This study was approved by the Ethics Committee of the Children’s Hospital of Zhejiang University School of Medicine. A total of 54 pediatric patients diagnosed with VEO-IBD in the Gastroenterology Department of Children’s Hospital of Zhejiang University School of Medicine from October 2005 to May 2017 were included in this study. According to Porto criteria[9], Crohn’s disease (CD) and CD-like intestinal inflammation present as noncontiguous aphthous or linear ulcers, primarily in the ileum or colon. Histologically, the disease is characterized by chronic focal inflammation with or without granulomas and perianal or fistulating disease. Ulcerative colitis (UC) or UC-like intestinal inflammation presents as continuous mucosal inflammation of the colon starting from the rectum, without small bowel involvement and without CD features. Patients with a definitive diagnosis of IBD with inflammation limited to the colon but without the typical features of UC and CD are classified as IBD of an unclassified type (IBDU).

The clinical characteristics of the pediatric patients, including sex, first symptom(s), time of onset of symptom(s), clinical features, surgical history, medication history and family history, were collected from medical records.

The following were criteria for genetic testing: (1) patients with disease onset before 6 mo of age; and (2) patients with disease onset beyond 6 mo accompanied with severe perianal disease, severe malnutrition or growth failure, or resistance to treatment.

Genomic DNA was isolated from peripheral blood of the patients. For whole exome sequencing (WES), whole exome library preparation was based on modifications of a protocol using the Agilent SureSelect XT Human All Exon Kit. Sequencing was performed using the Illumina HiSeq 2500 platform.

For targeted gene panel sequencing (TGPS), target sequences were enriched using customized capture probe chips (SureSelect Inherited Disease Panel; Agilent Technologies), which included 4,503 genes known to cause inherited disorders. DNA probes were designed for exons and flanking 10 bp intronic sequences. Briefly, 1 μg of genomic DNA was fragmented into 200-300 bp lengths by the Covaris Acoustics system. The DNA fragments were subsequently processed by end-repairing, A-tailing and adaptor ligation, 4-cycle pre-capture polymerase chain reaction (PCR) amplification, and targeted sequence capture. The captured DNA fragments were eluted and amplified by 15-cycle postcapture PCR. The final products were sequenced with 150-bp paired-end reads using the Illumina HiSeq X Ten platform according to the standard manual.

Raw data were processed by the Illumina pipeline (version 1.3.4) for image analysis, error estimation, base calling and generating the primary sequence data. For quality control, Cutadapt (https://pypi.python.org/pypi/cutadapt) and FastQC (https://http://www.bioinformatics.babraham.ac.uk/projects/fastqc/) were used to remove 3′-/5′- adapters and low-quality reads, respectively. Clean short reads were mapped to the human genome (hg19) using Burrows Wheeler Aligner software (http://sourceforge.net/projects/bio-bwa/). SOAPsnp software (http: //soap.genomics.org.cn/) and SAM tools Pileup software (http://sourceforge.net/projects/samtools/) were used to detect single-nucleotide polymorphisms and small insertions and deletions.

Variants were annotated using ANNOVAR software (http://www.openbioinformatics.org/annovar/), which includes function implications (gene region, functional effect, mRNA GenBank accession number, amino acid change, cytoband) and allele frequencies in dbSNP, 1000 genomes (http://www.1000genomes.org), ESP6500 (http://evs.gs.washington.edu/EVS/) and ExAc (http://exac.broadinstitute.org/). Damaging missense mutations were predicted by SIFT (http://sift.bii.a-star.edu.sg/), PolyPhen-2 (http://genetics.bwh.harvard.edu/pph2/) and MutationTaster (http://www.mutationtaster.org/). Interpretation of variants was based on recommended standards from the American College of Medical Genetics and Genomics[1,2], and all variants were categorized as pathogenic, likely pathogenic, variants of unknown significance, likely benign or benign.

To validate the variants based on next-generation sequencing (NGS) and WES, PCR amplification and Sanger sequencing were performed for the sequence centers on variant genes.

The categorical variables were presented as number (percentage). The continuous variables of normal distribution were presented as mean ± SD, otherwise as median ± interquartile range (IQR). The normality test was performed by Shapiro-Wilk test. The difference in variables with normal/abnormal distribution between nonmonogenic and monogenic IBD were estimated by Student’s t-test and Mann-Whitney test, respectively. The differences in categorical variables between nonmonogenic and monogenic IBD were assessed using the chi-square test; however, if total sample size was less than 40 or the number in one cell was less than 5, Fisher’s exact test was used for testing the difference. P < 0.05 was considered to be a statistically significant difference. All statistical analyses were conducted with SPSS 22.0 statistical software (SPSS Inc., IBM Corp., Armonk, NY, United States).

