Introduction

One of the most frequent complications in the clinical management of preterm infants is the persistence of the patent ductus arteriosus (PDA).1 A persistent PDA in preterm neonates increases the risk of pulmonary congestion and decreased blood flow to vital organs that could lead to various co-morbidities.2 PDA occurs in about 40–80% of premature infants with low birth weight.2 The closure of PDA can be achieved effectively with cyclooxygenase (COX) inhibitor administration leading to permanent ductal closure in 60–80% of infants.3 Nonsteroidal anti-inflammatory drugs (NSAIDs), such as indomethacin (INDO), are the standard pharmacological therapy during the neonatal period for prevention and closure of persistent PDA.4 However, treatment with INDO is associated with many adverse effects in neonates, mostly affecting the gastrointestinal tract with decreased blood flow and a higher risk of developing necrotizing enterocolitis (NEC) with intestinal perforation.5,6

In the last two decades, physicians have used other NSAIDs, namely ibuprofen (IBU), as medical therapy to prevent and treat PDA.7 Clinical trials have demonstrated that IBU was as effective as INDO in closing the PDA in premature infants.8,9,10 Regarding adverse effects, IBU has been associated with milder side effects than INDO on renal and cerebral circulation9 but a greater risk of triggering severe pulmonary hypertension.10 While there is still no clear consensus about the best route of administration of IBU in preterm infants with respect to the efficacy and safety profile,11 treatment with IBU in preterm infants has been associated with gastrointestinal complications, such as spontaneous intestinal perforation,12,13 suggesting that IBU is not safer than INDO for the premature bowel, which highlights the need of further studies to elucidate those adverse effects.

Recently, we have demonstrated that INDO exerts deleterious effects on the immature human intestine.14 We also identified that several different metabolic pathways were modulated by INDO such as glycolysis/gluconeogenesis, oxidative phosphorylation, and oxidoreductase activities, resulting in impaired glucose metabolism and mitochondrial function that could lead to a disruption of the intestinal epithelial barrier. In the present study, we combined our model of the mid-gestation human intestine in serum-free organ culture with global gene expression analysis to investigate the direct impact of IBU on the overall physiology of the immature small intestine and compared these effects with those observed previously with INDO. Our results suggest that IBU also exerts deleterious effects on the immature intestinal mucosa, and based on the pathway involved, the sequelae of this treatment could favor bacterial colonization. Consequently, we propose that IBU is not safer than INDO for the immature intestine.

Materials and methods

Tissues

Small intestinal (ileum) tissues were obtained from 16 fetuses ranging in age from 17 to 20 weeks following legal or therapeutic pregnancy termination with informed patient consent. No tissues were collected from cases associated with known fetal abnormalities or intrauterine fetal demise. Studies were approved by the Institutional Review Committee for the use of human material of the “Centre Hospitalier Universitaire de Sherbrooke/Faculté de Médecine et des Sciences de la Santé”.

Serum-free organ culture

Small intestinal tissues obtained from 16 fetuses were prepared at the same time and under the same conditions as previously described.14 Briefly, each intestinal tissue was cut into several 5 × 5-mm2 explants which were maintained in organ culture dishes (Falcon Plastics, Los Angeles, CA) at the interface of a 95% air-5% CO2 gas mixture and culture medium for up to 2 days (37 °C). For each intestinal tissue, four culture dishes containing approximately 6–9 explants each were prepared, two for each experimental condition (untreated and IBU treated). IBU (Sigma-Aldrich, St-Louis, MO) was added at a concentration of 1 nM to 300 µM for prostaglandin E2 (PGE2) production studies, and at 100 µM for gene expression studies, respectively (85% inhibition of intestinal PGE2 production and in the range of circulating levels in treated preterm babies). Explants were maintained in culture for 2 days and media were changed daily. Studies with INDO were conducted under similar conditions as described previously.14

Measure of intestinal PGE2 levels

The levels of PGE2 released in the culture medium were measured after 48 h of culture in the presence or absence of various concentrations of IBU (1 nM to 300 µM). Culture media were harvested and the concentration of PGE2 was determined using a PGE2 enzyme-immunoassay kit (Correlate-EIA, Assay Designs, Ann Arbor, MI) following the manufacturer’s protocol.

