Secondary Logo

Journal Logo

Original Articles: Gastroenterology

Distinctive Colonic Mucosal Cytokine Signature in New-Onset, Untreated Pediatric Crohn Disease

Sylvester, Francisco A.*; Draghi, Andrew; Menoret, Antoine; Fernandez, Marina L.*; Wang, Zhu*; Vella, Anthony T.

Author Information
Journal of Pediatric Gastroenterology and Nutrition: November 2014 - Volume 59 - Issue 5 - p 553-561
doi: 10.1097/MPG.0000000000000480
  • Free


Inflammatory bowel disease (IBD) encompasses a spectrum of chronic inflammatory disorders that affect the intestinal tract, commonly grouped under 2 conditions: Crohn disease (CD), which can affect any part of the digestive tract, and ulcerative colitis (UC), which is restricted to the colonic mucosa. In IBD, genetically susceptible individuals influenced by environmental factors exhibit a failure to contain their intestinal microbial ecosystem and develop chronic intestinal inflammation (1). Experimental models in rodents suggest a prominent role for TH17 and TH1 cells (and their prototypical cytokines interleukin [IL]-17 and interferon-gamma [IFN-γ]) in the initiation of CD-like colitis (2,3). Clinical trials testing specific neutralizing antibodies to IFN-γ and IL-17, however, do not improve and, in some cases, worsen human CD (4,5), suggesting that the immune response in CD may differ from early events that induce colitis in mice. It is possible that the inflammatory response in CD may involve cytokines and chemokines that modify responses to IFN-γ and IL-17 or that there may be differences between human CD and experimental colitis in rodents (3,6).

Although genome-wide association studies performed in patients with CD have identified susceptibility loci that point to dysregulation of the mucosal immune response, there is a large knowledge gap between the earliest, inductive inflammatory events well described in experimental colitis in rodents versus immune events present in early CD (7). To address these gaps in knowledge, pediatric patients with newly diagnosed, untreated CD are ideal to study compared with adults because of their short disease duration, lack of associated health problems and their medical treatments, and abstinence from smoking, recreational drugs, and alcohol consumption, which can influence the course of the disease (8,9). Our primary goal was to define cytokine responses in the colon of children with newly diagnosed, untreated CD. Using colonic biopsy tissue we defined a novel cytokine network that appeared in patients newly diagnosed as having CD, which was significantly downregulated in patients with CD in the medication-induced remission. Hierarchical clustering and scatterplot proximity maps demonstrated clear separation of patients with new CD versus those in remission. Interestingly, IL-17 was significantly increased, but pathway analysis showed that tumor necrosis factor (TNF)-α and IL-6 were central pathway nodes. A striking feature of the cytokine network was the presence of innate cytokines composed of a core of chemokines that are likely important for trafficking innate cells into the colon. Thus, the use of our cytokine network will allow new approaches to treatment. One therapeutic approach is to target specific chemokines, which may prevent the trafficking of innate cells to the site of inflammation in the colon and their activation.



Eligible patients included children 7 to 18 years of age scheduled to undergo colonoscopy indicated by their treating gastroenterologist at the Connecticut Children's Medical Center to evaluate for suspected CD, for digestive symptoms (eg, abdominal pain, diarrhea, blood in the stool), or for reevaluation of known CD. Exclusion criteria included the use of systemic corticosteroids in the 3 months before the study and history of colonic resection. Potential subjects were screened by reviewing the chief complaint before the clinic visit, and, if eligible, invited to participate by a study coordinator (M.L.F.) in the office setting. We collected demographic information (age, sex, and date of CD diagnosis) and described the disease phenotype of each patient with CD according to the Paris modification of the Montreal classification (10). In addition, we used the physician global assessment, endoscopic evaluation, and the intensity of histological inflammation to grade CD activity. We excluded subjects with CD with UC, IBD-undefined (IBD-U), and subjects with polyps, eosinophilic gastroenteritis, parasites, or known infections (eg, Clostridium difficile). We recorded the medication history of each subject, with a focus on medical therapy at the time of colonoscopy. Patients underwent an intestinal lavage procedure with polyethylene glycol 3350 without electrolytes, clear fluids on demand, and no solid food the day before the procedure. Usual medications for CD were allowed during the colonoscopy prep. Patients fasted for 6 to 8 hours before colonoscopy. The day of the procedure patients received an intravenous peripheral catheter, intravenous saline, and propofol for sedation. During colonoscopy we obtained 6 biopsies from the most distal site of visible inflammation in the colon. If ulcers were present, biopsies were obtained from the margin of the ulcers. If inflammation was not grossly present in the colon, then the mid-descending colon was biopsied. Biopsies were obtained from neighboring colonic mucosa for regular histopathology. The biopsies were put in tubes containing cold RPMI 1640, and the tubes were put in a Styrofoam box with cold packs and transported to our laboratory at the University of Connecticut Health Center (typically a 15-minute drive).

