Factors Associated With Macrophage Migration Inhibitory Factor in Neonates: a Pilot Study

Objective: The study was to conduct an investigation of associated factor with macrophage migration inhibitory factor (MIF) in neonates. Results: MIF was measured by Enzyme-linked Immunosorbent Assay (ELISA) using sera obtained from sick neonates on the median postnatal 2.0 nd day. Clinical details were reviewed from medical record and grouped into preterm and near or full-term neonates based on 34 weeks of gestation at birth. Statistical analyses were performed between MIF concentration and clinical factors. In total, 77 neonates consisted of preterm (n = 42) and term neonates (n = 35) were included. The median value of plasma MIF was higher in preterm neonates (7,037.6 pg/mL, IQR: 3,285.2-12,267.4) than in the others (3,968.8 pg/mL, IQR: 2,566.4-6,995.2, P = 0.013). Among 42 preterm neonates, those with necrotizing enterocolitis (NEC) in later had higher MIF concentration (12,170.9 pg/mL, IQR: 8,353.3-23,537.7, n = 9) compared with the others without NEC (5,189.6 pg/mL, IQR: 3,220.2-9,097.9, n = 33, P = 0.016). Associated factors of higher MIF in neonates were preterm neonates and NEC. P Abbreviations; PROM, prolonged 5`-AS, Apgar Score at RDS, respiratory distress syndrome; PDA, patent ductus arteriosus; IVH, intraventricular hemorrhage; NEC, necrotizing enterocolitis; BPD, bronchopulmonary dysplasia. from blood, urine or cerebral uid. P values were obtained using Mann-Whitney U or Fisher’s exact test. Abbreviations; PROM, prolonged rupture of membrane; 5`-AS, Apgar Score at 5-minute; RDS, respiratory distress syndrome; PDA, patent ductus arteriosus; IVH, intraventricular hemorrhage; NEC, necrotizing enterocolitis; BPD, bronchopulmonary dysplasia.


Introduction
Macrophage migration inhibitory factor (MIF) is the potent in ammatory cytokine and exists in the blood with a low concentration in healthy adults [1,2]. Preformed MIF exists already in cytoplasm and rapidly releases into the blood in response to microbial, proliferative and hypoxic stimuli [3,4]. It has been studied in sepsis, cancer, autoimmune, and metabolic diseases in adults [4]. In neonates, the plasma level of MIF is higher than in adults and a higher level of MIF can play a major role in innate immune response [5]. However, there has been a small number of study of MIF in preterm neonates until now. Preterm neonates have higher susceptibility to infection than full-term neonates due to detrimental effect of immaturity of the organ and immune system [6,7]. However, vulnerability of preterm neonates to infection is not con ned to the immature immune system or organs, and it is unclear how the set point of the immune system is regulated [8]. In addition to vulnerability to infection, preterm neonates are susceptible to respiratory distress syndrome (RDS), necrotizing enterocolitis (NEC), or bronchopulmonary dysplasia (BPD). Though immaturity could be the main factor, a number of immune or in ammatory responses could work as a part of pathogenesis in the various diseases of preterm neonate. On the aspect of different diseases and microbial susceptibility between preterm and term neonates, we hypothesized that MIF of preterm neonates would be different from that in term neonates and the difference of MIF might associate with the development of prematurity-associated diseases.

Subjects and blood samples
Clinical data and blood samples of 77 neonates with parental consents were obtained from Biobank of Korea. Major congenital anomalies or chromosomal abnormalities were excluded. The subjects were classi ed into 2 groups as neonates prematurely born at ≤ 34 (n = 42) and the others born after 34 weeks of gestation (n = 35). Clinical information was reviewed as follows.
1. Obstetric problems were reviewed under following de nitions; prolonged rupture of membrane (PROM) as leakage of amniotic uid over 18 hours prior to giving birth; maternal hypertensive disease as underlying essential hypertension, or pregnancy-induced hypertension, preeclampsia, eclampsia, or abruption of placenta caused by hypertension; preterm labor pain as pain caused by uterine contraction which occurred before 37 weeks of gestation and could not be controlled by medication; prenatal steroid as injection of betamethasone to mother prior to giving birth.
3. Postnatal clinical factors were investigated as following de nitions; RDS as ground glass opacity in both lungs on chest X-ray with respiratory di culties; patent ductus arteriosus (PDA) as an existence of hemodynamically signi cant shunt from aorta to pulmonary artery on echocardiography [9]; intraventricular hemorrhage (IVH) as intraventricular or intracerebral hemorrhage on ultrasonography of the brain [10]; Sepsis as positive result of culture from blood, urine, cerebrospinal uid or other bodily uid; NEC as stage ≥ 2 of modi ed Bell's classi cation [11]; BPD as a pulmonary condition which supplemental oxygen was necessary at 36 weeks of gestation or postnatal 28 days [12]; laboratory ndings were reviewed on the day of sample collection by the Biobank of Korea.
After clinical laboratory tests from neonate with the 2nd day of life (range: 1-9), remnant blood was collected and stored at -80℃ till analysis.

