Very Early Onset Inflammatory Bowel Disease: An Integrated Approach

Abstract

Purpose of the review:

Immune dysregulation disorders are among the most rapidly growing set of inborn errors of immunity. One particular subset is the category where early-onset inflammatory bowel disease is the most common manifestation. These disorders are being increasingly appreciated although there has been minimal effort to articulate a unified approach to their diagnosis and management. This review will cover current thinking and strategies related to diagnosis and management of very early-onset inflammatory bowel disease.

Recent findings:

There is an expanding set of monogenic causes of early-onset inflammatory bowel disease. In many cases, the precise genetic cause dictates management. Lessons learned from the management of these monogenic conditions can sometimes be extrapolated to other refractory cases of inflammatory bowel disease.

Summary:

An integrated approach to diagnosis, risk analysis and management can include diagnostic approaches not often utilized for traditional inflammatory bowel disease such as whole exome sequencing. Management can also include non-traditional approaches such as targeted biologics or hematopoietic cell transplantation.

INTRODUCTION

Inflammatory bowel disease (IBD), includes Crohn’s disease, ulcerative colitis and inflammatory bowel disease-unspecified. The phenotype ranges from mild disease to severe refractory disease [1*]. There can be benefits to establishing a multidisciplinary clinic to evaluate and manage patients with very early-onset IBD (VEO IBD), especially when the phenotype is severe or unusual. While not all patients will require an integrated approach, many patients with complex or refractory disease will benefit from it. The simple act of establishing a diagnosis can be impressively challenging as non-inflammatory causes are common (Table 1) [2*]. One of the advances in the past 5 to 10 years has been the increasing recognition of the genetic contribution to VEO IBD. Indeed, nearly all IBD is thought to have a genetic component. The difference is the high frequency of monogenic or Mendelian causes in VEO IBD. This review will cover strategies to integrate care for patients, genetic evaluations, and multimodal approaches.

Table 1:

Differential Diagnosis

Consideration	Notes	Potential for confusion with VEO-IBD
Anatomical
Diversion colitis	The excluded large bowel can show changes similar to IBD.	Medium
Diverticular colitis	Sigmoid disease usually in older adults with diverticulitis	Low
Ischemic colitis	The anatomic distribution is distinctive although the pathologic appearance can resemble IBD	Low
Radiation colitis	Fibrosis and vascular changes are characteristic in addition to chronic inflammation	Low
Infection
HIV related colitis	HIV testing and extensive pathogen testing required	Low
Lymphogranuloma venereum	Testing for syphilis can distinguish this feature and is typically seen in the setting of HIV	Low
Tuberculosis	Large granulomas	Medium
Yersinia	Abundant granulomas with central necrosis	High
Medication
Non-steroidal anti-inflammatory use	Epithelial cell apoptosis prominent	Low
Mycophenylate mofitil	Pathology indistinguishable from IBD	Low
Immune-mediated
Autoimmune enteropathy	Distinct autoantibodies. Common in primary immunodeficiencies	Low
Behcet’s disease	Lymphoid aggregates	Low
Graft versus host disease	Pathologically is nearly indistinguishable from IBD although apoptotic crypt epithelial cells are prominent	Low
Microscopic colitis (collagenous colitis or lymphocytic colitis)	These are not uncommon in primary immune deficiencies but respond to alternative therapies	Low
Congenital diarrhea
Congenital diarrhea syndromes	Defects in epithelial nutrient/electrolyte transport, polarity, and entero-endocrine function	Medium

REVIEW

This review will attempt to address common questions that arise in the consideration of a clinic dedicated to VEO IBD. For the purposes of this review, we will consider VEO IBD to be children with age of onset less than six years of age. This is aligned with a recent effort to standardize the nomenclature [3].

Epidemiology of IBD and VEO IBD

The incidence of IBD is increasing worldwide. A more recent recognition is that industrialized countries have begun to plateau whereas emerging economies continue to have an increasing incidence [4*]. Immigrants rapidly acquire the risk profile of their new country [4]. A seminal study from Canada demonstrated that in an industrialized country early age of onset is the fastest growing subset of patients [5*,6]. One study suggests that early onset IBD is now also beginning to plateau [7] although others found a continued rise [8,9].

