Neurodevelopmental Outcome of Infants who develop Necrotizing Enterocolitis: The Gut-Brain Axis

Abstract

Necrotizing enterocolitis (NEC) poses a significant risk for neurodevelopmental impairment in extremely preterm infants. The gut microbiota shapes the development of the gut, immune system, and the brain; and dysbiosis drive neonatal morbidities including NEC. In this chapter, we delineate a gut-brain axis linking gut microbiota to the adverse neurological outcomes in NEC patients. We propose that in NEC, immaturity of the microbiome along with aberrant gut microbiota-driven immaturity of the gut barrier and immune system can lead to effects including systemic inflammation and circulating microbial mediators. This nexus of gut microbiota-driven systemic effects further interacts with a likewise underdeveloped blood-brain barrier to regulate neuroinflammation and neurodevelopment. Targeting deviant gut-brain axis signaling presents an opportunity to improve the neurodevelopmental outcomes of NEC patients.

Introduction

Necrotizing enterocolitis (NEC) is one of the most prevalent gastrointestinal morbidities disproportionately affecting infants born before 32 weeks of gestational age (GA). Severe NEC is a medical emergency in the neonatal intensive care unit (NICU), moreover, long-term neurological complications may persist in those who develop NEC¹. Healthcare costs of medical and surgical NEC survivors through early childhood (up to three years old) are substantially higher than those of matched infants without a diagnosis of NEC during the initial birth hospitalization². Thus, options for improving early intervention strategies and long-term outcomes associated with NEC have significant economic implications and are paramount for care providers (NICU and parents) to ensure better treatment and quality of life. Mechanisms for neurodevelopmental impairment (NDI) in infants who develop NEC are difficult to identify and are likely multifactorial. The etiology of NEC remains elusive, but the known risk factors for the development of neonatal NEC are prematurity, bacterial colonization, enteral feeding, hypoxia, and/or intestinal ischemia^3,4. The bacterial communities representing the large portion of the gut microbiota acquired around birth are required for the development and maturation of the intestinal epithelium and immune system of the host⁵. Furthermore, gut microbiota maturates to a stable community concurrently with nervous system development during the first 2–3 years of life and has increasingly been shown to influence the development of brain and host behavior^6–10. Recent research on the interaction between the gut microbiota and brain function known as the gut-brain axis has raised questions about whether the immature intestinal microbiota in NEC patients can lead to aberrant gut-brain axis signaling and affect central nervous system (CNS) functions. A working hypothesis is that microbiota-driven immaturity of the gut barrier and immune system either as a single risk factor or with intricate interaction leading to systemic effects might contribute to the adverse neurological outcome in NEC. In this review, we will discuss the current understanding of factors contributing to NDI in NEC patients with an emphasis on the potential role of the gut microbiome in CNS functions.

Neurodevelopment outcomes associated with NEC

Overall, based on meta-analyses including studies from 1970 to 2006, it has been estimated that more than 40% of medical and surgical NEC patients develop NDI^11–23. Odds ratios for having NDIs were significantly higher in infants that developed NEC compared to infants without NEC. Furthermore, all NDI outcomes were worse in neonates with Bell’s stage III NEC or those surgically treated (surgical NEC) when compared to those managed non-surgically (medical NEC). The most prevalent NDIs identified in the NEC patients among very low birth weight (VLBW, <1,500g) or extremely low birth weight (ELBW, <1,000g) preterm infants were cerebral palsy (20% of neonates with NEC), cognitive impairment (36%) and psychomotor impairment (35%). A lesser proportion of neonates with NEC also developed visual (3%) and hearing impairment (3%). Of the studies included in the meta-analysis studies, Sonntag et al.¹³ revealed that almost 55% of newborns with NEC were severely compromised compared to only 22.5% of preterm infants without NEC. A large multicenter cohort study of the NICHD (Neonatal Research Network) evaluated the outcomes in 1100 ELBW and identified a particular higher probability of impairment of psychomotor development and increased chances of cerebral palsy¹⁴. Hintz et al.¹⁵ demonstrated in a study of over 2948 patients that among ELBW children at 18 to 22 months’ postmenstrual age (PMA), surgical NEC patients are at significant risk for Mental Development Index (MDI) less than 70, Psychomotor development index (PDI) less than 70, and overall impairment of neurological development compared to preterm infants without NEC. Later studies with follow-up at 18 to 22 months of age²⁴ or at school age^25–27 further suggested surgical treatment as a predictive factor for long-term cognitive and motor development in surviving NEC patients.

