Experiments complied with the National Health and Medical Research Council (NHMRC) of Australia guidelines for the care and use of animals for scientific purposes and were approved by Monash Medical Centre Animal Ethics Committee A. The experiments have been reported in compliance with the ARRIVE guidelines (Animal Research: Reporting in Vivo Experiments).
Surgery to induce FGR and ventilation of lambs
Procedures to induce FGR and delivery followed by ventilation in preterm lambs have been previously described . In brief, surgery for single umbilical artery ligation (SUAL) was performed at ~ 88 days’ gestation to induce early onset FGR in one fetus of twin-bearing Border-Leicester Merino crossbred ewes. The other fetus acted as the control. At ~ 126 days’ gestation, the pregnant ewe was anesthetized using thiobarbital followed by gas anesthesia and a cesarean section was undertaken. With the lamb exposed within the uterus but not yet delivered, a flow probe (size 4; ADInstruments, Castel Hill, Australia) was inserted around the carotid artery, and the lamb then delivered and the umbilical cord clamped and cut. The lambs were dried, weighed, and transferred to an infant warmer (Fisher and Paykel, Auckland, NZ) where each lamb was intubated (4.0-mm cuffed endotracheal tube), lung liquid passively drained, and gentle ventilation commenced. Umbilical vein and artery catheters were immediately inserted and secured using silk sutures. A pulse oximeter probe (Masimo, Irvine, CA) was placed on the lamb’s tail for the measurement of transcutaneous oxyhemoglobin saturation levels (SpO2). Near-infrared spectroscopy (NIRS; Fore-Sight Tissue Oximeter, CAS Medical Systems Inc., Branford, CT) was used for continuous recording of cerebral oxygenation, via placement of probes over the frontoparietal head region and covered with a lightproof dressing. Cerebral oxygenation was expressed as tissue oxygenation index (TOI, %) at 0.5 Hz . Ventilation of the preterm FGR and AGA lambs was initiated using assist control ventilation (Babylog 8000+, Dräger, Lüberk, Germany) with an initial peak inspiratory pressure (PIP) of 30 cmH2O and positive end-expiratory pressure of 5 cmH2O for the first 10 min. The inspired oxygen fraction (FiO2) commenced at 0.3, but was adjusted to maintain SpO2 between 91 and 95% after initial resuscitation. All lambs received prophylactic surfactant via the endotracheal tube (100 mg kg−1, Curosurf; Chiesi Pharma, Parma, Italy) at 10 min after birth.
Lambs were ventilated for 24 h on volume guarantee mode with a tidal volume (VT) set at 5–7 mL kg−1. Throughout ventilation, lambs were sedated by continuous infusion of Alfaxan (3 mg kg−1 min−1; CenVet, Lynbrook, Australia) through the umbilical vein catheter. Lamb physiological well-being was assessed via umbilical arterial blood gas parameters measured at 1, 6, 12, and 24 h post delivery. Lambs were euthanized at 24 h by intravenous pentobarbital sodium overdose (100 mg kg−1 I.V.; Valabarb, Rutherford, Australia). All ventilation and physiological data was digitally acquired using Powerlab (1 kHz) and Lab Chart 8 software (ADInstruments, Castle Hill, Australia).
Cell collection and preparation
The umbilical cord blood was collected at cesarean section from a separate cohort of healthy term lambs (144–145 days’ gestation). The umbilical cord was clamped, and the blood from the placental side was collected into heparinized syringes. UCBCs were then isolated from the buffy coat layer by centrifuging the blood at 3100 rpm for 12 min at room temperature, with no brake, and excess red blood cells removed using red blood cell lysis buffer. The cells were resuspended in fetal bovine serum with 10% DMSO (Merck, Darmstadt, Germany) and cryopreserved in liquid nitrogen. The cells were thawed immediately before administration and DMSO removed by washing cells with media (DMEM/F12, 10% FBS, 1% antibiotics). Cell yield and viability were assessed using trypan blue dye exclusion before administration. Cells were labeled with carboxyfluorescein succinimidyl ester (CFSE) to enable tracking of the cells within the brain . For this study, allogeneic umbilical cord blood mononuclear cells (25 million/kg) were suspended in 2–3 ml of sterile saline and were given intravenously (via the umbilical vein) to the preterm ventilated lambs at 1 h of life.
After euthanasia, cerebrospinal fluid (CSF) was collected with an 18G needle and a 3-ml syringe and the brain removed and weighed. The left brain hemisphere was divided into four sections (frontal, middle (× 2), occipital) and frozen for analysis. The right brain hemisphere was coronally cut into 5-mm slices and fixed in formalin for 48–72 h and then embedded in paraffin (ProSci Tech, Thuringowa, Australia) for histological and immunohistochemistry analysis.
Molecular assessment and cytokine assays
Concentrations of intracellular adhesion molecule (ICAM), vascular cell adhesion molecule (VCAM), interleukin (IL) 3, IL6, neuron-specific enolase (NSE), decorin, interferon (IFN) γ, IL17A, IL21, IL8, IP10, monocyte induced by gamma interferon (MIG), secreted frizzled-related protein (sFRP) 3, tumor necrosis factor (TNF) α, and vascular endothelial growth factor (VEGF)-A were assessed in the lamb serum and CSF samples using a human 5-plex and ovine 10-plex quantibody array following the manufacturer’s instructions (Crux Biolab, Scoresby, Australia). Ovine brain-derived neurotrophic factor (BDNF), nerve growth factor (NGF), and amyloid precursor protein (APP) ELISA assays were conducted on homogenized brain white matter tissue. Lastly, endothelin-1 receptor antibody (EDNRA) ELISA (Crux Biolab, Scoresby, Australia) was conducted using the manufacturer’s instructions on serum samples at 1, 6, and 12 h after birth.
