Lung (2002) 180:229±239 DOI: 10.1007/s004080000097

E€ects of Granulocyte Colony-Stimulating Factor on Hyperoxia-Induced Lung Injury in Newborn Piglets L. I. Wolko€,2,3 C. R. Levine,1,2 H. C. Koo,1 E. F. LaGamma,4 S. Pollack,5,6 D. Chester,2 J. Bashore,1,2 and J. M. Davis1,2 1 CardioPulmonary Research Institute, Winthrop-University Hospital, SUNY Stony Brook School of Medicine, Mineola NY 11501, USA 2 Department of Pediatrics (Neonatology), Winthrop-University Hospital, SUNY Stony Brook School of Medicine, 259 First Street, Mineola NY 11501, USA 3 Department of Pediatrics, South Nassau Communities Hospital, SUNY Stony Brook School of Medicine, Oceanside, NY 11572, USA 4 The Regional Neonatal Center Westchester Medical Center, New York Medical College, Valhalla, NY 10595, USA 5 Biostatistics, Winthrop-University Hospital, SUNY Stony Brook School of Medicine, Mineola NY 11501, USA 6 Department of Decision Sciences, St. John's University, Jamaica, NY 11439, USA

Abstract. Granulocyte colony-stimulating factor (G-CSF) increases the concentration and activation of neutrophils in the peripheral blood and has been used to prevent late-onset infection in premature infants. However, if G-CSF also augmented the in¯ammatory response in the lung, the incidence and severity of acute and chronic lung injury might be expected to increase. Using a newborn piglet model of acute lung injury, we examined the e€ects of rhG-CSF (recombinant-metHuG-CSF) on lung injury. Thirty-three newborn piglets were studied as follows: 1) Unventilated controls; 2) normally ventilated (PaCO2 = 35±45 torr) with room air(RA) for 48 h; 3) normally ventilated with RA for 48 h and received rhG-CSF (10 lg/kg/dose IV) at 0, 12, 24, and 36 h; 4) hyperventilated (PaCO2 = 15±25 torr) with 100% O2 for 48 h; 5) hyperventilated with 100% O2 for 48 h and received rhG-CSF (10 lg/kg/dose IV) at 0, 12, 24 and 36 h. Complete blood counts and and di€erentials were performed at 0, 24, and 48 h. Animals were sacri®ced at 48 h, lungs were removed en bloc, and bronchoalveolar lavage (BAL) was performed. Total blood white blood cells and neutrophil counts increased signi®cantly over 48 h in animals who Correspondence to: Carolyn Robbins Levine, MD, Department of Pediatrics, Winthrop-University Hospital, e-mail: [email protected]

