Abdelhamid, Adel E.*; Chuang, Shu-Ling†; Hayes, Peter‡; Fell, John M.E.*
See “Cow's-Milk Protein as a Specific Immunological Trigger of Necrotising Enterocolitis—or Food Protein–induced Enterocolitis Syndrome in Disguise?” by Murch on page 3.
Necrotising enterocolitis (NEC) is a devastating neonatal intestinal inflammatory condition, affecting predominantly preterm infants, around 5% to 10% very-low-birth-weight infants (<1500 g) (1–5). It remains a major cause of neonatal morbidity and mortality, with the increasing survival of the extremely premature infants (2). Despite decades of work, the aetiology and pathogenetic mechanisms of neonatal NEC remain elusive with many major questions still unanswered. The bulk of the present understanding is derived from epidemiological studies or animal experimental models with limited research in affected human infants. The epidemiology has highlighted the main risk factor to be prematurity. Other major postulated risk factors are enteral feeding, in particular, formula milk feeding, intestinal hypoxia/ischaemia, and bacterial translocation (4–6).
The present consensus is that NEC represents an exaggerated, inappropriate, and dysregulated inflammatory response in the immature developing gut and immune system to an otherwise “normal” luminal environment and innocuous antigens (7), following the initial trigger. Once initiated, a cascade of immunological events ensues, with the release of a multitude of pro- and anti-inflammatory mediators (8–10), leading to an alteration in the luminal environment, uncontrolled inflammation, progressive mucosal disruption, and damage (11).
Enteral feeding and the type of milk feeds appear to play a role in the pathogenesis of NEC. NEC occurs mostly in infants who have received enteral feeds (4). Epidemiological studies have indicated an increased risk of NEC in infants fed with cow's milk formula compared with those receiving breast milk (maternal or donor) (12,13). The association of NEC with the nature of nutrition has been investigated primarily in terms of the protective effect of breast milk and influence on gut bacteria flora in relation to feed types (6). The role of formula milk in terms of its direct harmful effects in the pathogenesis of human neonatal NEC remains to be evaluated, although animal models of intraluminal formula milk/casein-induced NEC-like injury were well described (13).
Cow's milk allergy is well recognised as a significant cause of morbidity in formula-fed term infants (14) and more recently, in preterm infants (15). The sequelae and enteral feed management in patients following NEC present further clinical challenge. There is a need for greater understanding of the role of dietary antigens in the recovery following profound mucosal injury.
We first reported evidence of cow's milk proteins (CMPs) antigen sensitisation in human neonatal NEC with demonstration of significant peripheral blood mononuclear TH1 (interferon-γ [IFN-γ])/TH2 (interleukin [IL-4], IL-5) cytokine responses to 2 major CMPs (β-lactoglobulin [β-lg], casein) in preterm babies with NEC during the acute phase of disease (16). We replicated and extended cytokine characterisation, in a second cohort of preterm infants with NEC, demonstrating a significant fall in IFN-γ, IL-4, and IL-10 but increased transforming growth factor-β1 (TGF-β1) spontaneous and dietary antigen responses at term.
In the present article, we report, in this second cohort of infants with NEC (17), the temporal changes in cytokine secretion, and dietary antigen sensitisation at different time-points during the natural history of NEC disease process, from the acute stage, through recovery, and until full enteral feeding is established. Cytokine secretion was again assessed at a single-cell level using the sensitive enzyme-linked immunospot (ELISPOT) methodology as before (16) on effector cytokines, IFN-γ (TH1) and IL-4 (TH2), and regulatory cytokines IL-10 and TGF-β1.
A cohort of 14 (10 boys and 4 girls) preterm babies admitted to the neonatal intensive care unit at Chelsea and Westminster Hospital, London, for 18 months who were diagnosed as having NEC stage II or III on the modified Bell staging (18) were recruited. Babies with lethal congenital malformations were excluded from the study. This cohort is the same group as we have previously reported with regards to their cytokine responses during the acute phase of NEC as compared with age- and gestation-matched healthy preterm controls (17) and at term. Following recruitment, babies were followed up at recovery (initiation of enteral feeding), and when full feeding was achieved. Two babies died during the acute stage of their disease from multiorgan failure; thus, longitudinal data until full feeding were obtained from 12 patients (8 boys and 4 girls).
