Necrotizing enterocolitis, gut microbes, and sepsis

ABSTRACT

Necrotizing enterocolitis (NEC) is a devastating disease in premature infants and the leading cause of death and disability from gastrointestinal disease in this vulnerable population. Although the pathophysiology of NEC remains incompletely understood, current thinking indicates that the disease develops in response to dietary and bacterial factors in the setting of a vulnerable host. As NEC progresses, intestinal perforation can result in serious infection with the development of overwhelming sepsis. In seeking to understand the mechanisms by which bacterial signaling on the intestinal epithelium can lead to NEC, we have shown that the gram-negative bacterial receptor toll-like receptor 4 is a critical regulator of NEC development, a finding that has been confirmed by many other groups. This review article provides recent findings on the interaction of microbial signaling, the immature immune system, intestinal ischemia, and systemic inflammation in the pathogenesis of NEC and the development of sepsis. We will also review promising therapeutic approaches that show efficacy in pre-clinical studies.

KEYWORDS:

• premature
• necrotizing enterocolitis (NEC)
• microbes
• sepsis

Introduction

Necrotizing enterocolitis (NEC) is a devastating gastrointestinal disease of premature neonates¹. This condition is characterized by ischemic necrosis of the small and large intestine, leading to the translocation of enteric organisms into the circulation, often resulting in overwhelming sepsis and death^2–4. NEC remains a leading cause of morbidity and mortality, especially in premature, low birth weight infants, and neonates receiving enteral feeds, who are at the highest risk^5–7. Breastfeeding is protective against NEC^8–10. However, mortality rates are reported to range from 15% to 45%^11–14. Those infants that survive suffer from severe sequelae, including short-bowel syndrome, growth restriction, and neurological deficits^15,16. The etiology of NEC is multifactorial and incompletely understood. Current thinking indicates that the pathogenesis of NEC involves the complex interplay between dysbiotic enteric microbes as defined by a reduction in bacterial diversity, increased immune reactivity of the intestinal mucosa, and genetic factors in certain cases^17–20.

Neonatal sepsis is characterized by the presence of bacteria, fungi, or viruses in the bloodstream, with preterm infants being most vulnerable, leading to a high rate of mortality and severe morbidity, including poor neurodevelopmental outcome^21,22. Frequent symptoms include high or low body temperature, pneumonia, apnea, lethargy, poor feeding, and bleeding manifestations. However, signs of sepsis in the premature period are often subtle and nonspecific, making the clinical diagnosis extremely challenging^22,23. The onset of neonatal sepsis is often associated with enteral dysbiosis and has features that parallel NEC as will be discussed below^24,25.

While not all infants with NEC develop sepsis, those with severe disease uniformly do. Specifically, neonates with advanced or surgical NEC (Bell’s Stage III) are twice as likely to have culture-proven bacteremia as those with suspected or medical NEC^21,26,27. In the setting of intestinal mucosal injury, bacterial translocation into the bloodstream commonly occurs, triggering NEC-associated sepsis, which has recently been reported to be associated with greater inflammatory response and hemodynamic support in preterm as compared with term infants²¹. Intravenous antibiotics that are effective against enteric microbes are a mainstay in treating NEC, even at early stages²⁸. Somewhat counterintuitively, the prolonged use of antibiotics is a risk factor for NEC^29–31. Although difficult to prove, this finding may be attributable to antibiotic-associated disruption in the microbiome, which has been linked to the development of NEC²⁸. Prolonged administration of antibiotics may also be associated with increased intestinal permeability, as well as impaired intestinal development^31,32. It is noteworthy that the development of sepsis may be a risk factor for NEC^33–36, although it is not clear that the intestinal injury that occurs in the setting of sepsis has the same clinical course as NEC in other settings^37,38.

Clinical presentation and diagnosis of NEC

The clinical presentation of NEC is highly variable. Infants may present with mild, nonspecific symptoms or with fulminant disease leading to sepsis and multiorgan system failure. This gastrointestinal disorder frequently begins after the initiation and exclusive administration of infant formula feeds in an otherwise stable infant³⁹. Common intestinal signs of NEC include abdominal distension and feeding intolerance, secondary to ileus, which can progress to abdominal wall discoloration, bilious aspirates, and hematochezia. The abdominal wall discoloration seen in NEC can vary from a shiny erythematous appearance to a violet appearance if bowel perforation has occurred^40–42. Male infants may additionally develop discoloration of the scrotum if peritoneal fluid, following intestinal perforation, herniates through the processes vaginalis⁴³. Additional abdominal physical exam findings may include abdominal tenderness and palpation of bowel loops. Infants who develop NEC may also foster nonspecific systemic symptoms. These include temperature instability, apnea, bradycardia, tachycardia, hypotension, hyperglycemia, hypoglycemia, and mottling. Laboratory studies are routinely obtained in these patients but remain nonspecific. Anemia, leukocytosis with a neutrophilic left shift, neutropenia, thrombocytopenia, metabolic acidosis, coagulopathy, elevated C-reactive protein level, and hyponatremia are most encountered^41,42,44. Findings of pneumatosis intestinalis on abdominal radiograph are pathognomonic for NEC, although the absence of this finding does not rule out NEC⁴². Radiograph signs for NEC may include portal venous air or free intraperitoneal air. Abdominal ultrasonography can be utilized to detect pneumoperitoneum, fluid collections, portal venous gas, bowel hypoperfusion, and pneumatosis intestinalis, although this technique is limited by the expertise of the sonographer⁴².