230

L. I. Wolko€ et al.

received rhG-CSF either with normoventilation (p < 0.0001) or hyperventilation with 100% O2 (p < 0.003), and did not change signi®cantly in the other experimental groups. However, there were no signi®cant di€erences in BAL total cell counts, neutrophil chemotaxis activity, total protein, or albumin concentrations among the groups. Despite signi®cantly increasing peripheral neutrophil counts, rhG-CSF did not potentiate acute lung injury or in¯ammation. This suggests that prophylactic administration strategies using rhG-CSF to prevent sepsis in premature infants should not increase the risk for developing acute and chronic lung disease. Key words: NeutrophilsÐIn¯ammationÐLung injuryÐG-CSF Introduction Bronchopulmonary dysplasia (BPD) and late-onset sepsis are two common problems encountered in the neonatal intensive care unit (NICU). BPD is a form of chronic lung disease that develops in newborn infants treated with oxygen and positive-pressure mechanical ventilation [10]. BPD is believed to begin as an acute in¯ammatory response in the lung characterized by signi®cant neutrophil in¯ux and activation which progresses to chronic lung disease [10]. With advances in neonatal care, more very low birth weight infants (VLBW < 1500 g) are surviving the newborn period and developing BPD. Similarly, these developmentally immature neonates remain in the NICU for prolonged periods of time and are continuously exposed to a wide variety of potential pathogens. Thus, late-onset neonatal sepsis is also a major cause of morbidity and mortality in the US affecting up to 30% of all VLBW neonates, an incidence that has not changed in 20 years [13, 23, 30, 34]. Moreover, recent epidemiological data also reveal a strong association between systemic infection and the development of chronic lung disease [8, 27, 28]. Though the cause of neonatal sepsis is a complex, multifactorial process, it is clear that the primary factor involved is the relatively immunode®cient state of the newborn which involves both the cellular and humoral arms of the immune system and is accentuated by increasing degrees of prematurity [17]. Recently, granulocyte colony-stimulating factor (G-CSF) has been studied in VLBW infants for the prevention or treatment of late-onset neonatal infection. G-CSF is the primary physiologic regulator of neutrophil production [32]. G-CSF induces maturation of committed myeloid progenitors, release of neutrophils from the bone marrow, and activation of neutrophil functions, all potentially bene®cial e€ects that may reduce the risk of late-onset neonatal sepsis [5, 14, 21, 25]. However, G-CSF is typically administered soon after birth while these VLBW infants are often receiving high inspired oxygen concentrations and ventilatory support. This could increase the number and activation of neutrophils in the lung as well as the peripheral blood and thus increase the incidence and severity of acute and chronic lung injury. We examined the relationship between rhG-CSF (recombinant-metHuGCSF) administration and the development of acute lung disease using a newborn

G-CSF and Lung Injury

231

piglet model of lung injury caused by hyperoxia and mechanical ventilation that has many similarities to that seen in early BPD. We have previously demonstrated that 48 h of hyperoxia and hyperventilation causes pulmonary in¯ammatory changes and acute lung injury in newborn piglets [7, 9, 11]. We hypothesized that rhG-CSF given to newborn piglets would increase neutrophil number and activation in the blood, but would not further potentiate in¯ammation in the lung from hyperoxia and mechanical ventilation. Methods Study Design Five groups of neonatal piglets, 1±3 days of life, weighing 1.0±1.8 kg (Archer Farms, Belcamp, MD) were studied as follows: 1) Controls; no mechanical ventilation, sacri®ced at 0 h (n = 5) 2) Normally ventilated (paCO2 35±45 Torr) for 48 h with room air (RA) and given IV placebo (n = 7) 3) Normally ventilated (paCO2 35-45 Torr) 48 h with RA and given IV G-CSF (n=5) 4) Hyperventilated (PaCO2 15±25 Torr) for 48 h with 100% O2 and given IV placebo (n = 7) 5) Hyperventilated (PaCO2 15±25 Torr) for 48 h with 100% O2 and given IV G-CSF (n = 9) Other than animals in the control group, all animals received either 10 lg/kg/dose of rhG-CSF, at a ®nal concentration of 15 lg/ml in D5W (Filgrastin; Amgen Corp., Thousand Oaks, CA) or the equivalent volume of placebo (normal saline) IV over 1 h at 0, 12, 24, and 36 h. This twice daily dosing regimen doubles the clinically used dose for rhG-CSF (10 lg/kg/dose given once/day).