Cerebral cellular apoptosis was assessed using activated caspase-3 (Cat# AF835, R&D Systems, Minneapolis, MN), astrocytes were assessed using glial fibrillary acidic protein (GFAP, Cat# G3893, Sigma-Aldrich, Castle Hill, Australia), inflammatory microglial cells were evaluated using ionized calcium binding adapter molecule 1 (Iba-1, Cat# 019-19741, Wako Pure Chemical Industries, Osaka, Japan), BBB permeability was assessed using albumin extravasation (Cat# S4265-2ML, Sigma-Aldrich, Castle Hill, Australia), and oxidative stress was assessed using 4-hydroxynonenal (4HNE, Cat# 393207-100ul, Merck, Germany). Ki67 (Cat# RM-9106-S, Thermo Fisher Scientific, Waltham, MA) was used to study cell proliferation, and endothelial cell proliferation was evaluated via expression of glucose transporter-1, which is present on mature endothelial cells (Glut-1; Cat# AB14683, Abcam, Melbourne, Australia). Standard immunohistochemistry protocols were followed. Briefly, for any single-label immunohistochemistry analysis, brain blocks at the level of caudo-putamen and the dorsal tegmental bundle containing the cortical gray matter (CGM), subcortical white matter (SCWM), periventricular white matter (PVWM), and hippocampus (hippo) and subventricular zone (SVZ) were sectioned at 10 μm. Serial sections were placed on SuperFrost glass slides and dewaxed in xylene followed by rehydration in serial ethanol solutions. Antigen retrieval was carried out by heating in citric acid buffer (pH 6) for 15 min (3 × 5 min) and then allowing the hot buffer to cool at room temperature for a further 20 min. Endogenous peroxidase activity was blocked by 0.3% hydrogen peroxide buffer in 50% methanol. Non-specific binding was blocked by animal serum (goat or rabbit serum in bovine serum albumin). Primary antibody was then added and sections incubated at 4 °C overnight. The following day, sections were incubated in corresponding secondary antibody followed by streptavidin horseradish peroxidase (HRP; Cat# GERPN1051-2ML, 1:200, Amersham Bioscience, UK). Staining was visualized using 3,3′-diaminobenzidine and coverslipping using mounting medium (DPX, Cat# 100579, Merck, Kilsyth, Australia).
Double-labeled fluorescent immunohistochemistry was conducted to determine the relationship of the desmin and smooth muscle actin (α-SMA), which labels the pericytes of the neurovascular unit. After dewaxing the sections, endogenous peroxidases were blocked with 0.3% hydrogen peroxide in 50% methanol, and sections were then washed with sodium borohydride (10 mg/mL) in PBS to reduce autofluorescence. Sections were subsequently treated with serum-free protein blocker (Cat# X090930-2, DAKO Australia, Campbellfield, Australia) and incubated with mouse monoclonal α-SMA (Cat# A522-200UL, 1/50; Sigma-Aldrich, USA) and rabbit polyclonal anti-desmin (Cat# D8281-1ML, 1/50; Sigma-Aldrich, USA) to identify the degree of pericyte coverage, as previously described by our group . Fluorescent secondary antibodies were used to study the interaction of the two proteins in the neurovascular unit.
Quantitative analysis of brain injury
Sections were viewed at a magnification of × 400 using light microscopy (Olympus BX-41, Japan) with the slides coded for blind analysis. Immunoreactive cell counts were assessed in three fields of view within regions of interest on two slides per animal, to give six fields of view per region per animal, from which an average was then calculated. Manual counts of immunopositive cells expressing GFAP (astrocytes), Iba-1 (activated or amoeboid microglia), caspase-3 (cell death), 4-HNE (oxidative stress), Ki-67 (cell proliferation), and Glut-1 (endothelial cell proliferation) were undertaken.
For cell proliferation, besides manual counting of Ki-67-positive cells in brain regions, we calculated the percentage of blood vessels with Ki-67-positive cells as compared to all blood vessels. Automated counts of percent area for Glut-1 staining were also assessed, using an automated program after tracing the blood vessel using a stylus and Glut-1 immunoreactivity assessed by the area covered. Fluorescent desmin and α-SMA colocalization (using Mander’s coefficients M1 and M2) was assessed by measuring the amount of co-localization between the red and green fluorescence within blood vessels at × 400 magnification. An image processing software (ImageJ- Fiji 2.0.0, National Institutes of Health, MD) was used for automated analysis using specific macros designed by us.
Statistical comparisons were carried out using GraphPad Prism (GraphPad Software v7, San Diego, CA) and SPSS (v25, IBM SPSS Statistics, Armonk, NY). Data are presented as mean ± standard deviation (S.D.) or as percentage or a fraction for semi-quantitative data (blood vessel proliferation, albumin extravasation). Group data were analyzed using one-way, two-way, or three-way ANOVA or repeated measures two-way or three-way ANOVA where appropriate and post hoc multiple comparisons test. Sources of variations for ANOVA included time factor, fetal growth (FGR vs. AGA), UCBC (FGR/AGA+UCBC vs. FGR/AGA) or a combination of these. Significance was accepted at p < 0.05.