Neutrophil and endothelial adhesive function during human fetal ontogeny

Sample collection and study population

All neonates included in the study were delivered by Caesarean section. Umbilical cord blood and a piece of the umbilical cord were collected immediately after delivery. Exclusion criteria were maternal HIV infection, congenital malformations, or suspected syndromes, vaginal delivery, and familial immune diseases. Peripheral venous blood from infants and healthy adult volunteers was drawn by venipuncture. Standard blood collection tubes (S-Monovette, Sarstedt, Nümbrecht, Germany) containing trisodium citrate were used for anticoagulation of blood samples.

Based on their gestational age, as estimated by the date of the last menstrual period and by ultrasound measurements, infants were grouped into extremely premature infants (≤30 completed weeks of gestation), moderately premature infants (30+1–36+6 weeks of gestation), and mature neonates (>37 completed weeks of gestation). Because of small sample volumes, especially in the extremely premature infant group, we were unable to perform all measurements on each individual. The exact number of subjects for each experiment is given in the figure legends. We recorded baseline characteristics, including gender and birth weight; clinical data, such as antenatal steroid treatment; and the reason for prematurity or Caesarean section, as well as laboratory values (cord blood pH, differential blood-cell count, C-reactive protein).

Isolation of PMNs

As leukocyte and differential counts vary largely among different neonatal age groups, we isolated PMNs from the umbilical cord or peripheral venous blood for standardization of experimental conditions. Whole blood was layered onto a Ficoll density gradient (LSM 1077; PAA Laboratories GmbH, Coelbe, Germany) and centrifuged (1200 g, 20 min, room temperature). The resulting erythrocyte-granulocyte pellet was washed twice in Dulbecco's PBS (1×, without Ca²⁺ and Mg²⁺; Invitrogen GmbH, Darmstadt, Germany) before lysis of erythrocytes by addition of hypotonic buffer (0.15 M NH₄Cl, 0.01 M NaHCO₃⁻, 0.001 M EDTA, in Aqua ad injetabilia) for 7 min. The remaining granulocytes were washed twice, resuspended in 1 ml PBS, and counted in a Neubauer chamber using Turks solution (Merck, Darmstadt, Germany). By this method, we achieved 76% (premature infants) to 96% (adults) purity and >90% viability of cells. For some experiments, whole blood from adult donors was pretreated with bethamethasone (0.02 μg/ml; Celestan, Essex Pharma GmbH, Munich, Germany) for 1 h before isolation of PMN.

Isolation and culture of HUVECs

ECs were isolated from the umbilical cord vein using collagenase A (1 mg/ml; Roche Diagnostics GmbH, Mannheim, Germany), as reported previously [16]. The cells were resuspended in EC growth medium, supplemented with 10% FCS, basic fibroblast growth factor (1.0 ng/ml), EC growth supplement/heparin (0.004 ml/ml), endothelial growth factor (0.1 ng/ml), hydrocortisone (1.0 μg/ml), and phenol red (0.62 ng/ml; all PromoCell GmbH, Heidelberg, Germany) and 1% penicillin/streptomycin (10,000 U/10 mg/ml; PAA Laboratories GmbH), and grown in standard cell culture dishes. When confluent, cells were harvested using trypsin/EDTA (0.5 mg/0.22 mg/ml; PAA Laboratories GmbH) and cultured further in Medium 199 (Invitrogen GmbH) containing 10% FCS. For flow cytometry, cells of Passage II were collected by use of EDTA (5 mM; Merck) alone. For flow chamber assays, cells of Passage II were transferred into microflow chambers (μ-Slide VI ibiTreat; ibidi, Martinsried, Germany) and grown to confluence overnight. In some cases, freshly isolated HUVECs were incubated with betamethasone (0.02 μg/ml; Celestan, Essex Pharma GmbH) for 12 h and processed further as described above.

