Amniotic fluid inhibits Toll-like receptor 4 signaling in the fetal and neonatal intestinal epithelium

Abstract

The fetal intestinal mucosa is characterized by elevated Toll-like receptor 4 (TLR4) expression, which can lead to the development of necrotizing enterocolitis (NEC)—a devastating inflammatory disease of the premature intestine—upon exposure to microbes. To define endogenous strategies that could reduce TLR4 signaling, we hypothesized that amniotic fluid can inhibit TLR4 signaling within the fetal intestine and attenuate experimental NEC, and we sought to determine the mechanisms involved. We show here that microinjection of amniotic fluid into the fetal (embryonic day 18.5) gastrointestinal tract reduced LPS-mediated signaling within the fetal intestinal mucosa. Amniotic fluid is abundant in EGF, which we show is required for its inhibitory effects on TLR4 signaling via peroxisome proliferator-activated receptor, because inhibition of EGF receptor (EGFR) with cetuximab or EGF-depleted amniotic fluid blocked the inhibitory effects of amniotic fluid on TLR4, whereas amniotic fluid did not prevent TLR4 signaling in EGFR- or peroxisome proliferator-activated receptor γ–deficient enterocytes or in mice deficient in intestinal epithelial EGFR, and purified EGF attenuated the exaggerated intestinal mucosal TLR4 signaling in wild-type mice. Moreover, amniotic fluid-mediated TLR4 inhibition reduced the severity of NEC in mice through EGFR activation. Strikingly, NEC development in both mice and humans was associated with reduced EGFR expression that was restored upon the administration of amniotic fluid in mice or recovery from NEC in humans, suggesting that a lack of amniotic fluid-mediated EGFR signaling could predispose to NEC. These findings may explain the unique susceptibility of premature infants to the development of NEC and offer therapeutic approaches to this devastating disease.

Necrotizing enterocolitis (NEC) is the leading cause of death from gastrointestinal disease in premature infants (1). Although the underlying etiology of NEC remains incompletely understood, recent studies have identified a critical role for the LPS receptor, Toll-like receptor 4 (TLR4) in its pathogenesis. TLR4 activation within the intestinal epithelium leads to increased mucosal injury through accelerated enterocyte apoptosis as well as reduced healing through impaired intestinal restitution and proliferation (2), and mice lacking TLR4 (2, 3) show reduced NEC severity as the result of reduced injury and enhanced healing (4). These findings suggest that NEC develops in part in response to exaggerated TLR4 signaling in the intestinal mucosa and, by extension, that strategies may exist within the intestine that can limit TLR4 signaling and the propensity for NEC development. Given that the premature intestine is bathed in amniotic fluid throughout its development and that an abrupt lack of exposure to amniotic fluid is a natural consequence of early delivery, we hypothesized that amniotic fluid may exert a restraining influence on TLR4 signaling and that the absence of the anti-TLR4 signaling effects of the amniotic fluid places the preterm infant at risk for NEC development. We further sought to identify the specific factor(s) within the amniotic fluid that could inhibit TLR4 within the intestinal epithelium of the newborn host.

Results

Amniotic Fluid Inhibits TLR4 Signaling in the Intestinal Epithelium in Utero.

We first used a backscatter ultrasound-guided microinjection approach to deliver LPS directly into the lumen of the fetal gastrointestinal tract (5) in the presence or absence of freshly harvested murine amniotic fluid. We used NF-kB–GFP reporter mice, which provide a highly sensitive readout of TLR4 activation (6). The delivery of LPS into the fetal gut resulted in a significant increase in the expression of GFP in the small intestinal mucosa 3 h later (Fig. 1 A, i, ii, vii, and viii) which correlated with a significant increase in the expression of the proinflammatory cytokines inducible NOS (iNOS) and IL-6 (Fig. 1B and Table S1), as compared with mice injected with either saline or amniotic fluid alone. Strikingly, the coinjection of amniotic fluid and LPS into the gastrointestinal tract markedly reduced the extent of TLR4 signaling, as manifested by a reduction in GFP expression (Fig. 1 A, iii and iv vs. ii, vii, and viii) and proinflammatory cytokines (Fig. 1B) within the intestinal mucosa at both 3 and 6 h. These findings demonstrate that amniotic fluid can inhibit TLR4 signaling in the fetal intestinal epithelium.

