Are EGF and TLR-4 crucial to understanding the link between milk and NEC?

No milk today, NEC would not go away (inspired by a Herman's Hesmits song)

Premature neonates are at risk for compromised postnatal outcomes such as poor nutritional uptake, failure to thrive, and subsequent postnatal growth restriction, which are all conditions with potential irreversible, lifelong consequences.^{1, 2} Moreover, prematurity is the most important independent risk factor for necrotizing enterocolitis (NEC), a severe gastrointestinal disease with a mortality of 20–40%, affecting 10% of very low birth weight infants.³ NEC is a multifactorial disease and predisposing or underlying disorders might have a prenatal transitional and postnatal origin.^{3, 4} A derailed inflammatory response against commensal bacteria by the immature intestine following mucosal injury is the leading hypothesis in the pathophysiology of NEC.^{3, 5} Since human breast milk protects infants from NEC,^{6, 7} a variety of studies focussed on unraveling the protective component(s) of breast milk and their mode of action.

Breast milk contains numerous components that have immunomodulatory and gut-maturating properties, including human milk oligosaccharides, caseins, and whey proteins, such as growth factors (insulin-like growth factor, heparin-binding epidermal growth factor-like growth factor (HB-EGF), transforming growth factor-β 2, and epidermal growth factor (EGF)).⁸ The first evidence for a protective role of EGF against NEC was suggested from a case, reporting the recovery of necrotic bowel disease in a critically ill 8-month-old girl after the infusion of EGF systemically.⁹ This was followed by the observation that infants with NEC had significantly diminished levels of salivary and serum EGF compared with postnatal age- and birth weight-matched babies.¹⁰ Furthermore, EGF is highest in early breast milk with ∼100 ng ml⁻¹ and decreases over lactation.^{11, 12} Interestingly, EGF levels in breast milk of mothers with extremely preterm babies are much higher than in term infants.¹¹ Numerous experimental studies in rat and mice NEC models confirmed the protective effect of EGF,^{13, 14} supporting the hypothesis that EGF in breast milk may be, at least in part, responsible for the protection against NEC. However, the exact mechanism for its beneficial effects needed to be clarified.

This problem has now been experimentally addressed in a rodent NEC model and in vitro by Good et al. in today’s issue.¹⁵ First, both in vitro (in IEC-6 enterocytes, expressing both Toll-like receptor-4 (TLR-4) and epidermal growth factor receptor (EGFR)) and in vivo (wild-type and transgenic neonatal mice) experiments demonstrated that breast milk inhibited TLR-4 signaling via EGF-mediated EGFR activation. Next, using IEC-6 cells, it was shown that the downstream target GSK3β, a critical component of the Wnt/β–catenin signaling pathway, mediates this EGFR-dependent TLR-4-driven immune activation, thereby improving enterocyte proliferation and inhibition of enterocyte apoptosis. Finally, newborn mice receiving breast milk were protected to develop experimental NEC via the inhibition of TLR-4, requiring EGF. For these experiments, they both used mice lacking intestinal EGFR, EGF-depleted milk and supplementation of recombinant EGF in milk. These observations provide a novel mechanistic link between EGF from breast milk and various innate immune receptors.

However, it is not clear whether EGF is the sole compound from breast milk that is of importance in the protection for NEC. Besner et al.^{16, 17} have shown that HB-EGF is protective for experimental NEC in a newborn rat model, in vitro and ex vivo organoid cultures. Reduction of NEC symptoms was dependent upon EGFR activation and mediated via the MEK 1/2 and PI3K signaling pathways, protecting intestinal stem cells from injury. Furthermore, HB-EGF was shown to preserve intestinal microvascular blood flow in rat pups with experimental NEC¹⁸ along with increasing enterocyte migration and proliferation.¹⁹ The human milk oligosaccharides (HMO) were also found to improve the survival and reduce the severity in rat pups with experimental NEC. In particular, the HMO disiallacto-N-tetraose was identified to be beneficial.²⁰ Unfortunately, both EGF and HB-EGF are absent from all commercially available infant formulas. Consequently, no studies have been performed in babies using EGF or HB-EGF.

Furthermore, recent studies identified more pathways that are activated by EGF or via its receptor, including the involvement of EGFR in lipopolysaccharide-induced cyclooxygenase-2 expression in enterocytes.²¹ Among these pathways is autophagy, a homeostatic process including the sequestration of cytoplasmic components in autophagosomes, which was reported to be activated in babies with NEC and in a rat model of NEC, which appeared to be also EGF dependent.²² EGF also alters the balance between regulators of apoptosis thereby reducing intestinal epithelial cell apoptosis and maintaining epithelial homeostasis.¹³ Another important role of EGF is its direct effect on neonatal gut development, as was shown by the exogenous infusion of EGF in utero in rabbits, leading to stimulated intestinal growth and increased maturation of intestinal enzyme activity.²³ Using transgenic mice and cultured enteroids, the group of Hackam²⁴ recently demonstrated that TLR-4 activation induces endoplasmic reticulum-stress mediated apoptotic cell death of intestinal stem cells and subsequent NEC development. At present, it remains unclear whether the protective effects of EGF rely on improved epithelial integrity, on anti-inflammatory aspects, or the combination of both phenomena.

The results of Good et al. need to be considered in the context that the development of the murine immune system does not resemble the situation in humans. This might explain the observed differences in the expression of pathogen recognition receptors in mice/rats on one hand and preclinical animal models and humans on the other hand and direct the subsequent discussion whether NEC development results from decreased or exaggerated immune activation. From preclinical large models and data in humans, it appears that the development of innate immune components is gestation dependent. In particular, TLR-4 mRNA levels are lower in preterm than in term pigs and highest at a developmental stage when NEC sensitivity is minimal.²⁵ In sheep, we showed relatively low intestinal TLR-4 and MD-2 mRNA levels at early gestational ages, which were further downregulated during endotoxin-induced chorioamnionitis.²⁶ In humans, increased TLR-2 and TLR-4 mRNA levels were shown in the intestine of immature babies when compared with children of older age, whereas its negative regulators (e.g., A-20 and TOLLIP) were downregulated in immature fetal intestine and NEC patients when compared with tissue from children.²⁷ Premature infants and neonates with NEC did not express MD-2 in their intestinal epithelium, whereas MD-2 protein was expressed in the healthy term neonatal and adult intestine. Moreover, NEC patients also lacked MD-2 staining in the abundantly recruited inflammatory cells.²⁸ In line with these data, Jenke et al. showed that human β-defensin-2 protein and mRNA levels and MD-2 and TLR-4 mRNA were almost absent in intestinal specimens from infants with severe NEC.²⁹ Collectively, these data from large animal models and humans suggest that the immune system of the immature gut inadequately responses to microflora which colonize the gut directly after birth.

The next step is to translate the new mechanistic insights by which breast milk, a multipotent drug, mediates protection against the development of NEC into the clinic for babies that are fed with formula. This translational approach can be conducted in large animal models including pigs, sheep, and primates since organ development, the developmental biology, and the physiology of such large animals closely resembles the human situation.^{30, 31} In addition, studies in babies were till recent hampered by lack of (noninvasive) markers for intestinal inflammation, damage, and maturation. Importantly, we recently reported on the availability of such human gut-specific biomarkers, which enable us to measure gut maturation and to early diagnose systemic and intestinal inflammation in NEC patients, which may be of help for the implementation of these findings into clinical practice.^{32, 33, 34, 35, 36}