RNA extraction

RNA was extracted with TRIzol (Invitrogen, Burlington, ON) according to the manufacturer’s protocol and stored at −80 °C. For each sample, RNA was quantitated and the quality was evaluated by determining RNA integrity values (RIN values > 7.0) as required for microarray experiments.

Microarray screening and data analysis

Probes for microarray analysis were generated from RNA isolated from two dishes of cultured explants for each experimental condition (untreated vs. IBU treated) tested on four sets of biological samples obtained from distinct fetuses for a total of eight samples (four IBU-treated ileums and four matched control ileums). The biological validity of these experimental conditions was recently established.15 The eight samples were processed at the microarray platform of the UHN Microarray Centre, University Health Network (Toronto, ON). One Illumina whole-genome Human HT-12 v4 expression beadchip (12 samples per beadchip) was screened, analyzed, and quantile normalized via the UHN Microarray Centre (data are accessible through GEO and are all MIAME compliant). A Wilcoxon test (p < 0.05) was used to identify genes expressed differentially (using TMEV 4.2 software). Microarray screening studies for INDO were performed as described previously14, and the original data are accessible through GEO series accession number GSE38406.

Ingenuity Pathway Analysis

Functional analyses were performed using Ingenuity Pathway Analysis (IPA) 8.8 software (Ingenuity Systems, www.ingenuity.com) as described previously.14,15 The reference lists containing differentially expressed genes (DEGs) with gene identifiers and corresponding expression values were uploaded into the IPA application. Each gene identifier was mapped to its corresponding gene object in the Ingenuity Knowledge Base. IPA allows filtering to consider only functions and interactions in protein networks and/or pathways that are known for the defined species and tissue or cell line range. The stringent filter set for human and relaxed filter for tissues and cell lines were used for the core analyses. Fisher’s exact test was used to calculate a p-value determining the probability that each biological function assigned to that data set would be due to chance alone.

Data validation by quantitative PCR

RNA from the same 32 samples used for microarray analyses was used for quantitative PCR (qPCR) confirmation. All reactions were performed in duplicate as previously described.14,15 Briefly, reactions were performed in an Mx3000P real-time PCR system (Stratagene, Cedar Creek, TX) starting with 10 min of Taq activation at 95 °C, followed by 40 cycles of melting (95 °C, 30 s), primer annealing at the temperature appropriate for each primer (55–60 °C, 45 s), and extension (72 °C, 45 s) ending with a melting curve analysis to validate the specificity of the PCR products. Fluorescence data were acquired after each annealing step. Amplification efficiencies ranged from 91% to 105%. Brilliant II SYBR®Green QRT-PCR Master Mix (Stratagene) was mixed with the appropriate primers and high-quality sterile water. The genes investigated were alanyl-tRNA synthetase (AARS), anterior gradient 2 (AGR2), apolipoprotein A1 (APOA1), apolipoprotein A4 (APOA4), apolipoprotein B (APOB), chloride channel accessory 1 (CLCA1), aspartyl-tRNA synthetase (DARS), fatty acid binding protein 2 (FABP2), Fc fragment of IgG binding protein (FCGBP), glycyl-tRNA synthetase (GARS), lipocalin 2 (LCN2), mucin 2 (MUC2), peptidase inhibitor 3 (PI3), regenerating family member 1 alpha (REG1A), and trefoil factor 3 (TFF3). Primers are listed in Table 1 and were generated using the primer formation software Primer3 (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi) with attention given to avoiding primer-dimer formation by stringent use of the maximum 3′ self-complimentarity function of the Primer3 program. Differences in gene expression were evaluated by comparing untreated vs. IBU- or INDO-treated samples for a given intestinal segment using the equation R = (Etarget) ∆Cttarget/ (Ereference) ∆Ctreference. The reference gene used was the ribosomal protein S3A (RPS3A).

Table 1 Primers used in this study

Western blot analysis

Analysis of protein markers in the culture medium was performed as previously described.16 Briefly, culture media were lyophilized by vacuum concentration (SpeedVac, Savant), resuspended in Laemmli buffer and evaluated for protein content. For each sample, 50 µg of proteins were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis under denaturing conditions and electrotransferred onto nitrocellulose for immunodetection using the primary antibodies anti-LCN2 (ab41105, Abcam, Toronto, ON), anti-REG1A (ab47099, Abcam), and anti-TFF3 (ab108599, Abcam), all diluted 1/1000 in 10% Blotto in PBS and detected using horse radish peroxidase-conjugated secondary antibodies (GE Healthcare-Amersham Bioscience, Baie d’Urfé, QC).