Study Protocol At the beginning of the experiment, the piglets were anesthetized with intraperitoneal pentobarbital sodium (25±30 mg/kg) and placed on a warmed operating table. They were intubated with a 3.5 ID endotracheal tube, and a 3.5 F umbilical arterial or venous catheter was inserted. Mechanical ventilation was initiated using a Bear Cub infant ventilator with positive end expiratory pressure at 3 cm H2O, inspiratory time 0.5 sec and a ¯ow rate of 8±10 1/min. Peak inspiratory pressure and ventilator rate were varied as necessary to maintain arterial PaCO2 at 15±25 torr in the hyperventilation group and 40±45 torr in the normal ventilation group. Arterial blood gases were measured every 3±4 h using a Corning 238 blood gas analyzer (Med®eld, MA.). Dextrose (5%) with 0.33 normal saline was administered at 80 cc/kg/day, and ampicillin (50 mg/kg) and gentamicin (2.5 mg/kg) were given every 12 h during the length of the study. Additional pentobarbital was administered as needed to maintain adequate sedation. Complete blood counts (CBC) with di€erential and platelet counts were measured at study entry and daily during the study. CBCs were performed electronically (Coulter STKS or GENS Counters; Coulter Electronics, Inc., Hialeah, FL) and di€erential counts were performed manually on Wrightstained blood smears. Arterial to alveolar (a/A) oxygen ratios were calculated on the arterial blood gas measured at the end of the experiment. At the conclusion of the 48 h study, the piglets were euthanized with additional pentobarbital (100 mg/kg). The chest was opened, the pulmonary artery was cannulated and the lungs were perfused with normal saline (60 ml ´ 3) to minimize red blood cell contamination of lung tissue. The trachea was isolated, cannulated and bronchoalveolar lavage (BAL) was performed using 75 ml (3 ´ 25 ml aliquots) of warmed saline. The BAL was immediately spun at 350 ´ g to remove cellular debris. The cell pellet was resuspended in 5 ml of saline and the number of cells were determined using a Coulter Counter HEMO-W (Biodynamics, Edison, NJ). The supernatant was subjected to additional cen-

232

L. I. Wolko€ et al.

trifugation at 5000 ´ g for 10 min. A portion of the supernatant was concentrated 20-fold by lyophilization and frozen at )70°C for subsequent analysis including total protein and albumin concentrations and neutrophil chemotactic activity (NCA).

Neutrophil Chemotaxis Activity (NCA) and Biochemical Assays To isolate neutrophils from unventilated control pigs, 40 ml of blood was placed in a 60 ml syringe containing 4 ml of acid citrate dextrose anticoagulant and 10 ml of dextran to agglutinate and sediment red blood cells. The white blood cell layer was drawn o€ and resuspended to a total of 35 ml in normal saline. Cells were sedimented at 350 ´ g for 10 min. Cell pellets were resuspended in 5 ml of bu€ered saline, mixed with an equal volume of Sepracell-MN (Sepratech Co, Oklahoma City, OK) and centrifuged at 3700 ´ g for 10 min followed by washing and lysing of any residual red cells. One milliliter neutrophils (3 ´ 107/ml) was labeled with Cr51, spun, washed and adjusted to a ®nal concentration of 5 ´ 105 cells/ml. The chemotaxis assay was performed as follows: A 48-well microchemotaxis chamber was used. Lower chambers held 30 ml of bu€er only (referenced as 0% activity, negative control), zymosanactivated serum (100% activity, positive control), or concentrated cell-free BAL ¯uid [11]. A single polycarbonate (pore size 8 lm) and nitrocellulose ®lter (pore size 3 lm) separated the lower from the upper chambers. To the upper wells, 50 ml of resuspended neutrophils were added. The chamber was incubated at 37°C in a humidi®ed tissue culture chamber for 50 min. Cells that penetrate the polycarbonate ®lter and were trapped in the nitrocellulose ®lter were counted in a gamma counter (LKB Instruments Inc., Gaithersburg, MD). The chemotactic activity of concentrated BAL ¯uids was calculated as a percentage of the positive control. All samples were assayed in triplicate. Concentration of total protein in the concentrated BAL was quanti®ed using the Pierce bicinchoninic microtitre technique, which has been previously described in detail [11, 18, 31]. Concentrations of albumin in concentrated BAL were measured by enzyme-linked immunosorbent assay, as has been previously described [11].

Statistical Analysis Data were analyzed using a 2 ´ 2 repeated measures analysis of variance (ANOVA). Multiple comparison testing was performed when necessary.