Fig. 1.

Amniotic fluid inhibits TLR4 signaling in fetal intestinal epithelium via the EGFR. (A) (i–vi) Confocal micrographs showing GFP (green) and E-cadherin (red) staining in whole-mount ileal sections from embryonic day 18.5 NF-kB–GFP fetuses that underwent gastrointestinal tract injection with saline (i); LPS (5 μg ) with saline (5 μL) (ii); amniotic fluid (AF) 3 h prior (iii) or 6 h prior (iv); cetuximab (Cet) 3 h prior (v); or control IgG (vi). (Scale bar: 100 μm.) (vii and viii) Quantification of NF-kB expression by qRT-PCR (vii) or fold change in fluorescence vs. control (Ctrl) (viii). (B) qRT-PCR expression of iNOS (i) or IL-6 (ii) as indicated. *P < 0.05 vs. saline; **P < 0.05 vs. LPS; ***P < 0.05 vs. LPS + amniotic fluid; by ANOVA. Data shown are representative of three separate experiments with three fetuses per experiment at each time point.

Amniotic Fluid Inhibits TLR4 Signaling in Cultured Enterocytes via EGF Receptor Signaling.

We next sought to determine whether amniotic fluid also could inhibit TLR4 signaling in cultured enterocytes and, if so, to determine the possible molecular mechanisms involved. We first examined the effects of amniotic fluid on TLR4 signaling in IEC-6 rat enterocytes known to express TLR4 (7). As shown in Fig. 2, treatment of IEC-6 cells with LPS significantly increased the extent of translocation of the p65 subunit of NF-kB from the cytoplasm to the nucleus 1 h later (Fig. 2 A, ii vs. A, i and F, i) as well as the induction of IL-6 6 h later (Fig. 2 F, ii and Table S1). Both NF-kB translocation (Fig. 2 A, iii vs. A, ii and F, i) and IL-6 expression (Fig. 2 F, ii and Table S1) were decreased significantly when cells were pretreated with amniotic fluid 60 min before LPS treatment.

Fig. 2.

Amniotic fluid inhibits TLR4 signaling in enterocytes via activation of the EGFR in a PPARγ-dependent manner. (A–D) Representative confocal micrographs of wild-type IEC-6 cells (A, i–iv) or IEC-6 cells deficient in EGFR (B, i–iv), PPAR-γ (C, i–iv), or transduced with scrambled shRNA (D, i–iv) that were treated with LPS (50 μg/mL for 1 h) (A, ii; B, ii; C, ii; and D, ii) or that were pretreated with amniotic fluid (100 μL/mL) (A, iii; B, iii; C, iii; and D, iii) or EGF (400 ng/mL) (A, iv; B, iv; C, iv; and D, iv). (Scale bars: 10 μm.) The p65 subunit of NF-kB is shown in green, and F-actin is shown in red. (E) RT-PCR showing EGFR (Upper) and PPAR-γ (Lower) with actin loading controls in wild-type, EGFR knock-down (EGFR-kd), PPARγ knock-down (PPAR-kd), or IEC-6 cells transduced with scrambled shRNA as indicated. (F) NF-kB translocation (i) and IL-6 qRT-PCR (ii) in the cell types and conditions indicated. AF^EGF−, amniotic fluid in which EGF was immunodepleted. *P < 0.05 vs. control; **P < 0.05 vs. LPS; by ANOVA. Results are representative of at least five separate experiments with more than 50 fields per experiment.