The Bell staging system was created 40 years ago and has been utilized to determine the severity of and guide treatment of NEC. This system utilizes clinical and radiographic findings to categorize infants into different stages. Ten years after this staging system was implemented, the Modified Bell staging criteria were introduced which increased the stages from 3 to 6 to further guide treatment. Although many studies and cohorts have adopted the Bell or Modified Bell staging criteria, this system has multiple limitations. While the tool was developed to determine the severity of NEC, it has become inappropriately utilized as a diagnostic tool. This has led to over-diagnosis of NEC in certain instances, as the symptoms reported in stage 1 are nonspecific and can be seen in normal very low birth weight infants. Additionally, spontaneous intestinal perforation (SIP) can be classified as Bell criteria stage 3, as there is pneumoperitoneum on abdominal radiograph. However, SIP is distinct from NEC with a very different epidemiological profile (i.e. earlier onset and minimal feedings). With these limitations, there have been accuracies in NEC datasets due to confounding diagnoses classified as NEC. There are ongoing efforts to come up with a more specific diagnostic criteria to exclude these confounders, such as the Vermont Oxford Network definition, Two of three rule, and Stanford NEC score⁴⁵.

Pathophysiology of NEC

The transition from the intra-uterine to the maternal and postnatal environment constitutes a complex interaction between the newborn and colonizing microbes that could serve as potential pathogens over the first few weeks of life^46,47. This transition is particularly challenging in the case of the premature and very low birth weight infants who display an immature and underdeveloped immune system⁴⁶. Unique characteristics of the neonatal immune system have been described to include changes in the cytokine, growth factor and hormone signaling pathways, nutritional effects of probiotics and breast milk, as well as distinct microbial colonization in the gut and other mucosa specific to the neonatal period^46,48–51. These transition processes have long-term effects on the immune system by creating immune activation or tolerance, which determines the development of NEC⁵² (Figure 1). Although the pathophysiology of NEC is still incompletely understood, the preponderance of evidence indicates that signaling in response to toll-like receptor 4 (TLR4) on the intestinal epithelium in response to a dysbiotic microbiome in the premature gut leads to impaired immune function and epithelial cell death by apoptosis, autophagy, and necroptosis^53–63. In addition, impaired intestinal mucosal restitution and reduced proliferation result in irreversible gut barrier injury^64–67, leading to the clinical manifestations of NEC (Table 1).

Figure 1. Pathophysiology of NEC.

Activation of the innate immune receptor toll-like receptor 4 (TLR4) by a dysbiotic microbiome on the intestinal epithelium of the premature gut leads to intestinal mucosal damage and impaired mucosal repair, leading to bacterial translocation and activation of TLR4 on the endothelium, resulting in mesenteric vasoconstriction, which leads to intestinal ischemia and NEC. Strategies to interrupt this signaling pathway or to restore a normal microbiome within the premature gut may prevent NEC development.

Table 1. Mechanistic factors in the development of NEC.

Pathophysiology	Mechanism	Consequence	Evidence
Exaggerated TLR4 signaling in the intestinal epithelium	increased epithelial apoptosis and reduced mucosal repair ↓ goblet + Paneth cells Thinner mucus secretion Impaired peristalsis Increased Th17 and reduced Tregs in intestinal mucosa	Baseline pro-inflammatory state ↓ tight junction ↑ enterocyte apoptosis ↓ enterocyte proliferation Global mucosal injury	↑ TLR4 leading to activation of NF-κB and release of IL − 6, IL − 1β, TNF-α, nitric oxide, IL − 17, and IL − 22 ↓ NETs Loss of enteric glia and intestinal dysmotility ↓eNOS, vasoconstriction and intestinal ischemia
Dysbiotic intestinal microbiota	↓ SCFA ↑ Proteobacteria phylum, Enterococcus, and Staphyloccus	Pro-inflammatory response	Phagocytosis and translocation of bacteria across intestinal mucosal barrier
Mesenteric vasculature ischemia	Increased resting vascular tone in the intestinal mesentery of the premature infant	TLR4 activation leading to vasoconstriction. ↓ intestinal perfusion and intestinal ischemia, bacterial translocation, and sepsis	↓ VEGF/VEGFR2 ↓ eNOS ↓ endothelial cell proliferation ↑ HIF − 1α and GLUT1 ↓ tight junction proteins, occluding, and ZO-1
Disrupted in utero signaling	Exaggerated baseline intestinal inflammation due to increased TLR4, increased Th17, and reduced Tregs	Alteration on intestinal transcriptional profiles and ↑ antibacterial peptides Chorioamnionitis	AHR for early postnatal immunity and ↓ TLR4 ↓ goblet + Paneth cells ↑ inflammatory markers