Plasma MIF
Plasma MIF was measured by an enzyme-linked immunosorbent assay (ELISA) using a human MIF ELISA kit (LSBio, Seattle WA, US). The concentration was compared between preterm and term neonates.
Comparisons of clinical characteristics were performed to investigate associating factor with MIF among 77 neonates. And then, MIF was statistically analyzed according to the presence of prematurity-related diseases such as RDS, PDA, IVH, NEC or BPD among the 42 preterm neonates.

Statistical analyses
Variables were reported as median and interquartile range (IQR), or number and percentage. Comparisons according to clinical factors were performed using Mann-Whitney U test or Fisher's exact test. The nonparametric Spearman test was used for assessment of correlation between continuous variables and the authors excluded one of them when the coe cient was more than 0.7. To investigate an associating factor with MIF in neonates, multivariate regression analysis was done among statistically signi cant factors on univariate comparisons. The P-value was set as < 0.05. Statistical analyses were performed using SPSS Statistics 25.0 (IBM, NY, US).

Clinical characteristics of included neonates
A total of 77 neonates with 34 6/7 weeks of gestation and weighed 2,340 g were included. Among them, 42 were born at 33 weeks of gestation (IQR: 31.9-34.3) and weighed 2,075 g (IQR: 1,690-2,270). Two prematurely born infants died after 5 and 12 months after birth because of sepsis and pneumonia, respectively. Clinical characteristics were described between the 2 groups at Table 1. The prenatal factors, sex, 5`-AS and breast milk feeding were not statistically different between the 2 groups. RDS, PDA, IVH, sepsis, NEC and hospital duration were signi cantly different between the 2 groups.

Plasma MIF and clinical factors
The median plasma MIF concentration was 4,534.1 pg/mL (IQR: 2,945.0-9,160.5), in total. MIF was not signi cantly different according to perinatal factors such as obstetric problems, use of prenatal steroid, number of gestation, delivery method, sex or feeding formula (Fig. 1). And it was not correlated with postnatal day (range: 1 to 9 days after birth, r = -0.101, P = 0.383). MIF was higher in preterm neonates (7,

Plasma MIF and laboratory ndings
To investigate whether MIF could be associated with laboratory ndings, speci c cell type and in ammatory marker were reviewed and statistical compared ( Table 1). The number of neutrophil, lymphocyte, or monocyte using Beckman Coulter DxH 800 (Beckman Coulter, CA, US) was not signi cantly different between the 2 groups. Plasma MIF showed a weakly negative correlation with monocytes (r = -0.256, P = 0.024).