Results Baseline blood WBC data were similar in all groups. WBC counts did not change in piglets who did not receive rhG-CSF over the 48-h study period (Fig. 1). As expected, animals who received rhG-CSF had signi®cant increases in total WBC counts over the 48-h study period. Animals who were normally ventilated with RA and received rhG-CSF demonstrated an increase in WBC counts and received rhG-CSF from a mean of 7,280 to 62,800 (p<0.0001), whereas animals who were hyperventilated with 100% oxygen demonstrated increases of WBC counts from 8,100 to 43,500 (p<0.003) (Fig. 1). Similarly, total neutrophil counts also increased over the 48-h study period in rhG-CSF-treated animals from 4,800 to 52,500 (p<0.0001) and from 5,480 to 34,000 (p<0.003). Total neutrophil counts did not change in those animals not treated with rhG-CSF. Hematocrit and platelet counts did not change signi®cantly over the 48-h study period in any experimental group (data not shown).

G-CSF and Lung Injury

233

Fig. 1. Changes in peripheral WBC counts over the 48-h study period. Animals were treated with either rhG-CSF (10 lg/kg/dose) or the equivalent volume of normal saline given at 0, 12, 24, and 36 h. WBC counts with complete di€erentials were done at 0, 24, and 48-h. Values are means ‹ SE. Animals who did not receive rhG-CSF showed no signi®cant change in WBC number over the 48-h study period. As compared to baseline values, WBC counts increased signi®cantly in response to rhG-CSF (+ p<0.0001 and *p<0.003, respectively).

Fig. 2. BAL total cell counts. Control piglets were unventilated and sacri®ced at time 0. All other animals were ventilated for 48 h and treated with rhG-CSF (10 lg/kg/dose) or the equivalent volume of normal saline at time 0, 12, 24, and 36 h. Values are means ‹ SE. Control animals had lower BAL WBC counts compared to the other 4 groups (ANOVA, p<0.006). There were no signi®cant di€erences in BAL WBC counts among the other experimental groups.

BAL was then analyzed for several in¯ammatory markers that indicate the presence of acute lung injury. BAL total cell count was signi®cantly lower in unventilated control animals compared to the other four experimental groups (p<0.006) as expected (Fig. 2). Most importantly, treatment with rhG-CSF did not appear to have an independent e€ect on BAL cell count. While animals who were exposed to 100% O2 appeared to have higher total cell counts compared to animals who were exposed to RA, the di€erences did not reach statistical signi®cance as in our previous reports (7,11) (Fig. 2). There were no signi®cant di€erences in BAL cell counts between the two oxygen treatment groups.

234

L. I. Wolko€ et al.

Fig. 3. BAL neutrophil chemotactic activity, expressed as % of positive control (zymosan-activated serum referenced as 100% and bu€er as 0%). Control piglets were unventilated and sacri®ced at time 0. All other animals were ventilated for 48 h and treated with rhG-CSF (10 lg/kg/dose) or the equivalent volume of normal saline at time 0, 12, 24, and 36 h. Values are means ‹ SE. There were no signi®cant di€erences in NCA among the 5 groups (ANOVA).

Fig. 4. BAL total protein concentration. Control piglets were unventilated and sacri®ced at time 0. All other animals were ventilated for 48 h and treated with rhG-CSF (10 lg/kg/dose) or the equivalent volume of normal saline at time 0, 12, 24, and 36 h. Values are means ‹ SE. There were no signi®cant di€erences in BAL protein concentration among the 5 groups (ANOVA).

Neutrophil chemotaxis activity (NCA) is an early indicator of pulmonary in¯ammation and lung injury. There were no signi®cant di€erences in BAL NCA in rhG-CSF-treated animals compared to untreated animals (Fig. 3). This would suggest that rhG-CSF did not promote pulmonary in¯ammation in this animal model of acute, hyperoxic lung injury. Similarly, BAL total protein concentration, a measurement of capillary leakage, was not signi®cantly higher in piglets exposed to rhG-CSF compared to those who received saline (Fig. 4). Furthermore, BAL total albumin concentration was also not signi®cantly higher in the piglets exposed to rhG-CSF (Fig. 5). Physiologic data such as heart rate, blood pressure, and a/A ratios were comparable in piglets exposed to rhG-CSF compared to those who received saline (data not shown).