The neonatal immune system and NEC

The adaptive and the innate components of the immune system are immature in all neonates, yet especially so in preterm infants. This immaturity is thought to be related to a lack of antigenic exposure, enhanced self-modulatory immunosuppressive mechanisms, reduced physical barriers and impaired function of most cell types^68–72. Furthermore, a lower number of goblet and Paneth cells, thinner mucus, reduced levels of IgG combined with low opsonic activity, reduced activity of the complement system and low numbers of neutrophils and monocytes have been linked to a higher risk for infections in this vulnerable cohort^73–86. Although there is evidence for early development of protective memory subsets, T cells are programmed to rapidly proliferate into cells that generate strong effector responses. Those responses are primarily confined to the mucosal sites of the respiratory and gastrointestinal tract, leading to exaggerated immune activity⁸⁷. In addition, it has been shown that preterm infants overexpress genes responsible for the negative regulation of IFN-γ production, T cell proliferation, and IL − 10 secretion⁸⁷. Differences in immune profiles of preterm and term infants have been reported to become even more exacerbated with maturity⁸⁷. The net effect is a skew toward a pro-inflammatory state in preterm infants, increasing the risk of sepsis compared to their term counterparts⁸⁷. As mentioned above, the rapid accumulation of effector responses results in impaired enterocyte tight junction, increased enterocyte apoptosis, and decreased enterocyte proliferation, ultimately leading to NEC^46,56.

Various mouse and neonatal tissue sampling studies have sought to investigate the underlying pathophysiology of the imbalance between injury and repair of the premature gut⁸⁸. The excessive immune response to microbial ligands in the premature infant results from an imbalance of CD4⁺ effector T cells, γδ T cells, Th17, Treg subsets, and their associated cytokines⁴⁶. Tregs are an anti-inflammatory subset, which are a major source of IL-10, and have been found to be decreased in the lamina propria in infants with NEC. Lipopolysaccharide (LPS), which is the main endotoxin in the cell walls of gram-negative bacteria like Escherichia coli and Helicobacter pylori, binds to the transmembrane (TLR4) on the intestinal epithelium, leading to an inflammatory response, which involves the activation of nuclear factor-κB (NF-κB). Activation of NF-κB induces the release of IL − 6, IL − 1β, TNF-α, along nitric oxide, as well as the release of IL − 17 and IL − 22 by proinflammatory Th17 cells^46,56,87. Vincent et al. described that following the induction of NEC, the formation of neutrophil extracellular traps (NETosis) was significantly increased in animals and human samples. The authors suggest that hyperinflammation observed in NEC is a NET-dependent process⁸⁹. However, Yost et al. showed that neutrophils isolated from term and preterm infants fail to form NETs, which may explain the increased susceptibility of preterm infants in particular to sepsis and infection^90,91. Furthermore, prevention of NETosis leads to reduced mortality, tissue damage, and inflammation in NEC samples, which might limit neutrophil-mediated epithelial damage and, thus, the consequences of NEC and sepsis⁸⁹.

Studies from our group have also revealed the important role of TLR4 in the regulation of intestinal development. We and others have shown that TLR4 expression is higher in the premature (and still developing) gut compared to the full-term gut⁸⁷, which reflects the nonimmune role of TLR4 in the regulation of intestinal stem cell differentiation through Notch activation. Importantly, the subsequent exposure of the premature infant to colonizing microbiota that are enriched in LPS leads to activation of TLR4, and its subsequent switch from a developmental to an inflammatory role, causing induction of NEC. We have further revealed that TLR4-mediated loss of enteric glia leads to intestinal dysmotility, which accounts for the abdominal distention and emesis that are encountered early in the pathogenesis of NEC¹⁸. TLR4 activation on the endothelium also induces loss of endothelial nitric oxide synthase (eNOS), vasoconstriction, and intestinal ischemia, which eventually results in a pattern of disease in clinical NEC^56,87.

The role of the intestinal microbiota in the pathogenesis of sepsis and NEC

The microbiome of the gut describes the collection of microorganisms, which reside in the small and large intestine and live in symbiosis with the human host²². Following birth, when the neonate is exposed to bacteria and other microorganisms for the first time, the infant gut is initially colonized with anaerobic species (Bacteroides, Bifidobacterium, and Clostridium spp.) and undergoes maturation with an increase in bacterial richness and diversity achieving an adult-like microbiome around 2 to 5 years of age, with a preponderance of Bacteroidetes and Firmicutes^22,92–94. Furthermore, reduced Bifidobacterium colonization with lower diversity in the microbiome is related to higher incidence of diseases, such as NEC and sepsis in infants^22,95,96.