One of the major proteins in amniotic fluid is EGF, which is critical for fetal intestinal growth and development (8). We next tested the possibilities that EGF also could inhibit TLR4 signaling within enterocytes and that the TLR4-inhibitory effects of amniotic fluid could be attributed to EGF. As shown in Fig. 2A, EGF caused a significant reduction in LPS-mediated translocation of NF-kB (Fig. 2 A, iv vs. A, ii and F, i), and reduced the LPS-induced expression of IL-6 at both 6 and 12 h after treatment (Fig. 2 F, ii). To investigate directly whether the inhibition of TLR4 by amniotic fluid required the EGF receptor (EGFR), we knocked down EGFR in IEC-6 cells through lentiviral transduction of EGFR shRNA (see Fig. 2E, Upper for evidence of knockdown of EGFR). Importantly, neither amniotic fluid (Fig. 2 B, iii) nor EGF (Fig. 2 B, iv) had any inhibitory effect on the extent of TLR4-mediated NF-kB translocation (see quantification in Fig. 2 F, i) or IL-6 expression (Fig. 2 F, ii) in EGFR-deficient cells. Amniotic fluid also inhibited the release if IL-6 from IEC-6 cells as determined by ELISA (Fig. S1A). Treatment of wild-type IEC-6 cells with amniotic fluid from which we previously had immunodepleted EGF did not inhibit either LPS-induced NF-kB translocation or IL-6 induction (Fig. 2F). Taken together, these findings demonstrate that amniotic fluid can inhibit TLR4 signaling in enterocytes via EGFR activation.

We next sought to evaluate how EGFR activation could inhibit TLR4 signaling. EGF had no effect on the expression of IL-1 receptor–associated kinase M, single Ig IL-1R–related molecule, or A20, which are known intracellular inhibitors of TLR4 (Fig. S1B and Table S1). However, because the TLR4 and EGFR signaling pathways converge at the transcription factor peroxisome proliferator-activated receptor (PPAR-γ), whose activation exerts anti-inflammatory effects within the intestine (9), we focused on a role for PPAR-γ and found that both amniotic fluid and EGF failed to protect TLR4 signaling in IEC-6 cells deficient in PPAR-γ (Fig. 2 C, iii vs. ii and iv vs. ii; quantification in Fig. 2 F, i and ii), which we had achieved via lentiviral transduction of PPAR-γ shRNA (Fig. 2E, Lower). The deletion of PPAR-γ did not affect the expression of EGFR by RT-PCR (wild-type IEC-6: 1, IEC-6–PPAR-γ knockdown: 1.2 ± 0.1; P = 1). Both amniotic fluid and EGF did inhibit LPS-mediated NF-kB translocation and IL-6 expression in IEC-6 cells transduced with scrambled (control) shRNA (Fig. 2 D–F). These findings illustrate that amniotic fluid inhibits TLR4 signaling in enterocytes in a manner that requires the EGFR and PPAR-γ.

Amniotic Fluid Inhibits TLR4 Signaling in the Fetal Intestinal Epithelium via the EGFR.

We next sought to evaluate whether inhibition of the EGFR could attenuate the inhibitory effects of amniotic fluid on TLR4 signaling in the fetal gastrointestinal tract. Therefore we delivered amniotic fluid and LPS into the fetal intestine along with the specific EGFR inhibitor, cetuximab or a control IgG (10). As shown in Fig. 1, cetuximab reversed the protective effects of amniotic fluid on TLR4 signaling within the fetal intestinal epithelium as manifested by a lack of reduction in the expression of NF-kB–GFP (Fig. 1 A, v vs. iii, iv, vii, and viii) and iNOS and IL-6 (Fig. 1B), whereas control IgG had no such effect (Fig. 1 A and B). These findings illustrate that amniotic fluid inhibits TLR4 signaling via the EGFR both within cultured enterocytes and within the developing intestine.

Amniotic Fluid Attenuates TLR4 Signaling in the Neonatal Intestinal Epithelium via EGFR Activation.