CD was defined by a combination of symptoms of active disease, discontinuous colonic inflammation on colonoscopy, and characteristic findings on histopathology (11). CD was considered in remission if patients had no symptoms of active disease and had normal colon on colonoscopy by visual inspection. Histology in patients with remission ranged from completely normal to a mild increase in lamina propria mononuclear cells (LPMCs) with preserved mucosal architecture. Controls (CTRL) were children evaluated for digestive symptoms but who had no evidence of inflammatory, neoplastic, anatomical, or metabolic disease after clinical, laboratory, endoscopic, and histologic assessment.

Lamina Propria Cell Isolation

We followed previously published methods used in our laboratory for cell isolation (12). Briefly, we placed 4 biopsies in Ca/Mg-free Hanks balanced salt solution (HBSS) containing 5 mmol/L ethylenediaminetetraacetic acid and 0.15 mg/mL dithioerythritol, and stirred at 37°C. Then, biopsies were incubated at 37°C in HBSS containing 1 mmol/L CaCl2, 1 mmol/L MgCl2, 0.3 mg/mL collagenase, and 0.1 mg/mL DNase I (Sigma, St Louis, MO). Supernatants from collagenase-treated tissue were poured over cell strainers and spun down at 500g. The cells were washed twice in RPMI 1640 with L-glutamine, 10% fetal bovine serum, and 1× antibiotic/antimycotic (GIBCO, Grand Island, NY), with LPMCs partitioning at the interface. We collected the LPMCs for subsequent analyses.

Flow Cytometry

The following antibodies were used: CD3 (immunoglobulin G1 [IgG1], clone UCHT1), CD4 (IgG2b, OKT4), CD8 (IgG1, HIT8a), TNF-α (IgG1, mAb11), IFN-γ (IgG1, 4s.B3), and IL-17a (IgG1, BL168), all from Biolegend (San Diego, CA). Similar procedures as described previously were followed (13). Briefly, LPMC were resuspended in staining buffer consisting of HBSS, 3% fetal bovine serum, and 0.1% sodium azide. Nonspecific binding was blocked with a solution containing human FCR block (Miltenyi Biotec, Auburn, CA). Blocking was followed by incubation with fluorescently conjugated monoclonal antibodies on ice for 30 minutes. After surface staining, cells were washed in staining buffer, fixed with fixation buffer (BD Cytofix/Cytoperm Fixation/Permeabilization Kit; BD PharMingen, San Diego, CA), and incubated in the dark at 4°C for 20 minutes. Cells were then washed twice with BD Permeabilization/Wash buffer and then stained with cytokine antibodies in BD Permeabilization/Wash buffer at 4°C for 30 minutes in the dark. Cells were then washed twice with BD Permeabilization/Wash buffer and resuspended in staining buffer. Flow cytometry was performed on a BD LSR II flow cytometer with human peripheral blood mononuclear cells for compensation controls. We analyzed data with FlowJo Software (Tree Star, Ashland, OR).

In Vitro Lymphocyte Stimulation

LPMCs were stimulated with phorbol 12-myristate 13-acetate (PMA) 50 ng/mL (Calbiochem, EMD Chemicals, Gibbstown, NJ) and ionomycin (1 mg/mL; Sigma) for 4 hours in the presence of GolgiPlug (BD Bioscience). Control cells were treated with GolgiPlug but not stimulated with PMA/Ionomycin. Additional controls were untreated cells. For intracellular cytokine staining LPMCs were fixed and permeabilized according to the manufactures instructions using BD Cytofix/Cytoperm fixation/permeabilization kit (BD PharMingen).

Biopsy Explant Culture

One biopsy was weighed and then placed into 1 mL of fresh RPMI 1640 in a 12-well plate. The plates were incubated at 37°C in 5% CO2 atmosphere for 72 hours. We then removed, aliquoted, and froze the supernatant at −80°C until used.

Multiplex Cytokine Assay

Biopsy culture supernatants were submitted to Myriad RBM (Austin, TX) to measure 45 analytes of interest (Human InflammationMAP 1.0). Following the manufacturer's instructions, analytes with ≥50% subjects with values below the lower assay limit (the lowest quantifiable point on the calibration curve) were considered undetectable and removed from further analyses. Furthermore, analytes with ≥50% values ≤5× least detectable dose (defined by Myriad RBM as the concentration interpolated by the mean + 3 standard deviations of 20 standard diluent blank readings) were also excluded from further statistical analysis.