Flow chamber experiments

To analyze PMN recruitment, we used an in vitro flow chamber assay. For this purpose, glass capillaries (2×0.2 mm; VitroCom, Mountain Lakes, NJ, USA) were assembled into microflow chambers, as described previously [17]. To assess PMN rolling, the flow chambers were coated with rhP-selectin (CD62-P, 5 μg/ml; R&D Systems, Wiesbaden-Nordensteadt, Germany) or rhE-selectin (CD62-E, 5 μg/ml) in PBS with 0.1% BSA (PAA Laboratories GmbH). For adhesion studies, a combination of selectin (rhP-selectin 10 mg/ml or rhE-selectin 4 mg/ml), chemokine rhCXCL8 (IL-8, 10 μg/ml; R&D Systems), and integrin-ligand rhICAM-1 (CD54, 4 μg/ml; R&D Systems) was used. After overnight incubation, the flow chambers were blocked with 5% casein (from bovine milk; Sigma-Aldrich, Munich, Germany) in PBS for 2 h at room temperature and flushed with PBS. For negative controls, we used four approaches: chemokine or selectin omission, addition of EDTA to the cell suspension, and coating with PBS/0.1% BSA alone. Before the experiments, the isolated PMNs were resuspended in DMEM (low glucose without glutamine; Invitrogen GmbH), containing 1% BSA, CaCl₂ (1 mM), and MgCl₂ (1 mM), to a calculated concentration of 250,000 cells/ml (rolling assay) or 500,000 cells/ml (adhesion assay). The cell suspension was flushed through the flow chamber using a high-precision syringe pump (Harvard Apparatus, Holiston, MA, USA) at a flow rate of 0.115 ml/min, resulting in a shear stress of ∼1 dyne/cm² [17].

In other experiments investigating the contribution of neonatal ECs to PMN recruitment, we stimulated HUVECs grown in ibidi flow chambers (3.8×4 mm, μ-Slide VI) with LPS (10 μg/ml, LPS from Escherichia coli 0111:B4; Sigma-Aldrich) for 4 h. After washing with HBSS (Pharmacy of the University Hospital Munich, Germany), the chambers were connected to a PE 50 tube (intramedic polyethylene tubing, inner diameter 0.58 mm; outer diameter 0.965 mm; Becton Dickinson, Franklin Lakes, NJ, USA) and perfused with PMNs from a healthy adult donor at a concentration of 250,000 cells/ml and a rate of 0.91 ml/min, resulting in a shear stress of ∼1.6 dyne/cm² based on the manufacturer's description (ibidi Application Note #11; www.ibidi.com).

Microscopy and data analysis

The flow chamber studies were performed on an upright microscope (BX51; Olympus, Hamburg Germany) with a saline immersion objective (XLUMPlanFl, 20×, 0.95 NA) or on an inverse microscope (IM35, Nikon Fluor 20×, 0.75 NA; Zeiss MicroImaging GmbH, Munich, Germany) for observation of ibidi chambers. Experiments were recorded for 10 min on a Super VHS recorder via a charge-coupled device camera (Model CF8/1; Kappa optronics GmbH, Germany) for later, off-line analysis. PMN rolling was quantified by counting the number of cells/FOV (740 mm×565 μm), moving at a velocity significantly lower than the center-line velocity for >30 s. Adherent PMNs were defined as cells that remained stationary for >30 s.

Flow cytometric analysis of adhesion molecule expression

Surface expression of adhesion molecules on PMNs and ECs was measured by flow cytometry. Isolated PMNs were incubated with primary mAb (all IgG1 mouse anti-human; BD PharMingen, Heidelberg, Germany) to LFA-1 (CD11a), Mac-1 (CD11b), PSGL-1 (CD162), or CXCR2 (CD182) for 45 min on ice. A PE-labeled goat anti-mouse antibody was used as a secondary antibody. Unstimulated and LPS-stimulated (4 h, 10 μg/ml) HUVECs of Passage II were stained with PE-labeled mAb (all IgG1 mouse anti-human; BD PharMingen) to E-selectin (CD62-E) or ICAM-1 (CD54) for 45 min on ice. For EC characterization, unstimulated HUVECs were treated with PE-labeled antibody to PECAM (CD31, IgG1; BD PharMingen) or with unlabeled primary antibody to vWf (IgG1; Dako, Eching, Germany) with PE-labeled goat anti-mouse secondary antibody. Measurements were performed on a FACSorter flow cytometer (Becton Dickinson) using the CellQuest Pro software (Becton Dickinson). The MFI for 10,000 cells/sample was obtained using log detection settings. All antibodies were normalized against isotype-matched controls.

Immunofluorescence staining

HUVECs from premature and mature neonates (for isolation procedure, see above) were grown to confluence in μ-Slides (ibidi) and fixed by addition of 3.7% paraformaldehyde (Applichem, Darmstadt, Germany) in PBS (Invitrogen GmbH) for 10 min at room temperature. After rinsing with PBS, cells were blocked with 1% BSA (PAA Laboratories GmbH) in PBS for 30 min at room temperature. Cells were incubated with primary monoclonal mouse anti-human VE-cadherin (CD144, IgG1; eBioscience, Frankfurt, Germany) for 2 h at room temperature, washed, and incubated with a secondary Alexa 546-labeled goat anti-mouse antibody (Invitrogen GmbH) for 2 h at room temperature. After repeated washing steps to remove unbound antibody, mounting medium (Permafluor; Beckman Coulter, Fullerton, CA, USA) was added to the channels, and VE-cadherin expression was detected by confocal laser-scanning microscopy (Leica TCS SP5; Leica Microsystems, Wetzlar, Germany) using a 40× oil immersion objective (HCX PLAPO CS 40×/1.25–0.75 oil) and Leica Application Suite software.