Protocol for the Management of NEC
Infants were managed for NEC according to the unit protocol. At diagnosis, feeding was discontinued and parenteral nutrition initiated. Infants received intravenous antibiotics (vancomycin and piperacillin/tazobactam). Enteral feeding was reinitiated when ileus had resolved (clinically satisfactory abdominal examination, absent or minimal gastric aspirate that is not abnormal [bilious or bloody]), resolution of radiological signs suggestive of NEC, and normal or near-normal inflammatory markers. Enteral feeding was commenced using expressed breast milk if available. Otherwise, special formulas (Neocate, Pregestimil), or cow's milk–based formula was used according to physicians’ choice, and progressively increased depending on the patient tolerance.
Recovery and Full Feeding
Recovery from the acute-stage NEC was defined by feeding reinitiation implying improvement of clinical, radiological, and laboratory parameters. Full feeding was defined as achieving an enteral feed volume of 150 mL · kg−1 · day−1.
Data were collected for the study cohort on the demographic criteria (birthweight in grams, postconceptional age [PCA] in weeks, postnatal age in days), and presampling feeding history (type of milk, advancement rate) at birth, at presentation, at recovery, and at full feeding, as well as the clinical data at the 3 stages.
Subjects were stratified according to the type of milk they were fed with nil by mouth (NBM); babies who are exclusively breast-(maternal or donor)fed (exclusively breast milk); babies who are predominantly breast-fed: if breast (maternal or donor) milk is >50% of the feed volume (predominantly breast milk); babies who are predominantly formula fed: if the formula constitutes >50% of the feed volume (predominantly formula milk); babies who are exclusively formula fed (exclusively formula milk), and babies fed on any special formula: elemental, lactose-free, or amino acid formula (eg, Pregestimil or Neocate).
Method for Assessment of Cytokine Secretion
ELISPOT measurements of cytokine secretion from peripheral blood mononuclear cells (PBMCs) undertaken at diagnosis (“acute phase”), “recovery,” and “full feeding” as previously reported (16,17) are described in brief.
Isolation of PBMCs
Approximately 0.5 mL of heparinised blood was collected from each infant at each of the 3 sampling points. All of the infants had all tests performed. PBMCs were isolated by density gradient centrifugation using standard procedures. The percentage of viable cells was determined by trypan blue exclusion, and found in all of the cases at all samples to exceed 95%. Average PBMC recovered was 3 to 10 × 106 cells/mL.
Plate Coating With Capture Antibodies
A nitrocellulose-bottomed microtitre plate (MAIPS 4510; Millipore, Watford, UK) was prewetted with 20 μL/well of 70% ethyl alcohol. Wells were washed with phosphate buffered saline (PBS) and incubated with cytokine-specific monoclonal capture antibody at least overnight. Just before use, unadsorbed antibody was decanted, and wells were washed with PBS and blocked with R10.
Incubation of PBMCs
Blocking medium was decanted. PBMCs were assayed in triplicate wells where cell yield permits, containing 0.5 × 105 cells per well. Cells were incubated for 20 to 24 hours in 5% CO2 humidified atmosphere at 37°C in the absence or presence of respective antigens or mitogens: β-lg (500 μg/mL), casein (500 μg/mL), which is a mixture of different casein fractions, (α, β, κ caseins), and phytohaemagglutinin (PHA) (10 μg/mL). All of the infants had all tests performed for the 4 cytokines. Stimulation with keyhole limpet haemocyanin, an irrelevant antigen not encountered by the infants, signifying a negative control to dietary antigen stimulation, induced negligible responses in patients with NEC at presentation (17) and was therefore not repeated in convalescent samples.
Detection of Cytokine-secreting Cells
Wells were washed with PBS/0.05% Tween 20 before filtered biotinylated cytokine-specific monoclonal antibody (MAb) was added and incubated for 4 hours at room temperature. Wells were washed with PBS/Tween and avidin-biotin peroxidase-complex (Vector Laboratories, Burlingame, CA) was added for 1 to 2 hours at room temperature. The wells were washed with PBS/Tween followed by PBS, and finally aminoethylcarbazole substrate (Sigma-Aldrich, Gillingham, UK) was added per well. Spots were developed at room temperature for 4 minutes before reaction terminated under running water and plate dried overnight. Spots were enumerated with an automated ELISPOT plate reader (Carl Zeiss, Welwyn Garden City, UK).