We next sought to evaluate whether amniotic fluid could attenuate TLR4 signaling in the postnatal intestine and, if so, whether amniotic fluid could have a protective effect on the development of NEC. In addition to inducing proinflammatory signaling, TLR4 activation in the newborn intestinal epithelium markedly reduces the proliferation of enterocytes that leads to impaired mucosal healing (4). As shown in Fig. 3, although LPS administration markedly reduced proliferating cell nuclear antigen (PCNA) expression in the intestinal crypts (Fig. 3 A, ii vs. i; quantification in C, i), the administration of amniotic fluid prevented this inhibition and restored enterocyte proliferation to levels similar to those in untreated mice (Fig. 3 A, iii vs. i, and C, i). Moreover, the enteral administration of amniotic fluid significantly reduced the extent of TLR4-mediated cytokine induction within the intestinal epithelium after both 6 and 12 h of LPS treatment (Fig. 3 C, ii and iii).

Fig. 3.

Amniotic fluid attenuates TLR4 signaling in the neonatal intestinal epithelium. (A and B) Representative confocal micrographs showing terminal ileal crypts from wild-type (A) or EGFR^ΔIEC (B) newborn mice treated with saline (A, i and B, i), LPS (5 mg/kg for 6 h) (A, ii and B, ii), LPS (5 mg/kg) plus amniotic fluid (50 μL/g, 1 h prior) (A, iii and B, iii), LPS plus amniotic fluid plus cetuximab (100 μg/d for 3 d prior) (A, iv and B, iv), or control IgG (equimolar, 2 d prior) (Inset in A, iv), stained for PCNA (green) and DAPI (blue). (Scale bars: 10 μm.) (B, iv) RT-PCR of EGFR in wild-type and EGFR^ΔIEC newborn intestine. (C and D) TLR4 signaling in wild-type (C) and EGFR^ΔIEC (D) mice. (C, i and D, i) Fold PCNA expression quantification. (C, ii and D, ii) qRT-PCR for IL-6. (C, iii and D, iii) RT-PCR for iNOS. AF^EGF−, amniotic fluid in which EGF was immunodepleted. (C, iv and v) IL-6 expression by serum ELISA (iv) and mucosal qRT-PCR (v). *P < 0.05 vs. saline; **P < 0.05 vs. LPS; ***P < 0.05 vs. LPS plus amniotic fluid; by ANOVA. (E) H&E of wild-type cells (i) or EGFR^ΔIEC cells 3 d after tamoxifen (ii) Results shown are representative of more than 100 fields, with at least five mice per group in five separate experiments.

We next sought to determine whether the inhibitory effect of amniotic fluid on TLR4 signaling in newborn mice also required EGFR activation. As shown in Fig. 3A, the administration of the EGFR inhibitor cetuximab 1 h before treatment reversed the protective effects of amniotic fluid on TLR4-mediated impaired enterocyte proliferation (Fig. 3 A, iv vs. iii and C, i) and cytokine release (Fig. 3 C, ii and iii) within the newborn intestinal epithelium to levels comparable to those in mice administered LPS alone, whereas the administration of control IgG had no such effect. Moreover, the enteral administration of amniotic fluid that was immunodepleted of EGF did not prevent the LPS-mediated induction in IL-6 or iNOS (Fig. 3 C, ii and iii). Further evidence that EGFR activation can inhibit TLR4 signaling in the newborn intestinal epithelium is shown in Fig. 3 C, iv and v, in which the extent of LPS-induced IL-6 expression within the serum (Fig. 3 C, iv) and the intestinal mucosa (Fig. 3 C, v) was reduced significantly when purified EGF was administered 1 h before the administration of LPS.

To confirm these pharmacological studies, we next generated a mouse line that selectively lacked EGFR within the intestinal epithelium by interbreeding EGFR-loxP mice with tamoxifen-inducible villin-cre-ERT2 mice (herein termed “EGFR^ΔIEC”) (Fig. 3 B, iv). Importantly, although LPS injection markedly reduced enterocyte proliferation (Fig. 3 B, ii vs. i and D, i) and increased cytokine expression (Fig. 3 D, ii and iii and Table S1) in EGFR^ΔIEC mice, amniotic fluid had no protective effect on either of these events in EGFR^ΔIEC mice (Fig. 3 B, iii vs. i and D, i–iii). It is noteworthy that, consistent with prior reports (11), the deletion of EGFR from the intestinal epithelium resulted in a slightly disrupted mucosal architecture (Fig. 3E). Taken together, these findings illustrate that amniotic fluid inhibits TLR4 in the newborn intestine via EGFR signaling.