From October 2005 to May 2017, 136 pediatric patients were diagnosed with IBD at our hospital, among which 54 were diagnosed with VEO-IBD; thus, VEO-IBD patients accounted for 39.7% of pediatric IBD patients in our cohort. Among the 54 VEO-IBD patients, 37 were males and 17 females, for a male to female ratio of 2.18:1. Thirty-one VEO-IBD patients (57.4%) had disease onset before the age of 2 years.

Of the 54 patients with VEO-IBD, 72.2% had a diagnosis of CD or CD-like intestinal manifestations, 7.4% had a diagnosis of UC or UC-like intestinal manifestations, and 20.4% had a diagnosis of IBDU. Only one patient was the offspring of a consanguineous union, and his older sister also had similar symptoms during the neonatal period, though without a definitive diagnosis. None of the remaining patients had a positive family history of IBD or PID.

The median age of disease onset was 14 mo (IQR: 0 to 72 mo), and the median age of diagnosis was 35.5 mo (IQR: 3 to 173 mo). The median time to diagnosis was 9 mo (IQR: 0.4 to 104 mo). The Z-scores for median weight-for-age, height-for-age and body mass index -for-age were -1.69 (IQR: -4.69 to 2.2), -1.41 (IQR: -6.43 to 2.51) (n = 51) and 14 (IQR: 0 to 72) respectively.

The overall mortality rate was 11.1% (6/54). The clinical conditions of these 6 patients are listed in Table 1. Among the 6 patients, 4 had a diagnosis of CD and 2 a diagnosis of IBDU. Moreover, 2 of the patients had disease onset before 2 years of age.

Nine patients (16.7%) were diagnosed with monogenic IBD; among them, 6 were diagnosed with CD and 3 with IBDU. We also compared the clinical manifestations of monogenic and nonmonogenic IBD patients. The monogenic group was predominately male, with a male to female ratio of 8:1. Four patients had disease onset during the neonatal period, and two patients presented with symptoms before 1 year of age. The median age of disease onset for the monogenic patients was earlier than that for the nonmonogenic patients at 1 mo (IQR: 0 to 72) and 19.5 mo (IQR: 0 to 72) respectively (P = 0.008). Similarly, the median age of diagnosis for the monogenic group was earlier than that for the nonmonogenic group, at 18 mo (IQR: 4 to 78) and 43.5 mo (IQR: 3 to 17) respectively (P = 0.021). However, there was no significant difference in the median time from disease onset to diagnosis between the two groups (P = 0.668; Table 2).

The incidence of perianal disease in the monogenic group was higher than that in the nonmonogenic group (P = 0.001); when comparing weight-for-age and height-for-age Z-scores, no significant differences between the two groups were found. Moreover, laboratory findings were similar between the two groups. Mesalazine was more commonly used in the monogenic than in the nonmonogenic group (P = 0.009). A greater number of patients underwent intestinal surgery in the monogenic group than in the nonmonogenic group; however, this difference was not statistically significant at P = 0.050. To date, more patients in the nonmonogenic group have died compared to the monogenic group (P = 0.000; Table 3).

A total of 16 patients underwent WES or TGPS, 6 and 12 respectively (2 patients were tested by both WES and TGPS). Two patients underwent genetic testing (IL10RA) at another hospital. Nine patients were diagnosed with monogenic IBD. Seven of these patients were diagnosed by TGPS, and two were diagnosed by both methods. Five patients were diagnosed with IL10R mutation, two patients were diagnosed with chronic granulomatous disease (CGD) with CYBB mutation, and one patient was diagnosed with XIAP deficiency. One patient was diagnosed with common variable immune deficiency (CVID) with TNFRSF13B mutation. The clinical information for the 9 monogenic patients are listed in Table 4, and the genetic phenotypes of the 18 patients are provided in Table 5.