G-CSF and Lung Injury

235

Fig. 5. BAL albumin concentration. Control piglets were unventilated and sacri®ced at time 0. All other animals were ventilated for 48 h and treated with rhG-CSF (10 lg/kg/dose) or the equivalent volume of normal saline at time 0, 12, 24, and 36 h. Values are means ‹ SE. There were no signi®cant di€erences in BAL albumin concentration among the 5 groups (ANOVA).

Discussion BPD is a form of chronic lung disease that develops in newborn infants following prolonged exposure to oxygen and positive-pressure mechanical ventilation [10]. Using a newborn piglet model of acute lung injury (induced by hyperoxia and mechanical ventilation), we sought to determine whether rhG-CSF, when used to prevent late-onset sepsis, would exacerbate the pulmonary in¯ammatory response that leads to the development of BPD. The term piglet has been extensively studied under a variety of ventilatory conditions, and has been shown to be morphologically similar to that of the preterm infant, with similar antioxidant enzyme pro®les [7, 9, 11]. We demonstrated that systemic administration of rhGCSF increased peripheral white cell counts 6±9-fold over baseline levels within 48-h. In addition, rhG-CSF increased the 48-h peripheral neutrophil counts 7±11-fold over baseline levels, consistent with data from many previous animal and human studies [2±4, 6, 12, 19, 20, 22]. However, rhG-CSF (in doses that are twice the routinely used neonatal dose) did not appear to potentiate pulmonary in¯ammation, as evidenced by similar BAL cell counts, NCA, and total protein and albumin concentrations in animals who received rhG-CSF or saline placebo. Although the pathogenesis of BPD is a complex, multifactorial process, previous studies have documented that hyperoxia and mechanical ventilation initiate an acute in¯ammatory response in the lung with attraction and activation of neutrophils and subsequent release of various cytokines, proteolytic enzymes (proteases, elastases), and oxygen radicals (superoxide) [10]. Once initiated, this in¯ammatory cascade, coupled with continuing exposure to hyperoxia and mechanical ventilation, leads to the development of acute lung injury which progresses to chronic lung disease (BPD). It is this prominent in¯ammatory cascade