Amniotic Fluid Attenuates the Severity of Experimental NEC Through Activation of the EGFR.

We next sought to determine whether amniotic fluid could attenuate the severity of experimental NEC and, if so, whether EGFR signaling was required. To do so, we established NEC in newborn mice with 4 d of hypoxia and formula gavage (12), which leads to the induction of iNOS (Fig. 4 A, ii vs. i and B, vii), reduced enterocyte proliferation (Fig. 4 A, vi vs. v), and disruption of the ileal mucosa (Fig. 4 B, ii vs. i) as compared with control mice that were allowed to breast feed (quantification in Fig. S2). Importantly, the enteral administration of amniotic fluid daily during the experimental model significantly attenuated NEC severity, as manifested by a reduction in the expression of iNOS within the intestinal mucosa (Fig. 4 A, iii vs. ii and B, vii), restoration in enterocyte proliferation (Fig. 4 A, vii vs. vi), preservation of mucosal architecture (Fig. 4 B, iii vs. ii), and, most significantly, a marked reduction in the histology scores as assessed by a blinded pathologist (Fig. 4 B, vi). Importantly, pretreatment of mice with the EGFR inhibitor cetuximab (Fig. 4 B, iv) or the enteral administration of amniotic fluid that was immunodepleted of EGF (Fig. 4 B, v) markedly reduced the salutary effects of amniotic fluid on NEC severity (Fig. 4 B, vi and vii). Taken together, these findings show that amniotic fluid protects against the development of NEC via the EGFR.

Fig. 4.

Amniotic fluid attenuates NEC severity through the activation of the EGFR. (A) (i–iv) Representative confocal micrographs of the terminal ileum of newborn mice that were either breast fed (i) or induced to develop NEC (ii–iv,) in the absence (ii) or presence of amniotic fluid (iii) or amniotic fluid and cetuximab (100 μg daily) (iv) stained for iNOS (red), e-cadherin (green), and DAPI (blue). (v–viii) Corresponding terminal ileum sections were stained with PCNA (green) and DAPI (blue). (B) (i–v) Histologic H&E-stained sections from the terminal ileum of newborn mice in the indicated treatment groups. (Scale bar: 100 μm.) AF^EGF−, amniotic fluid in which EGF was immunodepleted. (vi) NEC severity scores. (vii) iNOS qRT-PCR in intestinal mucosa under the indicated conditions. *P < 0.05 vs. breast fed (BF); **P < 0.05 vs. NEC; by ANOVA. Results are representative of more than 100 fields with at least 10 mice per group in five separate experiments.

Reduction in the Expression of the EGFR in the Intestinal Epithelium Is Associated with the Development of NEC in Mice and Humans.

We next sought to define whether the premature infant may be at particular risk for the development of NEC because of a reduction in the expression of EGFR and thus an inability to benefit from the protective anti-TLR4 effects of amniotic fluid. As shown in Fig. 5, EGFR expression was increased significantly in fetal mice compared with postnatal mice (Fig. 5A and Table S1). Importantly, mice in which NEC was induced showed a marked reduction in the expression of EGFR in the intestinal epithelium, as revealed by SDS/PAGE (Fig. 5B) and confocal microscopy (Fig. 5 E, ii vs. i; see quantification in Fig. S2). The enteral administration of amniotic fluid to mice restored the expression of EGFR levels in the newborn intestinal epithelium (Fig. 5 B and E, iii vs. ii), consistent with a reduction in NEC severity shown in Fig. 4B. The administration of either amniotic fluid that was immunodeficient in EGF or cetuximab prevented the restoration of EGFR expression in mice with experimental NEC, indicating that EGFR signaling is required for the reduced EGFR that occurs in NEC (Fig. 5 B, i and ii). Furthermore, the reduction in EGFR that was observed in experimental NEC occurred before the induction of mucosal inflammation as manifested by iNOS expression (Fig. 5C), suggesting that EGFR reduction did not occur simply in response to the inflammatory process.

