Neuroprotection strategies in preterm encephalopathy

Abstract

Advances in neonatology have led to unprecedented improvements in neonatal survival such that those born as early as 22 weeks of gestation now have some chance of survival, and over 70% of those born at 24 weeks of gestation survive. Up to 50% of infants born extremely preterm develop poor outcomes involving long-term neurodevelopmental impairments affecting cognition and learning, or motor problems such as cerebral palsy. Poor outcomes arise because the preterm brain is vulnerable both to direct injury (by events such as intracerebral hemorrhage, infection and/or hypoxia), or indirect injury due to disruption of normal development. This neonatal brain injury and/or dysmaturation is called “encephalopathy of prematurity”. Current and future strategies to improve outcomes in this population include prevention of preterm birth, and pre-, peri-, and postnatal approaches to protect the developing brain. This review will describe mechanisms of preterm brain injury, and current and upcoming therapies in the antepartum and postnatal period to improve preterm encephalopathy.

Introduction

Preterm birth (birth before 37 weeks of gestation) accounts for 35% of neonatal deaths and is the second most common cause of death under 5 years of age.^{1, 2} The rates of preterm delivery vary by country, and in general, are inversely related to its wealth. Of the 1.2 million preterm births that occur in high income countries, 42% occur in the United States, where the preterm birth rate was 9.85% in 2016.³ In 2005, the U.S. annual economic burden caring for preterm infants up to age 5 years was estimated to be $26 billion by the National Academy of Medicine.⁴

Advances in neonatology have led to unprecedented improvements in preterm survival disproportionately in high-income countries such that 70% of those born at 24 weeks of gestation, and over 90% of infants born at 27 weeks or later survive.⁵ With these successes, the limit of viability has continued to decrease, and is now considered to be 22–23 weeks of gestation in most developed countries. Survival of these increasingly smaller and sicker babies has led to a corresponding increase in the prevalence of neurodevelopmental disabilities. Forty to sixty percent of survivors born at less than 28 weeks of gestation still develop moderate or severe impairments (cerebral palsy (CP), intellectual disability, deafness, blindness), with the smallest children at highest risk.⁶ Preterm survivors are also at higher risk of autism spectrum disorder, attention deficit hyperactivity disorder, and psychiatric disorders than their term-born peers. The brain injury experienced by preterm survivors has changed as neonatology has evolved: severe intracranial hemorrhage and cystic white matter injury (WMI) were much more prevalent in past decades. Modern preterm brain injury is more likely to include dysmaturation of neuronal and glial progenitor cells.⁷ Despite this more subtle injury, overall developmental outcomes have changed little over the decades. Our challenge now is to develop neuroprotective approaches to improve these outcomes in preterm survivors.

Brain Development in the Third Trimester

Preterm infants are born during a time of active brain development with a tripling of surface area occurring between 25 and 40 weeks as gyrification occurs.⁸ Key events occurring during this period include continued birth of neurons in the subventricular zone and ganglionic eminence migrating to the thalamus, basal ganglia, and deeper cortical layers, new active growth of axons in periventricular regions, development of the oligodendrocytes, astrocytes and microglia, synaptogenesis, myelination, and finally, pruning (Figure 1). During this critical period of development, the brains of infants born preterm are exposed to exogenous and endogenous stimuli including hypoxia, hypoxia-ischemia, hyperoxia, inflammation, excitotoxicity, and an excess of free-radicals. Nutritional deficiencies, (both macro and micro), exposure to pain, light, noise, drugs and other factors in the neonatal intensive care environment also play a role in modifying extra-uterine development.

Figure 1. Schematic of human brain development.

(A) Sequence of brain maturation from conception to adulthood. (B) Human neuronal cell development during the life time. (C) Cortical Development in the third trimester of preterm and term infants. MZ: marginal zone, CP: cortical plate, SP: sub plate layer, IZ: intermediate zone, WM: white matter, SVZ: subventricular zone, VZ: ventricular zone.

Oligodendrocytes are the myelinating cells in the brain, and they play a critical role in the maturation of white matter. Oligodendrocytes develop and mature between 24 to 32 weeks of gestation, a period in which they are particularly vulnerable to injury. Oligodendrocyte progenitor cells (OPCs) give rise to preoligodendrocytes (preOL), which develop into mature myelinating oligodendrocytes (OL).⁹ Developing oligodendrocytes are highly vulnerable to injury caused by glutamate, free radicals, and inflammation. Current data suggests that injury to the OPCs and preOLs results in proliferation of OPCs, but due to a block in maturation, there remains a deficit in mature OLs with myelinating capability, which is associated with axonal injury and WMI.^10,11 Mechanisms postulated to contribute to the developmental arrest of OPC and preOLs include dysregulation of Notch and WNT-beta catenin signaling. This is mediated by microglial activation and subsequent hyaluronidase production with resulting hyaluronan break down products that are specific inhibitors of OL maturation.^12–14

Microglia are vital to synaptogenesis and pruning, inflammation and tissue repair. These cells reach peak abundance in cerebral white matter in the third trimester. These cells can have both beneficial and injurious effects and are often noted to be activated in areas of oligodendrocyte injury.