Results

Determination of IBU concentration for inhibition of intestinal PGE2

The dose of IBU used clinically in PDA closure generally aims to use the lowest possible concentration to achieve the desired outcome. Therefore, we first performed a dose–response curve study to determine the appropriate working concentration of IBU to use in organ cultures of mid-term intestinal tissue. Levels of PGE2 were measured as an indicator of COX activity, the target of NSAID inhibition. As seen in Fig. 1, increasing concentrations of IBU markedly decreased the production and release of PGE2 by small intestinal explants compared to controls. We observed that IBU significantly inhibited PGE2 production by almost 85% at concentrations of 100 µM, confirming that intestinal cultures were responsive to IBU. In preterm infants, studies showed that optimum PDA closure rates with IBU could be achieved with plasma levels varying between 90 and 210 µM.17,18 Based on these data, IBU was used at 100 µM concentration for the studies in organ culture.

Fig. 1
figure 1

Inhibition of PGE2 production by IBU in the human small intestine at mid-gestation. Changes in intestinal PGE2 levels after 48 h of culture in the presence of increasing concentrations of IBU. Data are expressed as percentage of inhibition of PGE2 compared to the corresponding untreated control segment. Values shown are the mean of three independent biological samples; *p < 0.05 and **p < 0.001 vs. untreated control segments

Global gene expression analysis of the effect of IBU on the mid-gestation human small intestine

To investigate the modulatory influence of IBU on gene expression profiles in the mid-gestation intestine, matched control and IBU-treated explants prepared from ileums of four different fetuses were cultured in serum-free medium for 48 h. Gene expression profiles were then determined using Illumina wholegenome expression beadchip microarrays providing coverage of more than 47,000 transcripts for each of the eight samples. Statistical analyses revealed that 444 genes were significantly differentially expressed (±1.5 fold change) when IBU was administered in the ileum, in comparison with controls (Supplemental Table S1 for gene list). Of the DEGs, 253 were upregulated and 191 were downregulated. The original data have been deposited into the National Center for Biotechnology Information’s Gene Expression Omnibus and are accessible through GEO Series accession number GSE108369.

Pathway analysis of the effect of IBU on the immature human ileum

All significantly DEGs identified in the mid-gestation ileum exposed to IBU were subjected to IPA software analysis which led to the identification of 37 canonical pathways modulated by IBU that can be classified as belonging to six major families of metabolic pathways (Table 2) including inflammation, fatty acid metabolism, amino acid metabolism, sugar metabolism, signaling and oxidoreductase activities. The list of genes associated with each canonical pathway is provided in Supplemental Table S2.

Table 2 Families of pathways modulated by IBU in the immature human intestine

Canonical pathways modulated by IBU and INDO in the mid-gestation human small intestine

To compare the effects of IBU and INDO in the mid-gestation small intestine, all significant DEGs modulated by either IBU or INDO were subjected to IPA comparative analysis of cellular, molecular, physiological and metabolic functions. Plotting the negative logarithm of p-values calculated by IPA for each of the functional categories found in IBU - against the negative logarithm of p-values of the corresponding categories found in INDO - allowed visualization of the functions that are more relevant to each treatment (Fig. 2), where each point corresponds to one of the 51 pathways. IBU and INDO were found to significantly modulate gene expression of 37 and 32 pathways, respectively. As illustrated with a Venn diagram (Fig. 2, inlet), 18 of these pathways were modulated by both treatments. Supplemental Table S3 lists these functions according to statistical significance in the variation of the expression of genes in each category. Among these common pathways were functions that were previously reported to be associated with NSAID treatment such as “glycolysis/gluconeogenesis”, “arginine and proline metabolism”, “fructose and mannose metabolism”, “LXR/RXR activation”, and “bile acid synthesis”. Is it noteworthy that numerous pathways were found to be exclusively modulated by either IBU or INDO. For instance, IBU modulated biological processes such as “fatty acid metabolism”, “glycerolipid metabolism”, “polyamine regulation”, and “aminoacyl-tRNA biosynthesis”, while pathways such as “role of IL-17A in arthritis”, “pentose phosphate pathway”, and “role of cytokines in mediating communication between immune cells” were found to be affected exclusively by INDO treatment, suggesting that IBU and INDO exert distinct effects on the immature intestine.