In Saudi Arabia, infantile or toddler age onset was strongly associated with a positive family history as well as consanguinity [10]. This supports the concept that earlier onset is associated with a stronger genetic contribution. The first monogenic form of IBD was reported in 2009: defects in the IL-10 receptor [11]. In the Middle East, a genetic approach to early-onset IBD identified enrichment of variants within the IL-10 and IL-12 pathways [12*]. In China, where the economy is more recently matured, diverse mutations in early-onset IBD have been identified, such as in the IL-10 and autoinflammatory pathways [13*,14*]. In a recent study from China, the Han ancestral group was found to have enrichment of the IL10RB R101W and T179T (a relatively common polymorphism that affects splicing) [15]. A total of 93 infantile-onset IBD patients had IL10RA variants identified with a phenotype that included 90% perianal disease, onset at a mean of 10 days of age, and oral ulcers. 28% had a positive family history of VEO IBD.

In adults, asthma has been recently identified as a risk factor for subsequent development of Crohn’s disease [16]. This finding may be confounded by the high rate of undiagnosed pulmonary disease in adults with IBD [17]. One could hypothesize that genetic or environmental factors that impact epithelial barrier function could drive both an increased risk of atopic disease as well as an increased risk of IBD. Somewhat counterintuitively, there are recent studies that suggest that poor oral health protects against IBD [18]. Thus, there is a great need to further explore the environmental contributions to the development of IBD.

Diagnostic delay

There are a number of causes of diarrhea in infants and children that represent potential diagnostic confusion [2]. These include defects in lipid trafficking, solute carrier defects, and epithelial dysplasia. However, the most common cause of diarrhea in childhood is certainly infectious. Therefore, it is perhaps no surprise that the diagnosis of IBD in young children can be delayed. Crohn’s disease and isolated small bowel disease are risk factors for delay [19**]. Nearly 2/3 of patients with Crohn’s disease diagnosed after more than two years of symptoms had at least one complication at the time of diagnosis compared to just one quarter of patients diagnosed within four months [20**]. These studies support efforts to increase awareness of early-onset IBD in the community as the single largest source of delay in diagnosis was the time to referral to a gastroenterologist.

Genetics and genomics

Since the seminal paper that identified age of onset as a predictor of Mendelian causality of IBD [21], there have been diverse and energetic efforts to identify, categorize, and understand the genetic basis of VEO IBD. The specific genes identified have been used to establish gene sequencing panels that are now often used diagnostically, as a targeted exome slice or as part of a larger comprehensive immunodeficiency panel [22*]. Table 2 offers a current listing of the genes clearly identified as associated with VEO IBD. These genes are pathophysiologically nearly identical to the genes identified in genome-wide association studies (GWAS) [23]. An overall model of IBD can be represented as a triangle with highly penetrant single gene defects at the tip and polygenic contributions, where each variant contributes modestly to the overall risk of developing IBD, at the bottom (Figure 1). Environmental contributions, gut microbiome, and epigenetics are envisioned to play significantly larger roles when the genetic risk is lower and are less impactful when the genetic risk is high. Having said that, it is known from murine studies that even highly penetrant genetic foundations of IBD often require a specific bacterial stimulus to manifest disease and can be impacted by other variables such as degradation of the mucus layer [24].

Table 2:

