ACTIVATION OF THE INNATE IMMUNE SYSTEM IN CHILDREN WITH IRRITABLE BOWEL SYNDROME EVIDENCED BY INCREASED FECAL HUMAN β-DEFENSIN-2

Abstract

Background & Aims:

The role of the innate immune system in functional gastrointestinal pain disorders is unclear. We investigated the role of β-defensin-2 and gut permeability in childhood irritable bowel syndrome (IBS) and functional abdominal pain (FAP) symptom generation.

Methods:

Fecal β-defensin-2 (and in a subset, gut permeability) was measured in children with IBS (n=116), FAP (n=33) and healthy control (HC) children (n=72). IBS and FAP patients were recruited from tertiary and primary care and HC from primary care.

Results:

β-defensin-2 concentration was greater in IBS (P=0.003) and FAP (P=0.03) than in HC. β-defensin-2 was greater in IBS girls than in HC girls (P=0.007) and in IBS girls vs IBS boys (P=0.036). There was no difference by sex in the FAP and HC groups. For the whole cohort β-defensin-2 correlated with multiple pain symptoms. In the IBS group, β-defensin-2 correlated with pain interference (P=0.014). No correlation with pain was found in the FAP or HC group. Gut permeability was greater in the IBS vs the FAP and HC (P=0.038). For the whole cohort permeability correlated with number of pain episodes (P=0.041) and interfering pain episodes (P=0.049). For the cohort there was correlation between β-defensin-2 and permeability (P=0.003) with borderline correlation in the IBS group (P=0.086). For the cohort and IBS and HC groups number of bowel movements was modestly inversely related to fecal β-defensin-2 concentrations.

Discussion:

Increased fecal β-defensin-2 concentration in children with IBS suggests activation of the innate immune system in some, which along with increased gut permeability, appears related to abdominal pain symptoms. Sex is an important variable in interpreting β-defensin-2 concentration in IBS.

INTRODUCTION

Functional abdominal pain disorders, now termed disorders of gut–brain interaction, affect up to 15-20% of school age children and adults world-wide.^1,2 Two common subtypes include irritable bowel syndrome (IBS) and functional abdominal pain (FAP). FAP is characterized by intermittent abdominal pain, often with significant disability, and when associated with an altered stooling pattern is denoted as IBS.³

While the etiology of IBS or FAP is not known, several factors are thought to contribute (e.g., impaired gut barrier function, visceral hypersensitivity, dietary factors, psychosocial vulnerability).^4-6 Recently, attention has been focused on antimicrobial peptides such as defensins that are involved in innate immune function and act at the mucosal surface conferring epithelial barrier function as an adjunct to the adaptive immune system.^{7, 8} Alterations in defensins have been described in mucosal GI diseases such as inflammatory bowel disease.⁹ Whether they are altered in IBS or FAP has received little attention. One study in adults observed that fecal β-defensin-2 was increased in both IBS and ulcerative colitis compared with healthy adults.¹⁰ Our primary aim was to investigate levels of fecal β-defensin-2 as a marker of innate immune system activation in children with IBS and FAP and compare them to healthy control children (HC). We also sought to explore the potential relationship of fecal β-defensin-2 to IBS and FAP symptoms and to identify possible relationships between β-defensin levels and changes in gut barrier function as measured by intestinal permeability.

METHODS

Participants and Measures

Children were recruited as part of a study (2013-2018) attempting to identify biomarkers to understand the pathophysiology of childhood IBS and FAP and to develop novel and innovative ways of defining IBS and FAP that are not solely symptom-based. Consequently, sample size was based on this convenience sample. As previously described, prepubertal children with IBS, FAP and HC were recruited simultaneously from a large healthcare network based in Houston, TX consisting of primary and tertiary care.¹¹ HC were recruited from the primary care groups within the network. Informed consent was obtained from parents and assent from the children. The study was approved by the Baylor College of Medicine Institutional Review Board.