An important recognition in the analysis of the “extremely low gestational age newborns” (ELGANs, less than 28 weeks) study²⁸ is that five abnormalities, cerebral white matter injury (WMI), bronchopulmonary dysplasia (BPD), retinopathy of prematurity (ROP), NEC and bacteremia, occur at a higher rate among ELGANs than among infants born close to term. Compared to medical NEC, infants with surgical NEC (Bell’s stage IIIB) are at increased risk of WMI, periventricular leukomalacia (PVL), ROP, and BPD. Although the sequential onsets of these morbidities are difficult to determine, a possibility is that instead of considering NEC as an individual entity parallel to other common preterm morbidities, NEC might contribute to the development of morbidities in multiple organs. In a study from the Canadian Neonatal Network²⁹, surgical NEC was associated with higher odds of a composite outcome of mortality or morbidity (BPD, ROP, PVL, or nosocomial infection). A recent study from a Spanish cohort revealed that NEC is significantly associated with BPD, late-onset sepsis (LOS), cystic leukomalacia, ROP, acute kidney injury, and death³⁰. The risk to develop these morbidities increased significantly when comparing infants with medical NEC and surgical NEC to those with no NEC.

In summary, NEC, in particular surgical NEC, poses a significantly greater risk for NDI than preterm infants without NEC. Understanding the pathogenesis of NEC might lead to better outcomes of the concurrent morbidities in NEC patients.

Gut-brain axis: microbiome-related mechanisms contributing to the neurodevelopmental outcomes of NEC

Prematurity remains the dominant risk factor for NEC development. The incidence of NEC is negatively correlated with GA and is most common in infants born less than 28 weeks of GA and occurs with the highest rate when the infants are less than 32 weeks of age^31,32. The pathogenesis of NEC is proposed to originate from the immaturity of the gastrointestinal tract as both physical and functional barrier immaturity predisposes preterm infants to develop NEC^33–36. The intestinal barrier is underdeveloped in preterm infants³⁷ and is more compromised in surgically-removed intestinal tissue from NEC infants than in preterm infants without NEC, even in grossly unaffected resection areas³³. The intestinal barrier represents the first line of defense against invading organisms and an immature gut barrier can allow opportunistic gut pathobionts to infiltrate the gut to act as triggers for both local tissue inflammation and systemic inflammatory processes^38–40. Gut microbiota shapes the normal development and maturation of the intestinal epithelium, immune system, and brain of the host^5,41–46. Thus, gut microbiota and its interplays with the immature “leaky” gut and immune system have been proposed to be the main contributing factors to the development of NEC and NDIs in NEC patients.

Immaturity of the microbiome

Numerous studies have attempted to investigate if there are aberrant microbial characteristics in infants who develop NEC that result in pro-inflammatory immune responses leading to NEC. In a large cohort of infants (n=1155) born between 23 and 27 weeks of GA, Martin et al demonstrated that infants who had surgical NEC with late bacteremia were at increased risk of psychomotor developmental impairment (PDI<70 on BSID), cerebral palsy and microcephaly when evaluated at 24 months corrected age⁴⁷. In a longitudinal study⁴⁸, infants who developed NEC (Bell’s stage 2 or 3) displayed a relative abundance of Gammaproteobacteria and a relative rareness of Negativicutes when compared with GA- and birth weight-matched controls. It appeared that NEC microbiome samples clustered separately from non-NEC samples before the onset of NEC and that NEC samples were characterized by reduced overall microbial diversity and an abundance of Gammaproteobacteria species^49,50. A meta-analysis study that included 14 studies further demonstrated that microbial dysbiosis preceding NEC in preterm infants is characterized by increased relative abundances of Proteobacteria and decreased relative abundances of Firmicutes and Bacteroidetes⁵¹. A difference in overall microbial diversity was not consistently observed between the NEC and non-NEC cohorts possibly due to the number of microbial species present in preterm infants being relatively low. In an attempt to develop rodent models of NEC, Jilling et al discovered that bacterial colonization with the gram-negative organisms Serratia marcescens and Klebsiella pneumoniae, and, to a lesser degree, a gram-positive bacteria Viridans streptococci were a prerequisite for the development of NEC in formula-fed, cold/asphyxia stressed neonatal rodents⁵². Interestingly, these species are among the pathogens identified in LOS in preterm infants⁵³. Overall, although no consensus on a single organism responsible for NEC development can be reached³¹, there is evidence that distinct characteristics of inappropriate colonization and bacterial postnatal development of the preterm infant are associated with the development of NEC.