Capture and detection MAbs to IFN-γ and IL-4 were obtained from MAbtech AB, Nacka Strand, Sweden, and those to IL-10 from BD Biosciences, San Diego, CA, and TGF-β1 capture Ab: recombinant Human TGF-β sRII/Fc Chimera and biotinylated anti-TGF-β1 antibody from R&D Systems, Minneapolis, MN.
Cytokine secretion assessed by ELISPOT was compared between the different time points: acute stage, at recovery, and then at full feeding for the 4 cytokines (IFN-γ, IL-4, IL-10, and TGF-β1). The results are expressed as the number of spot-forming cells (SFCs)/105 mononuclear cells. Where in vitro stimulation was performed, the response to stimulation is presented as ΔELISPOTs. This represents the number of SFCs postantigen/mitogen stimulation minus the number of spontaneous (unstimulated) SFCs. The frequencies of cytokine-secreting cells at each stage were compared with those at the 2 other stages using Wilcoxon matched pairs test (data distribution being nonparametric). Results are expressed as median and range or interquartile range. Observations are presented in unadjusted P values in this explorative work.
A total of 14 infants were initially recruited. Subsequently, 2 died during the acute stage for multiple organ system failure; thus, follow-up data were available from 12 patients with a median PCA at birth of 27, range 24 to 35 weeks and median birthweight of 845 g (range 560–2400). The PCA of these patients when NEC diagnosis was made ranged between 29 and 36 (median 32.5) weeks. According to modified Bell classification (18), 2 patients were stage IIA, 5 were IIB, and 5 were IIIB.
C-reactive protein was increased in all of the patients at diagnosis, median: 182 mg/L (interquartile range 118–241), 9 (2–134) mg/L at recovery, and 2 (2–70) mg/L at full feeding.
Acute Surgical Treatment
Surgical intervention was needed in 8 (66.7%) patients, with stoma formation in 5 patients (jejunostomy in 1, colostomy in 1, and ileostomy in 3). The gut resection was minimal in 5 patients, 18 cm in 1, and 45 cm in 1. Extensive small bowel and significant large bowel resection was performed in 1 patient. The ileocaecal valve was lost in 2, but preserved in 6 patients.
Feeding History and Progress
All of the preterms with NEC had been fed cow's milk formula before diagnosis. At recovery feeding was reinitiated on breast milk in 7 patients, exclusively in 5, and predominantly in 2. Exclusive formula milk was used for feeding reinitiation in 2 patients, and breast milk was gradually introduced in one of them to make up half of the feeds by the time full feeding was achieved. Feeding was reinitiated on a special feeding formula (hydrolysed casein or amino acid formula) in 3 patients and continued until full feeding.
Following diagnosis, enteral feeds were withheld for 7 to 16 (median 11.5) days until recovery, when feeding was reinitiated. Blood samples for ELISPOT assay were collected just before the first feed was introduced. Median PCA of second sampling point was 34 (range 32–38) weeks.
Overall, full enteral feeding (full feeding) was established between 6 and 72 (median 11) days postfeed reintroduction, at PCA ranging from 33 to 48 (median 35.5) weeks. Two infants, who had significant small bowel resections, were expectedly much slower in establishing full enteral feeding (at 40 and 48 weeks PCA, respectively). The remaining 10 patients achieved full feeding at median 35 (range 33–37) weeks PCA.
PBMC ELISPOT Responses at Diagnosis Compared With Recovery and Full Feeding
ELISPOTs were performed for the 4 cytokines: IFN-γ, IL-4, IL-10, and TGF-β1 on PBMCs. In the present study population, we have previously compared acute-stage cytokine secretion with that at term (17), whereas in the present study, we measured the changes in cytokine secretion over the time course of the NEC disease process from the acute stage to recovery and then to full feeding.