Subplate neurons are a transient population of excitatory neurons located just below cortical layer 6.¹⁵ These cells, which are most prominent between 24–32 weeks of gestation, play a critical role in the establishment of cortical lamination and thalamocortical connectivity by serving as an intermediary synaptic contact.¹⁶ After 33 weeks of gestation when thalamocortical axons establish synapses with cortical plate layer IV neurons, the subplate neurons undergo apoptosis. Damage to developing subplate neurons has been postulated to disrupt thalamocortical white matter axons, as well as disrupt cortical neuronal connectivity, which can lead to long term motor and cognitive deficits.¹⁷

Mechanism of Preterm Encephalopathy

There are many antenatal and postnatal factors (Figure 2) that can contribute to injury of the developing brain. Events such as intraparenchymal hemorrhage can damage specific structures that previously developed normally, but most factors interfere with the normal unfolding of development. Since brain development is an intricate unfolding of many events simultaneously, damage to one aspect can adversely affect others.

Figure 2. Flow diagram outlining risk factor and underlying mechanism contributing to preterm encephalopathy.

Mentioned risk factors lead to hypoxia-ischemia, inflammation and oxidative stress, which causes white and gray matter injury by different mechanisms and ultimately lead to long term cognitive and motor impairment.

White Matter Injury (WMI)

Periventricular WMI is the hallmark of preterm brain injury. This ranges from focal cystic necrosis, to focal microscopic necrosis, to diffuse non-necrotic lesions. The large necrotic cystic periventricular leukomalacia that used to be quite common is now seen less frequently, whereas diffuse, microcystic WMI is more common.¹⁸ Axonal degeneration occurs in focal necrotic WMI, but is generally spared in diffuse WMI.¹⁹ Focal necrotic WMI often contains axonal spheroids and dystrophic axons, which degenerate during the early phase of coagulative necrosis,²⁰ while diffuse WMI is characterized by persistent astrogliosis and microgliosis.^{9, 21, 22} Factors associated with WMI include perinatal hemodynamic instability, inflammatory exposures, as well as exposures to pain, hypoxia-ischemia, hypoxia, hyperoxia, hypocarbia, hypo and hyperglycemia. Other clinical experiences such as multiple attempts at intubation also are associated with brain injury.²³

Gray Matter Injury

Human neuroimaging studies have identified reduction in cerebral cortex volume in preterm survivors. Studies have also shown varying degrees of abnormal cortical folding, reduced gyral complexity, and increased ventricular and extra-axial volume.²⁴ Cerebellar structures and subcortical gray matter including basal ganglia, thalamus, and hippocampus are often adversely affected.^25–29 During weeks 24–32 of gestation, dendritogenesis and experience-driven synaptogenesis is active, and largely responsible for the four-fold increase in cortical volume that occurs during this period. These maturational processes are highly vulnerable to injury. Preterm neurons have variable susceptibility to hypoxia-ischemia and other insults. Vulnerability is dependent on the type of neuron, extent of prematurity, severity of the insult, and severity of WMI. Immature axons are susceptible to glutamate-mediated excitotoxicity at sites of contact with OL processes, and axon degeneration identified by fraction immunostaining is often seen in regions of periventricular leukomalacia. Dean et al. in their sheep model of preterm injury showed significant reduction in dendritic arborization and spine densities rather than loss of neurons as a cause of gray matter mass.³⁰ These changes correlate with an increase in functional anisotropy (FA), which usually decreases as the complexity of gray matter in the cerebral cortex increases.³⁰ In addition to altered arborization, there may be a reduction in cortical neurogenesis. Malik et al showed a reduced number of Tbr2+ progenitor cells and increased number of Sox2+ radial glia, suggestive of reduction in glutamatergic neurogenesis.³¹ A similar reduction in neurogenesis is also observed in GABAergic interneurons.

Dysmaturation and maturational arrest are emerging as key factors in preterm encephalopathy. Further studies are needed to understand the mechanisms leading to maturational arrest and whether there are pathways that can be modified to provide neuroprotection in preterm infants.

Cerebellar Injury

The cerebellum undergoes a 5-fold increase in volume between 24 and 40 weeks of gestation, with an estimated 30-fold increase in surface area as foliation occurs.³² Like the cerebral cortex, the cerebellum is subject both to injury (hemorrhage or infaction) and abnormalities of development.^{32, 33} Recent data suggest that up to 19% of premature infants born before 32 weeks of gestation have cerebellar abnormalities,³⁴ with higher rates among extremely low birth weight infants.³⁵ Cerebellar injury is now being recognized as one of the important complications of preterm birth leading to poor neurodevelopmental outcome. Cerebellar injury most commonly occurs as direct injury in the form of cerebellar hemorrhage.

Four categories of cerebellar hemorrhage have been described: 1) Primary cerebellar hemorrhage, 2) Venous infarction 3) Extension of intraventricular of subarachnoid blood into cerebellum, and 4) Traumatic laceration of cerebellum or rupture of major veins or occipital sinus. The range of severity of primary cerebellar hemorrhage includes mild punctate lesions, focal unilateral lesions, to extensive bihemispheric and vermian hemorrhages. Cerebellar atrophy can be seen as soon as 2 months after hemorrhage. Associated reduction in contralateral cerebral volumes occurs most likely due to impairment of transynaptic trophic effects.^{36, 37} Factors that are associated with white matter injury (exposure to hypoxia-ischemia, oxidative stress and infection/inflammation) are also associated with cerebellar underdevelopment, and similar pathologic findings are seen (neuronal loss and gliosis).³⁸ Cerebellar injury is associated with cognitive deficits, language impairment and autism spectrum disorder.^{39, 40} Cerebellar underdevelopment and injury can be an important contributer to preterm encephalopathy, and additional studies are needed to further define the long term effects in preterm infants.