Since a variety of pathogens have been linked to NEC possibly due to geographic location, NICU environment, antibiotic exposure, enteral feeding, and exposure to breastmilk³¹, an alternative approach is to evaluate the role of preterm microbiome as a whole on host physiology. For example, using random forest-based modeling, a microbiota for age has been defined by 25 age-discriminatory taxa in fecal samples of healthy infants at 6 to 18 months of age. Microbiota immaturity of undernourished infants defined by delayed microbiota for age transmitted a stunted growth phenotype when transfered to germ-free (GF) mice⁵⁴. There is also a temporal developmental pattern in which the preterm infant microbiome differs from the microbiome of full-term infants at two weeks of age but shifts towards the pattern of a full-term infant microbiome at six weeks of age and later, confirming an age-dependent maturation of the preterm infant microbiome⁴⁹. A concept has been proposed to consider microbiome maturity as a predictive factor for outcomes such as neurodevelopment. Microbial samples (postmenstrual age (PMA) 29 to 36 weeks) from human preterm infants born from 27 to 34 weeks GA were given to pregnant C57/BL6J germ-free (GF) dams to investigate the brain development of the offspring under the “isolated” influence of the microbiome maturity. Maturation of microbiota, defined by PMA, PMA-depended bacterial diversity, dominance of the key taxa, and circulating metabolites, led to improved associative learning and memory in adult mice⁵⁵. The current advanced multi-omics analysis platforms represent an opportunity for future studies to model microbiome maturity as a function of disease risk such as NEC as well as neurodevelopment.

Immature immune system

Immature immune system response to an immature gut microbiome in preterm infants can also contribute to the development of NEC⁵⁶. The immune system evolved primarily for host defense against infection by means of the innate and adaptive arms of the immune system. The connection between innate immunity and NEC was first demonstrated in epithelial cells of NEC rodent pups in which Toll-like receptor 4 (TLR4) is upregulated, while TLR4-deficient mice are protected from NEC⁵². Studies further confirmed the marked increases in the mucosal expression of TLR4 in the small intestine of human, surgically treated NEC infants and of mice pups with NEC⁵⁷. TLRs belong to the family of pattern recognition receptors (PRRs) in the innate immune system that are highly conserved receptors to sense a distinct pathogenic or endogenous ligand. In particular, TLR4 is a specific receptor for the gram-negative bacterial cell wall component lipopolysaccharide (LPS)⁵⁸. LPS can bind to TLR4 and activate the downstream nuclear factor-κB (NF-κB) pathway to elicit a proinflammatory response in the host⁵⁹. Immature human enterocytes have a propensity to exaggerated inflammatory responses and subdued anti-inflammatory mechanisms possibly due to the increased expression of TLR4 and its dysregulated signaling pathway^60,61. Compared to higher severity of NEC in wild-type TLR4 mice, TLR4-mutant mice had reduced enterocyte apoptosis and intestinal inflammation and increased enterocyte proliferation and healing^52,57. In addition, TLR4 mediates intestinal ischemia in NEC through eNOS expression⁶², reduces goblet cells differentiation via Notch signaling⁶³ and decreases proliferation, and increases apoptosis in intestinal stem cells⁶⁴. Another family of PRR is nucleotide oligomerization domain-like receptors sensing intracellular ligands. Genetic mutations in one of the receptors, nucleotide-binding oligomerization domain 2 (NOD2), have been implicated in inflammatory bowel diseases such as Crohn’s disease⁶⁵. NOD2 can be activated by muramyl dipeptide (MDP), a peptidoglycan motif present among both Gram-positive and Gram-negative bacteria to execute both inflammatory responses and anti-microbial functions. VLBW infants with known NOD2 loss-of-function mutations have an increased risk for developing NEC⁶⁶ potentially due to impaired function of Paneth cells in producing bacterial clearing molecule α-defensin.