Genetic causes of VEO-IBD^*

Gene (Inheritance pattern)	Frequency of IBD among those with gene defect	History	Physical examination	Immunologic features
High likelihood of presenting with VEO-IBD
ADAM17 (AR)	>80%	Staphylococcal infections	Psoriasiform erythroderma, pustules, broken hair, abnormal nails	Slightly low TNF response to LPS in PBMCs and T cells.
ARPC1B	>80%	Vasculitis prominent	Vasculitis	Microthrombocytes with aberrant spreading
FOXP3 (XL)	>80% enteropathy	Onset near birth	Dermatitis	Low regulatory T cells
GUCY2C (AD GOF)	>80% diarrhea, IBD less common	Ileal obstruction, adhesions, esophagitis, electrolyte abnormalities	Dilation of small intestine	Not assessed
IL10 (AR)	>80%	IBD onset near birth	Folliculitis, perianal disease	Normal
IL10RA (AR)	>80%	IBD onset near birth	Folliculitis, perianal disease	Lack of IL-10 suppression of LPS response
IL10RB (AR)	>80%	IBD onset near birth	Folliculitis, perianal disease	Lack of IL-10 suppression of LPS response
MALT1	>80%	Respiratory infections, sepsis, CMV	Dermatitis	Poor response to vaccines, poor antigen-specific proliferation, low switched memory B cells
PLA2G4A	>80% stenosing enteritis	Early strictures, Platelet dysfunction	-	None reported
SKIV2L (AR)	>80% enteropathy	IUGR, FTT	Trichorrhexis nodosa, frontal bossing	Villous atrophy, low immunoglobulins, low T cells
SLC02A1 (AR)	>80% ileal ulceration (females)	No response to TNF inhibitors	Digital clubbing in males	Normal
SLC9A3	>80%	Prenatal onset diarrhea	Abdominal distension	Normal or Low IgG, high fecal sodium
TTC37 (AR)	>80% enteropathy	IUGR, FTT	Trichorrhexis nodosa, frontal bossing	Villous atrophy, low immunoglobulins, low T cells
TTC7A (AR)	>80%	Intestinal atresia	Dermatitis, alopecia	Low T cells
Intermediate likelihood of presenting as VEO-IBD
COL7A1 (AR)	16% in childhood	Epidermolysis bullosa	Bullous lesions	Normal or low IgG
FERMT1 (AR)	25% in adults	Infections	Bullous lesions, progressive poikiloderma, gingival fragility	-
IKBKG (XL)	>50% by mid-childhood	Infections	Hypodontia, poor sweat, thin hair, frontal bossing	Poor titers, low NK function, abnormal TLR responses
IL2RA (AR)	30–50%	Infections	-	Poor T cell responses
IL2RB (AR)	30–50%	Infections	-	Poor T cell responses
POLA1	50% in males	FTT, corneal disease	Reticular pigmentation, photophobia	High interferon signature, amyloid in papillary dermis
RTEL (AR)	30–50%	IUGR, FTT, microcephaly	Fine hair	Low NK cells, apoptosis in biopsy
SLC37A4	30%	Hypoglycemic episodes	Hepatomegaly	Poor neutrophil funciton
STXBP2 (AR)	<50%	HLH	-	Poor NK function, low IgG
XIAP (XL)	30–50%	HLH, IBD does not respond to medications	Splenomegaly	Low IgG
Lowest likelihood of presenting with VEO-IBD
BTK (XL)	<10% in early life	Infections	Small tonsils	Very low B cells and immunoglobulins
CTLA4 (AD)	<10% in early childhood	Infections, interstitial pneumonitis	-	Low immunoglobulins, low switched memory B cells, low CD4 T cells
CYBA (AR) P22phox	<10% in early life	Infections	-	Low DHR, reduced switched memory B cells, low T cells
CYBB (XL) Gp91phox	<10% in early life	Infections, maternal discoid lupus	-	Low DHR, reduced switched memory B cells, low T cells
DKC1 (XL)	<10% in early childhood	Microcephalic, IUGR	Small, nail dystrophy	Low B cells. NK cells
DOCK8 (AR)	?	Cutaneous viral infections, other infections	Eczema	Low T cells, high eosinophils
G6PC3 (AR)	<10%	Cardiac anomalies, urogenital defects, IUGR	Superficial vessels enlarged	Neutropenia, intermittent thrombocytopenia, lymphopenia in severe forms
HPS1 (AR)	<10% in early childhood	Pulmonary fibrosis, founder effect in Puerto Rico and Swiss Alps, bleeding	Oculocutaneous albinism	Pigmented macrophages, lymphoid aggregates on biopsy