Medical charts were reviewed by trained research coordinators for ICD-9 codes for abdominal pain and IBS or well child visits (e.g., immunizations, school check-ups) for children 7-12 years of age. Secondary review of medical and laboratory records was performed, and subjects then were screened via telephone and initially classified (IBS, FAP or HC) using a modified pediatric Rome III questionnaire.^{12, 13}

Children were excluded if chart review or screening revealed a significant chronic medical condition (e.g., celiac disease, cystic fibrosis), chronic vomiting, unexplained weight loss, hematochezia, major GI tract surgery, significant developmental delay, an organic GI disorder, use of antibiotics or probiotics within the prior month, and/or menarche (in order to study a more developmentally homogeneous group).

Study participants fulfilling these screening criteria maintained daily validated pain diaries for 2 weeks.¹⁴ During this time they also recorded Bristol stool form.¹⁵ Children recorded the number and severity of pain episodes (the latter on a 0-10 scale with 10 being the most severe) and whether the pain interfered with activities (0-4 scale with 4 the most interference). Mean maximum pain was calculated as the average of the worst pain rating on each of the 14 days. IBS subtype was identified for each participant using the Bristol Stool Form Scale as previously described.¹⁶ The percent of children having constipation was based on those having ≤ 2 stools per week.¹⁷ Differentiation of IBS, FAP and HC was based on the 2-week pain and stooling diary as we have described.¹³

The families received detailed instructions on how to carry out the urine and stool collection using a sterile, self-sealing container that fit over the toilet. The stool collection for β-defensin and the permeability test (urine collection) were carried out on the same weekend which was during the time the participants kept the pain and stool diary. The stool and urine samples were maintained at −20°C and transferred to our laboratory by courier within 24 hours.

Defensin Analyses

Stools were analyzed in a blinded fashion for β-defensin-2 using ELISA. Stools (15 mg) first were extracted using 1.5 mL universal extraction buffer (30-EZEX-100, ALPCO Inc., Salem, NH). The supernatant was diluted 1:2 with wash buffer and 100 μL was added per well using appropriate standards and controls (30-BD2HU-E01, ALPCO Inc., Salem, NH). Absorption was determined at 450 nm (30-BD2HU-E01, ALPCO Inc., Salem, NH). Each β-defensin-2 kit was of the same lot number. Some samples were analyzed in multiple runs in order to assess any run-to-run variability. In such cases, the mean of the values from the runs was used in the analysis.

Intestinal Permeability Testing and Sugar Analyses

The permeability data was published previously as a cohort of a larger sample.¹⁸ The methodology has been described in detail previously.^11,27 In brief, participants fasted overnight and upon arising they voided and then ingested the permeability sugar solution that contained lactulose (5g) and mannitol (1g).^{19, 20} They then drank an additional 240 mL of water. They continued to fast for three hours during the urine collection.

Percent urinary recovery of sugar from 0-3 hours was determined by multiplying the concentration of sugar in the urine by the urine volume. The amount of excreted sugar then was divided by the mg of sugar ingested to calculate the urinary lactulose/mannitol ratio. Sugars were analyzed as we have described using the method of Catassi et al.²¹ The assay is sensitive to 1 μg/mL for lactulose and mannitol. All samples were validated against LC/MSMS technology from ARUP Laboratories (Salt Lake City, UT) with a correlation of r>0.9.

Data Analysis

Data are presented as mean ± SD. Because the permeability data were not normally distributed, they were log transformed for analysis but the raw data are provided. Differences in fecal β-defensin and the lactulose/mannitol ratio among and between groups were tested using ANOVA followed by Student's t test. Associations between fecal β-defensin concentrations, the lactulose/mannitol ratio, and clinical data were evaluated using Pearson correlation or Spearman’s rho as appropriate. Chi-square was used to evaluate categorical variables. P values < 0.05 were considered significant.