236

L. I. Wolko€ et al.

involving the in¯ux of neutrophils into the lungs of infants that raises concerns regarding the prophylactic use of rhG-CSF to prevent late-onset sepsis. VLBW infants are known to have a 25%±50% incidence of late-onset sepsis [8, 23, 33]. Systemic infection itself has been found to be associated with the subsequent development of chronic lung disease [8, 28]. Though causality has not been proven, many studies have identi®ed an association between both increased levels of in¯ammatory mediators and infection with the development of increased microvascular permeability and ultimately BPD [16, 27, 28]. This may have special relevance regarding those infants born to mothers with chorioamnionitis who have a higher risk for BPD despite having minimal initial lung disease [27, 28]. G-CSF is a hematopoetic growth factor that promotes the growth and differentiation of myeloid cells and their release from the bone marrow neutrophil storage pool (NSP) to the blood [25]. It has been shown to augment neutrophil functions including oxidative metabolism and superoxide production, chemotaxis, phagocytosis, and microbial killing, as well as increased expression of surface adhesion glycoproteins such as neutrophil C3bi [5, 12, 15, 32]. In animal studies, G-CSF has been shown to act synergistically with antibiotics to reduce the 1 morbidity and mortality in models of neonatal sepsis [3, 4]. In numerous studies involving septic and/or neutropenic infants, rhG-CSF has proven to be safe as well as ecacious (increasing neutrophil counts) [2, 6, 12, 22, 24, 26]. Prophylactic use of rhG-CSF has also met with some success [19]. Importantly, none of these studies demonstrated an increase in either acute or chronic lung injury in those who received rhG-CSF. Ahmad et al. [1] found no di€erences in BAL neutrophil counts, IL-6, or IL-8 levels in rhG-CSF-treated VLBW newborns compared to controls despite a several fold rise in the peripheral neutrophil count. Taken together, these studies suggest that rhG-CSF, administered to VLBW infants, enhances neutrophil function and number without signi®cant pulmonary side e€ects. To date, only one study has shown a trend toward a higher incidence of chronic lung disease among rhG-CSF recipients [29]. However, concerns still exist over whether constant up-regulation of the immune system by rhG-CSF will worsen BPD in VLBW infants. Our study is the ®rst to speci®cally investigate whether rhG-CSF potentiates lung injury in a newborn animal model of hyperoxic lung injury. The data indicate that rhG-CSF increases the number (and presumably the activation) of leukocytes in the systemic circulation but not in the lungs. This suggests that the acute e€ects of rhGCSF are restricted to the systemic circulation. Alternatively, it is possible that this animal model of acute lung injury may not have been severe enough to generate the necessary in¯ammatory mediators needed to cause activated neutrophils to enter the lungs from the peripheral blood. We have extensively studied this lung injury model in the past and have consistently demonstrated that 48 h of hyperoxia and mechanical hyperventilation causes signi®cant in¯ammatory changes in the lungs compared to both room air and unventilated controls [7, 9, 11]. In the present study, even though there were increased leukocytes in the lungs of piglets exposed to hyperoxia and mechanical ventilation compared to the room air groups, variability within the individual group may have negated any signi®cance. Despite attempts to minimize

G-CSF and Lung Injury

237

this (animals were added to the groups), variability persisted and signi®cance was not achieved. From the data in this study, it appears that rhG-CSF did not signi®cantly increase the in¯ammatory response in the lung from hyperoxia and barotrauma. This indicates that the peripheral blood and alveolar space seem to function as separate compartments and prophylactic administration of rhG-CSF to prevent late-onset sepsis in VLBW infants should not increase the incidence or severity of acute lung injury or BPD. As systemic infection itself is an independent risk factor for BPD, preventing infection with prophylactic rhG-CSF may have a derivative e€ect of reducing the incidence of BPD. Future placebo-controlled studies investigating the prophylactic use of rhG-CSF in VLBW infants are needed to con®rm these ®ndings and are currently in progress. [FDA IND #5606, Prophylaxis of Pre-eclampsia Associated Neutropenia (EFL)]. Acknowledgments. Many thanks to Shereen Dearr for her help with the data, to Doreen Wynter for her expertise in preparation of the manuscript, and to the hematology laboratory at WinthropUniversity Hospital for their prompt analysis of the piglet blood samples. This study was supported in part by a grant from Amgen Corp., Thousand Oaks, CA.