Fig. 2
figure 2

Cellular, molecular, and physiological functions modulated by IBU and INDO in the human small intestine at mid-gestation. The negative logarithm of p-values (Fisher’s test), calculated by IPA, for each of the functional categories affected by IBU was plotted against the negative logarithm of p-values of the corresponding categories affected by INDO in the small intestine. Insets: Venn diagram showing the 51 molecular and cellular functions identified in IBU and INDO. Of the 37 pathways found in IBU, 18 were also found in INDO. Numbers and letters indicate shared and exclusive functions belonging to IBU and INDO, respectively (listed in the inserted table). Thresholds (dotted lines) denote p = 0.05 [−Log (0.05) = 1.3]

Validation of microarray results by qPCR

As a first step to validate the microarray results, representative genes from two of the common pathways were tested by qPCR. As expected from previous observation with INDO14,15 IBU treatment induced a significant reduction in the expression of genes of the glycolysis/gluconeogenesis pathway ALDOA, ENO1, HK2 and PGK1, as well as an increase in CYP3A4 and a decrease in GPX2 and NOS2 for oxidoreductase activity (Fig. 3).

Fig. 3
figure 3

Differential effects of IBU on markers of glycolysis and oxidoreductase activities in the human small intestine at mid-gestation. Real-time quantitative PCR analysis of transcript levels of a glycolytic enzymes (ALDOA, ENO1, HK2, and PGK1), and b oxidoreductase activities (CYP3A4, GPX2, NOS2) after 48 h of culture in the presence of IBU (100 µM) in the mid-gestation human ileum. Samples were normalized to RPS3A and data are expressed as ratios of IBU treated over untreated segments, expressed on a Log2 scale. Values shown are the mean of three to four independent biological samples; *p < 0.05 vs. corresponding untreated control segments

We then focused on pathways that appear to be exclusively modulated by IBU using “fatty acid metabolism” and “aminoacyl-tRNA biosynthesis” as archetypes and tested genes representative of these functions by qPCR. As shown in Fig. 4a, we confirmed that apolipoproteins APOA1, APOA4, APOB, and FABP2 were significantly upregulated by IBU in intestinal tissue, while only APOA1 was increased by INDO. Another biological function that was identified by IPA as affected by IBU is “aminoacyl-tRNA biosynthesis”. IBU was found to increase the expression of two of the three tested genes (AARS and GARS) while INDO had no effect on their expression (Fig. 4a).

Fig. 4
figure 4

Differential effects of IBU and INDO on lipid metabolism, aminoacylation, antimicrobial response, and mucus production in the human small intestine. Real-time quantitative PCR analysis of transcript levels of several genes involved in a lipid metabolism (APOA1, APOA4, APOB, FABP2), b aminoacyl-tRNA synthetases (AARS, DARS, GARS), c antimicrobial response (LCN2, PI3, REG1A), and d mucus production (AGR2, CLCA1, FCGBP, MUC2, TFF3) after 48 h of culture in the presence of IBU (100 µM) and INDO (1 µM) in the developing human ileum. Samples were normalized to RPS3A and data are expressed as ratios of IBU-treated over untreated segments and INDO-treated over untreated segments. Values shown are the mean of three or four independent biological samples; *p < 0.05 vs. corresponding untreated control segments

Intestinal-related functions modulated by IBU

We also used some of the other available bioinformatics tools such as David Bioinformatics Resources 6.8 (https://david.ncifcrf.gov/home.jsp) and KEGG GENES Database (http://www.genome.jp/kegg/genes.html) for further mining of the microarray data generated from IBU and INDO treatments and confirmed that the two treatments distinctly affect the immature intestine. Interestingly, this complementary approach led to the identification of two specific intestinal-related functions, “antimicrobial response” and “mucus production”, that appear specifically modulated by IBU. Further analyses of specific genes that belong to these pathways were then carried out in both IBU- and INDO-treated explants for comparison. For the antimicrobial response, we observed that three genes involved in this category, namely LCN2, PI3, and REG1A, were significantly downregulated by IBU in the immature Ileum (Fig. 4b), while expression of genes involved in mucus production such as AGR2, CLCA1, FCGBP, MUC2, and TFF3 were all downregulated by IBU. It is noteworthy that none of these genes was found to be modulated by INDO in the immature intestine (Fig. 4b).

