nature publishing group Articles Clinical Investigation Neutrophil and monocyte toll-like receptor 4, CD11b and reactive oxygen intermediates, and neuroimaging outcomes in preterm infants Fiona M. O’Hare1–4, William Watson2,3, Amanda O’Neill2,3, Tim Grant2,3, Chike Onwuneme1–4, Veronica Donoghue1,5, Eoghan Mooney6, Paul Downey6, John Murphy1, Anne Twomey1 and Eleanor J. Molloy1–3,7–9 BACKGROUND: Activated leukocytes and infection are impli- who died with white matter injury (5) (see Supplementary cated in neonatal brain injury. Leukocyte surface receptors References 11–14 online). Postnatal infection also contributes are increased in stroke models and may be targets for future to the development of PVL and CP (6,7). Inhibiting inflamma- adjunctive therapies. tory responses may decrease secondary brain injury following METHODS: Serial blood samples were analyzed from preterm infection or hypoxia–ischemia (8) and recently tertiary brain infants (n = 51; <32 wk gestation) on days 0, 1, 2, and 7 of life. injury has been described as a possible mechanism of preterm Monocyte and neutrophil activation were evaluated via flow brain injury with persistent long-term inflammation (9). We cytometry at baseline and following endotoxin stimulation ex were interested in the systemic inflammatory response in new- vivo by measuring CD11b (activation), toll-like receptor 4 (TLR- born infants at risk of brain injury by examining markers of 4; endotoxin recognition) expression, and intracellular reactive activation of monocytes and neutrophils. oxygen intermediate (ROI) production (function). Reactive oxygen intermediate (ROI) generation is essential RESULTS: Control preterm infants with normal neuroimaging for neutrophil intracellular killing of invading microorganisms had elevated baseline CD11b and TLR-4 expression and ROI following phagocytosis. ROIs are a major mechanism of innate production compared with adults as well as a robust immune antimicrobial host defense (see Supplementary Reference 15 response following endotoxin stimulation. Preterm infants online) but can cause damage by oxidizing membrane phos- with abnormal neuroimaging had increased neutrophil TLR-4 pholipids, proteins, nucleic acids, and nucleotides (10). We and ROI compared with all controls. studied ROI production as a marker of immune cell func- CONCLUSION: Preterm infants have a robust immune tion. CD11b is a receptor on the cell surface that is important response compared with adults. Increased TLR-4 expression for neutrophil and monocyte migration to sites of infection/ in preterm infants with abnormal neuroimaging is similar to inflammation. Neonatal neutrophil migration is decreased at findings in adult stroke. In addition, ROI production may cause birth due to decreased total cell content of CD11b and issues tissue injury. The modulation of these responses may be ben- related to its cell surface translocation. Although baseline eficial in preterm inflammatory disorders. CD11b expression is reported to be similar to that of adults, neonates are unable to upregulate CD11b expression to the same magnitude following lipopolysaccharide (LPS) stimula- reterm infants are susceptible to inflammatory disorders tion, especially in preterm infants (11). The key receptor for Presulting in multiorgan dysfunction (1). Systemic inflam- recognizing endotoxin on the immune cell surface is toll-like mation may be the final common pathway for insults caused receptor 4 (TLR-4). Healthy neonates have similar basal TLR-4 by both hypoxia–ischemia and infection in these infants and expression to adults. Both term and preterm neonates increase may be associated with brain injury (2) (see Supplementary TLR-4 expression in response to LPS. Responses to LPS are Reference 1 online). Many studies demonstrate an asso- determined by the level of TLR-4 expressed, and overexpres- ciation between maternal/fetal infection and periventricu- sion can lead to uncontrolled inflammation resulting in dam- lar leukomalacia (PVL) detected on cranial ultrasound (3) age to healthy tissues. Shen et al. (12) showed a rapid increase (see Supplementary References 2–5 online), or the later in TLR-2 and TLR-4 expression over the first month of life but development of cerebral palsy (CP) (4) (see Supplementary no parallel increase in LPS-induced cytokine production. References 6–10 online). Elevated cytokines have been We hypothesized that markers of neutrophil and monocyte detected histologically in the brains of preterm infants activation may be altered in preterm infants with abnormal 1Department of Paediatrics, National Maternity Hospital, Dublin 2, Ireland; 2University College Dublin School of Medicine and Medical Sciences, Dublin, Ireland; 3University College Dublin Conway Institute of Biomolecular