Patient 1 was a male offspring from a consanguineous union, with disease onset during the neonatal period. The first manifestation was a perianal abscess followed by recurrent fever, bloody stools, anal fistula and anal abscess formation, and skin infections. Colonoscopic findings included strictures of the colon near the splenic flexure, extensive polypoid hyperplasia, irregular ulcers, and inflammatory tags. Esophagogastroduodenoscopy was normal. Preliminary immune tests and the serum IL10 level were normal. The patient underwent a colostomy at the age of 1 year and 1 mo due to recurrent diarrhea and perianal abscesses. At 2 years and 5 mo of age, he underwent transverse partial colectomy, small bowel internal fistula resection and anastomosis. The pathological evaluation revealed transmural inflammatory infiltrate and linear serpentine ulcerations. The patient also underwent incision, drainage and seton placement for multiple abscesses. He was resistant to treatment with antibiotics, steroids, total parenteral nutrition (TPN), enteral nutrition (EN) and infliximab. At the time of writing this manuscript, he is in clinical remission on a combination treatment of azathioprine and thalidomide. Genetic testing by both TGS and WES revealed the following homozygous IL10RB mutation: Chr21:34660499: c.737G>A, p.W246X. This mutation has not been reported in HGMDpro.

Patient 2 was a female with VEO-IBD of neonatal onset, with first symptoms of recurrent watery diarrhea and blood and mucous in stools followed by a perianal abscess, fistula formation and recurrent fever. The patient was initially treated with elemental formula due to suspected cow’s milk allergy. She had her first colostomy at 1 year and 11 mo of age. Colonoscopy showed a cobble stone appearance and linear ulcerations in the colon. The pathological evaluation revealed classic CD features (strictures, transmural inflammatory infiltrate, linear serpentine ulcerations and noncaseating granulomas). Preliminary immune tests were normal, and the serum IL10 level was slightly elevated. She was resistant to treatment with steroids, mesalazine, antibiotics, immunosuppressants and infliximab. At 3 years and 1 mo of age, she underwent repair for a small bowel perforation, pancolectomy and ileal pouch-anal anastomosis. TGPS revealed the following compound heterozygous IL10RA mutations: Chr11:117864125: c.537G>A, p.T179T and Chr11:117860269: c.301C>T, p.R101W. The two mutations were derived from the mother and father, respectively. Both mutations are reported in HGMDpro[10,11].

Patient 3 was a male with disease onset at 1 mo of age. The patient had symptoms of recurrent diarrhea with mucous and blood, oral ulcers, anal fistula, obvious malnutrition and growth failure. Initial immune tests were normal, and the IL10 level was elevated at 85.5 pg/mL (normal range from 2.6 to 4.9 pg/mL). Colonoscopy showed deep ulcers in the rectum. The pathology evaluation showed nonspecific inflammation of the colon. The traditional treatments of antibiotics, TPN, EN and steroids were unsuccessful. Genetic testing at an outside hospital detected the following compound heterozygous IL10RA mutations: Chr11:117860269, c.301C>T, p.R101W, Chr11:117993268: c.470A>G, p.Y157C. The two mutations were from his mother and father, respectively. His older sister is a hereditary carrier. At the time of this writing, he is awaiting hematopoietic stem cell transplantation (HSCT). The pathogenic mutation c.301C>T is reported in HGMDpro[10], and c.470A>G was predicted by software mentioned in the methods section to be the likely pathogenic gene.

Patient 4 was a male with first symptoms of recurrent perianal abscess followed by watery diarrhea with scant fresh blood, rash and severe malnutrition. Immune tests were normal, and the IL10 level was 47.2 pg/mL. Colonoscopy showed multiple irregular ulcers at the sigmoid colon and rectum with polypoid hyperplasia and ulcers of the anus. Pathological examination revealed nonspecific findings. TGPS revealed the following homozygous mutation of IL10RA: Chr11:117860269: c.301C>T, p.R101W. The mutation was verified in his mother. The mutation is reported in HGMDpro to be pathogenic[10].

Patient 5 was a male with disease onset during the neonatal period. He initially presented with bloody diarrhea and was diagnosed with necrotizing enterocolitis; however, his symptoms persisted after surgery. This patient subsequently presented with a perianal fistula and severe malnutrition. Colonoscopy showed polypoid hyperplasia at the colon and ulcers at the rectum. This patient was treated with mesalazine and enemas. Genetic testing at an outside hospital revealed the following compound heterozygous IL10RA mutations: Chr11:117860269: c.301C>T, p.R101W, c.350G>A, p.R117H. The two mutations are reported in the literature to be pathogenic[10,12].