Statistical Analysis

Concentrations for each analyte were first log transformed to obtain a normal distribution. Analysis of variance and subsequent Tukey post hoc tests were then applied to multiple group comparisons. Statistical significance was defined as P < 0.05. Analytes found to be significant among groups were further analyzed for their ability to differentiate patient groups. Principal component analysis was performed using only analytes selected from the P < 0.05 level of significance and the resulting case-wise scores of the top 2 components were plotted onto a scatterplot proximity map (14). In these plots only the internal relations are relevant. The scale of the map represents the area required to accurately show the internal relations. Samples that are similar across their analyte measures are located near one another, whereas samples more dissimilar appear farther apart. Hierarchical clustering analysis was also conducted without 1-way analysis of variance pretest. For identification of networks we used Ingenuity Pathway Analysis (IPA) software according to the manufacturer's instructions. The IPA database includes a large collection of individual relations curated from the literature, which can be used to generate causal networks and create mechanistic hypotheses that explain the expression changes observed in datasets (15).

Ethical Considerations

The project was approved by the Connecticut Children's Medical Center institutional review board in October 2010. All of the subjects and caregivers provided informed consent for participation in the study.


Patient Characteristics

The characteristics of the 36 subjects studied: CD-New (n = 12), CD-Remission (n = 11), and CTRL (n = 13) are shown in Table 1. There were no significant differences between the mean ages of the groups (age range 7–17 years). A subgroup of consecutive patients had analysis of cytokine potential in LPMC (7 CD-New, 8 CD-Remission, 10 CTRL) and did not defer in mean age either.

Demographic characteristics of study patients

Flow Cytometry Shows No Differences in IFN-γ, TNF-α, or IL-17 Intracellular Cytokine Potential

Given previous reports of the role of TH1 and TH17 cells in experimental colitis and advanced CD in adult patients (16), we first examined the potential to secrete IFN-γ, TNF-α, and IL-17 in stimulated LPMCs isolated from colonic biopsies from children with CD-New, CD-Remission, and CTRL. As shown in Figure 1, the percentage of CD4+ IFN-γ+, TNF-α+, and IL-17+ lymphocytes was the same in all 3 study groups. We also gated on other regions of the side and forward scatter flow cytometry plots to examine cytokine potential in nonlymphocyte cells but found no difference in any of the cytokines among groups either. Therefore, after analyzing 25 consecutive samples we concluded that T-cell cytokine potential did not discriminate between subjects in our 3 study groups. To gain a more thorough picture of the ongoing inflammatory response in CD, we studied colonic mucosal explants ex vivo.

IFN-γ, IL-17, and TNF-α potential of colonic LPMCs is the same in CD-New, CD-Remission, and CTRL. A single-cell suspension obtained from the lamina propria of 4 colonic biopsies was stimulated with PMA/I and labeled as described in the Methods section. Flow cytometry was used for analysis. A, CD4+ cells were gated to examine the presence of cells positive for each intracellular cytokine. Cytokine antibody histograms (blue line) were overlaid upon their respective isotype control (red line) to determine cytokine potential in nonstimulated versus PI-stimulated cells. We did not detect any cytokines in LPMCs not treated with PMA/I. B, Each scatterplot contains the percent positive cells for each cytokine (log transformed). CD-New = new, untreated Crohn disease; CD-Remission = CD in remission; CTRL = controls; IFN = interferon; IL = interleukin; LPMC = lamina propria mononuclear cells; PMA/I = phorbol 12-myristate 13-acetate/ionomycin; TNF = tumor necrosis factor.

Innate Cytokines/Chemokines Distinguish Between CD Disease States and CD-New and CTRL

As a method to visualize the continuation of the mucosal inflammatory process, colonic biopsies were explanted into an in vitro culture without any form of stimulation and then supernatants were analyzed. RBM Myriad multiplex analysis of 45 analytes was used to measure differences that may occur between the patient groups. All of the biopsies were obtained with standard colonoscopy forceps, and the mean weight of cultured biopsies was not different among groups (8 mg in all groups). A subset of 13 analytes was significantly different between CD-New (untreated) versus CD-Remission (with normal colonoscopy, on medications), followed by CD-New versus CTRL with 10/13 differences (Table 2 and Fig. 2). There were no differences in any of the analytes between CD-Remission and CTRL.

