INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The respiratory epithelium is the principal cellular barrier between the environment and the internal milieu of the airways. Upon contact with exogenous stimuli such as invading pathogens, the epithelium can modulate local responses by releasing proinflammatory mediators. Chemokines are a class of proinflammatory mediators that recruit and activate circulating leukocytes via discrete, receptor-mediated interactions. Recent studies have focused on the role of chemokines in respiratory diseases caused by viral pathogens (1). Respiratory syncytial virus (RSV), a nonsegmented single-stranded RNA virus of the family Paramyxoviridae, is one such pathogen, currently recognized as a major cause of significant morbidity in both pediatric and institutionalized elderly populations (2). The aim of this study was to determine the chemokine response pattern of respiratory epithelium when infected with RSV.

Several groups have shown, in vitro, that bronchial epithelial cells respond to RSV infection by increased production of interleukin-8 (IL-8) (3), a neutrophil chemoattractant and member of the CXC family of chemokines. Interestingly, RSV-mediated expression of the IL-8 gene occurred in the absence of viral replication (5), and has been shown to be mediated at least in part by consensus binding sites for the transcription factors, nuclear factor (NF)-kappa B and NF-IL-6 present in the IL-8 gene promoter (7, 8). The CC chemokine, RANTES, a chemoattractant for monocytes, eosinophils, and basophils (9), is also produced by respiratory epithelial cells in response to infection with RSV (10), although the mechanism by which this occurs remains to be clarified.

Macrophage inflammatory protein-1- (MIP-1-) is a small, pleiotropic chemoattractant of the CC chemokine family which has been shown to recruit and/or activate monocytes, eosinophils, basophils, and several lymphocyte subpopulations (13). In this work, we show that both respiratory epithelial cells and fibroblasts in culture respond to RSV infection by producing MIP-1-, a response that displays an absolute dependence on the presence of active, replicating viral particles. Additionally, we have examined lower respiratory secretions of critically ill, mechanically ventilated pediatric patients diagnosed with RSV bronchiolitis, and have observed an analogous pattern of chemokine responses. The presence of immunoreactive and biologically active eosinophil secretory proteins in these secretions provides evidence for the role of eosinophils in the pathogenesis of this disease, and suggests a role for MIP-1- as an eosinophil chemoattractant in vivo.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cell Culture

HEp-2 (human laryngeal carcinoma), WI-38 (human embryonic lung fibroblasts), and MRC-5 (human embryonic lung fibroblasts) cell lines were obtained from American Type Culture Collection (ATCC, Rockville, MD) and cultured in Eagle's modified essential medium (EMEM; Life Technologies GIBCO BRL, Gaithersburg, MD) supplemented with 10% fetal bovine serum, 2 mM glutamine, and penicillin/streptomycin (Life Technologies GIBCO BRL) at 37° C in 5% CO₂.

Preparation of Viral Suspensions

RSV-B (ATCC VR-1401) was obtained from ATCC. The viral suspensions used in these experiments were prepared as follows: The ATCC isolate of RSV was used to inoculate 180 cm² flasks containing semiconfluent monolayers of HEp-2 cells cultured as described previously. When cytopathic effect reached ~ 80% (at 72 to 96 h), the culture supernatants were harvested and cellular debris was removed by centrifugation at 500 g. Aliquots of the RSV viral suspension were flash frozen and stored at 80° C. Infectivity, determined by quantitative shell vial assay (17, 18) ranged from 10⁵ to 1.5 × 10⁶ infectious units (inf units)/ml. An infectious unit is defined as the component of the viral suspension that results in the infection of a single target cell as detected by immunofluorescence staining for viral antigens; infectious units have been shown to correspond to plaque-forming units (pfu) over a wide range of viral dilutions (18). A control suspension (-RSV) was prepared from uninfected HEp-2 cell cultures in an otherwise identical fashion. Ultraviolet (UV) inactivation was achieved by exposing the RSV viral suspensions to direct UV light (wavelength = 254 nm) for 3 to 5 min.

Infection of HEp-2, MRC-5, and WI38 cells with RSV

When cell monolayers were ~ 80% confluent (15 × 10⁶ cells per 30 ml, or 5 × 10⁵ cells/ml), the medium was replaced (30 ml) and the cells were inoculated with 500 µl of the RSV suspension described previously. An uninfected flask of each cell line at the cell density indicated was included as a control. Aliquots (0.5 ml) of culture supernatants were removed at the time of infection (t = 0), and then every 24 h thereafter. Cellular debris was removed by centrifugation (400 g × 5 min) and aliquots were stored at 80° C.