Statistical analysis

For multiple comparisons between groups, a Kruskal-Wallis ANOVA on ranks was used, followed by Student-Newman-Keuls or Dunn's post hoc test, as appropriate. To test for nonlinear correlation, Spearman's rank correlation coefficient was determined. A P value <0.05 was considered statistically significant. All statistical analyses were carried out with SigmaStat 3.5 (Systat Software GmbH, Germany). Data are presented as mean ± sem.

Study approval

The study was approved by the local Medical Ethical Committee of the Ludwig-Maximilians-Universität (Project 249-08), and informed, written consent was obtained from the mothers of all participants during pregnancy prior to inclusion into the study.

Study population

From 10/2008 to 12/2009, a total of 58 newborns were included in this study. Infants were classified according to their gestational age into extremely premature infants (n=20), moderately premature infants (n=18), and mature neonates (n=19). Reasons for prematurity were placental insufficiency, premature labor and/or premature rupture of membranes, pre-eclampsia, HELLP syndrome, pathologic Doppler flow, suspected amnion infection, and twin pregnancy. Baseline characteristics and laboratory data of the patient groups are given in Table 1. As children were included consecutively, the gender distribution among the experimental groups shows some variation. As expected, premature neonates had a significantly lower birth weight than mature neonates. The white blood-cell count and the number of PMNs in whole blood, as well as the arterial pH value, also differed significantly among the groups.

Rolling and adhesion of neonatal PMNs in relation to gestational age

To assess whether gestational age influences PMN recruitment under dynamic conditions, we quantified rolling and adhesion of neutrophils obtained from umbilical cord blood of premature and mature infants in a microflow chamber system at a shear stress of ∼1 dyne/cm², resembling a physiological shear stress found in postcapillary venules in vivo [18]. When looking at PMNs from adult healthy controls, we observed a continuous increase in the number of rolling cells in P-selectin (Fig. 1A)- and E-selectin (Fig. 1B)-coated flow chambers during the 10-min observation period. Compared with adult controls, the number of rolling PMNs was significantly lower in mature neonates and even further reduced in premature infants (Fig. 1A and B). Neutrophils from extremely premature infants showed almost no rolling on immobilized P- or E-selectin, suggesting that neutrophil rolling is developmentally regulated during fetal life.

Next, we studied PMN adhesion in flow chambers coated with a combination of P- or E-selectin, the CXCR2 chemokine IL-8, and the integrin ligand ICAM-1. We observed that the ability of neutrophils to adhere under conditions of flow strongly and significantly depends on gestational age. Whereas PMNs from extremely premature infants rarely showed adhesion (Fig. 2A and B), there were some adherent cells in the moderately premature infant group. A further increase in adhesion was seen when PMNs from mature neonates were perfused through the chambers. However, even mature neonates did not reach the levels seen in the adult control group (Fig. 2A and B). When plotting the number of adherent cells at 10 min against the gestational age of each individual, we found a highly significant correlation (Fig. 2C). To test whether PMN adhesion is mediated specifically by the immobilized molecules used in the experiment, we performed control microflow chamber experiments, where the selectin or the chemokine was omitted during coating, chambers were coated with PBS/BSA alone, or EDTA was added to the perfusion medium to eliminate divalent cations necessary for specific binding. We did not observe significant numbers of adherent cells from adult healthy donors in any of the control chambers (Fig. 2D), ruling out nonspecific PMN adhesion in our assay.

Recruitment of adult PMNs on HUVECs from premature and mature neonates

For the successful recruitment of PMNs in vivo, neutrophils need to interact with adhesion molecules expressed on ECs [10]. It has been shown that age-dependent variations in neutrophil recruitment to fetal porcine ECs exist, with a relative inability of ECs from midgestation fetuses to interact with neutrophils under conditions of flow [19]. We assessed whether ECs isolated from the umbilical cord of premature and mature infants are capable of supporting recruitment of PMNs isolated from adult healthy controls using the microflow chamber assay. As expected, we did not observe significant rolling or adhesion of adult PMNs on unstimulated HUVECs in any group (Fig. 3A and B). Upon EC stimulation with LPS for 4 h, PMNs started to roll and adhere on the HUVEC layer in all experimental groups. However, the extent of PMN rolling and adhesion differed significantly between the groups and depended on the gestational age of the neonates (Fig. 3C and D). After 10 min, the number of cells rolling on HUVECs from mature infants was nearly fivefold above the numbers found on HUVECs from extremely premature infants (Fig. 3A and C), and the number of adherent PMNs was nearly threefold higher (Fig. 3B and D).

