Gene expression profiling in necrotizing enterocolitis reveals pathways common to those reported in Crohn's disease

Abstract

Background: Necrotizing enterocolitis (NEC) is the most frequent life-threatening gastrointestinal disease experienced by premature infants in neonatal intensive care units. The challenge for neonatologists is to detect early clinical manifestations of NEC. One strategy would be to identify specific markers that could be used as early diagnostic tools to identify preterm infants most at risk of developing NEC or in the event of a diagnostic dilemma of suspected disease. As a first step in this direction, we sought to determine the specific gene expression profile of NEC.

Methods: Deep sequencing (RNA-Seq) was used to establish the gene expression profiles in ileal samples obtained from preterm infants diagnosed with NEC and non-NEC conditions. Data were analyzed with Ingenuity Pathway Analysis and ToppCluster softwares.

Results: Data analysis indicated that the most significant functional pathways over-represented in NEC neonates were associated with immune functions, such as altered T and B cell signaling, B cell development, and the role of pattern recognition receptors for bacteria and viruses. Among the genes that were strongly modulated in neonates with NEC, we observed a significant degree of similarity when compared with those reported in Crohn's disease, a chronic inflammatory bowel disease.

Conclusions: Gene expression profile analysis revealed a predominantly altered immune response in the intestine of NEC neonates. Moreover, comparative analysis between NEC and Crohn's disease gene expression repertoires revealed a surprisingly high degree of similarity between these two conditions suggesting a new avenue for identifying NEC biomarkers.

Fig. 1

Most significant functional pathways identified in necrotizing enterocolitis. a The negative logarithm of p-values (Fisher’s test), calculated by IPA, for each of the top 12 most significant canonical pathways over-represented in ileal NEC samples compared to control non-NEC samples. ([-Log (0.05) =1.3]) and the corresponding lists of genes associated with each functional pathway. b Biological function enrichment analyses associated with NEC. Activation z-score calculated by IPA for biological function enrichment represents the level of activation (red) or suppression (blue) of a function

Fig. 2

Differential expression of innate immune inflammatory response genes in human necrotizing enterocolitis. Real-time qPCR analysis of transcript levels of selected target genes related to intestinal innate immunity. Ct values of selected genes were normalized using B2M as reference gene and data are expressed as r∆Ct values (reverse ∆Ct: Ct_{reference gene}-Ct_{gene of interest}) in order to display direct variation in NEC vs non-NEC controls (CTRL). Horizontal line represents the median value of r∆Ct values for CTRL and NEC samples. *: p <0.05 between CTRL and NEC samples. Numbers indicated represent the fold variation between NEC and CTRL

Fig. 3

Comparative analysis of functional enriched pathways between necrotizing enterocolitis and Crohn’s disease. a The negative logarithm of p-values (Fisher’s test), calculated by IPA, for each of the functional categories over-represented in NEC samples was plotted against those modulated in Crohn’s disease according to published data [–25]. Canonical pathways represented by colored squares indicating the top 12 functional pathways identified in NEC are listed in Fig. 1. As shown, 11 of them are shared between NEC and CD. Insert: Venn diagram showing the 131 canonical pathways between NEC and CD. Of the 70 pathways found in NEC, 44 were also found in CD. Thresholds (dotted lines) denote the limit of statistical significance (p = 0.05 [-Log (0.05) =1.3]). b Venn diagram showing ToppCluster enrichment analysis associated with NEC patients and adult CD using phenotype terms. c Heatmap of some common genes found in the intersection of the ToppCluster enrichment analysis between NEC and adult CD as detailed in Additional file 6: Table S6

Fig. 4

Comparative analysis of functional enriched pathways between necrotizing enterocolitis and pediatric Crohn’s disease. a The negative logarithm of p-values (Fisher’s test), calculated by IPA, for each of the functional categories over-represented in NEC samples was plotted against those modulated in pedCD according to published data. [26] Canonical pathways represented by colored squares indicate the top 12 functional pathways identified in NEC as listed in Fig. 1. As shown, 9 of them are shared between NEC and CD. Insert: Venn diagram showing the 153 canonical pathways identified in NEC and pedCD. Of the 70 pathways found in NEC, 47 were also found in pedCD. Thresholds (dotted lines) denote the limit of statistical significance (p = 0.05 [-Log (0.05) =1.3]). b Venn diagram showing ToppCluster enrichment analysis associated with NEC patients and pedCD using phenotype terms. c Heatmap of some common genes found in the intersection of the ToppCluster enrichment analysis between NEC and pedCD as detailed in Additional file 8: Table S8

Fig. 5

Differential expression of antiviral response genes in human necrotizing enterocolitis. Real-time qPCR analysis of transcript levels of selected target genes related to the antiviral response. Ct values of selected genes were normalized using B2M as reference gene and data are expressed as r∆Ct values (reverse ∆Ct: Ct_{reference gene}-Ct_{gene of interest}) in order to display direct variation in NEC vs non-NEC controls (CTRL). Horizontal line represents the median value of r∆Ct values for CTRL and NEC samples. *: p <0.05 between CTRL and NEC samples. Numbers indicated represent the fold variation between NEC and CTRL