Chemokines and cytokines that distinguish any of the 3 study groups (mean of log2 of pg/mL)
Analytes used in IPA as a result significant differences among study groups. Concentration of each analyte in picograms per milliliter was log2 transformed and plotted using GraphPad Prism 6 (GraphPad, La Jolla, CA). Means were compared by 1-way analysis of variance followed by multiple comparisons with Tukey posttest (* P < 0.05). Analytes are presented in alphabetical order. B2M = β2 microglobulin; CCL = chemokine (CC motif) ligand; CRP = C-reactive protein; CSF = colony-stimulating factor; ICAM1 = intercellular adhesion molecule 1; IL = interleukin; IPA = Ingenuity Pathway Analysis; TNF = tumor necrosis factor; TNFRSF = tumor necrosis factor receptor superfamily; VCAM1 = vascular cell adhesion molecule 1.

The following cytokines were undetectable in explant culture supernatants: brain-derived neurotrophic factor, eotaxin-1, factor VII, fibrinogen, haptoglobin, IL-2, IL-3, IL-4, IL-5, IL-7, IL-12 (p40 and p70), IL-15, IL-18, IL-23, matrix metallopeptidase (MMP)-2, stem cell factor, and TNF-β. The following factors were detectable, but we found no significant differences among the 3 study groups: α1 antitrypsin, α2 macroglobulin, chemokine (C-C motif) ligand 2, complement C3, ferritin, IL-1 receptor antagonist, IL-1α, IL-1β, IL-8, IFN-γ, MMP-3, MMP-9, tissue inhibitor of metalloproteinase-1, von Willebrand factor, vascular endothelial growth factor, and vitamin D binding protein. Thus, an unbiased cytokine approach revealed a cytokine network composed of innate cytokines and chemokines that distinguish CD-New patients from those in medication-induced remission and controls.

Pathway Analysis Reveals a Fit Between Observed Cytokines/Chemokines and Known Networks Involved in Inflammation

IPA produced 5 distinct results. First, there was an excellent fit between factors overexpressed in biopsies from children with CD-New compared with CD-Remission and physiological pathways associated with cell-to-cell signaling and interaction, cellular movement, hematological system development, and function (IPA score of 33). Second, IPA revealed a number of functional relations among the 13 analytes that we found to be different between colonic biopsies of CD-New and CD-Remission. Among these factors, TNF-α and IL-6 had the highest number of interactions with other factors, followed by colony-stimulating factor-2 (CSF-2). These results show that pathway analysis can bridge known factors involved in CD, such as TNF-α, with new targets discovered from biopsy data and computational analysis speeding the process for drug discovery (Fig. 3A). Third, the top 3 canonical pathways in IPA included differential regulation of cytokine production in macrophages, T-helper cells (P = 1.11 × 10−22), and epithelial cells (P = 2.01 × 10−18) by IL-17a and IL-17f, and communication between innate and adaptive immune cells (P = 2.12 × 10−16). Oxidative stress (P = 2.15 × 10−15) was the top toxic pathway. Fourth, the 13 analytes fit into known pathogenic pathways with high probability (presented in Table 3 using IPA pathway nomenclature). Finally, IPA also predicted the activation of specific transcription factors (5 top upstream regulators with corresponding P values were NFκβ1 [nuclear factor kappa beta] 6.15 × 10−19; HMGB1 [high-mobility group box 1] 1.47 × 10−18; STAT3 [signal transducer and activator] 2.49 × 10−16; RELA [NFκβ transcription factor p65] 1.07 × 10−14; and CEBPB [C/AATT enhancer-binding protein β] 9.71 × 10−13] (Fig. 3B).

Pathway analysis using IPA based on the concentration of analytes in colonic biopsy culture supernatants. We input the log2 ratio and P value of the difference between means of 13 factors that were different between CD-New and CD-Remission into IPA software. Panel A shows a causal network and panel B shows transcription factors predicted to be activated or inhibited in the network of colonic biopsies of CD-New. The boxes under each plot represent activation gradients (from less to more). B2M = β2 microglobulin; CCL = chemokine (CC motif) ligand; CD-New = new, untreated Crohn disease; CD-Remission = CD in remission; CRP = C-reactive protein; CTRL = controls; GM-CSF = granulocyte macrophage colony-stimulating factor; ICAM1 = intercellular adhesion molecule 1; IL = interleukin; IPA = Ingenuity Pathway Analysis; TNF = tumor necrosis factor; TNFRSF = tumor necrosis factor receptor superfamily; VCAM1 = vascular cell adhesion molecule 1.
IPA results of analyte differences between CD-New and CD-Remission

Proximity Maps and Hierarchical Clustering Analysis Segregated Groups of Clinically Relevant Patients Based on Measured Analytes

To determine whether different patients groups would segregate based on cytokine production, a principal component analysis was completed. Scatterplot proximity maps and hierarchical clustering analysis based on significantly different biopsy supernatant analytes among study groups revealed clear separation of CD-New and CD-Remission, with occasional overlap of CTRL together with CD-New (Fig. 4A and B). In contrast to reliance on a single cytokine for disease diagnosis, the mucosal cytokine network allows greater confidence in diagnosis of remission patients from those with ongoing disease and, more important, reveals which specific factors are expressed at higher levels in individual new-onset patients.