HPS4 (AR)	<10% in early childhood	Pulmonary fibrosis, bleeding	Oculocutaneous albinism	Pigmented macrophages, lymphoid aggregates on biopsy
HSP6 (AR)	<10% early childhood	Bleeding	Oculocutaneous albinism	Pigmented macrophages, lymphoid aggregates on biopsy
ICOS (AR)	<10%	Infectious enteritis, founder effect along the Danube river	Splenomegaly	Absent class switched memory B cells, low TFH, poor germinal centers in lymph nodes, low IgG, nodular lymphoid hyperplasia of GI tract
ITGB2 (AR)	<10% in early childhood	Severe infections, delayed separation of umbilical cord	Gingivitis, scarring (poor wound healing)	High WBC/ANC, low CD18 expression, reduction of factor XIIIa+ DC in lymph node
LRBA (AR)	<10% in early childhood	Infections, interstitial pneumonitis	-	Low immunoglobulins, low switched memory B cells, low CD4 T cells
MEFV (AR)	<10%	Serositis, founder effect in Mediterranean	-	-
MVK	<10% in early life	Nausea and fever episodically, abdominal pain	Adenopathy, oral ulcers, arthritis	-
NCF1 (AR) P47phox	<10% in early life	Infections	-	Low DHR, reduced switched memory B cells, low T cells
NCF2 (AR) P67phox	<10% in early life	Infections	-	Low DHR, reduced switched memory B cells, low T cells
NPC1	<10% in early life	Airway infections, developmental delay	Splenomegaly, ataxia	Foam cell macrophages on biopsy
PIK3CD (AD GOF) P100	<10% in early life	Infections, PSC, herpes, lymphoma	Bronchiectasis, adenopathy, HSM	High IgM, low IgG, low CD4/CD45RA, nodular lymphoid hyperplasia, EBV viremia
PIK3R1 (AD LOF) P85	<10% in early life	Insulin resistance	Short stature	High IgM, low IgG, low CD4/CD45RA, nodular lymphoid hyperplasia,
PLCG2 (AD)	<10% in early childhood	Pleomorphic inflammation, may be cold induced	Dermatitis, evaporative cooling may induce urticaria	Low immunoglobulins, low switched memory B cells
PTEN (AD)	<10% in early life	Thyroiditis, autoimmune hemolytic anemia, hamartomas	Adenopathy, large tonsils, macrocephaly, developmental delay	Nodular lymphoid hyperplasia, low IgG
SLC26A3 (AR)	<10% in early life, but all have diarrhea	Onset of secretary diarrhea at birth	-	Inflammation occurs later in life
SLC37A4 (AR)	<10% in early life	Hypoglycemic episodes	Hepatomegaly	Neutropenia
STAT1 (AD GOF)	<10% in early life but enteropathy common	Pleomorphic autoimmunity, candida, other infections	Poor growth	Low NK cells, low IgA
STAT3 (AD GOF)	<10% in early life but enteropathy common	Pleomorphic autoimmunity: diabetes, thyroid	Poor growth, eczema	Low regulatory T cells, low IgG
TGFBR1 (AD)	<10% early in life	Aneurysms,	Cleft palate/uvula, hypertelorism, arachnodactyly, pectus, joint laxity	Eosinophilic colitis, high IgE, high eosinophils
TGFBR2 (AD)	<10% early in life	Aneurysms,	Cleft palate/uvula, hypertelorism, arachnodactyly, pectus, joint laxity	Eosinophilic colitis, high IgE, high eosinophils
WAS (XL)	<10% early in life	Thrombocytopenia	Eczema	Poor T/B/NK function
ZBTB24 (AR)	<10%	Developmental delay	Frontal bossing, short stature	Progressive decline in antibodies and T cells
Unknown frequency of IBD
ALPI (AR)	?	None noted	-	-
ANKZF1 (AR)	?	Infantile onset IBD	Perioral and oral inflammation	Lymphocytic colon infiltrates, T/B/NK lymphopenia (some)
CD55	?	Protein losing enteropathy, thrombosis		Complement deposition on biopsy
DUOX2	?	Hypothyroidism	-	High lymphocyte count
IL21R (AR)	?	Cryptosporidia, cholangitis	-	Low switched memory B cells
IL21 (AR)	?	Infections	-	Low B cells, low switched memory B cells, low IgG
NCF4 (AR) P40phox	?	Infections	-	DHR slightly low
NLRC4 (AD)	?	HLH, episodic inflammation	-	-
TGFB1	?	Encephalopathy, eosinophilic esophagitis	Posterior leukoencephalopathy	High IgG and IgE, poor proliferative responses
TRIM22	?	Early infancy onset, infections	Severe perianal disease	-