Adaptive immunity also plays a role in the development of NEC. Regulatory T-cells (Tregs) of lymphocytes and TLR4-mediated polarization of IL-17 producing lymphocytes are involved in NEC development as evidenced by studies demonstrating that restoring Tregs and neutralizing IL-17 reduced the severity of NEC⁶⁷. Impaired tight junctions, increased apoptosis, and reduced proliferation of enterocytes by IL-17 provide the underlying mechanisms by which irregular lymphocyte function might contribute to NEC⁶⁷. Furthermore, intraepithelial lymphocytes (IELs) are also important in managing bacterial clearance and maintaining barrier integrity in the event of mucosal inflammation⁶⁸. CD8⁺ γδ IELs, a predominant subset of IELs, are depleted in preterm infants with NEC compared to control infants⁶⁹. Reduced expression of occludin in the tight junction might be responsible for reduced γδ IELs migration into the epithelium in NEC patients. Although further investigation of other components of the immune system is warranted, the profound effects of bacterial sensing PRRs and lymphocytes revealed in these studies demonstrate clear evidence of dysregulated host-microbial interactions via immature immunity in the pathogenesis of NEC.

As a result of immature immunity, heightened gut microbiota-driven intestinal immune responses can result in local inflammation, releasing mediators including cytokines, and chemokines, along with other inflammatory factors such as prostaglandins, leukotrienes, complement, and immune cells to the circulation to have systemic effects on distal organs including the brain⁷⁰. In fact, recent studies have suggested that gut microbiota can shape the brain’s immune system. This was revealed when GF mice displayed global defects in microglia maturation and an immature phenotype, leading to impaired brain innate immune responses⁷¹. Gram-negative infections via TLR-4 activation can mediate brain injury through microglial activation and white matter injury^72–74. Microglia activation can target the preoligodendrocyte in diffuse white matter injury through the release of reactive oxygen and nitrogen species⁷³. In the context of NEC, brain injury observed in mice with NEC is mediated through activation of TLR4 on microglial cells, and TLR4-knockout mice were protected from NEC-induced brain injury⁷⁵. Furthermore, gut-derived CD4(+) T lymphocytes were required for the development of NEC-associated brain injury since inhibition of T lymphocyte influx into the brains of neonatal mice reduced neuroinflammation and prevented demyelination⁷⁶. In a model of stroke, gut IL-17+ γδ T cells can traffic from the gut to the brain and enhance ischemic neuroinflammation by secreting IL-17. This can in turn increases chemokine production in the brain parenchyma and subsequent infiltration of neutrophils⁷⁷. These recent discoveries strongly demonstrate an axis of gut immunity-driven-brain immunity that is involved in both local inflammation and neuroinflammation and could play a role in the neurological outcomes in NEC.

Microbiota-driven Systemic inflammation

Gut microbiota-derived systemic inflammation through activation of both innate and adaptive immunity in NEC patients may be an important mechanism for the development of long-term NDI. Aberrant gut microbiota characterized by decreased bacterial diversity, abundant Staphylococci, and Enterobacteriaceae and a delay of maturation to obligate anaerobe dominance has been shown to precede LOS in preterm infants^78–80. Neonatal sepsis is an independent risk factor for NDI in extremely preterm infants^81,82. It has been estimated that ~21% of VLBW infants develop at least one episode of LOS confirmed by blood culture, with a high rate of mortality and severe morbidity, including poor neurodevelopmental outcomes⁸³. Furthermore, systemic inflammation has also been linked to neonatal complications in other organs. Both in human and animal models, systemic inflammation has been associated with increased risk for BPD^84,85 and ROP^86–89 that commonly clustered with NEC in preterm infants.

In NEC patients, the acute development of necrosis of the small and large intestine is usually followed by the development of systemic sepsis⁹⁰. ELBW with Stage 2 or 3 NEC had more culture-proven sepsis, in particular Gram-negative infections when compared to matched infants without NEC⁹¹. Sustained systemic inflammation was reported as infants with NEC had elevated levels of CRP, SAA, IL-6, and IL-8 on days 7 and 14 after birth⁹². In infants with proven NEC, systemic inflammation is marked by significantly higher serum concentrations of IL-6, IL-8, and IL-10 when compared with age-matched healthy controls⁹³.