RESULTS

Clinical Measures

The sexes, ages, and race/ethnicities were well-matched among the groups (Table 1). As anticipated, abdominal pain symptoms were greater in the IBS and FAP groups versus HC (Table 2). There was a small but significant difference in the number of stools passed during the two weeks, being greater in the IBS group (Table 2). However, the percentage of participants with constipation (≤ 2 stools passed per week) was similar among the groups (Table 2). The percentages of the Bristol Stool Form Scale stool types were similar among the groups except for Type 6 which was more frequent in the IBS group (Table 2). Within the IBS group, 50 (43.1%) were IBS-constipation, 11 (9.5%) were IBS-diarrhea, 5 (4.3%) were IBS-mixed, and 50 (43.1%) were IBS-unspecified based on Bristol Stool Form.¹⁶

Table 1.

Demographics of Participants

	IBS	FAP	Healthy Controls
Number	116	33	72	P Value
Sex
Female	69 (60%)	20 (61%)	35 (49%)	0.29
Male	47 (40%)	13 (39%)	37 (51%)
Age (yrs.)*	9.5 ± 1.4	9.4 ± 1.6	9.7 ± 1.5	0.50
Race/Ethnicity				0.44
White	85	21	53
Black	18	7	12
Asian	8	0	0
Native
American/Hawaiian	0	1	0
Mixed or unidentified	5	4	7
Hispanic	31	7	20	0.81

^*Mean ± SD

Table 2.

Clinical Characteristics of Irritable Bowel Syndrome, Functional Abdominal Pain, and Healthy Control Groups

	IBS n=116	FAP n=33	HC n=72	P Value+	IBS vs FAP	IBS vs HC	FAP vs HC
Abdominal Pain
Number of pain episodes in 2 weeks	12.7 ± 9.0*	11.4 ± 12.0	0.4 ± 0.6	<0.001	0.58	<0.001	<0.001
Mean pain rating (0-10)	3.1 ± 1.3	3.0 ± 1.4	1.2 ± 2.0	<0.001	0.57	<0.001	<0.001
Number of interfering pain episodes	6.0 ± 6.6	6.2 ± 8.6	0.2 ± 0.5	<0.001	0.88	<0.001	<0.001
Pain interference rating (0-4)	0.56 ± 0.46	0.66 ± 0.60	0.18 ± 0.49	<0.001	0.36	<0.001	<0.001
Mean of maximum daily pain rating	3.3 ± 1.5	3.3 ± 1.6	1.2 ± 2.0	<0.001	0.80	<0.001	<0.001
Stooling
Number of bowel movements	13.2 ± 5.0	10.9 ± 4.6	11.6 ± 4.4	0.015	0.015	0.024	0.44
Percent with constipation†	3.5	6.1	4.2	0.80
Type 1#	7.8 ± 14.3	5.4 ± 13.3	3.9 ± 8.5	0.11
Type 2	19.9 ± 18.9	16.9 ± 20.7	20.2 ± 23.0	0.72
Type 3	30.8 ± 23.0	33.9 ± 26.5	37.3 ± 28.3	0.23
Type 4	20.1 ± 20.2	26.7 ± 26.4	24.9 ± 26.2	0.27
Type 5	11.5 ± 17.2	10.7 ± 14.8	10.8 ± 18.5	0.95
Type 6	7.5 ± 11.3	5.9 ± 14.7	2.3 ± 5.8	0.005	0.56	<0.001	0.19
Type 7	2.4 ± 9.4	1.7 ± 9.5	0.7 ± 3.0	0.33

^*Mean ± SD; values in bold are significant

⁺ANOVA

^†≤ 2 stools passed per week

^#Bristol Stool Form Rating (Percent); Type 1 and 2 are considered constipation; Types 3-5 are considered normal; Types 6 and 7 are considered diarrhea for purposes of defining IBS subtypes (i.e., constipation, diarrhea, mixed, unspecified). See references 15 and 16.

Fecal β-Defensin Concentrations

Fecal β-defensin concentration differed among the groups (P=0.015) being greater in the IBS versus HC group (77.4 ± 55.8 vs 57.3 ± 35.2, respectively; P=0.003) but not the FAP group (80.9 ± 55.3, P=0.75). Fecal β-defensin concentration was greater in the FAP vs HC group (P=0.03).