Proximity maps (A) and hierarchical clustering (B) analyses show relations between study groups based on the concentration of 13 analytes: In A, each icon represents a sample and each color a study group. In B, the more intense the color, the higher the concentration of a specific analyte in the sample. The red line represents CD-New (CDN) and the blue line CD-Remission (CDR). CTRL = controls; PCA = principal component analysis


To our knowledge, this is the first report that characterizes inflammatory pathways in newly diagnosed, untreated CD based on an unbiased analysis of proteins released by colonic biopsies. The number of samples studied was small. The differences between CD-New and CD-Remission in 13 different factors were, however, highly significant. It is possible that with additional samples we may have detected more differences among study groups, but this is unlikely given the small variance observed in factors that were not significantly different among groups. An important strength of our approach is the use of endoscopic biopsies from children with a new diagnosis of CD, as opposed to tissue from adults with a longstanding disease or resection specimens that likely reflect end-stage CD. Our analysis is based on released proteins and not messenger RNA transcripts, which may not necessarily track with protein release (17). Our observations confirm central factors such as TNF-α and IL-6, suggest novel therapeutic targets (eg, CCL5), and may provide a rationale to choose specific molecules for targeting based on the number of interactions with other factors. Ultimately our data point to an inflammatory process in colonic CD that involves complex interactions between immune cells and cytokines with similarities to the inflammatory response in rheumatoid arthritis. This is not surprising, given that rheumatoid arthritis and CD share genetic susceptibility factors (18) and medications such as anti-TNF-α antibodies (19–22) and an anti-IL-6 receptor antibody (23,24) treat both conditions effectively.

Our pathway analysis suggests a central role for TNF-α in colonic CD in children. Our findings may explain the marked therapeutic response to anti-TNF-α blocking agents in children with CD (25,26), despite the presence of a myriad of additional factors in the inflamed intestine. A second key node in the colonic inflammatory pathway, based on the number of interactions with other factors, appears to be IL-6. In addition to the anti-IL-6 receptor antibody tocilizumab, other IL-6 inhibitors are being tested to target this pathway. CSF-2 is a third node and is intriguing because treatment with CSF-2 may be effective in a subset of patients with CD (27), and anti-CSF-2 antibodies present in serum predict severe CD in children (28). On the contrary, IL-17a, although part of the CD pathway, has a paucity of interactions with other factors, which may explain why a monoclonal antibody against IL-17a, secukinumab, was ineffective against moderate-to-severe CD (5). We therefore propose that the number of interactions in the colonic inflammatory network may predict the clinical response to therapeutic agents that target factors in the pathway and may in the future allow for precision targeting of therapeutic agents to specific patients.

Our pathway analysis identified 2 subnetworks, one containing activators of innate immune cells (CCL3, CCL4, CCL5) and the other containing vascular adhesion molecules (ICAM1 and VCAM1), suggesting that innate immune activation and increased leukocyte adherence to blood vessels are important in the pathogenesis of colonic CD. Among the transcription factors predicted to be activated, NFκβ, RelA, and Stat3 are known to be activated in CD (29,30). HMGB1, on the contrary, is a target for perinuclear anti-neutrophil antibody, a diagnostic marker for UC (31). HMGB1 can be released by damaged cells (32) and act as an alarmin (33) to mediate inflammation and cellular migration (34), whereas fecal HMGB1 is a marker of active IBD in children (35). Nevertheless, our analysis uncovered CEBP as a potential mediator, but its function in CD is unknown.

Some of the factors that we detected may mediate repair responses in the inflamed colon. For example, CRP is involved in recognition of both pathogens and damaged cells and their clearance (36); CSF-2 can boost neutrophil phagocytic activity (37), enhance colonic epithelial cell repair (38), and expand regulatory T cells (39). Moreover, some factors in our network may have dual roles, depending on the cell type that secretes them. For instance, TNFR2 expression is induced in human IBD in the colonic epithelium (40) and may play a role in disease pathogenesis by impairing epithelial cell barrier function (41,42). In the lamina propria TNFR2 may be anti-inflammatory by helping to maintain a population of FoxP3+ regulatory T cells (43). Other factors such as TNF-α may have protective roles at steady-state concentrations, whereas they may be pathogenic at higher concentrations (44). Therefore, gut immune responses are complex, and immune cell function and their products are likely determined by the context in which they appear and by the presence of other factors that are also present. Perhaps the best recent example of this case is the presence of IL-6 with or without TGF-β, which determines TH17 versus Tregs (45,46). Thus, the presence of a cytokine network and their individual levels are likely key to personalizing treatment for patients.