and Biomedical Sciences, Dublin, Ireland; 4National Children’s Research Centre, Dublin, Ireland; 5Department of Radiology, Children’s University Hospital, Dublin, Ireland; 6Department of Pathology, National Maternity Hospital, Dublin 2, Ireland; 7Department of Neonatology, Our Lady’s Children’s Hospital, Dublin, Ireland; 8Department of Paediatrics, Royal College of Surgeons in Ireland, Dublin, Ireland; 9Department of Paediatrics, Trinity College Dublin, University of Dublin, Dublin, Ireland. Correspondence: Eleanor J. Molloy ([email protected]) Received 12 February 2014; accepted 5 January 2015; advance online publication 20 May 2015. doi:10.1038/pr.2015.66 82 Pediatric REsEarcH Volume 78 | Number 1 | July 2015 Copyright © 2015 International Pediatric Research Foundation, Inc. Preterm immune function and brain injury Articles neuroimaging. We examined markers of monocyte and neu- Table 1. Demographic variables trophil function and activation (CD11b (activation), TLR-4 Preterm infants’ expression (LPS recognition), and ROI production (function)) neuroimaging outcome serially over the first week of life and correlated our findings Characteristics Normal Abnormal P value with neuroimaging. Neonatal Gestational age (weeks+days) 28+6 ± 1+6 28+3 ±1+6 0.519 RESULTS (mean ± SD) Patient Demographics Gestational age, <28/40 (N (%)) 12 (32) 6 (43) 0.487 Fifty-one preterm infants born <32 wk gestation were included, and three infants died (normal neuroimaging n = 40; Gestational age, >28/40 (N (%)) 25 (68) 8 (57) 0.487 abnormal neuroimaging/ Death (RIP); n = 11). A total of 404 Birth weight (grams) (mean ± SD) 1,178 ± 324 1,117 ± 305 0.542 samples were processed. There were no differences in gender Sex (N (%)) (male) 27 (73) 11 (79) 0.492 distribution, preeclampsia, prolonged rupture of membranes, SVD (N (%))(mean ± SD) 10 (28) 3 (21) 0.646 maternal pyrexia, histological chorioamnionitis, surfactant Emergency CSxn (N (%)) 26 (70) 11 (79) 0.741 treatment, gestational age, birth weight, mode of delivery, doses of antenatal steroids received, Apgar scores, cord or admission Antenatal steroids, one dose (N (%)) 8 (22) 4 (29) 0.737 blood gas parameters, nasal continuous positive airway pres- Antenatal steroids, two doses (N (%)) 28 (76) 10 (71) 0.737 sure (CPAP) hour, or duration of intubation, between preterm Apgar score at 1 min (mean ± SD) 6 (3) 7 (2) 0.512 neonates with normal and abnormal neuroimaging (Table 1). Apgar score at 5 min (mean ± SD) 8 (2) 9 (1) 0.092 There was a statistically significant difference in mortality (n Cord pH 7.33±0.05 7.31±0.09 0.395 (%)) observed between the two groups: normal vs. abnormal neuroimaging 0 (0) vs. 3 (21) (P = 0.04). Cord base excess −2.7 ± 2.8 −5.6 ± 3.1 0.088 All infants had serial cranial ultrasounds, and 29 infants Admission gas, pH (mean ± SD) 7.27 ± 0.10 7.24 ± 0.14 0.315 had a magnetic resonance imaging (MRI) brain at term cor- Admission gas, base excess −3.7 ± 4.1 −5.5 ± 4.1 0.196 rected age. MRI was scored according to the Inder criteria (6) (mean ± SD) independently by a consultant Pediatric Radiologist. Twenty Admission gas, lactate 4.5 ± 1.4 6.4 ± 2.5 0.502 infants had completely normal imaging (white matter (WM) (mean ± SEM) score = 5–6; gray matter (GM) score 3–5). Nine infants had Chorioamnionitis (N (%)) 7 (19) 0 (0) 0.080 evidence of WM abnormality (mild = 4: WM score 7–9; mod- Surfactant (N (%)) 13 (35) 5 (36) 0.824 erate = 3: WM score 10–12; severe = 2: WM score 13–15). One nCPAP hours (mean ± SD) 0.688 infant had both WM and GM abnormality (GM score 6–9; 197 ± 43 162 ± 81 Tables 2 and 3). Twenty-two infants had only cranial ultra- Hours intubated (mean ± SEM) 162 ± 63 88 ± 25 0.432 sound imaging. No abnormality was detected in 14 infants. CLD (N (%)) 8 (20) 1 (9) 0.401 Evidence of intraventricular hamorrhage (IVH) was detected NEC (N (%)) 3 (8) 1 (9) 0.880 in five infants (grade 1 = 4; grade 2 = 1). Three infants had RIP (N (%)) 0 (0) 3 (21) 0.040 increased echogenicity in the periventricular WM. CLD, chronic lung disease; CSxn, caesarean section; nCPAP, nasal continuous positive There were no other significant differences between the airway pressure; NEC, necrotizing enterocolitis; SVD, spontaneous vaginal delivery. study groups with respect to chronic lung disease, patent duc- tus arteriosus, necrotizing enterocolitis, retinopathy of prema- turity, late-onset sepsis (LOS), and number of septic episodes ROI production following ex vivo LPS stimulation. Higher lev- and antibiotic days. There were no significant differences in els of ROI production were seen in both preterm groups com- duration of intubation, intermittent positive pressure ventila- pared with adults following stimulation which was statistically tion, nasal CPAP or nasal prong oxygen hours required,