Antenatal Therapies

Docosahexaenoic acid (DHA)

Long chain polyunsaturated fatty acid (LC-PUFA), including docosahexaenoic acid (DHA) and arachidonic acid (AA), are incorporated into membrane phospholipids and play an important role in brain development, including roles in neurogenesis and neurotransmission. DHA is derived from α-linolenic acid, which is an essential omega-3 fatty acid. The accumulation of DHA in the brain takes place during the brain growth spurt in the third trimester and continues up to two years of age. The high levels of DHA in the brain are maintained throughout life. The Organization of the United Nations agrees that pregnant and lactating women should have a minimal intake of 200–300 mg of DHA per day. A study in 2017 by Nordgren et al showed that in United States on average, DHA intake was only 66% of recommended in pregnant women, and only 9% of women took DHA supplements during pregnancy.⁴¹ The neonatal dietary source for DHA is breast milk or formula. DHA content in breast milk varies with maternal diet and obesity. Studies have shown that supplementing DHA in maternal diet during pregnancy increase DHA levels in breast milk. Donor breast milk DHA content varies, but is significantly lower than the fetal accretion rate of 42–67mg/day in the third trimester. As accretion of DHA occurs in third trimester, preterm infants become deficient in DHA over time, particularly if nourished by total parenteral nutrition and intralipids, as they are deficient in DHA.⁴² A randomized trial of different doses of DHA in preterms showed a DHA dose of 120mg/kg/day prevented the drop of DHA in preterm infants.⁴³ Early studies have shown association of DHA levels and cognitive outcomes. A small study by Tam et al showed an association between higher DHA levels and a decrease incidence in intraventricular hemorrhage (IVH).⁴⁴ A recent randomized controlled trial (DINO trial) evaluated infants of mothers on high DHA and standard DHA diet until 40 weeks corrected age.⁴⁵ At 18 months there was no overall significant difference between group, but girls showed significant improvement in Mental development index (MDI) scores. Follow up of these infants at 7 years of age did not showed any significant difference between groups. In contrast, a randomized trial done in Norway, which supplemented DHA in milk for 9 weeks showed significantly higher performance on the problem solving section of the Ages and Stages Questionnaire in treated infants compared to control infants.⁴⁶ In 2017 Cochrane review analyzed infants with LCPUFAs supplementation (median DHA concentration of 0.33%).⁴⁷ The author concluded that premature infants did not obtain a benefit from receiving formula milk supplemented with LCPU- FAs. However, the doses used were different and the optimal dose for neuroprotection is still to be determined.

Choline

Choline is an essential component of the phospholipid, phosphatidylcholine, and sphingomyelin membranes. It is an important nutrient for methyl metabolism, acetylcholine synthesis and cell signaling. Choline plays an important role in neurogenesis and neural migration during fetal development, and deficiency is associated with spinal bifida.⁴⁸ Animal studies have shown that choline supplements during pregnancy increases neuronal size and cell proliferation, reduces hippocampal apoptosis, and improves learning and memory in the offspring.⁴⁹ Choline is actively transported to the fetus via placenta, and choline concentration in fetal plasma is 3 to 10 times higher than in plasma of pregnant women.⁵⁰ Wu et al conducted a prospective cohort study of 154 mother-term infant pairs. Maternal plasma choline levels at 16 and 36 weeks were measured and neurodevelopment assessments were done at 18 months. Significant positive associations were found between infant cognitive test scores and maternal plasma free choline.⁵¹ A double blinded randomized controlled trial was done comparing nutritional supplement (DHA (1% of total daily estimated fatty-acid intake), eicosapentaenoic acid, arachidonic acid, choline, UMP, cytidine monophosphate, vitamin B12, zinc, and iodine) to placebo. The supplements were given for up to 2 years. No significant difference in cognitive composite score of the Bayley Scales of Infant Development, Third Edition was shown.⁵² Further studies are needed to evaluate the optimal dose and effect of choline on neurodevelopmental outcome of preterm infants.

Antenatal corticosteroid

The effectiveness of antenatal corticosteroids for neuroprotection was established early in 1972 by Liggins and colleagues. They demonstrated a reduction in respiratory distress syndrome and IVH among preterm infants given antenatal glucocorticoids.⁵³ Over 20 years and 12 randomized controlled trials later, the National Institute of Health published a consensus statement recommending use of antenatal corticosteroids for mothers at risk for preterm birth between 24 and 34 weeks.⁵⁴ A Cochrane review in 2006 evaluated 21 studies (3885 women and 4269 infants) and showed an overall reduction in neonatal death (relative risk (RR) 0.69, 95% confidence interval (CI) 0.58 to 0.81, 18 studies, 3956 infants), respiratory distress syndrome (RR 0.66, 95% CI 0.59 to 0.73, 21 studies, 4038 infants), cerebroventricular hemorrhage (RR 0.54, 95% CI 0.43 to 0.69, 13 studies, 2872 infants), necrotizing enterocolitis (RR 0.46, 95% CI 0.29 to 0.74, eight studies, 1675 infants), respiratory support, intensive care admissions (RR 0.80, 95% CI 0.65 to 0.99, two studies, 277 infants) and systemic infections in the first 48 hours of life (RR 0.56, 95% CI 0.38 to 0.85, five studies, 1319 infants).⁵⁵ Despite mounting evidence, it has taken a long time to change clinical practice such that antenatal steroids are routinely given when preterm delivery is likely. The long-term outcomes for infants whose mothers received antenatal corticosteroids vary. In a 30-year follow-up of the Auckland steroid trial, there was no difference in cognitive outcomes between infants exposed to antenatal steroids and those exposed to placebo.⁵⁶ A more recent meta-analysis, which included 14 studies on long term neurodevelopmental outcome showed reduction in CP in infants less than 34 weeks of gestation.⁵⁷ In contrast, a recent Cochrane review looking at antenatal and intrapartum interventions showed a low level of evidence in reduction of CP in infants born to women received antenatal corticosteroid.⁵⁸ The use of antenatal steroids globally is controversial as a large cluster randomized trial in six low- and middle-income countries showed increased perinatal mortality, possibly due to maternal and neonatal infection.⁵⁹