Spontaneous and Mitogen (PHA)-elicited Responses
We have previously shown that the spontaneous secretion of cytokines IFN-γ, IL-4, IL-10, and TGF-β1 from PBMCs is significantly greater in this cohort of infants with NEC during the acute phase when compared with age- and gestation-matched septic and healthy controls, an effect that was further augmented following stimulation with PHA (17). Now looking at recovery, the spontaneous secretion of all 4 cytokines is further increased (by a factor of 2) (P < 0.005). At full feeding, secretion returns to, or below, the acute-phase level for IFN-γ, IL-4, and IL-10, whereas for TGF-β1, secretion increases further (P < 0.005) (Table 1). A similar pattern of cytokine secretion, increasing at initial recovery, and falling back at full recovery (full feeding) for IFN-γ, IL-4, and IL-10, yet rising further at full recovery for TGF-β1, is also observed following stimulation with PHA (Table 2).
Antigen-specific Response to β-lg and Casein
β-lg stimulation elicited significantly enhanced responses in the patients with NEC in the acute phase for the 4 cytokines: IFN-γ, IL-4, IL-10, and TGF-β1. These responses are further enhanced (on the order of 50%) at recovery (P < 0.005), but at full feeding, there is a decline in the IFN-γ, IL-4, and IL-10 response, to a level below that was recorded at the acute phase, whereas TGF-β1 secretion in response to β-lg increases at recovery (P < 0.005), and increases further at full feeding (P < 0.005) (Fig. 1).
This pattern of response is also seen with casein. Although in general the numbers of cytokine-secreting cells are fewer than for β-lg, once again the increase from acute presentation to recovery is observed for the 4 cytokines (P < 0.005) with a decline when fully recovered (fully fed) to a level at or below the acute-phase values for IFN-γ, IL-4, and IL-10. TGF-β1 secretion, however, increases further form recovery to full feeding (Fig. 2).
Looking at the cytokine responses when fully fed, the nature of the feed does not seem to influence the pattern of cytokine secretion observed. Infants fed with breast milk do appear to have greater secretion of IFN-γ, IL-4, and IL-10, but these infants also tended to have greater cytokine secretion before they were fed (at the recovery stage) (Figs. 1 and 2).
The present study documents the temporal changes in PBMC cytokine secretion, to mitogen and CMP stimulation in a cohort of preterm infants following acute NEC presentation, through initial recovery before enteral feeds recommencement and at full recovery with full enteral tolerance. We observed further increment in spontaneous IFN-γ, IL-4, IL-10, and TGF-β1–secreting cell frequency at “initial recovery,” significantly amplified by mitogen (PHA) and CMP antigen (β-lg and casein) stimulation. At full recovery, although IFN-γ, IL-4, IL-10 cytokine responses returned to or below acute-phase levels, TGF-β1 remains elevated and enhanced by CMP stimulation.
The high level of IFN-γ and IL-4 secretion during the acute stage of NEC representing a mixed TH1/TH2 cytokine activation profile in NEC and sensitisation to CMPs has been previously reported by our group (16,17) along with concomitant rise in regulatory cytokines (IL-10 > TGF-β1) (17). It is, however, interesting and intuitively unexpected to find that the cytokine responses and milk antigen reactivity should continue to “peak” at recovery, 7 to 16 (median 11.5) days after the initial presentation despite apparent clinical improvement. This is perhaps not dissimilar to postburns hyperinflammatory responses, which persist for weeks longer than expected (19). It has been shown in experimental models that induction of barrier loss activates immunoregulatory processes and consequent increase in IFN-γ and IL-10–secreting lamina propria mononuclear cells and CD4+CD25+ T cells (20). None of the infants had been fed between the acute phase and recovery; thus, no further dietary antigen challenge would have taken place. It is plausible that this may be attributed to a delayed rise in immunomodulatory TGF-β1 responses relative to the other proinflammatory cytokines in the acute inflammation cascade.
Somewhat speculative, but some of the changes may be due to changes in the amount of stimulating antigen (cow's milk proteins) at different time points. High-dose dietary antigen exposure following the massive mucosal disruption may directly affect immune tolerance/suppression pathways and subsequent systemic and local responses. The initial antigenic overload may result in partial tolerance due to antigen excess; the effect is temporarily lost when oral feedings are stopped, which may result in the observed increase of cytokine secretion. It is also possible that following the initial dietary antigen exposure, the sensitised effector cells remain primed, mounting an exaggerated overcompensatory response upon dietary antigen rechallenge in vitro, in the absence of orchestrated counterregulatory mechanisms (21).