Exposure of HEp-2 Cells to UV-inactivated Viral Suspension or Irradiated RSV Particles

Purified gamma-irradiated RSV (Chemicon International, Temecula, CA) was determined to be ~ 10-fold more concentrated than the RSV viral suspensions (1.5 × 10⁶ pfu/ml) by Western immunoblotting using a horseradish peroxidase-conjugated goat anti-RSV polyclonal antibody (Accurate Chemical and Scientific Corporation, Westbury, NY). No cytopathic effect was observed in control (-RSV) tissue cultures, cultures inoculated with either the UV-inactivated RSV suspension, nor with the commercially prepared gamma-irradiated RSV. HEp-2 cells (9 × 10⁶ cells in 30 ml complete media, or 3 × 10⁵/ml) were either left uninfected, inoculated with 500 µl RSV suspension, with 500 µl control (-RSV) suspension, with 500 µl UV-inactivated RSV suspension, or with 50 µl of gamma-irradiated RSV at t = 0, followed by additional 150 µl aliquots at t = 16, 24, and 40 h (4 exposures to antigen over time). Aliquots (0.5 ml) were removed at the intervals indicated for determination of chemokine (IL-8, RANTES, and MIP-1-) or myeloperoxidase (MPO) concentrations by quantitative ELISA (R&D Systems, Minneapolis, MN). All measurements were performed on undiluted samples. Sensitivities of the ELISAs, as reported by the manufacturer, are as follows: IL-8 = 10 pg/ml, MIP-1- = 6 pg/ml, MPO = 2 ng/ml, RANTES = 5 pg/ml.

Collection and Preparation of Lower Respiratory Tract Specimens

Mechanically ventilated patients between the ages of 0 and 24 mo were eligible for study enrollment. Ten consecutive patients with RSV requiring mechanical ventilation were enrolled during the study period. For study purposes, patients with RSV bronchiolitis (defined as upper respiratory symptoms and apnea, wheezing, or pneumonia) were required to have a positive RSV ELISA (Abbott Laboratory, North Chicago, IL; sensitivity 94.3%, specificity 95.3%), with confirmation of RSV by standard roll tube culture. Patients with clinical bronchiolitis and a negative RSV ELISA were excluded from the study. A convenience sample of 10 control subjects without bronchiolitis was enrolled during the study period. Virologic testing for RSV and control patients was performed in the Clinical Virology Laboratory at the State University of New York Health Science Center (SUNY HSC) at Syracuse. Control subjects were excluded if they had a positive viral culture. All enrollees were patients in the pediatric intensive care unit at the SUNY HSC at Syracuse between November 1997 and April 1998. Informed consent was obtained and thorough diagnostic testing for respiratory viral pathogens was performed.

To obtain the samples, normal saline (2 ml) was instilled into the endotracheal tube, a catheter inserted and suction applied as the catheter was withdrawn. Samples were collected daily as long as the patient was mechanically ventilated. The suctioned lower airway secretions (routinely 0.5 to 1.0 ml) were diluted with an equal volume of phosphate-buffered saline containing aminoethyl benzene sulfonyl fluoride (AEBSF) and 2% mucocil. The specimen was clarified by centrifugation and stored in aliquots at 80° C. Total protein concentration was determined by the Bradford microassay (Bio-Rad, Richmond, CA) against bovine serum albumin standards (Sigma Chemical Corporation, St. Louis, MO). Eosinophil cationic protein (ECP) concentrations were determined in duplicate on undiluted lower respiratory tract specimens using a commercially available radioimmunoassay (sensitivity 2 µg/L; Pharmacia AB, Uppsala, Sweden). Chemokine and MPO concentrations were determined in duplicate as previously described; results are expressed as concentration of chemokine, MPO, or ECP per mg total protein.