The factors measured in biopsy explant culture may originate from the lamina propria, the colonic epithelium, or both. Detailed and well-controlled flow cytometry analysis from stimulated LPMCs in our study showed no differences in IFN-γ, IL-17, and TNF-α among the 3 study groups. This suggests that some IFN-γ, TNF-α, and IL-17 present in the biopsy culture supernatant originated from the inflamed colonic epithelium, which would explain the differences found among groups in the concentration of these cytokines in biopsy explant culture (47–49). Our flow cytometry results differ from previous studies performed in tissues from adults with longstanding IBD, suggesting that the inflammatory response may evolve in IBD as a result of age, time, treatment, and other factors.

There was some overlap in the concentration of factors between controls and children with new CD, which affected the separation of CD-New from CTRL in our principal component and hierarchical clustering analyses. Although CTRL met the definition for functional gastrointestinal disorders, recent evidence suggests that immune activation occurs in the mucosa of patients with irritable bowel syndrome (a functional gastrointestinal disorder) (50,51), so it is likely that at least some of our control patients had a degree of mucosal immune activation in their colon that we detected with our analysis. It is, however, unethical to obtain colonoscopic biopsies from healthy children with no medical indication to refine the control population. Other possible controls such as unaffected segments of colon from cancer patients have their own limitations, such as immunosuppression from the neoplastic process, and are not readily available in pediatric patients.

In conclusion, our results show that colonic immune responses at the time of CD diagnosis are complex and characterized by upregulation of vascular adhesion molecules, factors that stimulate innate immune cells, cytokines that perpetuate immune responses, attempts at tissue repair, and downregulation of novel anti-inflammatory transcription factors. The treatment for CD results in effective suppression of innate and adaptive immune responses, and in patients with remission this is reflected in mucosal immunosuppression. It is possible that cytokine networks present at the diagnosis of CD may predict the likelihood of responding to medications. There is a critical need to perform longitudinal studies starting at CD diagnosis in which cytokine networks in biopsy explant culture supernatants can be traced to effective therapy prospectively. Thus, our approach may help clinicians to tailor anti-inflammatory therapies to the individual patient by combining colonic biopsy cytokine network analysis with the available treatments.


We thank Dr Jeffrey S. Hyams for critical reading of the manuscript. We acknowledge our patients who participated in this study and our colleagues (Drs Zev Davidovics, Karan Emerick, Jeffrey S. Hyams, Franziska Mohr, Wael Sayej, Bella Zeisler, Donna Zeiter) who helped collect biopsy and blood samples from study volunteers.