^*Excluding leaky severe combined immunodeficiency genes- all of which could theoretically be associated with VEO-IBD.

Abbreviations: DHR= dihydrorhodamine, HLH= hemophagocytic lymphohistiocytosis, IUGR= intrauterine growth retardation, FTT= failure to thrive, HSM= hepatosplenomegaly, and PSC= primary sclerosing cholangitis. Inheritance is given as AR= autosomal recessive, AD= autosomal dominant, XL= X-linked, GOF= gain of function mutation.

Figure 1. The genetic basis of VEO IBD.

A schematic diagram illustrating the contributions to the development of IBD. Environmental effects are most pronounced in polygenic disease where the effect size of any individual gene is low and collectively, the genetic contribution to disease is less than what is seen for single gene causality. Environment and microbiome are nearly always a contribution, in the monogenic cases of VEO IBD, their relative contribution is less than what is seen in polygenic disease. Age of onset somewhat tracks with single gene causality but increasingly adults are being recognized with significant inborn errors of immunity causing IBD.

Genome-wide association studies

Genome wide association studies (GWAS) paved the way for a functional understanding of the genomic architecture of IBD. GWAS continue to be performed, refining risk variant gene assignment and adding additional loci. There are now over 230 individual loci associated with IBD. Recent studies have highlighted the role of integrin-associated variants [25], a concept aligned with our current understanding of IBD pathogenesis and highlighted by the success of integrin-directed therapy (vedolizumab). An additional area for fertile study is the recognition that some loci may include variants that lie within microRNAs (miRNAs) [26]. MiRNAs regulate the transcriptome by binding to messenger RNA and modulating either translation or stability. The recent identification of variants that impact miRNAs suggests that this may be a future area of emphasis. Less frequently investigated have been studies evaluating the epigenome in IBD. One study identified differential DNA methylation at inflammatory genes in pediatric IBD that were associated with dysregulated expression of miRNAs [27]. In another study, the authors compared intestinal epithelial cells in newly diagnosed pediatric patients with IBD and found changes in DNA methylation and transcriptome patterns [28**]. These studies suggest that in spite of the maturation of the GWAS studies for IBD that additional mechanisms of disease will be revealed with additional refinement. The recently developed Human disease methylation database (DiseaseMeth) has provided a comprehensive DNA methylation repository of human disease and will serve to expand the study of epigenetics in IBD [29].

New genes identified as associated with early onset IBD

Multiple new genes have been identified in the past two years [30-33*, 34-35*, 36*, 37*, 38*, 39* 40*, 41]. Additional studies defining phenotypic differences and therapeutic interventions directed at gene-stratified patient populations are eagerly awaited. The gene variants currently implicated in VEO IBD are highlighted in Table 1. Many classical primary immunodeficiencies are associated with IBD or enteropathy [42*]. The list in Table 1 is restricted to genes where VEO IBD has been identified in multiple cases.

Mechanisms of disease

The two areas most intensively investigated over the past two years are the microbiome and the role of T cells in IBD. None of these studies specifically target VEO IBD but the lessons are likely extrapolatable. The interactions of the maturing microbiome and T cells in childhood is just beginning to reveal the critical interfaces. An excellent recent review portrays the pathologic pathways in VEO IBD graphically [43*].

Microbiome

Significant differences in the microbiome have been recognized in patients with IBD compared to healthy controls [44]. These differences have been attributed to both a cause of IBD and an effect of IBD. The microbiome develops between birth and 3 years of life suggesting a vulnerable window of time. In IBD, decreased relative abundance of bacteria within the phyla, Firmicutes and Bacteroidetes, and enrichment of Proteobacteria and Actinobacteria have been seen. Host genetics and environmental factors (smoking, diet, air pollution, and antibiotic use) have been shown to influence the microbiome in IBD. A recent study in mice demonstrated that antibiotics administered at the time of delivery can alter the offspring’s intestinal microbiome and leads to a long-term increased susceptibility to colitis [45*]. Given the high rate of antibiotic use to prevent group B streptococcal sepsis in newborns, this represents a significant finding. A second murine study was directed at defining trajectories of change as a strategy to understand the relationship between alterations to the microbiome and the interacting T cells. Increased responses to bacterial ligands preceded the clinical development of IBD [46*]. At the time of clear disease activity, increased T cell activation was identified [47].