We also evaluated the expression of some of these markers at the protein level by measuring their release from the explants into the medium over the last 24 h of culture using Western blot. For these studies, we were able to specifically detect LCN2, REG1A, and TFF3 (Fig. 5a). Using five independent cultures grown under control conditions or in the presence of IBU or INDO, we observed a consistent reduction of immunoreactive LCN2 and REG1A in the media of explants treated with IBU (Fig. 5b, c), while TFF3 levels remained comparable to control conditions (Fig. 5d).

Fig. 5
figure 5

Differential effects of IBU and INDO on the release of LCN2, REG1A, and TFF3 proteins into the culture medium. a Representative Western blot analysis of LCN2, REG1A, and TFF3 immunodetected in 48 h culture media from explants cultured in the presence of 100 µM IBU or 1 µM INDO vs. control conditions (CTRL). Relative amounts of LCN b, REG1A c, and TFF3 d were calculated from independent cultures. Data were expressed relative to CTRL; *p < 0.05 vs. corresponding untreated control using one-way Anova test

Discussion

IBU and INDO demonstrate equal effectiveness in closing PDA,8,9,10 but, the wider use of IBU over INDO appears to come from its apparent milder side effects on renal, cerebral, and mesenteric circulation.8,9 Furthermore, INDO was associated with a higher risk of developing NEC with intestinal perforation.5 However, IBU also has been associated with gastrointestinal complications in preterm infants including intestinal perforation,12,13 emphasizing the need for further research on prenatal intestinal physiology. We showed previously that INDO exerts a deleterious influence on immature small intestinal functions when tested in organ culture.14 In the present study, we determined the global gene expression profiles of immature human intestinal explants cultured in the presence of IBU in the same serum-free organ culture and compared the direct effects of each NSAID on the human immature intestine.

As expected from our previous study with INDO, analysis of the gene expression profiles revealed that IBU affects a broad range of biological pathways in the mid-gestation human intestine. A large part of these functions was related to inflammation such as “arachidonic acid metabolism”, “acute phase response signaling”, and “differential regulation of cytokine production in intestinal epithelial cells by IL-17” as well as to various basic metabolisms (sugar, amino acid, glycerolipid, and linoleic metabolisms).

We then proceeded to a direct comparison of the DEGs from small intestinal explants treated with IBU and INDO to generate the pathway analysis. It is noteworthy that as for INDO14, the IBU concentration used in the organ culture system was based on reported plasma levels in preterm babies treated with IBU17 and corresponded to an ~85% inhibition of PGE2 production from the mid-gestation small intestine. Interestingly, it was found that about half of the significant functional pathways identified were common to both NSAIDs. Among them were several metabolic clusters involved in glucose metabolism such as “fructose and mannose metabolism” and “glycolysis/gluconeogenesis” including the genes ALDOA, ENO1, HK2, and PGK1, consistent with our previous observations.14 IBU has been shown to impair gluconeogenesis from lactate and fructose in primary cultured hepatocytes.19 These results suggest that, like INDO, IBU could alter glucose metabolism and impair mitochondrial function in the intestinal mucosa of preterm infants treated for PDA. Other interesting metabolic pathways shared by the two drugs were involved in oxidoreductase activity, namely “xenobiotic metabolism signaling” and “production of nitric oxide and reactive oxygen species in macrophages”. At the gene expression level, we observed that the immature intestine exposed to IBU induced a significantly increased expression of CYP3A4, one of the most common CYP450 genes in the intestine that is involved in the metabolism of several drugs used in clinics.20 Moreover, expression of the NOS2 transcript was downregulated in the presence of IBU in the immature ileum. We have recently shown that NOS2 plays a pivotal role in the inflammatory response of the immature intestine being targeted either by anti- or pro-inflammatory effectors.15

We next analyzed specific functional pathways that were exclusively modulated by IBU in the immature intestine. As mentioned above, there were a number of significantly modulated metabolisms in this category. Fatty acid metabolism was among these pathways. At the individual gene level, expressions of various apolipoproteins, namely APOA1, APOA4, APOB, as well as FABP2, were found to be specifically upregulated by IBU. The mechanism leading to an upregulation of these apolipoproteins after IBU treatment is not known but may involve the activation of the peroxisome proliferator-activated receptor isoforms, PPARα and PPARγ.21,22 It is also interesting to note that the binding of lipophilic drugs such as IBU to FABP2 can facilitate its transport into intestinal cells.23