Fig. 5.

Decreased expression of the EGFR in the intestinal epithelium is associated with NEC development. (A) EGFR RT-PCR in the mouse intestine. *P < 0.05 vs. day −4; **P < 0.05 vs. day −2; Student’s t test. Results shown are from three separate experiments. (B) (i) SDS/PAGE showing EGFR (Upper) and β-actin (Lower) in newborn mice treated as indicated. AF, amniotic fluid; AF^EGF−, amniotic fluid immunodepleted of EGF; BF, breast fed; Cet, cetuximab. (ii) EGFR densitometry relative to actin. *P < 0.05 vs. control; **P < 0.05 vs. NEC; by ANOVA. (C) RT-PCR for EGFR (black bars) and iNOS (red bars) on the day indicated. *P < 0.05 vs. day 1 (black); **P < 0.05 vs. day 1 (red); by ANOVA. Results shown are from three separate experiments. (D) SDS/PAGE for EGFR in human intestinal samples from a 24-wk aborted fetus (Fetal), a 26-wk premature infant with NEC (NEC), and a healed newborn; EGFR by densitometry relative to actin: *P < 0.05 vs. fetal; *P < 0.005 vs. NEC; by ANOVA. Results shown are from three separate samples per group. (E and F) Representative confocal micrographs of terminal ileum in conditions indicated stained for EGFR (red), E-cadherin (green), and DAPI (blue). (Scale bar: 100 μm.) Results shown are representative of more than 100 fields.

We next investigated EGFR expression in the human intestine before, during, and after the development of NEC. The expression of EGFR was significantly greater in human fetal intestine than in the intestine of a premature infant at a comparable gestational age with NEC, as revealed by SDS/PAGE (Fig. 5D) and confocal microscopy (Fig. 5 F, ii vs. i; see quantification in Fig. S2). These findings are consistent with the findings in mice. Strikingly, the expression of EGFR in the intestine of human infants in whom NEC had resolved—as determined at the time of stoma closure—showed a marked recovery of EGFR expression (Fig. 5 D and F, iii vs. ii). Taken together, these findings indicate that the reduced expression of EGFR within the intestinal epithelium is associated with the development of NEC in mice and humans and also provide insights into the unique susceptibility of the premature infant for the development of NEC.

Discussion

We show here that amniotic fluid has an ability to inhibit TLR4 signaling within the intestinal mucosa of the fetal and neonatal mouse. We recently demonstrated that TLR4 expression rises within the small intestine during embryonic development and that activation of TLR4 in response to microbial colonization of the postnatal gut leads to a proinflammatory response that leads to NEC (5). Although amniotic fluid is considered to be sterile, it often is colonized with microbes, and bacterial colonization of the amniotic fluid is a frequent finding in the setting of early parturition (13). Therefore, it is tempting to speculate that the TLR4-inhibitory properties of amniotic fluid may serve to limit the capacity of microbes from activating the immune system of the fetal intestine in order to maintain homeostasis, a failure of which may be an important contributor to premature birth. By extension, in the absence of the TLR4-inhibitory properties of amniotic fluid in the postnatal period, the premature infant may exhibit unrestrained TLR4 signaling within the intestinal mucosa upon its microbial colonization, a situation that we have found leads to the development of NEC (14). We also have identified a role for EGFR activation as a pathway that leads to TLR4 inhibition within the fetal and neonatal gastrointestinal tract and through immunodepletion of EGF from the amniotic fluid, we have shown that EGF is an important ligand in the amniotic fluid that mediates this pathway. It could be speculated that several of the previously reported effects of EGFR activation on intestinal mucosal injury and repair also may reflect, in part, the inhibitory effects of EGFR on TLR4. In this regard, it is a curious coincidence that EGF can reverse several processes that can result from TLR4 activation during inflammatory states, namely a reduction in enterocyte proliferation and migration and increased enterocyte apoptosis (15, 16), suggesting perhaps that the salutary effects of EGF on these barrier functions also may reflect an inhibition of TLR4 signaling. Furthermore, several groups have reported that the administration of EGF and its homolog HB-EGF can treat NEC (17, 18). The current findings emphasize the role of the amniotic fluid, and, in particular, its ability to activate the EGFR, serve as a TLR4 inhibitor in enterocytes and raise the possibility that unraveling the anti-TLR4 signaling properties of the amniotic fluid and the EGFR may provide further therapeutic insights into this devastating disease.