References 1. Ahmad M, Kocherlakota P, Niu J, Parton L, Munchi U, LaGamma EF (1997) Does rhG-CSF increase the risk of BPD or modulate tracheal cytokines in ELBW neonates? Pediatr Res 41:127A 2. Bedford Russel AR, Davies EG, Ball SE, Gordon-Smith E (1995) Granulocyte colony-stimulating factor treatment for neonatal neutropenia. Arch Dis Child 72:F53±F54 3. Cairo MS, Mauss D, Kommareddy S (1990) Prophylactic or simultaneous administration of recombinant human granulocyte colony-stimulating factor in the treatment of group B streptococcus sepsis in neonatal rats. Pediatr Res 27:612±616 4. Cairo MS, Plunkett J, Mauss D, van de Ven C (1990) Seven-day administration of recombinant human granulocyte colony-stimulating factor to newborn rats: modulation of neonatal neutrophelia, myelopoesis, and group B streptococcus sepsis. Blood 76:1788±1794 5. Cairo MS, van de Ven C, Toy C, Mauss D, Sender L (1989) Recombinant human granulocytemacrophage colony-stimulating factor primes neonatal granulocytes for enhanced oxidative metabolism and chemotaxis. Pediatr Res 26:395±399 6. Carr R, Modi N, Dore CJ, El-Rifai R, LiUndo D (1999) A randomized, controlled trial of prophylactic granulocyte-macrophage .colony-stimulating factor in human newborns less than 32 weeks gestation. Pediatrics 103:796±802 7. Davis JM, Dickerson B, Metlay L, Penney DP (1991) Di€erential e€ects of oxygen and barotrauma on lung injury in the neonatal piglet. Pediatr Pulmanol 10:157±163 8. Davis JM, Narula P, Aronson J, Hirsch DS (1997) Severe sepsis in infants and children. In: Fein 4 AM, Abraham EM, Balk RA, Bernard GR, Bone RC, Dantzker DR, Fink MP (eds) Sepsis and multiorgan failure, Williams and Wilkins, Maryland, pp 297±309 9. Davis JM, Penney DP, Notter RH, Metlay L, Dickerson B, Shapiro DL (1989) Lung injury in the neonatal piglet caused by hyperoxia and mechanical ventilation. J Appl Physiol 67:1007± 1012 10. Davis JM, Rosenfeld W (1994) Chronic lung disease In: Neonatology: pathophysiology and management of the newborn, 4th ed. Avery G, Fletcher MA, MacDonald MG (eds) JB Lippincott Co, Philadelphia, pp 453±477 11. Davis JM, Whitin J (1992) Prophylactic e€ects of dexamethasone in lung injury caused by hyperoxia and hyperventilation. J Appl Physiol 72:1320±1325