Detection of Eosinophil Granule Ribonucleases

Heparin-sepharose resin (100 µg; Pharmacia, Piscataway, NJ) was added to lower respiratory tract secretions containing 1 mg total protein, and rotated end-over-end at 4° C overnight. The resin was harvested by centrifugation and resuspended in 40 µl sodium dodecyl sulfate-polyacrylamide gel electrophresis (SDS-PAGE) loading buffer. Western blotting with 1:300 dilutions of either rabbit polyclonal anti- eosinophil-derived neurotoxin (anti-EDN) or anti-ECP (pretreated to remove cross-reacting species) followed by 1:1,000 dilutions of alkaline phosphatase-conjugated goat anti-rabbit IgG antibody (Bio-Rad) was performed as previously described (19). Blots were developed with nitroblue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate (Bio-Rad).

Ribonuclease Assay

The ribonuclease assay was performed as described (19). Briefly, reactions were initiated with 40 µg of yeast transfer RNA (tRNA) substrate (Sigma) added to 0.8-ml reactions containing 40 mM sodium phosphate, pH 7.0, and 10 µg of lower respiratory tract specimens described previously. Reactions were stopped at given time points by the addition of ice cold 3% perchloric acid with 40 nM lanthanum nitrate, and acid-soluble ribonucleotides remaining in the supernatant fraction after centrifugation were quantified spectrophotometrically (260 nm). All time points were evaluated in triplicate, and data were evaluated using Microsoft Excel 5.0 software.

Statistical Analysis

In vitro experiments comparing two conditions over time (central tendencies expressed as mean ± standard error) were evaluated by two-way analysis of variance (ANOVA). Fisher exact test was employed for categorical data. Unpaired t tests were used to compare continuous data. Statistical significance was set a priori at p < 0.05. Pearson's correlations were performed for paired sets of continuous data.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

MIP-1- Is Produced by RSV-infected HEp-2 Respiratory Epithelial Cells

MIP-1- production by HEp-2 human respiratory epithelial cells in response to RSV infection (0.75 × 10⁶ inf units) was assessed by ELISA performed on supernatant fractions harvested just after inoculation (t = 0), and on Days 1 through 5 thereafter (Figure 1A). MIP-1- was detected on Day 3 after inoculation, and increased steadily through Day 5, with 2,600 pg/ml culture supernatant detected at this final time point. MIP-1- production was also assessed in response to inoculation with RSV suspension after inactivation of the virus by exposure to UV light (see METHODS). No syncytia were detected in cultures inoculated with UV-treated suspension up to and including Day 7 postinoculation, in contrast to the suspension containing active RSV, in which syncytia formation was routinely detected by Day 3 (data not shown). No MIP-1- was detected through Day 5 after inoculation, demonstrating that this chemokine is produced in direct response to the (replicating) virus, not to an otherwise unrecognized component of the viral suspension. Similarly, no MIP-1- was detected in culture supernatants from cells inoculated with control (-RSV) suspension, not in supernatants from uninfected cells.

View larger version (15K): [in this window] [in a new window]   Figure 1.   MIP-1- detected in culture supernatants of respiratory cells in response to infection with RSV. (A) MIP-1- detected in supernatants from HEp-2 respiratory epithelial cells (3 × 105/ml, 80% confluence at t = 0) inoculated with 1.5 × 106 inf units/ml RSV (black circles) or with 1.5 × 106 inf U/ml RSV inactivated by UV irradiation (shaded circles) (p < 0.001, 2-way ANOVA). Lines identical to that shown for the UV-irradiated virus were obtained from cells inoculated with control (-RSV) suspension and from uninfected cells. Each point represents the mean of four independent experiments with samples assayed in duplicate. (B) MIP-1- (pg/ml) detected in supernatants from MRC-5 and WI-38 human respiratory embryonic fibroblast cells (5 × 105/ml, 80% confluence at t = 0) inoculated with RSV (1.5 × 106 inf U/ml) or remaining uninfected. Each point represents the mean of four independent experiments, with samples assayed in duplicate (p < 0.001, 2-way ANOVA). Mean concentration ± standard error of the mean are shown.

MIP-1- Is Produced by Other Cells Susceptible to RSV Infection

MIP-1- production by MRC-5 and WI-38 human respiratory fibroblasts was assessed as described previously for the HEp-2 cells (Figure 1B). MIP-1- was detected in culture supernatants from cells inoculated with the RSV suspension, but not in supernatants from uninfected cells.