A constellation of studies have revealed that the intestinal microbiota plays an important role in the development and function of the immune system^97–100. The function of the bacterial component of the microbiome includes roles in the synthesis of mediators, including peptidoglycans and short-chain fatty acids (SCFA) that serve to regulate the mucosal immune system and to contribute to the host defense against infections^{46,99,101,102}. Observational studies of patients with sepsis have shown a link between microbiota composition and diversity and adverse outcomes, including increased infections and mortality^99,103–105. For example, several studies have found that the gut microbiota of patients with sepsis is relatively depleted in anaerobic fermenters, with reductions in fecal SCFA levels. In addition, the dysbiotic microbiome in septic patients was observed to be enriched with Proteobacteria phylum as well as Enterococcus and Staphylococcus^99,104.

Intestinal microbes can communicate with immune cells in extra-intestinal organs via soluble mediators as well as through extracellular vesicles, neurotransmitters, and hormones⁹⁹. Bacterial dysbiosis can disrupt the normal, homeostatic host-microbial signaling, and lead to impaired intestinal epithelial integrity, weakened mucus-associated defense, and activation of resident immune cells^106–109. In neonates, dysbiosis has been shown to be associated with the development of inflammatory processes in the gastrointestinal tract, which can lead to NEC^106,110.

Interaction mechanisms among NEC, gut microbes, and sepsis

Clinical studies have shown that the gut microbiota of preterm infants is immature and less diverse as opposed to full-term infants.^46,111–114. Younge et al. have reported that extremely premature infants with growth failure have a microbiome that resembles a metabolic state analogous to fasting, in spite of adequate caloric intake¹¹⁵. Above all, the gut microbiome among premature infants with NEC is also different from the microbiome of those without NEC. Several studies analyzing fecal microbiota of premature infants found an increase in abundance of Proteobacteria and a decrease in abundance of Firmicutes and Bacteroidetes before the onset of NEC^116–121.

Dysbiosis among premature infants is not only comprised of an immature immune system and immature intestine but also other risk factors for sepsis, such as prolonged use of central catheters, delayed initiation of enteral feeding and increased duration of ventilator support^22,122. An association between an aberrant gut microbiome development with low bacterial diversity, including higher Proteobacteria and Firmicutes, and delayed colonization by obligate anaerobic species has also been linked to sepsis onset^{22,96,123–125}. In more detail, a variety of gram-negative enteric bacteria (Lebsiella spp., Pseudomonas spp., and E. coli) and gram-positive bacteria (Streptococcus spp., Enterococcus spp., and coagulase-negative Staphylococci) have been shown to cause sepsis in preterm infants^124,125. Interestingly, Stewart et al. have recently demonstrated that the causative organism of sepsis was even abundant in the gut at the time of diagnosis and the presence of Bifidobacteria was linked with protection against gut epithelial translocation⁹⁶.

Proteobacteria contain large quantities of LPS in their cell wall, which initiates a pro-inflammatory response following the activation of TLR4 via NF-κB signaling leading to phagocytosis and translocation of these gram-negative bacteria across the intestinal mucosal barrier. On the other hand, Firmicutes are gram-positive microbes and also capable of producing SCFA, which are important for colonic epithelial integrity¹¹⁸. In terms of Bacteroidetes, Bacteroides fragilis mediates the conversion to Treg that produce IL − 10, which is important for immunoregulation and homeostasis^87,118.

Ischemic injury and barrier dysfunction in the pathogenesis of NEC

Several studies have suggested that the interplay between bacteria in the intestinal lumen and the underlying immature vascular endothelial network can lead to impaired intestinal perfusion, which results in the development of intestinal ischemia and NEC⁵⁶. For example, Aski et al. performed a Cochrane Systematic Review (2017) and prospective meta-analysis collaboration (2018) comparing the effects of lower (85%−89%) and higher (91%−95%) oxygenation saturation in 4965 infants born before 28 weeks’ gestation. There was no significant difference between a lower and higher oxygen target range on death and major disability. However, the lower oxygen saturation range was associated with a significantly higher incidence of NEC^126,127. Specifically, Yazji et al. has demonstrated that activation of TLR4 on the mesenteric endothelium by bacteria that have translocated across the intestinal barrier results in reduced eNOS expression, which causes mesenteric vasoconstriction and the development of intestinal ischemia¹²⁸. In addition, Bowker et al. have shown a downregulation of pro-angiogenic signaling pathways such as vascular endovascular growth factor (VEGF)/VEGF receptor (VEGFR) 2 in premature infants before the intestinal microvasculature sufficiently develops, while the lack of VEGFR2 signaling predisposes to the development of NEC in mice¹²⁹. Perinatal stress, like perinatal inflammation, further reduces VEGFR2 signaling and endothelial cell proliferation, which contributes to ischemia and necrosis¹³⁰. Chen et al. have reported that hypoxia markers, hypoxia-inducible factor 1α (HIF-1α), and glucose transporter 1 (GLUT1) were elevated NEC¹³¹.