Patient 6 was a male with disease onset during the neonatal period. He presented with recurrent diarrhea, bloody stools and a perianal abscess. Preliminary immune tests were normal. Colonoscopy showed pancolitis and a perianal abscess. TGPS revealed the following hemizygotic mutation of CYBB: ChrX:37658209: c.676C>T, p.R226X. The mutation originates from his mother and is reported in HGMDpro to be pathogenic. The mutation is a termination mutation, which can alter the function of the encoded protein[13].

Patient 7 was a male diagnosed with intestinal malrotation during the neonatal period. He presented with pneumonia, perianal abscess and liver dysfunction; he developed diarrhea and fever during hospitalization. The immunoglobulin level and CD3 (46.81%), CD4 (29.86%), natural killer cell (7.79%), CD4/CD8 (1.86) levels were normal. Colonoscopy showed a longitudinal ulcer at the rectum. Soon thereafter, the patient developed multiple organ failure due to severe infection. TGPS revealed the following hemizygotic mutation of CYBB: ChrX:37642741: c.142-2A>G splicing. The mutation is reported as pathogenic in HGMDpro[14].

Patient 8 was a male who presented with jaundice and recurrent respiratory infections during the neonatal period. Immunoglobulin (Ig)G and IgM levels were below normal. He was treated with antibiotics and intravenous immunoglobulin. He gradually developed splenomegaly. At the age of 9 mo, this patient presented with recurrent bloody stools. Colonoscopy showed ulcers and polypoid hyperplasia, which eventually spread to involve the entire colon. Pathological examination of the intestinal biopsy showed nonspecific inflammatory changes. The patient underwent laparotomy, which revealed intestinal swelling and with fissure ulcers, and ulcers involving the tunica muscularis. Treatments with steroids, TPN, EN, mesalazine, infliximab and multiple immunosuppressants were unsuccessful. Both WES and TGPS showed the following compound heterozygous mutations in TNFRSF13B: Chr17:16843819: c.452C>T, p.P151L and Chr17:16852132:c.365G>A, p.R122Q. These mutations are reported to be associated with CVID, though they have not been reported in HGMDpro. Further tests are required to determine the effects of these mutations.

Patient 9 was a male, aged 5 years and 11 mo, who presented with recurrent bloody stools, fever and abdominal pain. He had been diagnosed with hemophagocytic syndrome at 4 years of age. Colonoscopy showed blunting and flattening of villi at the terminal ileum. The entire colon showed irregular deep ulcers and aphthous ulcers. Capsule endoscopy revealed sporadic jejunal ulcers. Immune tests were normal. He did not respond to exclusive enteral nutrition therapy and oral steroids. He was treated with infliximab and 6-Mercaptopurine and is currently in clinical remission. His colonoscopy findings have also improved. TGPS revealed the following mutation associated with XIAP deficiency: ChrX:123022501: c.910G>T, p.G304X. The mutation was not acquired from his parents. This mutation has not been reported in HGMDpro.

Very-early-onset inflammatory bowel disease (VEO-IBD) patients show a close association with primary immunodeficiency diseases, defined as monogenic IBD. More than 50 VEO-IBD related genes have been reported to date. Nonetheless, the incidence of monogenic IBD in Chinese population remains unknown.

Most reports regarding monogenic IBD were based on small population or case report, with only a small number of genes investigated. This study reports the largest cohort of genetically screened patients with VEO-IBD from China.

The objective of this research is to characterize monogenic IBD phenotypically and genotypically via genetic testing and to analyze clinical differences between monogenic and nonmonogenic VEO-IBD patients.

A retrospective analysis of children aged 0 to 6 years diagnosed with VEO-IBD in a tertiary hospital in southern China from 2005 to 2017 was performed. Clinical data for VEO-IBD patients were collected, and their genetic characteristics were analyzed using whole exome sequencing or target gene panel sequencing.

Nine patients (16.7%) were identified to have monogenic IBD by genetic testing. Five patients were shown to have IL10R mutation, two patients had chronic granulomatous disease, one patient had common variable immunodeficiency disease, and one patient had X-linked inhibitor of apoptosis deficiency.

A high proportion of monogenic IBD was observed among the VEO-IBD group, especially with disease onset before the age of 6 mo. IL10RA was the predominant mutation in this cohort. Monogenic IBD and nonmonogenic IBD demonstrated similar clinical features. Next-generation sequencing played an important role in the diagnosis of monogenic IBD.

Next-generation sequencing revealed a high proportion of monogenic IBD in our VEO-IBD cohort. Multicenter prospective studies are expected to determine the incidence of monogenic IBD in the Chinese VEO-IBD population and to investigate the genetic characteristics of monogenic IBD in China.