Chemokine Production by HEp-2 Cells in Response to Active and Inactivated Forms of RSV

Chemokine production by the HEp-2 cells in response to ongoing RSV infection was compared with that elicited by challenge with replication-incompetent forms of the virus. Expression of IL-8 was effectively elicited by the active RSV suspension, and by both UV-inactivated RSV suspension and multiple aliquots of gamma-irradiated RSV, consistent with previous observations (5) (Figure 2A). Active RSV infection also induces the production of the chemokine RANTES, analogous to that described in primary culture (12). The RANTES response was also elicited by multiple aliquots of gamma-irradiated RSV (Figure 2B). In contrast, production of MIP-1- occurred only in response to ongoing RSV infection, and could not be elicited by exposure to either inactivated form of the virus (Figure 2C). No MIP-1- was detected in supernatants from uninfected cells or from cells inoculated with inactivated virus up to and including Day 6 postinoculation.

View larger version (17K): [in this window] [in a new window]   Figure 2.   Chemokine concentrations detected in supernatants from uninfected and RSV-infected HEp-2 cells, and from HEp-2 cells inoculated with UV-irradiated RSV or gamma-irradiated RSV on Day 4 postinoculation. (A) Interleukin-8, (B) RANTES, and (C ) MIP-1-, all in pg/ml, detected in supernatants from cells inoculated at an initial concentration of 3 × 105/ml (80% confluence). The data in panel B have been corrected for baseline production of RANTES by uninfected HEp-2 cells. Bars represent averages of three independent experiments, with samples assayed in duplicate. Mean concentration ± the standard error of the means are shown. Absent bars indicate that no chemokine was detected under that condition.

Detection of MIP-1-, RANTES, IL-8, ECP, and MPO in Lower Respiratory Tract Secretions

Control and RSV-infected patients were similar in terms of age and weight as determined by unpaired t test (p > 0.05). MIP-1-, RANTES, IL-8, and MPO produced in response to RSV infection in vivo were assessed by quantitative ELISA, and ECP by radioimmunoassay. As shown in Table 1, MIP-1-, RANTES, and IL-8 were detected in lower airway secretions from patients diagnosed with RSV bronchiolitis. MIP-1- was detected in all samples from all patients infected with RSV; the concentration of MIP-1- varied both between individuals and at different time points within a single individual, but ranged from 13 to 1,076 pg/ml/mg protein. MIP-1- was not detected in lower airway secretions from any of the control patients (Fisher exact test p < 0.0001) In contrast, RANTES was detected in nine of 10 samples from RSV-infected patients and in only one of 10 control patients (Table 1) (Fisher exact test p = 0.0001). ECP concentrations were higher in samples from RSV-infected patients when compared with control subjects (unpaired t test p < 0.001), and the ECP concentration correlated with the MIP-1- concentration (Figure 3A, r² = 0.868) IL-8 concentrations in samples from RSV- infected patients were also significantly different from concentrations detected in control samples (unpaired t test p = 0.001). IL-8 and MPO concentrations correlate with one another (Figure 3B, r² = 0.515). No correlation between ECP and RANTES concentrations was observed (Figure 3C).

View larger version (11K): [in this window] [in a new window]   Figure 3.   (A) Bivariate scattergram of MIP-1- concentration (pg/ml/mg total protein) versus ECP concentration (mg/L/mg total protein), (B) IL-8 concentration (pg/ml/mg total protein) versus MPO (ng/ml/mg total protein) concentration, and (C ) RANTES (pg/ml/mg total protein) versus ECP (mg/L/mg total protein) concentration in 48 lower respiratory tract samples. Regression lines with Pearson coefficients are indicated.

Detection of Eosinophil Granule Proteins in Lower Respiratory Tract Secretions

The eosinophil granule ribonucleases eosinophil-derived neurotoxin (EDN) and ECP were detected in heparin-sepharose concentrates prepared from 1 mg protein samples of lower airway secretions from patients diagnosed with RSV bronchiolitis (Patients 1-10; lanes 1-10) on Western blots probed with polyclonal anti-EDN and anti-ECP antisera (Figure 4). In contrast, no immunoreactive EDN or ECP was detected in concentrates prepared from lower airway secretions from patients with unrelated diagnoses (Patients 11-20; lanes 11-20), demonstrating the eosinophils are recruited to the lower airways in response to specific signals generated by infection with RSV rather than as a response to inflammation in general. The results presented in Table 2 demonstrate enhanced ribonucleolytic activity in the lower airway secretions of patients infected with RSV (3- to 5-fold over uninfected controls, p < 0.01), suggesting that EDN, the major eosinophil ribonuclease with characterized direct toxicity against extracellular virions of RSV (17) has been released from eosinophils in biologically active form.