Strong evidence has demonstrated that systemic inflammation can regulate brain function through neuroinflammation in the CNS^94–99. Neuroinflammation can lead to common neonatal brain injuries such as PVL and hypoxic-ischemic encephalopathy¹⁰⁰. These neuroinflammation-induced brain injuries in early life can contribute to long-term adverse neurological outcomes such as cerebral palsy, autism and schizophrenia^101,102, and motor and cognitive deficits^103,104. Animal models of systemic inflammation have demonstrated that sub-optimal gut microbial communities can result in systemic inflammation, and neuroinflammation and negatively regulate brain development^75,105–108. Systemic inflammation induced by LPS can lead to astrocyte and microglia activation¹⁰⁹ and astrocyte gene transcription towards a proinflammatory state¹¹⁰. Furthermore, neonatal infection modeled by Escherichia coli colonization or LPS challenge can lead to memory deficit in adult life^111,112. In mice with commensal microbes, peripheral IL-1β-induced systemic inflammation can result in neuroinflammation and delay neuron and oligodendrocyte progenitor cell development, and increased astrocyte and microglia activation in the young brains¹¹³. In particular, studies of the preterm infant microbiome revealed that preterm infant microbiome with a poor growth outcome can result in increased local intestinal inflammation evidenced by NF-κB activation and systemic inflammation by increased circulating levels of IL-1β, TNF, and IFNγ in a “humanized” GF mouse model¹⁰⁸. Furthermore, preterm microbiome associated with poor growth led to neuroinflammation, impaired neuron development and delayed myelination of oligodendrocytes, and altered gene expression in multiple neurotransmission pathways as well as neurotransmitter transporters and ion channels¹⁰⁷. Notably, brain abnormalities in NEC are characterized by reduced myelination in the white matter^114,115. Thus microbial-associated systemic inflammation might have a significant role in the adverse neurological outcomes in NEC.

Microbial metabolites

Gut Microbiota-derived metabolites have been proposed to be another mechanism by which the gut microbiota can regulate CNS function^44,45. Several studies have attempted to characterize the metabolic profiles related to NEC development. In NEC patients, NEC diagnosis corresponded to fecal microbial metabolites in steroid hormone biosynthesis, linoleate metabolism, leukotriene metabolism, and prostaglandin formation from the arachidonate pathway¹¹⁶. Urine alanine was positively associated with early-onset NEC that was preceded by Firmicutes dysbiosis and histidine was negatively associated with late-onset of NEC preceded by Proteobacteria dysbiosis¹¹⁷. Compared to healthy controls, enzymes involved in tryptophan metabolism, biotin synthesis, and HMOs degradation were depleted in fecal samples of NEC infants¹¹⁸. The enzyme lactate dehydrogenase (LDH) was increased in preterm infants before NEC development¹¹⁸. The severity of NEC shifted the plasma metabolite profile in a pig model of NEC¹¹⁹. Pigs with severe NEC had higher serum levels of alanine, histidine, and Myo-inositol, and lower levels of 3-hydroxybutyric acid and isobutyric acid. More studies are needed to define the microbial metabolic profile in NEC and to understand the role of microbial metabolites contributing to neurological outcomes since the current studies have displayed diverse profiles possibly due to the small sample size and sample resources. Circulating microbial metabolites would provide a crucial element to link the effects of gut microbiota with brain development and NDI of NEC patients.

Blood-brain barrier

As described above, current studies of the gut-brain axis suggest the gut microbiome can influence CNS function via pathways of systemic inflammation, microbial mediators, and immune surveillance¹²⁰. A common trait of these pathways is the release of microbial mediators (i.e. cytokines, metabolites, activated immune cells) into the systemic system. Whether or what microbial mediators reach the CNS to influence brain function is dependent on systemic communication between peripheral blood and the tightly regulated CNS barrier known as the blood-brain barrier (BBB). In normal physiological conditions, CNS is an “ immune-privileged” site due to the BBB comprised of endothelial cells, pericytes, astrocytes, and a basement membrane with inter-endothelial TJ proteins¹²¹. The contribution of BBB integrity in the development of NEC has not been explored, however several aspects of BBB development and function suggest a potential role, possibly related to microbiota.