Magnesium sulfate

Magnesium has been used in obstetrics as a tocolytic and for prophylaxis of eclampsia. An early case control study showed that very low birth weight (less than 1500 grams) infants whose mothers were treated with magnesium sulfate for maternal indications had a lower rate of CP compared to healthy matched controls.⁶⁰ Several subsequent observational studies and randomized trials showed a similar trend in reduced incidence of CP. A Cochrane review published in 2009 evaluated 6145 children (5 prospective randomized trials). The effect of antenatal magnesium sulfate as a neuroprotective agent was studied in patients at risk of preterm birth. No difference in infant mortality was noted, but a 31% decrease in the overall incidence of CP among infants whose mothers were treated with magnesium sulfate (RR 0.68: 95% CL 0.54 to 0.87) was found. The risk of motor dysfunction was decreased by 39% in those infants whose mother received magnesium sulfate (RR 0.61, 95% Cl 0.44–0.85). There was no difference in IVH, WMI, deafness, intellectual impairment, bronchopulmonary dysplasia at 28 days or 36 weeks, or length of hospital stay. There was also no difference in maternal mortality or obstetric complications.⁶¹ The estimated number of pregnant women at risk of preterm birth who need to be treated with MgSO4 to prevent one case of CP (number needed to treat, NNT) is 63 women (95% Cl 43 – 155).⁶²

Mechanisms of magnesium sulfate neuroprotection are still under investigation. One proposed mechanism includes modulation of N-methyl-D-aspartate(NMDA) glutamate channels thus reducing NMDA receptor mediated injury.⁶³ There is also evidence of it stabilizing cell membranes, inhibiting oxidative stress and reducing secondary inflammation. A recent study of preterm infants using near infrared spectroscopy (NIRS) showed that antenatal magnesium sulfate treatment reduced cerebral fractional tissue oxygen extraction in the first 24 hours of life. The authors speculated that the neuroprotective effect was mediated by reduced cerebral metabolic demand.⁶⁴

Doyle and colleagues followed infants from the ACTOMgSO4 trial at school age. Effects of magnesium on rate of abnormal motor function and CP were not significant at this age. The authors speculated that additional therapeutic interventions in the control group may have improved the outcomes in these patients⁶⁵

Overall evidence supports magnesium sulfate as a neuroprotectant in preterm infant, thus, the American College of Obstetricians and Gynecologists (ACOG) with the Society for Maternal Fetal Medicine issued a statement in support of magnesium sulfate for neuroprotection in women at risk for delivery under 32 weeks’ gestation.⁶⁶

Perinatal Therapies

Delayed cord clamping

The optimal timing for umbilical cord clamping has long been controversial, going back to the time of Hippocrates. It is estimated that one third of the fetal-placental blood volume resides in the placenta, and immediate cord clamping prevents the newborn from receiving this blood volume, resulting in increased early physiologic instability, and decreased iron stores later. Delayed cord clamping (DCC) has been of increasing interest in the past decades, as evidence of benefit to newborns accumulates. DCC has been variably defined as 30 seconds up to 1 or 5 minutes after delivery. ACOG and the American Academy of Pediatrics published a Committee opinion in 2012, revised in 2017, recommending DCC up to 60 seconds for vigorous term and preterm infants.⁶⁷ It was further recommended that future research was needed to evaluate placental transfusion for preterm infants requiring resuscitation, continuous placental transfusion during active resuscitation and long-term effects of potential transfer of regenerative cells. A meta-analysis in 2012 reviewed 15 eligible studies, which included 738 infants born between 24 weeks and 36 weeks’ gestation. DCC was associated with lower incidence of IVH (RR, 0.59; 95% CI, 0.41–0.85), necrotizing enterocolitis (RR, 0.59; 95% CI, 0.41–0.85) and transfusion for anemia (RR, 0.61; 95% CI, 0.46–0.81). Small prospective follow up studies showed long term benefits in preterm infants with DCC compared to early cord clamping at 18 to 22 months.⁶⁸ This suggests that placental transfusion from DCC has short term and long term benefits in preterm infant.

Umbilical cord milking (UCM) is an alternative to DCC. In UCM, the umbilical cord is stripped toward the infant several times and then clamped. This procedure can be performed within 20 seconds. A meta-analysis of 6 studies, which included 292 preterm infants with UCM and equal number of DCC concluded that similarly to DCC, umbilical cord milking decreases IVH, necrotizing enterocolitis, death and decreased the incidence of transfusion, with a pooled RR of 0.74 (95% Cl 0.61 – 0.90).⁶⁹ A recent prospective study looking at neurodevelopmental outcome at 2 and 3.5 years of age showed no difference between infant who underwent DCC and infant who underwent UCM.⁷⁰