Another pathway exclusively modulated by IBU is “amino acid metabolism”. We have noted that the expression of two aminoacyl-tRNA synthetases (aaRSs), namely AARS and GARS, were significantly increased in the intestine exposed to IBU. aaRSs are responsible for the aminoacylation reaction which fuses each amino acid to its cognate tRNA, an essential step for accurate protein synthesis.24 Interestingly, non-canonical functions of aaRSs have been recently highlighted in diverse biological processes such as inflammation, immune response, and angiogenesis.25 To our knowledge, this is the first report of the effect of any NSAID on the expression of aaRSs, and suggests a potential role for these proteins on the effect of IBU in intestinal physiology.

In an ongoing effort to better understand the adverse effects of IBU in the immature intestinal mucosa, we further screened DEGs modulated by IBU with other bioinformatic tools leading to the identification of other gene clusters of interest for intestinal homeostasis. Among them was the “antimicrobial response”. Analysis of representative genes from this group disclosed the negative impact of IBU on the expression of three antimicrobial molecules suggested to play a preventive role on chronic intestinal inflammation such as LCN2 and PI326,27 or on intestinal injury such as REG1A.28 In support of these observations, we also observed herein a significant reduction in LCN2 and REG1A protein released into the culture medium. Taken together, these findings suggest that the downregulation of the expression of genes of the antimicrobial response could be involved in the specific adverse effect of IBU on the intestine of premature infants.

Another group of genes involved in intestinal epithelial cell protection is “mucus production”. All analyzed genes of this category, namely AGR2, CLCA1, FCGBP, MUC2, and TFF3, were significantly downregulated by IBU, a phenomenon not observed with INDO. These proteins are produced by goblet cells and are major components of the intestinal mucus.29 AGR2 has been reported to be essential for mucus production and its deficiency led to a greater susceptibility to colitis in an experimental mouse model.30 The goblet cell-derived protein CLCA1, despite not having a role in mucus synthesis or structure, has a possible signaling role in the early immune response in colitis.31 FCGBP is overexpressed in ulcerative colitis32 while downregulated in the normal colon.33 Interestingly, it has been reported that FCGBP can link to TFF3 via a disulfide bridge and that this FCGBP-TFF3 heterodimer is an important structural component of intestinal mucus, in addition to interacting with MUC2.34 TFF3 is a member of the trefoil factor family and is expressed by goblet cells of the small and large intestine.35 The protective role of TFF3 in inflammatory bowel disease has been demonstrated by the maintenance of epithelial barrier function through inducing expression of tight junction proteins.36 Mucus secretion and mucus layer formation act as a protective barrier for the intestinal epithelium from luminal threats. This is exemplified by mice lacking the Muc2 gene developing colitis similar to patients affected with ulcerative colitis.37 These data suggest a drastic effect of IBU on mucus production and assembly, which may result in an enabling environment that could weaken mucus layer formation and favor abnormal microbial colonization of the immature intestine. The lack of reduction of TFF3 released into the culture medium is surprising considering its significant decrease at the transcript level and may indicate that IBU targets immature goblet cells (actively transcribing but not yet secreting) while the regulated secretion of mature goblet cells is not directly affected by IBU. Further study is, however, needed to test this hypothesis.

In summary, comparison of the gene expression profiles of both INDO and IBU in the immature human intestine confirmed that these NSAIDs target a number of metabolic and biological pathways such as glucose metabolism and xenobiotic metabolism that are detrimental to the immature intestine. More importantly, these data demonstrated that each drug specifically targets a distinct subset of biological functions at least in the organ culture context. While important metabolic pathways exclusively modulated by IBU were identified, such as fatty acid metabolism and protein synthesis, our findings that IBU specifically triggers a significant reduction in the expression of genes involved in the antimicrobial response and mucus production suggest that this deleterious effect of IBU on intestinal primary defenses may at least in part enhance susceptibility to intestinal infection and inflammation in the immature intestine, thus emphasizing the potential risk of IBU treatment in preterm infants in neonatal intensive care unit (NICU). Nevertheless, considering the inherent limitation of an ex vivo approach such as organ culture, further studies are required to investigate this possibility in the more general context of NICUs.