The microbiome is also a living community and metabolites also impact the immune response. Commensal bacterial products stimulate cytokines and release of antimicrobial peptides, such as RegIIIƔ [48,49]. For example Bacteroides fragilis produces short chain fatty acids, inhibiting activity of histone deacetylases in colonocytes and immune cells leading to down-regulation of pro-inflammatory cytokines and induction of Treg cells [50]. Perhaps pertinent to the rise in incidence in VEO IBD, recent studies suggest that metabolites can be maternally transferred to the offspring and affect the developing immune response [51*].

One of the confounding aspects in the study of microbiota and IBD is the vast diversity of species and genera observed. The microbial dysbiosis index collapses this diversity into a manageable variable, a strategy that may enhance analyses [52]. Vedolizumab response, a monoclonal antibody to T cell integrins required for gastrointestinal migration was associated with high bacterial diversity at baseline [53*]. Branched-chain amino acid synthesis was significantly higher at baseline in those achieving remission as well. Another metabolic feature, bacterial urease production and nitrogen flux, have been implicated in the development of intestinal dysbiosis and subsequent immune-mediated colitis [54*]. The importance of these findings is that microbial metabolic pathways can be targeted through medical management [47].

One additional study on the microbiome highlights the systemic effects of Crohn’s disease. Pediatric patients with Crohn’s disease were found to have a completely different subgingival microbial structure in children with Crohn’s disease compared to controls [55]. Patients who responded to therapy normalized their subgingival microbiota suggesting that subclinical inflammation throughout the gastrointestinal tract impacts the microbial community.

T cells

T cell immunodeficiencies have a high rate of IBD [56*]. Recent data have focused on atypical T cells as well as conventional T cells and their functional responses. The gastrointestinal tract has been termed the largest immune organ in the body. The richness of the cell types and their intracellular location have posed methodologic problems and only recently have advances been made in understanding the diversity and function. Long recognized are the innate lymphoid cells (ILC). ILC have been demonstrated to mold responses to commensal bacteria by killing highly responsive T cells [57]. The highly responsive T cells are largely polarized to produce γ-interferon, IL-17, and IL-22 [58*]. In the setting of VEO IBD, patients with mutations in the IL-10 receptor have enhanced production of IL-17 [59*]. This is paired with increased macrophage expression of IL-1β [60]. IL-17 inhibition is not felt to be therapeutic and in fact can be harmful by disrupting T cell interactions with barrier function, however, IL-1β inhibition has been used therapeutically and was found to be specifically useful in patients with IL-10 receptor deficiency [60].

In murine models, the effect of the glycosylation pathway has been recently investigated. Mucosal T cells in patients with ulcerative colitis have a deficiency in branched chain glycosylation. This was modeled in mice to show that the branched chain glycosylation pathway is critical for T cell function [61**]. Indeed, supplementation with glycosylation precursors was able to restore T cell function. This study directly relates to VEO IBD because oral GlcNAc is available for human utilization and has been described as promoting mucus production in children with treatment-resistant IBD [62]. A second glycosylation defect in epithelial cells was identified in mice suggesting that glycosylation is a growth area for understanding IBD [63].

An additional study in mice identified an atypical T cell utilizing the γδ T cell receptor [64*]. The location of the γδT cells at the tips of villi is dependent on the microbiota and they exhibit rapid movements in response to changes in the microbial community. These T cells regulate the intestinal barrier via epithelial tight junction proteins. Therefore, the concept of a highly dynamic interplay between the microbial community, conventional T cells, and unconventional T cells at the epithelial surface and within the lamina propria is developing.

Providing integrated care to patients with VEO IBD

An integrated program can have significant benefits to the patient by simplifying and unifying the approach. As is true for all conditions, patient care is improved at centers with high volume and expertise. Physicians also benefit by having real time interactions regarding patient care. Nevertheless, there are some challenges to developing an integrated care approach. This section will attempt to address common hurdles for the development and implementation of an integrated approach.

Which patients should be invited to participate in an integrated clinic?