Postnatal Therapies

Breast milk

Human breast milk (HM) has essential nutrients and other bioactive molecules including lactoferrin, lysozyme, and secretory IgA. Numerous studies have showed beneficial effects of HM on preterm outcomes. The mechanism by which HM provides neuroprotection is potentially via direct effects of high concentration of LCPUFAs and other micronutrients like choline. HM improves outcome indirectly by reducing preterm morbidities including sepsis and necrotizing enterocolitis. The National Institute of Child Health and Human Development (NICHD) Neonatal Research Network study of 1035 infants showed that children in the HM group were more likely to have a Bayley MDI greater than or equal to 85, higher mean Bayley PDI, and higher Bayley Behavior Rating Scale percentile scores for orientation/engagement, motor regulation, and total score. Multivariate analyses, adjusting for confounders, confirmed a significant independent association of HM on all four primary outcomes: (1) the mean Bayley MDI, (2) PDI, (3) Behavior Rating Scale, and (4) incidence of rehospitalization.⁷¹ A follow up study of this cohort at 30 months showed persistent significant improvement in HM group versus non-HM group.⁷² Analysis of the EPIPAGE cohort, demonstrated that the odds for mild and severe cognitive deficiency at 5 years of age are decreased in children who received HM at hospital discharge.⁷³ A recent study by Belfort et al studied 180 infants born < 30 weeks of gestation and showed that HM feeding in the first 28 days of life was associated with a greater deep nuclear gray matter volume at term equivalent age and better IQ (0.5 points/d; 95% CI, 0.2–0.8), mathematics (0.5; 95% CI, 0.1–0.9) working memory (0.5; 95% CI, 0.1–0.9), and motor function (0.1; 95% CI, 0.0–0.2) at 7 years of age in very preterm infants.⁷⁴

Caffeine

Caffeine, a selective adenosine A2a receptor antagonist and non-selective adenosine A1 receptor antagonist, is used to treat apnea of prematurity. Preclinical studies have shown that caffeine is neuroprotective by modulating adenosine A1 receptor, which leads to improvement in hypoxia-induced WMI and ventriculomegaly.^75–77 Caffeine also potentiates synaptic plasticity by modifying neural synapses and increasing the size of dendritic spines. Several studies have shown short term neuroprotective effects of caffeine.⁷⁸

Small studies have shown that caffeine-treated infants have improvement in auditory processing.⁷⁹ Other studies reported that caffeine treated infants had enhanced cerebral cortical activity in brain.⁸⁰ In a study of over 2000 infants with birth weights 500–1,250 g, the Caffeine for Apnea of Prematurity trial (CAP trial) showed that in addition to reducing apnea and the need for mechanical ventilation, caffeine treatment decreased the outcome of death or survival with one or more of the following impairments: cognitive deficit, deafness, CP and blindness. Caffeine use nearly halved the rate of CP.⁸¹ A follow-up study of 1640 infants enrolled in CAP trial, at the age of 5 years, showed no significant difference between the two groups. Further analysis showed a significant reduction of the incidence of developmental coordination disorder in the caffeine treated group (11.3% vs 15.2% adjusted OR: 0.70, 95%CI: 0.51–0.95).⁸² While caffeine has been shown to be beneficial at usual doses (10–20 mg/kg), a pilot study looking at the efficacy of high dose caffeine (80 mg/kg) vs standard dose showed a higher incidence of cerebellar injury with subsequent alterations in early motor performance.⁸³

Family Centered Developmental Care

Fetal maturation is disrupted due to preterm birth. Preterm infants in the neonatal intensive care unit (NICU) are exposed to excessive sensory stimuli which are not present in utero. They are continuously exposed to fluctuations of sound, light, temperature, and oxygen. They experience a multitude of painful procedures, and their sleep is constantly interrupted. During this time, the developing brain continues to modulate synaptic neuronal connections, both predetermined, and affected by environmental stimuli. Epigenetic changes (DNA methylation, acetylation and microRNA production) occur as a result of both environmental stimuli, leading to specific patterns of RNA and protein production, and ultimately to the end product of brain development. Thus, encouraging normal patterns of development, minimizing the abnormal sensory stimuli in this critical period of brain growth can lead to improvement in long-term neurodevelopmental outcomes. Consistent and effective neurodevelopmental care given to preterm infants from moment of birth is important in order to support the optimal brain development and protect it from abnormal external stimuli.

One approach to neurodevelopmental care is based upon the Synaptic Theory of Development.⁸⁴ The core of this approach is individualized, cue-based caregiving, plus interventions designed to reduce stress and promote behavioral organization and physiological stability. Many developmental care strategies have been implemented in NICU to improve neurodevelopmental outcomes. The most studied and implemented model is “Seven Core Measures of Neuroprotective Family-Centered Developmental Care” developed by Altimier and Phillips.⁸⁵ Core measures include: Safeguarding Sleep, Positioning and Handling, Minimizing Pain and Stress, Protecting Skin, Nutrition, Partnering with Families, and Healing Environment. Different family centered neuroprotective strategies have been shown to improve health outcomes, improve family satisfaction, decrease length of stay and hospital costs.^86–88 Kangaroo Mother Care (skin to skin contact between parent and infant) is an important part of family centered care, relevant in both resource sufficient and low resource settings. Short term benefits include decreased maternal anxiety and depression, improved breastfeeding, increased physiologic stability, increase pain tolerance, improved growth and decreased mortality.^89–94 Long term benefits of kangaroo mother care are being recognized, with studies showing improved attention and quality of movements⁹⁵ and enhanced child cognitive developmental and executive functions for up to 10 years.⁹⁶