238

L. I. Wolko€ et al.

12. Gillan ER, Christensen RD, Suen Y, Ellis R, van de Ven C, Cairo MS (1994) A randomized, placebo-controlled trial of recombinant human granulocyte colony-stimulating factor administration in newborn infants with presumed sepsis: signi®cant induction of peripheral and bone marrow neutrophilia. Blood 84:1427±1433 13. Gladstone IM, Ehrenkranz RA, Edberg SC, Baltimore RS (1990) A ten-year review of neonatal sepsis and comparison with the previous ®fty-year experience. Pediatr Infect Dis J 9:819±825 14. Glaspy JA, Baldwin GC, Robertson PA, Souza L, Vincent M, Ambersley J, Golde DW (1988) Therapy for neutropenia in hairy cell leukemia with recombinant human granulocyte colonystimulating factor. Ann Int Med 109:789±795 15. Goldman S, Ellis R, Dhar V, Cairo MS (1998) Rationale and potential use of cytokines in the prevention and treatment of neonatal sepsis. Clin Perinatol 25:699±710 16. Groneck P, GoÈtze-Speer B, Oppermann M, Ei€ert H, Speer CP (1994) Association of pulmonary in¯ammation and increased microvascular permeability during the development of BPD: a sequential analysis of in¯ammatory mediators in respiratory ¯uids of high-risk preterm neonates. Pediatrics 93:712±718 17. Haeney M (1994) Infection determinants at extremes of age. J Antimicrob Chemother 34:A1±A9 18. Hsu SM, Raine L, Fanger H (1981) A comparitive study of the peroxidase-antiperoxidase method and an avidin-biotin complex method for studying polypeptide hormones with radioimmunoassay antibodies. Am J Clin Pathol 75:734±738 19. Kocherlakota P, La Gamma E (1998) Preliminary report: rhG-CSF may reduce the incidence of neonatal sepsis in prolonged preeclampsia-associated neutropenia. Pediatrics 102:1107±1111 20. Kocherlakota P, La Gamma EF (1997) Human granulocyte colony-stimulating factor may 2 improve outcome attributable to neonatal sepsis complicated by neutropenia. Pediatrics 100:(1)e6 21. Kocherlakota P, LaGamma EF, Ahmad M, Breen C, Golightly MG, Fleit HB (1998) Neonatal sepsis-an unsolved problem. In: Marstyn G, Dexter TM, Foote MA (eds) Filgrastim in clinical practice. Marcel Dekker, Inc, New York, pp 469±490 22. LaGamma EF, Alpan O, Kocherlakota P (1995) E€ect of granulocyte colony-stimulating factor on preeclampsia-associated neonatal neutropenia. J Pediatr 26:457±459 23. LaGamma EF, Drusin LM, Mackles AW, Machalek S, Auld PAM (1983) Neonatal infections: an. important determinant of late NICU mortality in infants less than1000 g at birth. Am J Dis Child 137:838±841 3 24. Leibovitz E, et al (June 1995) A pilot, prospective trial of recombinant human granulocyte colonystimulating factor (rhG-CSF, Filgrastim) in the treatment of newborn infants with presumed sepsis and neutropenia (abstract). Int Congress of Chemotherapy, Montreal. 25. Lord BI, Bronchud MH, Owens S, Chang J, Howell A, Souza L, Dexter TM (1989) The kinetics of human granulopoiesis following treatment with granulocyte colony-stimulating factor. Proc Natl Acad Sci 86:9499±9503 26. Makhlouf RA, Doron MW, Bose CL, Price WA, Stiles AD (1995) Administration of granulocyte colony-stimulating factort neutropenic low birth weight infants of mothers with preeclampsia. J Pediatr 126:454±456 27. Pierce MR, Bancalari E (1995) The role of in¯ammation in the pathogenesis of bronchopulmonary dysplasia. Pediatr Pulmomol 19:371±378 28. Rojas MA, Gonzalez A, Bancalari E, Claure N, Poole C, Silva-Neto G (1995) Changing trends in the epidemiology and pathogenesis of neonatal chronic lung disease. J Pediatr 126:605±610 29. Schibler KR, Osborne KA, Leung LY, Le TV, Baker SI, Thompson DD (1998) A randomized, placebo-controlled trial of granulocyte colony-stimulating factor administration to newborn infants with neutropenia and clinical signs of early-onset sepsis. Pediatrics 102:6±13 30. Siegel J, McCracken G (1981) Sepsis Neonatorum. N Engl J Med 304:642±647 31. Smith PK, Krohn RI, Hermanson GT, Malliaa AK, Gartner FH, Provenzano MD, Fujimoto EK, Goeke NM, Olson BJ, Klenk DC (1985) Measurement of protein using bicinchoninic acid. Anal Biochem 150:76±85 32. Sreenan C, Osiovich H (1999) Myeloid colony-stimulating factors. Use in the newbom. Arch Pediatr Adolesc Med 153:984±988

G-CSF and Lung Injury

239

33. Stoll BJ, Gordon T, Karones SB, Shankaran S, Tyson JE, Bauer CR, Fanaro€ AA, Lemmons JA, Donovan EF, Oh W, Stevenson DK, Ehrenkranz RA, Papile LA, Verter J, Wright LL (1996) Late-onset sepsis in very low birth weight neonates: a report from the National Institute of Child Health and Human Development Neonatal Research Network. J Pediatr 129:63±71 34. Thompson PJ, Greenough A, Hird MF, Philpott-Howard J, Gamsu HR (1992) Nosocomial bacterial infections in very low birth weight infants. Eur J Pediatr 151:451±454 35. Watterberg KL, Demers LM, Scott SM, Murphy S (1996) Chorioamnionitis and early lung in¯ammation in infants in whom bronchopulmonary dysplasia develops. Pediatrics 97:210±215 36. Yasuda H, Ajiki Y, Shimozato T, Kasahara M, Kawada H, Iwata M, Shimizu K (1990) Therapeutic ecacy of granulocyte colony-stimulating factor alone and in combination with antibiotics against Pseudomonas aeruginosa infections in mice. Infect Immun 58:2502±2509 Accepted for publication: 16 July 2002