Gastrointestinal epithelial cells throughout the GI tract provide a physical barrier against the translocation of bacteria, bacterial antigens, digestive enzymes, and digested food¹³². Following bacterial colonization, TLR4-induced intestinal inflammation leads to impairment of the intestinal epithelial barrier, bacterial translocation, and ultimately to sepsis^132,133. Epithelial barrier integrity is maintained by the formation of tight junctions (TJ) between adjacent cells, which are composed of cytoplasmic and transmembrane proteins, as well as active transport mechanisms^46,132,134.

In preterm infants, some tight junction proteins are downregulated resulting in alterations in the function of intestinal tight junction^135,136. As demonstrated by Yu et al., intestinal tissue obtained from germ-free mice, which was colonized with microbiota from preterm infants showed a lower expression of occludin and protein ZO − 1 combined with disorganization in TJ protein assembly¹³⁷, resulting in impaired tight junctions. Furthermore, Bein et al. examined specimens from NEC patients and found significant downregulation of tight junction genes in addition to an upregulation of HIF-1A. Overall, hypoxic conditions seemed to initiate the destructive process of the epithelial barrier¹³⁵. Enterocyte migration is also impaired as a consequence of TLR4 activation, with an increase in apoptosis, autophagy, reduced cell proliferation, and impaired cell regeneration, adding to the impairment of the barrier of the epithelium^61,138.

In utero effects on the development of NEC

Various studies indicate a potential role of the maternal environment on the pathogenesis of NEC,^139,140, including the role of the maternal microbiome and maternal infections⁴⁶. Gomez de Agüero et al. have demonstrated that the exposure of E. coli to germ-free mice during pregnancy results in pups with an increase in intestinal type 3 innate lymphoid cells (ILC3) and F4/80 mononuclear cells. Maternal antibodies protect the neonate not only through pathogen neutralization but also via microbial molecular transfer, altering the intestinal transcriptional profiles and increased production of epithelial antibacterial peptides. Ligands for the aryl hydrocarbon receptor (AHR) ligand, a conserved sensor that integrates environmental, microbial, metabolic, and endogenous signals into specific cellular responses, can be derived from the maternal microbiota and shape the composition and function of early postnatal immunity, thereby driving ILC3 expansion and limiting adult bacterial translocation^141,142. In addition, Lu et al. have demonstrated that administration of a diet rich in AHR ligand indole -3-carbinol, prevents NEC in newborn mice by reducing TLR4 signaling in the newborn gut, which was also confirmed in human samples. Breast milk was also found to be rich in AHR ligands, highlighting the significance of breast milk in the prevention of NEC¹⁴³. Another study investigated the effects of chorioamnionitis on intestinal immune development in piglets following intra-amniotic administration of LPS 3 days before delivery. At birth and up to 5 days postnatally, LPS-exposed pigs showed higher intestinal endotoxin, neutrophil/macrophage density, and shorter villi¹⁴⁴. Elgin et al. injected pregnant mice on embryonic day 15.5 with LPS to stimulate exposure to maternal inflammation induced by chorioamnionitis followed by a second dose of LPS to pups on day 5. Interestingly, chorioamnionitis did not affect the growth of the small intestine; however, numbers of both goblet and Paneth cells were significantly decreased with concomitant increase in inflammatory markers in an IL − 6 dependent manner. The authors suggest that these changes could potentially explain the higher incidence of NEC in preterm infants exposed to chorioamnionitis prior to birth^145,146. Taken together, a focus on the role of the prenatal period in the development of NEC could elucidate new mechanisms and identify new potential therapeutic approaches.

Prevention and treatment of NEC

Currently, there is no specific treatment for NEC, and current therapies include antibiotics, cessation of oral feeds, and surgical resection of the necrotic bowel according to the severity of NEC⁵⁶. We will therefore mainly focus on strategies that have been shown to be effective in the prevention of NEC, in animal models and in clinical studies (Table 2).

Table 2. Therapeutic and preventative factors in necrotizing enterocolitis.