For the entire cohort, concentration was greater in girls than in boys (78.2 ± 56.4 vs 62.9 ± 43.8, P= 0.022). This difference appeared to be driven by girls in the IBS group where fecal β-defensin concentration was greater in girls than in boys (86.3 ± 60.2 vs 64.2 ± 46.1, P= 0.028; Figure) and greater in IBS girls compared with HC girls (P=0.007, Figure). There was no difference by sex in the FAP or HC groups. No other differences were noted between groups by sex (Figure).

FIGURE. Fecal β-defensin-2 Concentrations in Children with Irritable Bowel Syndrome, Functional Abdominal Pain and Healthy Controls.

Fecal β-defensin-2 concentration was greater in the IBS versus HC group (77.4 ± 55.8 vs 57.3 ± 35.2, respectively; P=0.003) but not the FAP group (80.9 ± 55.3, P=0.75) which differed from HC (P=0.03). For the entire cohort, concentration was greater in girls than in boys (78.2 ± 56.4 vs 62.9 ± 43.8, P= 0.022). In the IBS group fecal β-defensin-2 concentration was greater in girls than in boys (86.3 ± 60.2 vs 64.2 ± 46.1, P= 0.036) but not in the FAP or HC groups. Fecal β-defensin-2 concentration was greater in IBS girls than in HC girls.

Two subgroups of the IBS cohort were examined – those with fecal β-defensin concentrations greater than or less than 2 standard deviations of the HC group. There was a trend for the pain interference rating to be greater in those with β-defensin concentrations greater than 2 standard deviations (n=20) compared with those who concentrations fell within 2 standard deviations of the HC (0.76 ± 0.55 vs 0.51 ± 0.43, respectively; P=0.07). No other differences were noted in pain symptoms.

There were too few children in the IBS-diarrhea and IBS-mixed groups to evaluate whether fecal β-defensin differed among all IBS subgroups. When comparing just the IBS-constipation and IBS-unsubtyped groups no differences were found (76.9 ± 54.7 ng/g vs 75.7 ± 75.0, respectively; P=0.92).

Intestinal Permeability

The sample size in each group that had permeability measured previously was: IBS, n=87; FAP, n=22; HC, n=49; total n=158). Similar to our previous report on permeability, the lactulose/mannitol ratio was greater in the IBS group than the HC (0.120 ± 0.181 vs 0.0720 ± 0.088, respectively; P=0.038).¹⁸ Similarly, it was greater in the IBS group than in the FAP group (0.075 ± 0.053, P=0.046).¹⁸ The lactulose/mannitol ratio did not differ between the FAP and HC groups (P=0.86). Similarly, lactulose/mannitol ratio did not differ by sex among the groups (data not shown).

Correlation of Fecal β-defensin-2 Concentration with Symptoms

For the entire cohort, fecal β-defensin-2 concentration correlated modestly but significantly with pain symptoms (Table 3). In the IBS group, β-defensin-2 concentration correlated modestly but significantly (P=0.014) with the pain interference rating (Table 3). In the FAP group β-defensin-2 concentration approached significance for mean of the maximum daily pain rating (P=0.057). Fecal β-defensin-2 concentration did not correlate with any pain symptom in the HC group (Table 3).

Table 3.

Correlation of Fecal β-defensin-2 with Pain Symptoms

	Whole Cohort	IBS	FAP	Healthy Controls
Clinical Measure	n=221	n=116	n=33	n=72
Number of pain episodes in 2 weeks	0.133 (0.047*)	−0.021 (0.821)	0.199 (0.266)	0.079 (0.509)
Mean pain rating (0-10)	0.177 (0.008)	0.147 (0.115)	0.248 (0.163)	0.009 (0.943)
Mean of maximum daily pain rating	0.198 (0.003)	0.151 (0.106)	0.334 (0.057)	0.016 (0.895)
Number of interfering pain episodes	0.209 (0.002)	0.143 (0.126)	0.2222 (0.215)	0.011 (0.926)
Pain interference rating (0-4)	0.224 (0.001)	0.228 (0.014)	0.289 (0.102)	−0.010 (0.936)

^*Pearson correlation (P value)

Values in bold are significant

For the subgroup of IBS with fecal β-defensin concentrations greater than 2 standard deviations of the HC group, there was no correlation between β-defensin and pain symptoms (data not shown); likely because of the smaller sample size (n=20).

For the entire cohort, the number of bowel movements was modestly inversely related to fecal β-defensin concentrations (−0.148, P=0.027). This same modest correlation was noted in the IBS group (−0.194, P=0.037) and in the HC group (−0.284, P=0.016) but not in the FAP group. There was no other correlation with stooling characteristics (i.e., percent constipation, normal, diarrhea or with Bristol Stool Scale type; data not shown).

Correlation of Intestinal Permeability with Symptoms

For the whole cohort the urinary lactulose/mannitol ratio correlated with the number of pain episodes (r=0.163, P=0.041) and the number of interfering pain episodes (r=0.157, P=0.049). There was a borderline correlation with the number of pain days (r=0.155, P=0.052). No other correlations with symptoms were noted. There were no significant correlations between intestinal permeability and symptoms in the individual IBS, FAP, or HC groups. Stooling characteristics did not correlate with the lactulose/mannitol ratio for the entire cohort or in any group (data not shown).

Correlation of Fecal β-defensin-2 Concentration with Permeability

For the whole cohort there was a modest correlation between β-defensin and the lactulose/mannitol ratio (r=0.163, P=0.003). For the IBS group there was a trend toward a correlation (r=0.181, P=0.086). There was no correlation between β-defensin and the lactulose/mannitol ratio for the FAP or HC groups (data not shown).

DISCUSSION

Our data demonstrate for the first time to our knowledge that fecal β-defensin-2 is elevated in children with IBS compared to children with FAP and HC; the elevation in the IBS group appeared to be due to differences between sexes in β-defensin-2 concentration; girls with IBS having higher fecal concentrations than boys. For the entire cohort β-defensin-2 concentration correlated positively with measures of abdominal pain; to a lesser degree this was true in the IBS group. Along these lines, there was a trend (P=0.07) for the pain interference rating to be greater in those with β-defensin concentrations greater than 2 standard deviations above the HC group.

Human defensins are expressed on nearly every surface of the human body and in leukocytes.^{7, 8} As part of the innate immune system, these antimicrobial peptides act at the mucosal surface conferring epithelial barrier function.^{7, 8} Defensins are categorized into three subfamilies based on region-specific disulfide bond linkages; both α- and β-defensin being found in humans.²² Whereas α-defensins primarily are found in high concentrations in neutrophils, β-defensins are expressed in epithelial cells (gut, respiratory tract, etc.).²³ To date, four β-defensins have been described.²² Many of the actions of the β-defensins are overlapping but others are unique to the subtype (e.g., β-defensin-2 activated by proinflammatory cytokines whereas β-defensin-1 constitutively expressed).⁸

We chose to focus on β-defensin-2, as a previous report demonstrated increased fecal concentrations in adult IBS and ulcerative colitis patients compared with HC; levels were comparable between IBS and ulcerative colitis.¹⁰ The levels of β-defensin-2 in our pediatric study in the IBS and HC groups are comparable to what was reported by Langhorst et al.¹⁰ However, in the adult report, associations between β-defensin-2 concentrations and IBS symptoms were not examined.¹⁰ Similarly, a possible role of sex on the expression of β-defensin-2 was not studied.

The finding of increased β-defensin-2 in children with IBS aligns with the evidence that in some patients the disorder may be associated with low grade gastrointestinal mucosal inflammation (immune dysfunction), often evidenced by increased mast cells or mast cell activation.^{4, 24-26} However, most studies have focused on the adaptive immune system. To our knowledge, the study by Langhorst et al. in adults and our study in children are the first to describe the potential involvement of the innate immune system in IBS as measured by β-defensin-2.¹⁰ In addition to their antimicrobial activity, defensins can recruit a multitude of immune cells with the potential of upregulating and downregulating adaptive immunity.^{8, 27} Thus, there is synergy between the innate and adaptive immune systems.⁸

In patients with IBS and gut immune dysfunction the etiology remains elusive. In some individuals, impaired gut barrier function may predispose to gut immune dysfunction, which may be related to symptom generation in IBS.^28-30 We as well as others have shown that gastrointestinal permeability is increased in some children and adults with IBS.^{4, 18, 31} Thus, we evaluated gut barrier function in our participants using the urinary recovery of lactulose and mannitol which generally is considered a measure of small intestinal permeability although that remains debated.^{32,33, 34}

As we have reported previously, the lactulose/mannitol ratio was greater in the IBS group compared with children with FAP and HC.¹⁸ Like fecal β-defensin-2, permeability correlated with some measures of pain symptoms. For the whole cohort there was a modest correlation between β-defensin-2 and intestinal permeability and a similar trend for the IBS group. These results suggest that changes in fecal β-defensin-2 and gut permeability may be related but the association is not strong.

The reason(s) for the differences between boys and girls in fecal β-defensin-2 concentrations is unclear although we previously have shown sex-related differences in immune markers (granins) and in gut permeability in children with IBS.^{18, 35} A study in adults suggested no sex-related difference in β-defensin-2 gene expression from skin epithelium.³⁶ The relevance of this finding to our report is unclear given the differences in age and the fact that, as noted above, β-defensin-2 is produced in multiple sites within the body. An in vitro study suggested that estrogen can inhibit the secretion of α-defensins 1-3 by isolated dendritic cells but whether this is true for other cell types and β-defensin-2 remains to be clarified.³⁷

There are some limitations to our study. Given the multiple comparisons testing the correlations between β-defensin concentration and abdominal pain symptoms, the results should be seen as exploratory. The sample size did not permit clear delineation as to whether IBS-subtype affected β-defensin level. Although we chose to examine β defensin-2 based on the previous report in adults, it is possible that other defensins may show a stronger (or lessor) relationship with IBS symptoms.

Strengths of the study include the rigorous phenotyping to define IBS, FAP, and HC and IBS-subtype and the prospective use of diaries for IBS symptoms and stooling characteristics which has been shown to be superior to questionnaires-based on recall.^{13, 38} In addition, we had a relatively large sample size in the IBS and HC groups and relatively robust sample size in the FAP group. Finally, we examined a developmental homogeneous group of children.

In summary, the results of our study are the first to suggest a role of the innate immune system in IBS symptom generation as measured by increased fecal β-defensin-2 concentrations although our findings require confirmation. Its role in the generation of symptoms in FAP appears less clear and a larger sample size may be needed to definitively address this question. Sex is an important variable to consider in assessing the role of fecal β-defensin-2 concentrations in IBS in children. The increase in β-defensin-2 concentrations may be associated with increased gut permeability but further studies are needed.

What You Need to Know:

Background

• Gut immune dysfunction and increased gut permeability may be related and play a role in the pathogenesis of irritable bowel syndrome (IBS) and possibly, functional abdominal pain (FAP)

• The role of antimicrobial peptides, such as β-defensins, a component of the innate immune system is unclear

Findings

• Fecal β-defensin-2 concentration and gut permeability are increased in children with IBS compared with children with FAP and healthy controls

• β-defensin-2 concentration correlates with abdominal pain symptoms

• There is a borderline positive relationship between fecal β-defensin-2 concentration and gut permeability

Implications for Patient Care

• Activation of the gut innate immune system as measured by fecal β-defensin-2 may play a role in generating abdominal pain symptoms in children with IBS and FAP

• We speculate that reducing gut permeability through pharmacological or dietary means may blunt gut immune dysfunction as measured by β-defensin-2

• Abbreviations:

IBS

irritable bowel syndrome

FAP

functional abdominal pain

GI

gastrointestinal

HC

healthy controls