Erythropoietin

Erythropoietin (Epo) is a 30.4 kDa cytokine originally recognized for its role in erythropoiesis. Epo and its receptor (EpoR) are required for normal brain development. Astrocytes are the main source of brain Epo, and many cell types express the receptor, including neurons, OL and epithelial cells. The EpoR is expressed in hippocampal and cerebral cortical neurons, and expression of both Epo and EpoR decline postnatally in brain but remain present at lower levels throughout adulthood.^{97, 98} Prolonged exposure to hypoxia leads to upregulation of EpoR gene expression. In the absence of Epo to bind these upregulated EpoRs, neurons and OL have a higher risk of apoptotic cell death.^{99, 100}

Mechanisms of Epo neuroprotection include phosphorylation of JAK2, activation of mitogen-activated protein kinase (MAPK), extracellular related kinase (ERK1/2), phosphatidylinositol 3-kinase (PI3K/Akt) and signal transducer and transcriptional activator 5 (STAT5) pathways, which are critical for cell survival.¹⁰¹ Acutely, Epo’s effects are anti-apoptotic, anti-inflammatory, neurotrophic, and anti-oxidant. Long term effects that may promote brain development and healing include angiogenesis, neurogenesis, and oligodendrogenesis. Recent preclinical studies have shown that Epo administration as late as 72 hours or one week after injury leads to improved behavioral outcomes, reduced WMI, enhanced neurogenesis and axonal sprouting, emphasizing the importance of the regenerative and reparative effects of Epo.¹⁰² Ferroptosis is a mechanism of regulated cell death triggered by free iron in cells. As an important effect of Epo is to increase erythropoiesis, which in turn increases iron utilization and decreases unbound free iron, Epo may also provide neuroprotection by modulating ferroptosis in neuronal cells.

Phase I and II studies using doses that vary from 400 U/kg three times a week to 3000 U/kg for 3 doses have demonstrated safety and the potential for neuroprotection.^103–106 A recent meta-analysis of neuroprotective effects of prophylactic Epo in preterm infant included 1133 patients from 4 randomized controlled trials. Despite differences in the dosage and timing of Epo treatment, they showed that prophylactic Epo administration decreased the incidence of infants with an MDI < 70 at 18–24 months, with an odds ratio of 0.51 (0.31 – 0.81), P <0.005. The number needed to treat was 14. Significance was not present in the subgroup analysis of infants < 28 weeks of gestation, however, there were significant size limitations, as the subgroup analysis included only 60 Epo treated and 57 placebo treated subjects.¹⁰⁷ To evaluate the effectiveness of Epo in the most vulnerable preterm infants (< 28 weeks of gestation) a large prospective, randomized, placebo controlled, double blind study of Preterm Erythropoietin neuroprotection (PENUT trial, ) is being conducted. 941 subjects of gestational age of 24–0/7 weeks to 27–6/7 weeks have been enrolled at 19 sites across the United States. All subjects will be evaluated at 22–26 months’ corrected age. Subjects were randomized to either Epo treatment (1000 U/kg/dose x 6 doses every other day), or placebo, and maintenance treatment (400 U/kg/dose three times a week) was continued until 32–6/7 weeks postmenstrual age. The primary outcome is neurodevelopment at two years of age, which consists of a standardized neuro exam, Gross Motor Function Classification Scale and standardized Bayley III. This study is adequately powered to provide definitive results regarding the neuroprotective potential of Epo in this population. Results should be available in early 2019.¹⁰⁸

Future Neuroprotective Strategies

Stem Cells

Stem cells have the ability to differentiate into multiple cell types and have self-regenerative properties. These properties allow them to repair, regenerate and modulate cell death and tissue damage. Stem cells can be of embryonic, fetal or adult in origin. Embryonic stem cells (ESCs) are pluripotent and regenerate but have tumorigenic potential. There are ethical issues in procuring embryonic stem cells from tissues. Adult stem cells include mesenchymal stem cells that can be obtained from a variety of sources (bone marrow, adipose tissue and dental pulp) and neural progenitor cells (NPCs) found in the subventricular zone of the brain. Lastly, fetal derived stem like cells can be obtained from placental tissue and umbilical cord blood (UCB). They include umbilical cord MSCs, endothelial progenitor cells (EPCs), hematopoietic stem cells (HSCs), and amnion epithelia stem cells. Stem cell therapy is an established first line or adjunctive therapy for a variety of neonatal diseases including inborn errors of metabolism. Stem cell treatments have been effective in providing significant neuroprotection and improvement in functional outcomes in animal models of neonatal hypoxia-ischemia, CP, and stroke.^109–111

In recent years, focus has shifted to fetal derived MSCs and umbilical cord blood. Preclinical studies have shown the effectiveness of stem cells in improving outcomes in models of preterm encephalopathy. These studies have elucidated potential mechanism by which MSCs exert neuroprotective effects. MSCs down regulated pro-inflammatory cytokines, reducing inflammation and apoptosis.^{112, 113} In a preterm brain injury model of IVH, UCB-derived MSCs suppressed the upregulation of cytokines IL-1α, IL-1β, IL-6, and TNF-α within the CSF and periventricular brain region. This resulted in reduction in edema and improvement in behavioral outcomes.¹¹⁴ MSCs also alter immune cell programming and proliferation. Preclinical studies have shown that MSCs do not engraft, but rather respond to signals of local injury by secreting trophic factors including fibroblast growth factor-2, Brain-derived neurotrophic factor (BDNF), epidermal growth factor, glial cell line-derived neurotrophic factor, and sonic hedgehog, thereby increasing progenitor cell proliferation and survival of neurons and neuronal stem cells.^{112, 113, 115} MSCs also stimulate axonal sprouting and proliferation of oligodendrocyte precursors, and promote mature OL fate.^{112, 116, 117}