Age-dependent expression of adhesion molecules on fetal PMNs and umbilical cord ECs

As a result of the age-dependent differences observed in PMN recruitment, we characterized the cells isolated from the umbilical vein of premature and mature neonates and quantified adhesion molecule expression by flow cytometry and confocal laser-scanning microscopy. Unstimulated cells expressed similar levels of CD31 and vWF, as well as VE-cadherin, demonstrating their endothelial nature (Fig. 4A and B). We then quantitated by flow cytometry the adhesion molecule expression on umbilical vein ECs from premature and mature neonates. ICAM-1 was expressed constitutively at low levels on untreated cells, whereas E- and P-selectin were not detected (Fig. 4C). After stimulation with LPS, expression of E-selectin increased significantly in all groups. However, the amount of up-regulation was lowest in extreme premature infants. Comparably, we found a significant LPS-stimulated up-regulation of ICAM-1 expression in term neonates, whereas ICAM-1 expression was less inducible by LPS in premature infants (Fig. 4C). In line with prior reports showing that LPS does not induce P-selectin surface expression in HUVECs [20], we were not able to detect any P-selectin on LPS-stimulated HUVECs by means of flow cytometric analysis (not shown).

Next, we measured the surface expression of adhesion molecules relevant for rolling and adhesion on neonatal and adult neutrophils. For this purpose, PMNs were quantified for the selectin-ligand PSGL-1, the chemokine receptor CXCR2, and the β-2 integrins LFA-1 and Mac-1 using flow cytometry (Fig. 4D). Whereas expression levels of CXCR2 and LFA-1 were identical among our experimental groups, we found that PSGL-1 and Mac-1 surface expression was reduced significantly in extremely premature infants compared with adult controls. The expression of these molecules subsequently increased with advancing gestational age (Fig. 4D), suggesting that reduced expression of PSGL-1 and Mac-1 contributes to reduced neutrophil rolling and adhesion during fetal development.

Influence of antenatal steroid treatment on PMN and HUVEC functions

Acceleration of fetal lung maturation by betamethasone treatment of pregnant women with impending preterm delivery is nowadays standard clinical practice [21, 22]. In our patient cohort, 89% of extremely premature infants and 72% of moderately premature infants were exposed to antenatal maternal steroid treatment. To exclude the possibility that such exposure is responsible for the reduction of PMN recruitment seen in premature neonates, we performed microflow chamber experiments using adult donor PMNs that were pretreated (1 h) with betamethasone at a concentration found in the umbilical cord blood of neonates [23]. Adhesion of betamethasone-pretreated adult PMNs in flow chambers coated with P-selectin, IL-8, and ICAM-1 was similar to adhesion observed for untreated adult control PMNs (Fig. 5A). In addition, we incubated freshly isolated HUVECs with betamethasone (12 h) before seeding into flow chambers. HUVECs pretreated with betamethasone supported LPS-induced adhesion of adult PMNs to the same extent as HUVECs without prior betamethasone incubation (Fig. 5B). Furthermore, flow cytometric analysis of LPS-stimulated HUVECs (Passage II) demonstrated that the LPS-induced up-regulation of E-selectin and ICAM-1 was independent of prior betamethasone treatment (Fig. 5C), suggesting that betamethasone does not interfere with neutrophil recruitment in our setup.

Postnatal development of PMN recruitment

Recent reports indicated that neutrophils are subjected to a postnatal maturation process [24–26]. This prompted us to investigate the development of PMN recruitment postnatally and evaluate whether exposure of infants to extrauterine factors changes the development of PMN function when compared with the situation in utero. To address this, we compared adhesion of isolated neutrophils from premature infants born at <30 weeks of gestation directly after birth and 4–6 weeks later using P-selectin-, ICAM-1-, and IL-8-coated flow chambers. Compared with the adhesion observed for PMNs isolated directly after birth, PMN adhesion increased significantly with advancing postnatal age of the infants (Fig. 6A). However, this increase in adhesion (extrauterine development) was similar to the adhesion observed for PMNs obtained from umbilical cord blood of premature infants born >30 weeks of gestation (intrauterine development), which were matched for the same postconceptional age (Fig. 6B). These findings indicate that extrauterine factors do not seem to have a relevant influence on the development of neutrophil-adhesive capacity, implying that neutrophil maturation during development follows an intrinsic program rather than depending on postnatal factors.