	(in addition to antibiotics, gut rest, intravenous nutrition)
Treatment	Mechanism	Evidence
Breast milk, HMO	Inhibition of TLR4 on the intestinal mucosa Modulation of microbiota and immune modulation ↑ mucin production	Reduce TLR4 signaling → ↓ inflammatory response ↓ intestinal permeability ↑ TJ ↑ absorption
Probiotics	Multiple studies show ↓ NEC morbidity and mortality	Reduction of severe NEC Reduced TLR4 signaling Probiotic DNA leads to ↑ TLR9 → ↓ TLR4 Reversal of FOXP3+Treg reduction ↓ intestinal permeability ↑ TJ
Amniotic fluid/exosomes	Development and maturation of fetus ↓ Tissue injury	↓ TLR4 ↓ cytokines ↑ TJ proteins

Breast milk and human milk oligosaccharides

The administration of breast milk has been shown to have tremendous benefits in preventing NEC. A systematic review and meta-analysis of 49 studies investigated the effect of human milk on very low birth weight infants and showed a possible reduction in late onset sepsis, as well as a clear protective effect against NEC, with an approximate 4% reduction in incidence¹⁴⁷. Furthermore, breast milk administration can reduce signaling via the LPS receptor TLR4, in part through its constituent oligosaccharides 2’-Fucosyllactose (2’-FL) and 6’-Sialyllactose (6’-SL), and therefore attenuates NEC development¹⁴⁸. Human milk oligosaccharides (HMOs) are a subset of highly abundant carbohydrates in human breast milk. Because of their complex structure, HMOs are highly bioactive and resist gastrointestinal digestion and are therefore not absorbed in significant amounts¹⁴⁹. HMOs make up about 10% of the dry weight of mother’s milk, making them the third most abundant solid component after lipids and lactose¹⁵⁰. There are over 100 different HMOs, with the five most prevalent HMOs being 2’-FL, 3-FL, 3’-SL, 6’-SL, and Lacto-N-tetraose (LNT)¹⁵¹. The mean levels for each of these HMOs are roughly 2.6 g/L for 2’-FL, 0.6 g/L for 3-FL, 0.3 g/L for 3’-SL, 0.4 g/L for 6’-SL, and 0.9 g/L for LNT. However, the levels of each vary depending on gestational age, lactation stage, the mother’s health, genetics, and environmental factors^152,153. Since 2’-FL is the most abundant HMO, found in levels up to 10 g/L of human milk, it is the most studied to date.

HMOs may serve as a prebiotic substrate for commensal bacteria, where they may shape the developing microbiome and innate immune system in the infant’s gut^154,155. Positive HMO effects include microbiota and immune modulation, improved brain development, and modulation of intestinal epithelial cell response^156,157. For example, HMOs have a beneficial effect during inflammation as they can directly bind to selectin receptors, thereby inhibiting rolling and adhesion of leukocytes^155,158. With HMOs serving as prebiotics and modulators of the intestinal immune response, they may play a critical role in protection from late-onset sepsis and necrotizing enterocolitis by guiding the immune system and microbiota in the right direction^159,160. In support of this possibility, 2’-FL was shown to reduce the inflammatory response to certain bacteria, while formula supplementation with 2’-FL reduced proinflammatory cytokines in infants^161,162. This protective effect could explain the finding that infants fed formula containing 2’-FL and LNT had lower levels of bronchitis and required fewer antibiotics¹⁶³. HMOs have also shown great efficacy in preventing NEC in animal models of the disease. For example, supplementing formula with pooled HMOs at 10 mg/mL in neonatal rats was shown to lead to improved survival (73% to 95%) and reduced histopathological NEC scores (1.98 to 0.45), while 2’-FL and 6’-SL prevent NEC in mice and piglet models by inhibiting TLR4 signaling^9,164. In humans, cohort studies have correlated HMO concentration in maternal milk with NEC-related parameters such as changes in microbiota¹⁶⁵. One multicenter prospective cohort by Autran et al. found that the level of disialyllacto-N-tetraose (DSLNT) was significantly lower in almost all breast milk samples in NEC cases compared with controls¹⁶⁶. Moreover, researchers found that the concentration of DSLNT could be used to identify NEC cases before their onset¹⁶⁷.

Apart from the effects of HMOs on the microbiota, several studies could demonstrate a modification of the extracellular glycosylation of epithelial cells, limiting the adhesion of pathogens (bacteria and viruses)¹⁵⁵. Furthermore, expression of mucin, which is the main protein that composes intestinal mucus, has been shown to be increased following exposure to HMOs leading to a decrease in intestinal permeability and strengthening of TJ^155,168,169. In addition, HMOs can modulate cell cycle preventing cell growth and inducing cell differentiation in villus enterocyte-like cells, which results in a better absorption of nutrients^{155,170–172}.

