 29.5 ± 15.8% of the airways during acute RSV infection and the mechanisms the eosinophils demonstrated specific binding of the virus underlying the possible development of subsequent airway hy­ to the cell membrane, as detected by fluorescent staining perreactivityare not completely understood. with an anti-RSV MAb. Superoxide production and leu­ Recent evidence suggests an important role for eosinophils in kotriene C4 release were measured as determinants of cell a variety of inflammatory states (3). Although the role of eosin­ activation. Using a cytochrome c reduction assay, super­ ophils in natural RSV infection is unclear, reports offield studies oxide could be detected in the supernatant 30 min after with a formalin-inactivated RSV vaccine carried out in the late exposure to RSV. Maximal release was reached at 3 h 1960s suggest a role for eosinophils in disease due to RSV. postexposure (5.88 ± 2.19 nmol cytochrome c reduction/S Lymphocytes and eosinophils predominated in the airways on X 105 cells). The virus-induced superoxide generation var­ autopsy of two patients who died as a consequence of severe ied in magnitude among different subjects and ranged from RSV bronchiolitis after vaccination (4), and children who had 0.6 to 11.5 nmol cytochrome c reduction/S x 105 cells. more severe bronchiolitis when subsequently infected with the RSV also appeared to prime eosinophils to the effects of natural virus demonstrated an increased blood eosinophilia com­ other known cell activators, as demonstrated by an increase pared to the controls (5). In addition, the release ofLTC4 in the in superoxide production upon subsequent stimulation of respiratory tract during RSV infection has been described (6, 7), RSV-primed cells with phorbol-12-myristate-l3-acetate and we recently showed higher levels of eosinophil cationic (21.4 ± 5.8 versus 9.4 ± 2.7 nmol cytochrome c reduction/ protein in nasopharyngeal secretions of children with natural 5 X 105 cells for primed and unprimed cells, respectively) RSV bronchiolitis in comparison with children affected only by (p < 0.04). RSV did not directly induce leukotriene C4 RSV upper respiratory tract infection or non-RSV respiratory release from the eosinophils but primed the cells to exhibit illness (8). The present study was therefore undertaken to ex­ a more vigorous response on subsequent challenge with the amine in vitro the nature of interaction between RSV and calcium ionophore A23187 (9.16 ± 1.04 versus 4.2 ± 1.3 eosinophils. 6 ng leukotriene C4/1 x 10 cells) (p < 0.005). These findings indicate that RSV can activate or prime eosinophils to MATERIALS AND METHODS release various inflammatory mediators. Such virus-in­ duced effects on inflammatory cells may playa role in the Human volunteers. Peripheral blood (120 mL) was obtained pathogenesis of RSV bronchiolitis and may also be critical in heparinized syringes from healthy adult volunteers who had a for the development of persistent airway hyperreactivity differential eosinophil count of 1-4%. The subjects had no after viral infections. (Pediatr Res 32: 160-164, 1992) respiratory disease for at least 3 wk and had not taken any medication for at least 3 d before sampling. Signed statements of informed consent were obtained. All subjects had serum Abbreviations antibodies to RSV as measured by ELISA (9). RSV, respiratory syncytial virus Preparation 01eosinophil suspensions. Percoll dilutions were PMA, phorbol-12-myristate-l3-acetate prepared as described previously, with modifications (10). PAF, platelet activating factor Briefly, a stock solution ofPercoIl was prepared by adding 1 part LTC4, leukotriene C4 HBSS lOx to 9 parts Percoll, The osmolality of this solution was HBSS, Hanks' balanced salt solution measured with a freezing-point osmometer (The Advanced Mi­ SOD, superoxide dismutase cro-Osmometer model 3MO; Advanced Instruments, Needham m.o.i., multiplicity of infection Heights, MA) and adjusted to 340 ± 1 mosmol/L. The density was measured with a glass 2-mL pyknometer (Ace Glass Incor­ porated, Vineland, NJ). From this stock solution (density 1.123 g/rnl.), five Percoll solutions with gradually increasing densities Received September 27, 1991; accepted March 12, 1992. (1.080, 1.085, 1.090, 1.095, and 1.100 g/ml.) were prepared by Correspondence: Dr. J. L. L. Kimpen, Department of Pediatrics, Academisch adding calculated amounts of HBSS ix. All densities were Ziekenhuis Groningen, 59 Oostersingel, 9700 RB Groningen, The Netherlands Supported by National Institute of Allergy and Infectious Diseases, National verified by measurement with a pyknometer. Manipulations of Institutes of Health Grant No. AI 15939-11. Dr. Kimpen is the recipient of a the solutions were carried out under sterile conditions and the NATO Research Fellowship. Percoll solutions were checked regularly for bacterial or fungal I Presented in part at the meeting of the Society for Pediatric Research, Los contamination by microscopic examination and culture. Angeles, 1990 (abstr 932). 