1. Sun L, Nava GM, Stappenbeck TS. Host genetic susceptibility, dysbiosis, and viral triggers in inflammatory bowel disease. Curr Opin Gastroenterol 2011; 27:321–327.
2. Kosiewicz MM, Nast CC, Krishnan A, et al. Th1-type responses mediate spontaneous ileitis in a novel murine model of Crohn's disease. J Clin Invest 2001; 107:695–702.
3. Spencer DM, Veldman GM, Banerjee S, et al. Distinct inflammatory mechanisms mediate early versus late colitis in mice. Gastroenterology 2002; 122:94–105.
4. Reinisch W, de Villiers W, Bene L, et al. Fontolizumab in moderate to severe Crohn's disease: a phase 2, randomized, double-blind, placebo-controlled, multiple-dose study. Inflamm Bowel Dis 2010; 16:233–242.
5. Hueber W, Sands BE, Lewitzky S, et al. Secukinumab, a human anti-IL-17A monoclonal antibody, for moderate to severe Crohn's disease: unexpected results of a randomised, double-blind placebo-controlled trial. Gut 2012; 61:1693–1700.
6. Kugathasan S, Saubermann LJ, Smith L, et al. Mucosal T-cell immunoregulation varies in early and late inflammatory bowel disease. Gut 2007; 56:1696–1705.
7. Jostins L, Ripke S, Weersma RK, et al. Host-microbe interactions have shaped the genetic architecture of inflammatory bowel disease. Nature 2012; 491:119–124.
8. Boyko EJ, Koepsell TD, Perera DR, et al. Risk of ulcerative colitis among former and current cigarette smokers. N Engl J Med 1987; 316:707–710.
9. Ernst A, Jacobsen B, Ostergaard M, et al. Mutations in CARD15 and smoking confer susceptibility to Crohn's disease in the Danish population. Scand J Gastroenterol 2007; 42:1445–1451.
10. Levine A, Griffiths A, Markowitz J, et al. Pediatric modification of the Montreal classification for inflammatory bowel disease: the Paris classification. Inflamm Bowel Dis 2011; 17:1314–1321.
11. Bousvaros A, Antonioli DA, Colletti RB, et al. Differentiating ulcerative colitis from Crohn disease in children and young adults: report of a working group of the North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition and the Crohn's and Colitis Foundation of America. J Pediatr Gastroenterol Nutr 2007; 44:653–674.
12. McAleer JP, Saris CJ, Vella AT. The WSX-1 pathway restrains intestinal T-cell immunity. Int Immunol 2011; 23:129–137.
13. McAleer JP, Zammit DJ, Lefrancois L, et al. The lipopolysaccharide adjuvant effect on T cells relies on nonoverlapping contributions from the MyD88 pathway and CD11c+ cells. J Immunol 2007; 179:6524–6535.
14. Han J, Kamber M, Pei J. Data preprocessing. In: Data Mining. 3rd ed. Boston: Morgan Kaufmann; 2012:83–124.
15. Kramer A, Green J, Pollard J Jr, et al. Causal analysis approaches in Ingenuity Pathway Analysis. Bioinformatics 2014; 30:523–530.
16. Strober W, Fuss IJ. Proinflammatory cytokines in the pathogenesis of inflammatory bowel diseases. Gastroenterology 2011; 140:1756–1767.
17. The Human Protein Atlas Published 2013. Accessed February 18, 2014.
18. Lee JC, Espeli M, Anderson CA, et al. Human SNP links differential outcomes in inflammatory and infectious disease to a FOXO3-regulated pathway. Cell 2013; 155:57–69.
19. Elliott MJ, Maini RN, Feldmann M, et al. Randomised double-blind comparison of chimeric monoclonal antibody to tumour necrosis factor alpha (cA2) versus placebo in rheumatoid arthritis. Lancet 1994; 344:1105–1110.
20. den Broeder A, van de Putte L, Rau R, et al. A single dose, placebo controlled study of the fully human anti-tumor necrosis factor-alpha antibody adalimumab (D2E7) in patients with rheumatoid arthritis. J Rheumatol 2002; 29:2288–2298.
21. Hanauer SB, Sandborn WJ, Rutgeerts P, et al. Human anti-tumor necrosis factor monoclonal antibody (adalimumab) in Crohn's disease: the CLASSIC-I trial. Gastroenterology 2006; 130:323–333.
22. Targan SR, Hanauer SB, van Deventer SJ, et al. A short-term study of chimeric monoclonal antibody cA2 to tumor necrosis factor alpha for Crohn's disease. Crohn's Disease cA2 Study Group. N Engl J Med 1997; 337:1029–1035.
23. Maini RN, Taylor PC, Szechinski J, et al. Double-blind randomized controlled clinical trial of the interleukin-6 receptor antagonist, tocilizumab, in European patients with rheumatoid arthritis who had an incomplete response to methotrexate. Arthritis Rheum 2006; 54:2817–2829.
24. Ito H, Takazoe M, Fukuda Y, et al. A pilot randomized trial of a human anti-interleukin-6 receptor monoclonal antibody in active Crohn's disease. Gastroenterology 2004; 126:989–996.
25. Hyams J, Walters TD, Crandall W, et al. Safety and efficacy of maintenance infliximab therapy for moderate-to-severe Crohn's disease in children: REACH open-label extension. Curr Med Res Opin 2011; 27:651–662.
26. Hyams JS, Griffiths A, Markowitz J, et al. Safety and efficacy of adalimumab for moderate to severe Crohn's disease in children. Gastroenterology 2012; 143:365–374.e2.