Discussion
MIF is a proin ammatory cytokine and plays a role in modulating diverse immune responses. It has been studied in diverse diseases in adults including sepsis, autoimmune disease, cancer and so on [4,13,14].
However, a few studies on the role of MIF in neonates have been reported and the clinical signi cance has been unclear. Recently, Roger et al reported that plasma MIF in healthy neonates was 10-fold higher than in adults, and MIF obtained from cord blood was lower in preterm neonates than in term [3].
Regarding to preterm birth, elevated MIF in cord blood and lower MIF in amniotic uid were reported previously [15,16]. In this study, MIF in preterm neonates was about 2-fold higher than that of term neonates (P = 0.013, Fig. 1). The previous reports had studied MIF from cord blood and amniotic uid [3,5,15,16], but we studied MIF in peripheral venous blood from the 2nd day-old neonates with diverse clinical conditions. In this study, higher MIF of preterm than term neonates was not associated with perinatal factors including obstetric problems, use of prenatal steroid, multi-gestation, mode of delivery, sex, or 5'-AS (Fig. 1). Although there was no statistically signi cant difference according to RDS, PDA, and IVH among preterm neonates (Fig. 2), unstable early clinical condition in preterm neonates compared to term neonates might affect to high MIF concentration in this study. There have been a few studies on the role of MIF in RDS and BPD [3,17,18]. It was emphasized that MIF could promote lung development, provoke hypoxia-induced lung injury in mice, and develop BPD [3,[17][18][19]. Apart from the previous studies, we found no signi cant difference in plasma MIF according to RDS or BPD among 42 preterm neonates (Fig. 2). The differences from the previous studies might be caused by different samples of cord blood or tracheal aspirates [3,17], character or number of inclusions [3,18], and species of mouse or human [17,19]. Since BPD could be developed in preterm infants caused by multifactorial factors including arrest of lung development, mechanical trauma, oxygen toxicity, infection, in ammation, and presence of PDA, some limitation might exist for prediction of BPD based on MIF level in the early period of preterm neonate [20]. Although there have been many reports of MIF in sepsis in adults, the role of MIF in neonatal sepsis has been unclear. Roger et al. reported MIF can play a role in sustaining the innate immune response of neonate [5]. In our result, plasma MIF of neonates with sepsis was higher than those without sepsis but was not signi cantly different (median 8,180.65 pg/mL, IQR 3,871.28-13,110.33 vs. 5,379.14 pg/mL, IQR 3,044.88-11,810.20, P = 0.252). Further study with larger number of inclusions would be necessary.
MIF in preterm neonates who experienced NEC in later was noted at about 2-fold higher than that of preterm without NEC in this study, albeit small number of patients with NEC (n = 9, Fig. 2). NEC is a systemic devastating disease in prematurely born infants not con ned to intestine [21]. Although the survival rate of preterm neonates has been increased, mortality or complication caused by NEC has not been decreased. The main risk factors of NEC are shown to be prematurity, bacterial colonization and formula feeds, which can disturb the in ammatory balance that consequently leads to bowel necrosis and systemic catastrophic state [22], [23]. However, the pathogeneses of NEC are still unclear. Recently, activation of Toll-like receptor 4 (TLR4) signaling has been focused on pathogenesis of NEC [24][25][26]. MIF could modulate a host immune response by regulating the expression of TLR4 to lipopolysaccharides in NEC [13,27]. On the aspect of activation of pro-in ammatory response and TLR4 signaling in NEC, the authors tentatively suggested that MIF could play a role on the initiation of NEC and high MIF concentration of preterm neonate in the early postnatal period could be related with occurrence of NEC in later.
In the present study, MIF was negatively correlated with monocyte count (r = -0.256, P = 0.024). Neonatal monocyte can play phagocytic function and kill bacteria intracellularly, but their ability of signal ampli cation and cytokine production are signi cantly reduced, compared to monocyte in adult [28]. The weakly negative correlation between MIF and monocytes in this study might re ect reduced neonatal monocyte function and secretion of MIF from various cells other than monocyte, indirectly [4].
The authors found plasma MIF was higher in preterm neonates than in term during transitional period, and reviewed clinical characteristics in detail and tried to investigate clinical factors associated with MIF, unlike the previous reports [4,15,17,19,20,29]. By comparing the concentration of MIF according to each prematurity-associated disease, higher MIF level in early postnatal period was statistically signi cant with the occurrence of NEC in later among preterm neonates. Further prospective research regarding the role of MIF in preterm neonates would be warranted.

Limitations
Retrospective and exploratory nature with small number of inclusions was the rst. The second, we could not obtain additional MIF in neonates with NEC when the disease occurred. Hence, the change of MIF before and after NEC was not available. The third, since the remnant blood after clinical laboratory tests were used, more detailed investigation of speci c neonatal immune cells using ow cytometry could not be performed because of insu cient amount of blood.

List Of Abbreviations
MIF, macrophage migration inhibitory factor; ELISA, Enzyme-linked Immunosorbent Assay; NEC, necrotizing enterocolitis; RDS, respiratory distress syndrome; BPD, bronchopulmonary dysplasia; PROM, prolonged rupture of membrane; GA, gestational age at birth; BW, birth weight; 5'-AS, Apgar score at 5-minute; PDA, patent ductus arteriosus; IVH, intraventricular hemorrhage; IQR, interquartile range; TLR4, Toll-like receptor 4 Declarations Ethics approval and consent to participate: The study was reviewed and approved by the Institutional Review Board of our Hospital (GNUH 2018-06-007) in accordance with the Declaration of Helsinki.
Consent to publish: Figure 1 Comparisons of the plasma level of macrophage migration inhibitory factor (MIF) according to the perinatal variables among total 77 neonates. Data were presented as median with interquartile range. * P = 0.009 and ** P >0.05 were obtained by Mann-Whitney U test according to each clinical variable.