'Present address for Drs. Ogra and Garofalo: Department of Pediatrics, UTMB Discontinuous Percoll gradients were prepared in 15-mL con­ at Galveston, Child Health Center, Galveston, TX 77550-2776. ical plastic tubes (Falcon Plastics, Los Angeles, CA), using a 160 EOSINOPHIL ACTIVATION BY RSV 161 peristaltic pump (Peristaltic Pump PI; Phannacia Fine Chemi­ addition of 1.5 mL of ice-cold methanol. Twenty ng of prosta­ cals, Uppsala, Sweden). Two and one half mL of 1.080 g/ml. glandin B2 were added as an internal standard for HPLC analysis. Percoll were underlayered subsequently with 3 mL of 1.085 g/ The resultant mixture was centrifuged at 1000 x g for 15 min at mL, 3 mL of 1.090 g/ml., 2.5 mL of 1.095 g/ml, and 1.5 mL of 4°C. The supernatant was applied to a CI8 reversed phase 1.100 g/ml, Percell. extraction cartridge (Sep-Pack; Millipore, Waters Associates, Mil­ Five parts blood were mixed with one part 4.5% dextran and ford, MA) that had been washed previously with 10 mL of left at room temperature for 45 min to sediment the red cells. distilled HPLC-grade water and activated with 10 mL HPLC­ Twenty-mL aliquots of the leukocyte rich supernatant were grade methanol. After the sample was placed on the cartridge, it underlayered with 10 mL Ficoll-Hypaque (1.076 g/rnl.) and was washed with 10 mL of distilled water and the leukotrienes centrifuged (450 x g, 20 min at room temperature). The cell were eluted with 10 mL of70% methanol. The methanol fraction pellet was collected and washed twice with HBSS Ix without wasevaporated under vacuum in a rotary evaporator(Rotavapor, calcium or magnesium and finally resuspended in HBSS Ix with Buchi, Switzerland) and the residue was redissolved in 500 ILL of 5% neonatal calf serum. A cell count was performed with an 30% methanol and stored at - 70T until testing. The samples automated counter (Coulter counter S+IV; Coulter Corporation, were injected into a C18 reverse-phase column (Beckman, San 6 Hialeah, FL), and the suspension adjusted to 20 x 10 cells/rnl., Ramon, CA) and eluted isocratically at a flow rate of I mL/min Two mL of this cell suspension were layered on top of each in a HPLC system (Pharmacia, Piscataway, NJ) with a solvent Percoll gradient and centrifuged (600 x g, 20 min at room system of methanol/water/acetic acid (70/28/2 vol/vol/vol) ad­ temperature). justed to pH 5.5 with ammonium hydroxide. The eluate was Cells from the three lower interfaces (1.100-1.095, 1.095­ monitored at 280 nm by a wavelength spectrometric detector 1.090, and 1.090- I.085) were collected and washed twice in (Phannacia). For identification and quantification of LTC4 in HBSS Ix with 0.1% gelatin. Residual red blood cells were lysed the samples, different mixtures of purified LTC and prostaglan­ by hypotonic shock. The number of cells in each band was 4 din B2 were prepared in HBSS and processed as described above. determined and a differential cell count was performed on cyto­ Each sample was tested in parallel with these standards. centrifuge slides (Cytospin 2; Shandon, Sewickley, PA) after Reagents. HBSS with and without calcium and magnesium, staining with a modified Wright's stain (Stat Stain, VWR Scien­ tific, Piscataway, NJ). RPMI 1640, and neonatal calf serum were purchased from Gibco Laboratories (Grand Island, NY). Gelatin was obtained from Normodense eosinophils obtained from the 1.095-1.100 in­ terface were used in subsequent experiments. The cell population Bio-Rad Laboratories (Richmond, CA). Dextran , Percoll, PMA, calcium ionophore A23187, fonnyl-methionyl-Ieucyl-phenylal­ was 87.8 ± 2.0% pure (n = 17). Viability of the cells after the separation was always more than 98% as determined by trypan anine, SOD, horse-heart ferricytochrome c (type VI), prostaglan­ blue exclusion. Contaminating cellswere exclusivelyneutrophils.