Numerous open-label clinical trials are ongoing or have been completed for treatment of children with established CP. A recent randomized double blinded placebo-controlled study had subjects with CP aged 10 months to 10 years divided in three groups: intravenously UCB + Epo, Epo only, and controls. Compared with the Epo and control groups, the UCB group + Epo had significantly higher developmental scores at 6 months. Diffusor Tensor Imaging (DTI) revealed significant correlations between the gross motor performance measure increment and changes in FA in the UCB + Epo group.¹¹⁸ A safety trial conducted at Duke University assessed the safety and feasibility of therapeutic hypothermia with concomitant intravenous administration of non-cryopreserved, autologous UCB cells for term infants with hypoxic ischemic encephalopathy (HIE). They concluded that transfusion of up to four doses of 1–5×107 cells per dose did not result in significant adverse reactions, cardiopulmonary compromise, or infections. Seventy four percent of UCB recipients were alive at 1 year with total Bayley scores >85, compared to only 41% in the concurrent cooled infants who received only therapeutic hypothermia.¹¹⁹ Ongoing phase I and II trials include; autologous cord blood and human placental derived stem cells in neonates with severe HIE (HPDSC+HIE) (), umbilical cord milking for neonates with HIE (), and neural progenitor cell and paracrine factors to treat HIE (). There are no clinical trials testing the safety and efficacy of stem cells for treatment of preterm encephalopathy. However, a recent phase 1 trial of intratracheal MSCs for treatment of broncopulmonary dysplasia appeared to be safe and with no adverse respiratory, growth, and neurodevelopmental effects at 2 years’ corrected age.¹²⁰

There remain many questions regarding the use of stem cells for preterm encephalopathy including the following: (1) The type of most effective stem cell type, (2) The optimal number, timing (early or late), duration and route of administration, (3) Selection of infants who would benefit from therapy.

Melatonin

Melatonin (N-acetyl-5-methoxytryptamine), a neurohormone derived from the amino acid tryptophan. It is a strong antioxidant capable of scavenging free radicals and stimulating several antioxidative enzymes including glutathione, glutathione reductase, peroxidase, and superoxide dismutase.¹²¹ Melatonin exerts its effect through both receptor-mediated and receptor-independent mechanisms. Melatonin can directly stimulate cell membrane G protein-coupled high-affinity melatonin receptors that activate numerous second messenger cascades, which vary in cell, tissue, and species-specific ways. Melatonin can also induce receptor-independent intracellular activities by targeting calcium-binding proteins, cytoskeletal and scaffold proteins, and components of mitochondrial signaling.¹²² Melatonin’s safety profile, antioxidant, anti-inflammatory and anti-apoptotic properties have made it an attractive neuroprotective candidate for treating neonates with preterm encephalopathy.^{123, 124}

Melatonin has been shown to be neuroprotective in preterm and near-term fetal sheep by reducing oxidative stress, apoptosis and decreasing inflammation.^{125, 126} Similar findings were seen in fetal brain with prophylactic systemic administration of melatonin in pregnant rats.^{127, 128} Maternal administration one hour before hypoxic-ischemic insult results in improved outcome, but melatonin administration > 3hours did not result in improvement. This suggests there is a critical window of time to administer melatonin in mothers for acute neuroprotection of the fetus. Similar to antenatal treatment, preclinical postnatal studies have shown improvement in short term and long-term outcomes of preterm encephalopathy. Melatonin had a dose-dependent protective effect on the developing white matter in a mouse model of excitotoxic WMI. Systemic administration of melatonin following germinal matrix hemorrhage in rats led to improved cognitive and sensorimotor dysfunction at juvenile age.

The pharmacokinetic profile of melatonin has been studied by Merchant et al.¹²⁹ The study showed that neonatal dosing of melatonin cannot be extrapolated from adults because the half-life and clearance rate of melatonin are prolonged in preterm infants, and its volume of distribution is decreased.

Melatonin also holds potential as an antenatal therapy that could be administered to pregnant mothers with at-risk fetuses since it appears safe, crosses the placenta,¹³⁰ and crosses the blood brain barrier.¹³¹ A multicenter trial ‘PREMELIP’ is ongoing to test the neuroprotective effects of melatonin administration in immediate peripartum period on preterm infants. The primary outcome will evaluate white matter damage detected by MRI at term equivalent age. A phase 2 exploratory; multi-center double-blinded randomized placebo-controlled 2-arm trial ‘MINT study’, evaluating the neuroprotective effect of melatonin in preterm infants less than 31 weeks’ gestation. In this trial melatonin (0.1 mcg/kg/hr for 2 hrs) was given for 7 days starting by 48 hours after birth, targeting adult melatonin levels. This trial did not show any difference in FA on term corrected MRI between two groups. Results from the above trial indicate that the optimal dose of melatonin for preterm infants still needs further investigation.

Melatonin and Epo may provide synergistic benefit when given in tandem, as melatonin may provide acute downregulation of inflammation and oxidative stress while Epo may provide long term brain repair, enhancing neuronal and oligodendroglia survival and differentiation after CNS injury. This combined therapy is now being studied in preclinical models with the potential for clinical studies for preterm encephalopathy.