The subsequent shift toward greater TGF-β1 secretion results in an alteration in the balance of proinflammatory and regulatory cytokine responses. This coincides with the improvement in clinical status, reduction in mucosal inflammation, and establishment of oral tolerance. The dramatic fall in effector proinflammatory cytokines against a background of increased TGF-β1–secreting cells with the termination of mucosal inflammation may reflect the complex interactions governing balance between pro- and anti-inflammatory responses within the immune regulatory network. There is considerable evidence that TH1 responses suppress both the expansion of TGF-β1–secreting cells and TGF-β1 signalling and, contrariwise, TGF-β1 interferes with IL-12 signalling. Hence, TH1 and TGF-β1 responses have a reciprocal feedback relation and appear to be mutually exclusive (22,23); however, there is some evidence that TH2 and TGF-β1 responses can coexist and that it requires higher levels of TGF-β1 to suppress a TH2 response than a TH1 response (23).
Both IL-10 and TGF-β1 play a role in the amelioration of inflammation (24). In the present study, a differential time-course of IL-10 and TGF-β1 cytokine activation over the NEC disease progression has been clearly demonstrated. IL-10 responses appear to be mounted early in the inflammatory process in parallel with the proinflammatory cytokines (IFN-γ, IL-4), whereas the TGF-β1 responses lagged behind somewhat. These findings may be relevant in respect to the differential roles and mechanisms of actions of these 2 regulatory cytokines in accordance with the large body of literature already established (22,23).
IL-10 is known to be an important negative regulator with main function in limiting cell inflammatory reaction. Mechanisms elucidated include inhibition of cytokine and chemokine production, induction of antagonists to key mediators of gut inflammation such as IL-β1 and TNF-α, and promoting tolerance to bacteria flora via actions on the innate immune cells. Intuitively, this is also relevant in NEC, where bacterial translocation, in the presence of mucosal injury or altered mucosal permeability, may trigger or exacerbate the local and systemic inflammation. Thus, the early induction of IL-10 response counterbalances the proinflammatory cytokines and permits the abrogation of established enterocolitis/inflammation and, in particular, inflammation induced by strong microbial stimulus, thereby avoiding extensive tissue damage (23). It is perhaps not surprising that declining IL-10 level is observed between the recovery and full feeding phase, once the acute inflammation has subsided.
TGF-β1 is a pleiotropic cytokine with diverse immunmodulatory functions, including suppression of proinflammatory immune responses and direct inhibition of effector cells. It plays a vital role in the promotion of gut homeostasis, induction of IgA expression, and mitigation of mucosal inflammation via various complex mechanisms such as negative regulation of the NF-κB activation (23). Its crucial role in the dampening of tissue-damaging immune responses has been investigated in a number of human and experimental intestinal inflammatory models (25,26). Deficiency of this potent immunosuppressive molecule affects adaptive and innate immunity, leading to fatal uncontrolled inflammation and multiorgan failure with endotoxin hypersensitivity as shown in TGF-β1 knockout mice (27). Furthermore, TGF-β1 also plays a major role in the restoration of barrier function and reconstitution of intestinal mucosa (21).
The patterns of activation following PHA and CMP stimulation were similar, although the magnitude of PHA responses was significantly greater than β-lg stimulation. This is not surprising because lymphocytic response to mitogen do not require earlier sensitisation, in contrast to antigens. The subject numbers are too small to allow for true comparison between the different feed groups, although the pattern of CMP sensitisation appears to be similar, not influenced by the nature of the feeds used after NEC. It is difficult to quantify the dietary antigen load, potential passage of CMP via breast milk, and dietary antigen uptake in presence of altered gut permeability (21).