Probiotics

The efficacy following the administration of probiotics to neonates at risk of NEC and neonatal sepsis has been demonstrated in several large trials but still provides a large area of research. For example, a meta-analysis that examined 45 trials with 12, 320 participants determined that the combined supplementation of Bifidobacterium plus Lactobacillus was associated with lower rates of mortality (risk ratio 0.56) and NEC morbidity (risk ratio 0.47) compared to placebo. Interestingly, Bifidobacterium plus prebiotic had the highest probability of having the lowest rate of mortality, and Lactobacillus plus prebiotic had the highest probability of having the lowest rate of NEC¹⁷³. Similar results have been reported in a large systematic review and network meta-analysis of randomized trials. The analysis of 56 trials including 12, 738 preterm infants has revealed that with high- or moderate-certainty evidence the combination of Lactobacillus spp and Bifidobacterium spp, Bifidobacterium animalis subspecies lactis, Lactobacillus reuteri, and Lactobacillus rhamnosus significantly reduced severe NEC (stage II or higher)¹⁷⁴. Sharif et al. have recently published a review focusing on very preterm (born more than 8 weeks’ early) and very low birth weight (less than 1.5 kg) infants (56 trials and 10, 812 infants) and suggested that probiotics (Bifidobacterium spp., Lactobacillus spp., Saccharomyces spp., and Streptococcus spp. alone or in combination) might reduce the risk of NEC. However, due to lack of clarity on methods to conceal allocation and mask caregivers or investigators as well as variation of formulation of probiotic findings are stated with a low to moderate level of certainty, emphasizing the need for further, large- high-quality trials to establish practice policies¹⁷⁵. Another Cochrane review also compared the efficacy and safety of prophylactic enteral probiotics (either Lactobacillus alone or in combination with Bifidobacterium) in preterm infants and found a reduction in the incidence of severe NEC (stage II or more) and mortality but no evidence of significant reduction in nosocomial sepsis¹⁷⁶. Good et al. have shown that the administration of probiotics leads to activation of TLR9 by bacterial DNA, which is an inhibitor of TLR4, and subsequent reduction in pro-inflammatory signaling in cultured enterocytes and samples of resected human ileum, suggesting a therapeutic mechanism in clinical NEC^177,178. Regarding immune response including lymphocyte dynamics in the pathogenesis of NEC, Liu et al. have demonstrated that the reduction in FOXP3⁺Treg cells in the intestine was reversed following the administration of the probiotic Lactobacillus reuteri in mice¹⁷⁹. In addition, Guo et al. have also investigated the protective effects of probiotics on TJ and intestinal permeability. Bifidobacterium infantis and Lactobacillus acidophilus could protect the intestinal barrier against IL-1β-induced NF-κB activation, which might explain a possible effect on preservation of gut permeability^180,181.

Synthetic amniotic fluid and exosomes

Amniotic fluid and breast milk both contain similar growth factors, including epidermal growth factor, and in-utero consumption of amniotic fluid contributes to the development and maturation of the fetal gut, responsibilities that are replaced by breast milk postnatally¹⁸². Good et al. injected amniotic fluid into the fetal gastrointestinal tract of mice and found an inhibition of TLR4 on the intestinal epithelium via EGF receptors. Those results were confirmed in piglets and humans with NEC¹³³. In recent years, there has been significant interest in lipid bilayer-enclosed milk extracellular vesicles, especially the subfraction of milk exosomes that only contain proteins and lipids, but also mRNAs, microRNAs, circular RNAs, and long non-coding RNAs¹⁸³. A variety of different studies have shown the therapeutic effects of exosomes. For example, the administration of human milk exosomes 6 h prior to induction of NEC revealed a less severe intestinal tissue injury compared to controls, lower levels of pro-inflammatory cytokines and higher levels of epithelial TJ proteins ZO − 1, claudin, and occludin¹⁸⁴.

Antibiotic administration

In general antibiotics, including ampicillin, gentamicin, vancomycin, and metronidazole, are administered intravenously to prevent sepsis in affected infants^185,186. Bury and Tudehope performed a systematic review evaluating the benefits and harms of enteral antibiotic prophylaxis for NEC in low birth weight and preterm infants. They found evidence that enteral antibiotics could have a prophylactic effect on the development of NEC but also raised concerns about potential harm, particularly related to the formation of resistant bacteria¹⁸⁷. Interestingly, in pigs, the beneficial effects against NEC were limited to oral and non-parenteral administration of antibiotics^188–190. In contrast, more recent studies investigating broad-spectrum antibiotics given intravenously have suggested that exposure of less than 3 to 5 days decreases subsequent risk for NEC^{37,188,191,192}. A Cochrane review from 2012 compared the efficacy of different antibiotic regiments on mortality and the need for surgery in neonates with NEC but could not identify a superior antibiotic class to be more beneficial in the treatment of NEC¹⁹³. Although human and animal studies seem to suggest that treatment with antibiotics may alter the future risk of NEC, the limitations of those studies often need to be carefully considered for interpretation¹⁸⁸. In particular, the administration of antibiotics still bears the risk of developing antibiotic resistance, growth of other pathogenic bacteria and even altering the intestinal microbiome, thereby possibly increasing the risk of NEC development^194,195. Nevertheless, treatment of medical NEC also includes gut rest, intravenous nutrition, and assessing bowel function¹⁹⁵. Further studies in humans and animals are necessary for a better understanding the effects of prophylactic administration of antibiotics on later development of NEC, specifically focusing on changes in intestinal immunity and in the gut microbiome¹⁸⁸.