First, the BBB is immature in preterm infants. Even though the presence of TJ proteins appears early in embryonic age in both humans and mice when the neocortex is first vascularized¹²², the pericyte coverage and junctional organization are not fully developed until birth^123–125. The dominant period of differentiation of astrocytes and the ensheathment of the brain vascular system with astroglial end-feet occurs in rodents in the first three weeks of postnatal age^126,127. Second, temporal transport mechanisms of the BBB, in particular, both influx and efflux families of transporters present in the brain, are developmentally regulated due to specific requirements for brain development at different stages of early life^127–132. Meanwhile, the sealing properties of the barrier only become fully mature when the intracellular pathway of transcytosis transport is developmentally downregulated by increased expression of the major facilitator superfamily domain-containing 2A (Mfsd2A) towards the late prenatal period^133–135. Immature developmental properties of the BBB may render developing brains of preterm infants more susceptible to circulating cytokines, immune cells, and microbial metabolites, contributing to cerebral damage and later neurological disorders¹³⁶.

BBB integrity can also be regulated by microbial communities. GF mice display increased BBB permeability due to reduced expression of the TJ proteins occludin and claudin-5 throughout the life course compared to mice with commensal gut microbiota. The difference in BBB is seen as early as E17.5, indicating an early effect of microbiota on BBB development¹³⁷.

In addition, studies have demonstrated that microbiota-related systematic inflammation can disrupt BBB integrity^137,138. Stonestreet et al demonstrated that IL-1β contributes to impaired blood-brain barrier function when systemic anti-IL-1β monoclonal antibody treatment decreased the uptake of IL-1β into the brain and reduced the permeability of BBB in ischemic fetal brain¹³⁹. Multiple systemic inflammatory cytokines can act on the BBB or across the BBB to trigger the activation of microglia, resulting in the WMI of PVL¹⁴⁰. There seems to be a specific time window in brain development when the BBB is more susceptible¹⁴¹. LPS-induced systemic inflammation-induced increased BBB permeability was only observed in rats before P20¹⁴¹, a stage of brain development equivalent to 22–40 weeks of gestation in humans¹⁴². Peripheral IL-1β can induce systemic inflammation and increased BBB permeability of P14, but not P28, days old mice¹¹³. Probiotics administration of Lactobacillus acidophilus and Bifidobacterium infantis (LB), has been shown to improve gut barrier function in preterm infants with NEC^143,144. These probiotics have also been associated with improved BBB function. Maternal LB supplementation inhibited peripheral IL-1β-induced systemic and neuroinflammation and restored BBB dysfunction in mouse offspring by regulating tight junction disruption, vascular injury, leukocyte recruitment, and extracellular matrix repair¹¹³.

Besides providing surveillance between CNS and peripheral blood, BBB actively regulates the delivery of energy metabolites such as lactate and ketone bodies and essential nutrients such as glucose and amino acids to the brain by highly specialized substrate-specific transport proteins¹⁴⁵. Mfsd2A is known as a key regulator of BBB integrity and transcytosis in CNS endothelial cells¹³⁵. Mfsd2A is further identified as the transporter for uptake of docosahexaenoic acid (DHA), an omega-3 fatty acid essential for normal brain growth and cognitive function^146–148. Genetic ablation of Mfsd2a results in significantly reduced levels of DHA in the brain accompanied by neuronal cell loss in the hippocampus and cerebellum, neurological and behavioral deficits, and reduced brain size. Early studies have shown that DHA supplementation reduced circulating LPS, epithelial TLR4, and NEC incidence and severity in rodent NEC models^149,150. Regulation of fatty acids uptake through BBB can provide another aspect of how BBB can contribute to brain outcomes.

Overall, these studies suggest that BBB immaturity of preterm infants and microbiota-mediated BBB function might contribute to the adverse neurological outcomes in NEC.

Conclusion

Preterm infants who develop NEC are at increased risk for neurodevelopmental impairment later in life. Immature gut microbiota drives morbidities including NEC in preterm infants and alters long-term neurological outcomes in animal models. Future studies of microbiota-mediated immunity, systemic inflammation, microbial metabolites, and BBB development might provide critical links between gut and brain in the context of neurological outcomes in infants who develop NEC.