The increase in TGF-β1 secretion in response to CMPs at a time when infants have shown their tolerance of enteral feeds is noteworthy. This response was found both in infants who had been re-exposed to cow's milk formula or in those that had not. Thus, the increased response cannot be explained by re-exposure to milk antigens. We have also shown, by comparison of patients with NEC with healthy preterm counterparts at term, that the upregulation was not explained by developmental maturational changes (17). There is considerable evidence that TGF-β1 plays a critical role in the induction of oral tolerance. The abrogation of intestinal inflammation requires not only suppression of immunity but also minimisation of stimulatory burden fuelling the inflammation (23). In the present study, we observed that in vitro CMP stimulation during the acute phase led to an induction of proinflammatory cytokines. The commencement of enteral feeds at recovery to full enteral tolerance, however, is associated with a change in profile of antigen-specific response (CMP) to one dominated by TGF-β1. It is unclear whether this reflects part of a more general process coincident with clinical improvement, and mucosal recovery (epiphenomenon) or, speculatively, enteral feeding leads to an induction of regulatory responses and active downregulation of system inflammation and reactivity, promoting “reestablishment” of tolerance to food antigens (21).
It is important to remember that the results presented here are from blood rather than from the mucosal site; thus, direct extrapolation to mucosal inflammatory events need to be interpreted with caution. It is speculated that the protective benefits of breast milk against NEC may be attributed in part to TGF-β, in particular TGF-β2, which is absent in formula milk. Maheshwari et al (28) reported decreased TGF-β1 and -β2 expression in human NEC mucosal resection (27). The group further demonstrated in mouse pups that disruption of TGF-β signalling worsened NEC-like intestinal injury, whereas enteral supplementation with recombinant TGF (TGF-β2 isoform) seemed to confer protection with attenuation of severity (28). In the same vein, evidence from several animal experimental models supports a role of milk-derived TGF-β in the induction of oral tolerance to allergens and favourable early immune programming (29,30).
TGF-β enhances fibrogenesis and promotes repair and tissue regeneration as part of normal wound healing after tissue injury in various organ systems, including the gut mucosa; however, excessive or persistent expression, in particular TGF-β1, has been implicated in excessive fibrosis and scarring including intestinal strictures in Crohn disease (31). Clinical experience suggests that intestinal stricture and/or adhesions manifest at 4 to 6 weeks after the initial NEC disease, in the same time frame as full feeding is established. The relation of the compensatory “excessive” increase of TGF-β1 expression noted in our patients to such long-term complications as stricture formation in NEC or retinopathy of prematurity, a vasoproliferative disorder, is also worthy of further evaluation (32).
In conclusion, our data further our understanding of cytokine regulation in the process of NEC disease evolution. There is temporal differential expression of inflammatory and regulatory cytokines following NEC upon clinical recovery and re-establishment of oral tolerance and to in vitro CMP challenge. The clinical implications of the observed in vitro CMP sensitisation on the disease process or in the longer term are unclear and will require further studies.
1. Hsueh W, Caplan MS, Qu X-W, et al. Neonatal necrotizing enterocolitis: clinical considerations and pathogenetic concepts. Pediatr Dev Pathol 2002; 6:6–23.
2. Uauy RD, Fanaroff AA, Korones SB, et al. Necrotizing enterocolitis in very low birth weight infants: biodemographic and clinical correlates. J Pediatr 1991; 119:630–638.
3. Bisquera JA, Cooper TR, Berseth CL. Impact of necrotizing enterocolitis on length of stay and hospital charges in very low birth weight infants. Pediatrics 2002; 109:423–428.
4. Petrosyan M, Guner YS, Williams M, et al. Current concepts regarding the pathogenesis of necrotizing enterocolitis. Pediatr Surg Int 2009; 25:309–318.
5. Garg P. Necrotising enterocolitis: newer insights. J Clin and Diagnost Res 2007; 1:90–103.
6. Neu J, Walker WA. Necrotizing enterocolitis. N Engl J Med 2011; 364:255–264.
7. Simmonds A, LaGamma EF. Addressing the “new” NEC: Part I: rediscovering the basics. Indian J Pediatr 2006; 73:1011–1018.
8. Eelson MB, Bagwell CE, Rozycki HJ. Circulating pro- and counterinflammatory cytokine levels and severity in necrotizing enterocolitis. Paediatrics 1999; 103:766–771.
9. Harris MC, D’angio CT, Gallagher PR, et al. Cytokine elaboration in critically ill infants with bacterial sepsis, necrotizing enterocolitis, or sepsis syndrome: correlation with clinical parameters of inflammation and mortality. J Pediatr 2005; 147:462–468.