View larger version (25K): [in this window] [in a new window]   Figure 4.   Immunoreactive eosinophil granule proteins detected in lower airway secretions. Western blots from lower airway secretions obtained from patients with RSV bronchiolitis (Patients 1-10; lanes 1-10) or unrelated diagnoses (Patients 11-20; lanes 11-20; diagnoses in accordance with Table 1) were probed with (A) polyclonal anti-EDN or (B) polyclonal anti-ECP antisera. +C refers to dilute human eosinophil lysate (~ 1 µg loaded).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We show here that human respiratory epithelial cells and fibroblasts in culture respond to RSV infection by synthesizing and secreting MIP-1-, a response that demonstrates an absolute requirement for ongoing viral replication. We go on to show that MIP-1-, RANTES, and IL-8 are detected in lower airway secretions of patients with RSV bronchiolitis, and that the production of MIP-1- is associated with the presence of the biologically active eosinophil degranulation products, EDN and ECP. Indeed, ECP concentrations are strongly correlated with MIP-1- concentrations, suggesting the importance of this chemokine in the recruitment and/or degranulation of eosinophils during RSV bronchiolitis.

Although MIP-1- was reported as absent in RSV-infected HEp-2, A549, and primary explanted respiratory epithelial cell culture supernatants in an earlier report (12), we (20) and others (21) have since reported production of this chemokine in response to infection with RSV in vitro. In this work, we have focussed on chemokine expression in clinical specimens, rather than explanted pulmonary epithelial cells as this more closely represents the biologic milieu in vivo. To provide some sense that it was the epithelial cells (and not contaminating tissue macrophages or fibroblasts) responding to replicating virus by upregulating the MIP-1- response, we opted for the preliminary in vitro work to be done in the HEp-2 cell line. Interestingly, MIP-1- production has been reported to be a response of isolated macrophages (22) and of epithelial cells (23) to infection with influenza A, and of both macrophages (24) and human T-cell lines (25) to infection with human immunodeficiency virus-1 (HIV-1). Similarly, messenger RNA (mRNA) encoding MIP-1- was detected in brain tissue isolated from mice infected with lymphocytic choriomeningitis virus (LCMV; 26), and in mononuclear cells isolated from lymph nodes of monkeys infected with the simian immunodeficiency virus, SIVmac251 (27). Although the role of MIP-1- in host defense against viral infection is not clear, Cook and colleagues (28) have demonstrated that mice engineered to be MIP-1--deficient (/) exhibited delayed clearance of influenza virus along with a markedly suppressed inflammatory response to both influenza and coxsackie viruses. In addition, MIP-1- is one of the CC chemokines shown to suppress transmission of HIV-1 (29). Two viral homologues of MIP-1- have been identified; the genome of Kaposi's sarcoma associated herpesvirus (KSHV, HHV-8) includes a gene encoding a functional homologue of MIP-1- (30, 31), and the genome of the poxvirus molluscum contagiosum virus (MCV), a gene encoding a chemokine-like protein that blocks MIP-1--induced chemotaxis (32). While the activities of these viral homologues remain unspecified, they are believed to function by subverting the normal host responses to viral infection; their existence suggests a complex role for MIP-1- in host defense against viral disease.

MIP-1- is a potent chemoattractant for human eosinophils (13), cells that have been associated with the inflammatory response to RSV (33). Our results demonstrate that RSV infection and MIP-1- production are associated with the presence of eosinophil degranulation products in the lower respiratory tract in vivo. Although eosinophils are generally perceived as villains in RSV disease, we have recently shown that isolated human eosinophils mediate the direct destruction of extracellular virions of RSV in vitro via the actions of their secretory ribonucleases, EDN and ECP (17). The presence of biologically active forms of EDN and ECP in the lower airway secretions of critically ill, RSV-infected children, and the high correlation between ECP and MIP-1- concentrations in these airway specimens suggest a role for MIP-1- as an eosinophil chemoattractant and/or inducer of eosinophil degranulation in vivo.