Therapies in preclinical stages

Exosomes

Exosomes are small extracellular nano-size vesicles (50–100 nm in diameter) from cells that contain cytoplasmic proteins, micro ribonucleic acid (miRNA), messenger RNA (mRNA) and cytokines. Exosomes are presumed to originate from multivesicular bodies of all cells. The transfer of exosomes between cells is recognized as an important cell-to-cell communication mechanism. Current evidence suggests that the therapeutic potential of stem cells is due to secretion of paracrine factors, mainly being mediated by these extracellular vesicles (EVs).¹³² Exosomes derived from various MSc have shown to have reparative and regenerative properties in different organs systems, including the brain.^133–135 A recent preclinical study showed exosomes derived from bone marrow MSCs improved neurodevelopment in preterm injured animals.¹³² The study showed that exosomes treatment significantly reduced inflammation induced neuronal cellular degeneration, reduced microgliosis and prevented reactive astrogliosis. MSCs exosomes also improved myelination and provided long lasting improvement in cognitive function. Exosome therapy has potential advantages over stem cells as the risk of grafting and potential malignant transformation is reduce because exosomes are non self-replicating, also exosomes are small in size thus has better potential to reach injured areas and it can be sterilized by filtration. Stem cells modify themselves according to the surrounding environment and their secretome is affected by the environment. In contrast, exosomes are static and have fixed cargo, but there is potential for in-vitro alteration of exosomal cargo by subjecting MSCs to the disease specific microenvironment. EVs like exosomes hold promise as a potential alternative to stem cell therapy.

Polyphenols

Polyphenols are natural molecules with variable phenolic structures that are enriched in foods like fruits, tea, vegetables, wine and other foods. Many known polyphenols have antioxidant, anti-inflammatory, and anti-apoptotic properties. Hence, they are ideal candidates to be evaluated as neuroprotective agents. They are classified into groups depending on the number of phenol rings, and chemical groups bound to the rings.

Resveratrol.

Resveratrol (3,5,4′-trihydroxystilbene) is a non-flavonoid polyphenolic compound consisting of two aromatic rings attached by a methylene bridge. Resveratrol is present commonly in grapes, soybeans and pomegranates.¹³⁶ Preclinical studies have shown resveratrol as a neuroprotective agent in models of ischemia, brain and spinal cord injury. Resveratrol provides neuroprotection through multiple pathways.^137–139 Several preclinical studies have demonstrated the efficacy of resveratrol in hypoxic-ischemic brain injury in neonatal rats, preserving neocortical and subcortical brain areas in the short term,^{140, 141} with long term improvement in motor and behavioral functions.¹⁴² Resveratrol has also been investigated as prenatal preventive treatment for neonatal brain injury, as it crosses the placenta and maternal supplementation can lead to reduction in tissue loss and decreased apoptosis in neonatal brains.^{143, 144}

Curcumin.

Curcumin (diferuloyl methane) is a polyphenol present in the rhizomes of Curcuma longa (zingiberaceae). It possesses many therapeutic properties including anti-inflammatory,^{145, 146} anti-oxidant,^{147, 148} and anti-cancer effects.¹⁴⁹ The neuroprotective effect are due to its modulation of neuronal apoptosis, oxidative stress and pro-inflammatory cytokines.^150–152 Curcumin is a potent inhibitor of NFkB signaling which leads a reduction in pro-inflammatory cytokines and enzymes.^{146, 152} In-vitro and in-vivo studies have shown that curcumin imparts anti-oxidative effects by direct scavenging of free radical with its phenolic structure, and modulation of Nrf2 pathway.¹⁴⁸ Curcumin also inhibits activation of NMDA, which is an important mechanism for excitotoxic brain injury.¹⁵³ Curcumin induces neurogenesis by modulating Canonical Wnt/β-Catenin Pathway, with reversal of cognitive deficits in Alzheimer’s disease.¹⁵⁴ Recently we showed curcumin-loaded nanoparticles were neuroprotective in a model of neonatal hypoxic ischemic encephalopathy. Curcumin loaded nanoparticles reduced microglial activation, infarct size, and improved gross injury scores.¹⁵⁵ Thus curcumin, like Epo and melatonin works on multiple pathways affecting brain injury and has a potential to be an effective treatment for neonatal brain injury.¹²⁹

Drug Delivery and Cell-specific Targeting.

A major limitation of polyphenols and other potentially protective compounds is poor bioavailability and cell specificity. Alternative delivery methods are being studied to circumvent this issue. Nanomedicine has emerged as an important field for delivering drugs to specific targets. Nanoparticles, such as polymeric dendrimers, can bind drugs, target their uptake by specific cell types, thereby increasing the drug concentration where it is needed, and they can also modulate drug delivery by modifying speed of release.¹⁵⁵ This can increase the bioavailability of drugs and decrease the dose needed for effect, thereby decreasing dose-dependent side effects. While this has been intensively studied for delivery of cancer therapeutic agents, attention is now turning to using these nanoparticles for other purposes such as neuroprotection. Curcumin-loaded nanoparticles have been shown to be more effective than free curcumin in ameliorating cerebral ischemic reperfusion injury in rats.¹⁵⁶ Polyamidoamine dendrimers-NAC, which localize in activated microglia and astrocytes, have been used to treat newborn rabbits with CP.¹⁵⁷ Nanomedicine has great potential to deliver drugs across the blood brain barrier targeting the site of injury and repair to improve neurologic outcomes.

Conclusions

The success of neonatology has led to an increased population of NICU survivors. The challenge for the next generation of neonatal researchers will be to improve the developmental outcomes of these survivors. Currently, 50% of children born < 1500 gm (~ 31 weeks of gestation) require special educational services, and 15% have repeated at least one grade in school.¹⁵⁸ For those born before 28 weeks of gestation, on average, 9% develop CP, 25% intellectual disability, 3% hearing loss, 2% blindness, and 10% autism. We can and must do better.