10. Markel TA, Crisostomo PR, Wairiuko GM, et al. Cytokines in necrotizing enterocolitis. Shock 2006; 25:329–337.
11. Martin CR, Walker WA. Intestinal immune defences and the inflammatory response in necrotising enterocolitis. Semin Fetal Neonatal Med 2006; 11:369–377.
12. Sullivan S, Schanler RJ, Kim JH, et al. An exclusively human milk-based diet is associated with a lower rate of necrotizing enterocolitis than a diet of human milk and bovine milk-based products. J Pediatr 2010; 156:562–583.
13. Di Lorenzo M, Bass J, Krantis A. An intraluminal model of necrotizing enterocolitis in the developing neonatal piglet. J Pediatr Surg 1995; 30:1138–1142.
14. Thorsdottir I, Thorisdottir AV. Whole cow's milk in early life. Nestle Nutr Workshop Ser Pediatr Program 2011; 67:29–40.
15. Dupont C, Soulaines P, Lapillonne A, et al. Atopy patch test for early diagnosis of cow's milk allergy in preterm infants. J Pediatr Gastroenterol Nutr 2010;50:463–4.
16. Chuang SL, Hayes PJ, Ogundipe E, et al. Cow's milk protein-specific T-helper type I/II cytokine responses in infants with necrotizing enterocolitis. Pediatr Allergy Immunol 2009; 20:45–52.
17. Abdelhamid AE, Chuang SL, Hayes P, et al. In vitro cow's milk protein-specific inflammatory and regulatory cytokine responses in preterm infants with necrotizing enterocolitis and sepsis. Pediatr Res 2011; 69:165–169.
18. Walsh MC, Kliegman RM. Necrotizing enterocolitis: treatment based on staging criteria. Pediatr Clin North Am 1986; 33:179–201.
19. Finnerty CC, Jeschke MG, Herndon DN, et al. Temporal cytokine profiles in severely burned patients: a comparison of adults and children. Mol Med 2008; 14:553–560.
20. Turner JR. Intestinal mucosal barrier function in health and disease. Nat Rev Immunol 2009; 9:799–809.
21. Chehade M, Mayer L. Oral tolerance and its relation to food hypersensitivities. J Allergy Clin Immunol 2005; 115:3–12.
22. Monteleone I, Vavassori P, Biancone L, et al. Immunoregulation in the gut: success and failures in human disease. Gut 2002;50 (suppl 3):III 60–4.
23. Izcue A, Coombes JL, Powrie F. Regulatory lymphocytes and intestinal inflammation. Annu Rev Immunol 2009; 2:313–338.
24. Mucida D, Park Y, Kim G, et al. Reciprocal Th17 and regulatory T cell. Science 2007; 317:256–260.
25. Serhan SN, Savill J. Resolution of inflammation: the beginning programs the end. Nat immunol 2005; 12:1191–1197.
26. Ohtsuka Y, Sanderson IR. Transforming growth factor-beta1: an important cytokine in the mucosal immune response. Curr Opin Gastroenterol 2000; 16:541–545.
27. McCartney-Francis N, Jin W, Wahl SM. Aberrant Toll receptor expression and endotoxin hypersensitivity in mice lacking a functional TGF-beta 1 signaling pathway. J Immunol 2004; 172:3814–3821.
28. Maheshwari A, Kelly DR, Nicola T, et al. TGF-β2 suppresses macrophage cytokine production and mucosal inflammatory responses in the developing intestine. Gastroenterology 2011; 140:242–253.
29. Oddy WH, Rosales F. A systematic review of the importance of milk TGF-β on immunological outcomes in the infant and young child. Pediatr Allergy Immunol 2010; 21:47–59.
30. Oddy WH, McMahon RJ. Milk-derived or recombinant transforming growth factor-beta has effects on immunological outcomes: a review of evidence from animal experimental studies. Clin Exp Allergy 2011; 41:783–793.
31. Cutroneo KR. TGF-beta-induced fibrosis and SMAD signaling: oligo decoys as natural therapeutics for inhibition of tissue fibrosis and scarring. Wound Repair Regen 2007;15(suppl 1):S54–60.
32. Sood BG, Madan A, Saha S, et al. Perinatal systemic inflammatory response syndrome and retinopathy of prematurity. Pediatr Res 2010; 67:394–400.