27. Korzenik JR, Dieckgraefe BK, Valentine JF, et al. Sargramostim for active Crohn's disease. N Engl J Med 2005; 352:2193–2201.
28. Gathungu G, Kim MO, Ferguson JP, et al. Granulocyte-macrophage colony-stimulating factor autoantibodies: a marker of aggressive Crohn's disease. Inflamm Bowel Dis 2013; 19:1671–1680.
29. Ellis RD, Goodlad JR, Limb GA, et al. Activation of nuclear factor kappa B in Crohn's disease. Inflamm Res 1998; 47:440–445.
30. Lovato P, Brender C, Agnholt J, et al. Constitutive STAT3 activation in intestinal T cells from patients with Crohn's disease. J Biol Chem 2003; 278:16777–16781.
31. Sobajima J, Ozaki S, Osakada F, et al. Novel autoantigens of perinuclear anti-neutrophil cytoplasmic antibodies (P-ANCA) in ulcerative colitis: non-histone chromosomal proteins, HMG1 and HMG2. Clin Exp Immunol 1997; 107:135–140.
32. Willingham SB, Allen IC, Bergstralh DT, et al. NLRP3 (NALP3, Cryopyrin) facilitates in vivo caspase-1 activation, necrosis, and HMGB1 release via inflammasome-dependent and -independent pathways. J Immunol 2009; 183:2008–2015.
33. Lamkanfi M, Sarkar A, VandeWalle L, et al. Inflammasome-dependent release of the alarmin HMGB1 in endotoxemia. J Immunol 2010; 185:4385–4392.
34. Wang H, Yang H, Tracey KJ. Extracellular role of HMGB1 in inflammation and sepsis. J Intern Med 2004; 255:320–331.
35. Vitali R, Stronati L, Negroni A, et al. Fecal HMGB1 is a novel marker of intestinal mucosal inflammation in pediatric inflammatory bowel disease. Am J Gastroenterol 2011; 106:2029–2040.
36. Volanakis JE. Human C-reactive protein: expression, structure, and function. Mol Immunol 2001; 38:189–197.
37. Han X, Uchida K, Jurickova I, et al. Granulocyte-macrophage colony-stimulating factor autoantibodies in murine ileitis and progressive ileal Crohn's disease. Gastroenterology 2009; 136:1261–1271.e1–3.
38. Bernasconi E, Favre L, Maillard MH, et al. Granulocyte-macrophage colony-stimulating factor elicits bone marrow-derived cells that promote efficient colonic mucosal healing. Inflamm Bowel Dis 2010; 16:428–441.
39. Bhattacharya P, Gopisetty A, Ganesh BB, et al. GM-CSF-induced, bone-marrow-derived dendritic cells can expand natural Tregs and induce adaptive Tregs by different mechanisms. J Leukoc Biol 2011; 89:235–249.
40. Mizoguchi E, Mizoguchi A, Takedatsu H, et al. Role of tumor necrosis factor receptor 2 (TNFR2) in colonic epithelial hyperplasia and chronic intestinal inflammation in mice. Gastroenterology 2002; 122:134–144.
41. Wang F, Schwarz BT, Graham WV, et al. IFN-gamma-induced TNFR2 expression is required for TNF-dependent intestinal epithelial barrier dysfunction. Gastroenterology 2006; 131:1153–1163.
42. Su L, Nalle SC, Shen L, et al. TNFR2 activates MLCK-dependent tight junction dysregulation to cause apoptosis-mediated barrier loss and experimental colitis. Gastroenterology 2013; 142:407–415.
43. Chen X, Wu X, Zhou Q, et al. TNFR2 is critical for the stabilization of the CD4+Foxp3+ regulatory T cell phenotype in the inflammatory environment. J Immunol 2013; 190:1076–1084.
44. Kaiser GC, Polk DB. Tumor necrosis factor alpha regulates proliferation in a mouse intestinal cell line. Gastroenterology 1997; 112:1231–1240.
45. Veldhoen M, Hocking RJ, Atkins CJ, et al. TGFbeta in the context of an inflammatory cytokine milieu supports de novo differentiation of IL-17-producing T cells. Immunity 2006; 24:179–189.
46. Mucida D, Park Y, Kim G, et al. Reciprocal TH17 and regulatory T cell differentiation mediated by retinoic acid. Science 2007; 317:256–260.
47. Ou G, Baranov V, Lundmark E, et al. Contribution of intestinal epithelial cells to innate immunity of the human gut-studies on polarized monolayers of colon carcinoma cells. Scand J Immunol 2009; 69:150–161.
48. Ohkusa T, Yoshida T, Sato N, et al. Commensal bacteria can enter colonic epithelial cells and induce proinflammatory cytokine secretion: a possible pathogenic mechanism of ulcerative colitis. J Med Microbiol 2009; 58 (pt 5):535–545.
49. Xue X, Ramakrishnan S, Anderson E, et al. Endothelial PAS domain protein 1 activates the inflammatory response in the intestinal epithelium to promote colitis in mice. Gastroenterology 2013; 145:831–841.
50. Camilleri M, Lasch K, Zhou W. Irritable bowel syndrome: methods, mechanisms, and pathophysiology. The confluence of increased permeability, inflammation, and pain in irritable bowel syndrome. Am J Physiol Gastrointest Liver Physiol 2012; 303:G775–G785.
51. Matricon J, Meleine M, Gelot A, et al. Review article: associations between immune activation, intestinal permeability and the irritable bowel syndrome. Aliment Pharmacol Ther 2012; 36:1009–1031.

children; colon; Crohn disease; immune responses; inflammatory bowel disease; innate

© 2014 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology,