Introduction

Inflammatory bowel disease (IBD) may also develop in neonatal or infantile periods in less than 1% of pediatric patients.1, 2 Clinical characteristics of infantile-onset, or very early-onset IBD (VEO-IBD) seem to be different from those of adult-onset or adolescent-onset IBD. Data about VEO-IBD reported severe clinical course and a high rate of resistance to immunosuppressive treatment.3, 4 VEO-IBD is thought to be important to investigate the pathophysiology of IBD, because IBD that starts earlier in life may have a strong association with genetic background. The Paris classification reflected clinical practice of pediatric IBD. It just subdivided the Montreal classification A1 (<17 years) into A1a, which represents 0–9 years of age, and A1b, which represents ages 10∼16 years.5

Genome wide association studies have revealed the genetic complexity of IBD and discovered 163 candidate loci.6 However, most of them are focused to the study of common variants and may overlook the rare but loss-of-function mutations. VEO-IBD patients are very select population, and they may carry low-frequency variants. Next generation exome sequencing can overcome the limitation of genome wide association study and can identify rare disease causing variants.7 Glocker et al.8 performed genetic linkage analysis and candidate gene sequencing to identify Mendelian mutations of interleukin-10 receptor A (IL-10RA) and IL-10RB in four children with VEO-IBD. Subsequently, mutations of IL-10RA and IL-10RB have been reported. However, there are still limited data, particularly in Asians.9, 10, 11, 12, 13 We recently identified IL-10RA mutations in three children with ‘intractable ulcerating enterocolitis in infancy’ and neonatal onset Crohn’s disease (CD).14

Interleukin-10 (IL-10) is an anti-inflammatory cytokine that is critical for immune homeostasis in the intestine. IL-10 signals are initiated by its binding to the IL-10 receptor, which consists of two α subunits of IL-10R1 (encoded by IL-10RA), and two β subunits of IL-10R2 (encoded by IL-10RB). This complex activates the Janus kinase (Jak) 1 and tyrosine kinase (Tyk) 2, leading to phosphorylation of signal transducer and activator of transcription 3 (STAT3) and gene transcription. IL-10-mediated signaling restricts inappropriate Th17 cell expansion, which is associated with CD.15 IL-10- and IL-10RB-deficient mice develop spontaneous chronic enterocolitis.16

The aim of this study is to identify monogenic defects in genes involved in the IL-10 signaling pathway in children with IBD, and to investigate whether such mutations contribute to the phenotypes of IBD, especially the difference observed between VEO-IBD and IBD with the onset of older age.

Materials and methods

Forty Korean children diagnosed with CD or ulcerative colitis (UC) before genetic study below the age of 10 years (Paris classification A1a) at Seoul National University Children’s Hospital were enrolled. Diagnosis of IBD was based on standard tools including clinical features, endoscopy with biopsies and adequate imaging of the small bowel. Thirty-seven children were diagnosed with CD, two with UC and one with ‘intractable ulcerating enterocolitis’. Follow-up periods were 2∼18 years. We subdivided groups into very early-onset IBD (Group I, defined as age of onset <1 year, n=14) and IBD with the age of onset 1 to <10 years (Group II, n=26). We obtained their clinical characteristics, including disease type, sex, age of onset, disease location, disease behavior and the presence of perianal disease, according to the Paris classification.5 Information pertaining to family history, medical treatment, surgical intervention—including the age at operation—and history of infections was also obtained.

We analyzed the peripheral blood samples from the patients for IL-10RA, IL-10RB and IL-10 upon informed consent. DNA was isolated using a G-DEXTM genomic DNA extraction kit (iNtRON Biotechnology, Seoul, Korea). The PCR primer pairs were used as described in Table 1. The PCR reactions were performed in a 20-μl volume containing 100 ng genomic DNA, 2.5 pmoles of each primer and AccuPower ProFi Taq PCR PreMix (Bioneer, Seoul, Korea). Initial denaturation was done at 94 °C for 5 min, 30 cycles of denaturation at 94 °C for 1 min, annealing at the temperature described in Tables 1 for 1 min, extension at 72 °C for 1 min and a final extension step at 72 °C for 10 min. The amplified fragments were separated by 1% agarose gel electrophoresis, purified with a Qiaquick PCR purification kit (Qiagen, Hilden, Germany) and sequenced by an ABI PRISM 3730XL Analyzer (Applied Biosystems, Foster City, CA, USA). Identified variants were also sequenced in 100 healthy controls of Korean ancestry, and in silico analysis was performed using the online tools PolyPhen-2 (http://genetics.bwh.harvard.edu/pph2)17 and SIFT (http://sift.bii.a-star.edu.sg/www/SIFT_seq_submit2.html).18

Table 1 Primer sequences for candidate genes
Full size table

Statistical analysis for the phenotypic correlation was performed by Mantel–Haenszel χ2 test, one-way analysis of variance and Kaplan–Meier test using SPSS version 20.0 (IBM, New York, NY, USA).

This study was approved by the Institutional Review Boards at Seoul National University Hospital (Seoul, Korea).

Results

Genotypes of IL-10RA, IL-10RB and IL-10

We found mutations in IL-10RA in 7 out of 14 children in Group I. None of 26 children in Group II had mutations in IL-10RA. No mutations of IL-10RB and IL-10 were detected in all 40 children. Overall detection rate of IL-10RA mutations was 17.5% (7/40) in total patients. Particularly, the rate was 50.0% (7/14) in Group I.

Identified IL-10RA mutations were p.(R101W), p.(Y91C), p.(R262C), p.(R117H) and p.(W69R). The most common mutations were c.(301C>T), p.(R101W) and c.(272A>G), p.(Y91C) in exon 3 in four children. p.(R262C) and p.(R117H) were found in two children, respectively, and p.(W69R) and p.(A10G) were also found. All children have compound heterozygous mutations, but one child has only one mutation.

Of them, c.(272A>G), p.(Y91C) and c.(205T>C), p.(W69R) were novel variants. We investigated 100 healthy controls of Korean ancestry and they were not found. The PolyPhen-2 predicted that p.(Y91C) and p.(W69R) would probably be damaging with scores of 1.000. The SIFT tool predicted that p.(W69R) might affect protein function with a score of 0.00 and p.(Y91C) might be tolerated with a score of 0.12.

Normal polymorphisms were also found. c.(1051A>G), p.(R351G) and c.(670A>G), p.(I224V) in IL-10RA, and c.(139 A>G), p.(K47E) in IL-10RB which were identified in 40, 20 and 31 children, respectively.

Phenotypic characteristics of IL-10RA-deficient children

Of total seven children with IL-10RA mutations, six children developed their symptoms within 1 month after birth, and one child at 6 months (median 10 days after birth). Colonoscopies with histologic examinations were consistent with CD in six of seven children. One child was diagnosed with ‘intractable ulcerating enterocolitis in infancy’ 15 years ago, which was first described by Sanderson et al.,19 and we recently found IL-10RA mutations in her blood sample and reported it.14 All seven children (100%) had perianal diseases including rectovaginal or perianal fistulae. Three had frequent oral ulcers. All were treated with immunosuppressants with corticosteroid, azathioprine and/or infliximab. All children except one (85.7%) underwent surgery because of the poor response to medical management including immunosuppressive therapy. Five children (71.4%) underwent bowel resections (age range 8 months∼5 years, at the time of first operation), and five children (71.4%) underwent ileostomy or colostomy (age range 3 months∼2 years). Five children underwent multiple surgical interventions.

Three had recurrent skin folliculitis, and three had recurrent infections such as pneumonia, otitis media, varicella infection, candidial fungemia or sepsis. One patient who has been followed up as ‘intractable ulcerating enterocolitis in infancy’ has a family history of ‘ulcerating enterocolitis in infancy’, whose two brothers20 died of severe colitis and sepsis during neonatal periods. All but one patient had no family histories. None had parental consanguinity. Characteristic of children with VEO-IBD with or without mutations were summarized in Table 2.

Table 2 Genotypic and phenotypic characteristics of children with very early-onset inflammatory bowel disease
Full size table

The IL-10RA mutations were associated with younger age of onset (<1 year) (P<0.001), perianal fistulae (P<0.001), surgical interventions due to poor response to medical managements (P=0.017) and early surgical interventions (P<0.001). Comparisons of phenotypes between Group I and Group II, and between IL-10RA mutations group and without mutations group were summarized in Tables 3, 4 and Figure 1.

Table 3 Comparisons between Group I and Group II
Full size table
Table 4 Comparison between children with IL-10RA mutations and children without mutations
Full size table
Figure 1
figure 1

Cumulative rate of surgical interventions in children with inflammatory bowel disease by age. Children with interleukin-10 receptor A (IL-10RA) mutations underwent early surgical interventions because of poor response to medical treatments than children wihtout a mutation in genes of the IL-10 signaling pathway (Kaplan–Meier curve, P<0.001).

Full size image

Discussions

Glocker et al.8 first indicated that IBD may be a monogenic disorder. Homozygous mutations of p.(W159X) in IL-10RB were found in Kurdish siblings with consanguineous parents, homozygous IL10-RA mutations of p.(G141R) in an Arabic child and p.(T84I) homozygous mutations in a Caucasian infant. Children in that report exhibited symptoms within the first year of life with severe colitis and perianal diseases. They also found IL-10 mutations in two infants.21 Kotlarz et al.9 reported IL-10RA mutations in three children with IBD, IL-10RB mutations in six children and IL-10 mutations in three children. Begue et al.13 investigated 13 children and reported one child with IL-10RA mutations and one child with IL-10RB mutations. Moran et al.10 studied 188 children with IBD and found mutations in IL-10RA in one child. More recently, Engelhardt et al.12 reported mutations in IL-10RA or IL-10RB in seven children with IBD. There are not enough data and most were limited to Whites, Turkish and Arab. Many of them, particularly children in Arab, had histories of parental consanguinity.

In this study, we identified mutations of IL-10RA in seven Korean children with VEO-IBD. Particularly, 50% of our children with VEO-IBD (onset within the first year of life) had IL-10RA mutations. This is the highest detection rate ever reported. No children had parental consanguinity. There might be ethnical differences compared with White who showed low detection rate. The onset of symptoms post the first year of life did not correlated with IL-10-related mutations. All children with IL-10RA mutations developed their symptoms along with severe colitis and perianal fistulae (behavior ‘p’ in Paris classification) during the infantile period. Poor response to immunosuppressive therapy in all patients except one (85.7%) resulted in multiple surgical interventions early in life. With respect to the one exception, the child, currently 5 years old, had a relatively short follow-up period. He also responded poorly to other medications, and has been treated with infliximab. One child with an IL-10RA mutation in one allele showed very similar phenotypes to other children with compound heterozygote mutations. She developed symptoms 10 days after birth with a rectovaginal fistula and underwent multiple surgeries in infancy. It is possible she also has a mutation in the other allele, which, however, failed to be detected by direct sequencing methods. Children who did not have a mutation required fewer surgical interventions, and underwent surgery only after 10 years of age. The IL-10RA mutations demonstrated a significant association with the onset of symptoms in infancy, the presence of perianal fistulae and the requirement of early surgical interventions. To our knowledge, present study is the first report of phenotype analysis. Our investigation identified distinct phenotypic characteristics in IL-10RA-deficient patients compared with children without mutations.

The Paris classification5 determined an age cut-off point of 10 years in pediatric IBD. The classification focused on the location or extent of disease,22 and the variation in serologic responses23 in CD. VEO-IBD represents a rare phenotype within the pediatric IBD. According to the EUROKIDS registry of pediatric IBD patients, only 21 of 2022 pediatric patients manifested their symptoms during infancy.24 Cannioto et al.4 reported severe disease courses in VEO-IBD including resistance to immunosuppressive treatment and colectomy. These findings can be explained by the high rate of IL-10RA mutations in children with VEO-IBD in our study. The differences of clinical characteristics in Group I and Group II were consistent with the differences between IL-10RA mutations group and without mutations group. In previous studies, the presence of perianal disease and disease behavior do not differ with respect to age of onset.25, 26 However, in our study, children with VEO-IBD with IL-10RA mutations were more likely to have perianal disease than IBD without an IL-10RA mutation. Our investigation supports the evidence to subdivide a separate group of children with IBD at a very early-onset (0∼1 year) in the Paris classification.

Most of our patients with IL-10RA mutations were clinically and endoscopically diagnosed with CD, and one patient with ‘intractable ulcerating enterocolitis in infancy’ shared same mutations in patients with CD. Another study reported that the IL-10RA polymorphisms were associated with early-onset UC.10 These findings indicate that it may be more important to identify IL-10R mutations than to classify the disease types (CD or UC) in VEO-IBD.

Three out of our seven children with IL-10RA mutations had histories of folliculitis and recurrent infections such as sepsis, pneumonia and otitis media. The immunologic status of one patient was evaluated before treatment, and results revealed a relatively low CD3 T lymphocyte level.14 Recently, Engelhardt et al.12 reported certain subtle and variable immunological abnormalities in patients with IL-10R mutations. These findings demonstrate an association between the altered IL-10 signaling pathway and immune defects. The occurrence of an IBD phenotype in Wiskott–Aldrich syndrome, Immunodysregulation, polyendocrinopathy, enteropathy, X-linked syndrome (IPEX) and, in particular, chronic granulomatous disease have been reported and it indicates that IBD may be the complex immunodeficiency disorder.27

Very early-onset IBD in infancy is phenotypically and genetically different from IBD with onset of older age. It has a strong association with IL-10 receptor gene. It might be needed to subdivide a group of children with very early-onset IBD (0∼1 year) in the Paris classification. We should consider genotyping for genes of the IL-10 signaling pathway including IL-10RA in patients with VEO-IBD in infancy, particularly in cases that are associated with an onset of perianal fistulae and severe colitis.