Materials and Methods

In Utero Microinjection of the Fetal Intestine.

Backscatter ultrasound-guided microinjection to deliver reagents directly into the fetal gut was performed as described (5). At embryonic day 18.5 the uterus of a pregnant NF-kB–GFP mouse (obtained from C. Jobin, University of North Carolina, Chapel Hill, NC) was exposed by laparotomy, and a fenestrated dish was placed over the uterus. A single embryo (one uterine saccule) was brought through the fenestration. The indicated reagent was injected directly into the fetal stomach, via a glass syringe, using a Vevo Imaging Station under ultrasound guidance (VisualSonics).

Determination of TLR4 Signaling in Enterocytes.

Wild-type, PPAR-γ– or EGFR-deficient IEC-6 cells were pretreated with amniotic fluid (100 uL/mL medium) or EGF (400 ng/mL medium) for 1 h before treatment with LPS (50 μg/mL) for 6 or 12 additional h and then were assessed for iNOS and IL-6 by quantitative RT-PCR (qRT-PCR). Translocation of NF-kB from the cytoplasm to the nucleus was determined as described (5). Immunohistochemistry, immunofluorescence, and SDS/PAGE were performed as described (19).

Induction of Endotoxemia and NEC in Neonatal Mice.

All experiments were approved by the Children’s Hospital of Pittsburgh Animal Care Committee. To generate mice lacking EGFR in the intestinal epithelium, Egfr^loxP/loxP mice (obtained from David W. Threadgill, University of North Carolina, Chapel Hill, NC) were mated with villin-Cre-ERT2 (provided by Sylvie Robine, Morphogenesis and Intracellular Signaling, Institut Curie-CNRS, Paris, France). To induce Cre-recombinase, neonatal (postnatal day 6) wild-type and villin-ER2-EGFR (Egfr^{loxP/loxP;vil-Er2}) littermates were administered tamoxifen (Sigma-Aldrich) i.p. (25 μL; 5 mg⋅mL⁻¹⋅d⁻¹ for 3 d). Neonatal pups were injected i.p. with LPS (5 mg/kg) after enteral pretreatment with amniotic fluid (50 μL/g body weight). Cetuximab was administered i.p (100 ug/d) for 3 d before the start of the model and then was coadministered with LPS. Equimolar IgG (1.5 μg/μL) was administered in parallel (Thermo Scientific). NEC was induced in 10-d-old mice (12) using formula gavage [Similac Advance infant formula (Abbott Nutrition):Esbilac canine milk replacer, 2:1] 50 μL/g body weight, five times per day for 4 d, and hypoxia (5%O₂, 95%N₂) administered for 10 min twice daily for 4 d using a hypoxia chamber (Billups-Rothenberg Inc.).

Statistical Analysis.

All experiments were repeated at least in triplicate, with more than 100 cells per high-power field. For endotoxemia, at least three mice per group were assessed; for NEC, more than 10 pups per group and litter-matched controls were included in all cases. Statistical analysis was performed using SPSS 13.0 software. ANOVA, χ², or two-tailed Student’s t test were used as appropriate. Significance was accepted at P < 0.05.