Outcomes and complications of NEC

Infants with NEC suffer significant complications, involving not just the intestine, but also the lungs and the brain^14,196,197. The mortality rates of NEC range from 15% to 45%^11–14, with a higher rate of death associated with lower birthweights and NEC totalis (defined as necrosis of the entire small intestine). There is a consistently higher mortality rate of NEC in infants of African American ethnicity, which requires further evaluation^198,199. Recurrence of NEC occurs in 4–10% of post-NEC patients^200–202. A recent meta-analysis showed that earlier re-initiation (<5–7 or median 4 days) of enteral feeding after diagnosis of NEC resulted in a lower risk for the combined outcome of recurrent NEC and/or post-NEC stricture²⁰³. In addition, several studies have focused on the antibiotic management of NEC^204–206. Scheifele et al. evaluated the complication rate in two different regimens (ampicillin + gentamicin vs cefotaxime + vancomycin) to treat NEC²⁰⁷. The recurrence rate was lower (10.9% vs 0%) in the latter group but did not show statistical significance.

Post-NEC intestinal stricture occurs in 9–36%^208,209, 80% of which occur in the left colon, presumably due to the relatively low vascular supply in the region of splenic flexure. Dukleska et al. reported the incidence of stricture to be more common in the non-surgical cases²⁰¹, whereas Heida et al. reported a higher incidence of post-NEC stricture in the surgical cases²⁰⁹. They also reported that higher CRP levels during the initial NEC were associated with developing post-NEC intestinal strictures.

Short bowel syndrome (SBS), defined as inadequate intestinal length to support independent enteral nutrition, is an important consequence of NEC and is associated with high morbidity and mortality. Amin et al. have reported that a threefold increase in mortality was seen in SBS infants in NICU compared to the control group without SBS²¹⁰. Sparks et al. reviewed 109 SBS patients to evaluate the achievement rate of enteral autonomy²¹¹. While 64.9% of post-NEC patients with SBS achieved enteral autonomy, only 29% of patients with a different primary diagnosis fully weaned from parenteral nutrition, with a mean PN duration of 15.3 months. The diagnosis of NEC was an independent predictor of achieving enteral autonomy, which indicates NEC-related SBS has a higher potential to achieve oral nutrition even after a long duration of parenteral supports.

Several studies have shown an important association between NEC and neurodevelopmental impairment^212–214. Especially in very low birth weight infants with NEC, physical and mental developments were significantly delayed compared to the age-matched controls without NEC²¹⁴. Systematic reviews show that survivors of stage II NEC or higher have and increased risk for long-term neurological impairment^15,215. Several authors have found that NEC infants are at higher risk of cerebral palsy, visual, cognitive, and psychomotor impairment^{15,213,215–217}. Zozaya et al. described a higher risk of neurodevelopmental impairment in SIP, NEC, or any other bowel perforation, compared to non-perforated intestinal disease, which supports the idea of surgical NEC as a risk factor for neurological impairment²¹⁸. A recent study showed a widened intra-parenchymal space with decreased myelination in the cranial MRI in the NEC group compared to the non-NEC group. In addition, the difference in the early EEG was seen between NEC and non-NEC patients, which can be a predictive tool in the early stage of post-NEC to evaluate for possible neurodevelopmental delay (220). Taken in aggregate, we support early surgical intervention to remove the necrotic bowel in order to prevent the subsequent development of brain injury, although it remains to be definitively proven as to whether such an approach will reduce long-term NEC-induced brain injury.

Summary

NEC remains a devastating disease of the premature infants that continues to cause death and disability in many premature infants. Recent discoveries have provided new insights into the pathogenesis of the development of NEC. A dynamic interface between bacteria and the immature immune system seems to be the main mechanism leading to severe acute effects at the intestinal epithelium, with secondary effects in distant systems including lungs and brain.

Sepsis leads to further derangements of the immature neonatal immune system, and suspicion in a neonate often leads to the initiation of empiric antimicrobial therapy. A central key regulator following bacterial colonization and subsequent dysbiosis includes TLR4, which is expressed at higher levels in the premature gut than in the full-term gut. Apart from clinical interventions, including the administration of antibiotics, intravenous fluids, and surgical procedures, a variety of investigators have discovered various different therapeutic approaches to prevent or treat the consequences of sepsis and NEC, with administration of breast milk and amniotic fluid leading the way. Further studies are necessary to continue investigating new potential therapeutic interventions to advance our ability to help preterm infants with NEC, thereby reducing the morbidity of this complex condition seen in this vulnerable population.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data of this study are openly available at http://dx.doi.org/10.1080/19490976.2023.2221470.

Additional information

Funding

DJH is supported by R35GM141956 from the National Institutes of Health; CML and DS are supported by T32DK007713.