There are three conceptual approaches to identify the constituency. The selection can be according to age criteria, severity, or geographic constraints. The VEO IBD clinic at The Children’s Hospital of Philadelphia is based largely on age of onset (<6 years of age at onset) but older patients with refractory or unusual disease can petition for evaluation. Approximately 500 patients are followed at this time and >90% had their onset at <6 years of age.

Which providers will be involved in the structure of the clinic?

The VEO IBD clinic at The Children’s Hospital of Philadelphia includes specialists in Gastroenterology, Immunology, Nutrition, and Social work at a minimum. Extensive nursing involvement supports the complex care and the nurses act as liaisons for the patients. The addition of Genetics and Rheumatology is patient-specific. The holistic approach ensures that patients receive coordinated care, have the best support through their difficult journey and receive a consistent message. Patients also appreciate the one-stop approach to care.

People who support the clinic but are not directly involved in patient care include a Rheumatologist, dedicated Geneticist, Bioinformatician, Study Coordinators, Study Researchers and research laboratory staff who perform flow cytometry, functional analyses and examine biomarkers and microbiome. This component is not essential for patient care but is often critical for variant validation.

What laboratory evaluation is critical to identify an inborn error of immunity?

In our clinic, very low switched memory B cells has the highest sensitivity for the identification of Mendelian inborn errors of immunity. Other commercially available studies have a low yield when used in an unselected fashion but are useful in specific settings. Infantile-onset disease with fistulas suggests an IL-10 pathway defect and an assay that detects lack of IL-10 inhibition of responses to lipopolysaccharide will detect receptor mutations (but not IL-10 ligand defects). The sensitivity of widely available flow cytometry assays, often the standard evaluation for inborn errors of immunity, is surprisingly low. For this reason, sequencing often represents a cost-effective approach.

Stratification of patients for sequencing

The “hit” rate varies across studies but overall about 20% of patients with VEO IBD have been found to have an inborn error of immunity through sequencing [65–67]. Using a gene panel and looking at a cohort of general pediatric IBD, only 4% had a clear causal gene variant (s) [41]. Using age of onset <10 years, a German study found 7% frequency of Mendelian gene variants [68*]. Conversely a cohort of infantile-onset IBD cases found 31% of cases had Mendelian disease [69]. These studies suggest that early age of onset enriches for the finding of monogenic causality but even adults with Mendelian causality have been described and CGD has a rate of IBD approaching 50% by adulthood [70*,71]. Due to the expense it is common to be asked if there is an approach to risk stratify other than age of onset. The most common inborn error of immunity in our cohort is CGD and we therefore screen all patients with a DHR. In some other cohorts, IL-10 receptor defects have been the most common and the screening test for this is available. Enrichment by selecting patients with infantile onset, from consanguineous parents, and growth failure has been observed in our clinic and has been reported in the literature [69,72*]. Our data suggest that low switched memory B cells may select out a group enriched for inborn errors of immunity. These findings have led to our current approach of opting for sequencing for most patients. There is a caution, however. In our cohort, trisomy 21 and Turner syndrome were found with a collective frequency of 19% among those with an identified Mendelian cause. Furthermore, 33% of those sequenced where a causal variant(s) was found had a variant in a gene not included in IBD panels. Lastly, nearly a third of our patients with causal or potentially causal variants require additional high-level validation approaches. Thus, sequencing should be undertaken with a vision of what is possible at each center and the patients should be educated regarding expectations.

Sequencing methodology

There are four options for sequencing and the costs, yield and turnaround time are different, not to mention the difficulty in getting insurance approval (Table 3). Labs with an interest in VEO-IBD will occasionally do the sequencing on a research basis. There are a number of potential pitfalls here. Data are typically delivered with little analysis requiring time and effort to analyze. Findings most often cannot be used for clinical decision making and any variants must therefore be validated through clinical-grade sequencing.

Table 3:

CLIA approved sequencing options

	Sanger sequencing	Whole exome “slice” or panel	Whole exome sequencing
Strengths	• High accuracy • Can include splice junctions • Often reflexes to deletion/duplication testing	• Affordable • Often more timely than WES • No off target findings	• Most comprehensive
Weaknesses	• Limited number of genes sequenced (i.e. best used from his pre-test probability) • Some genes are incompletely sequenced and only “hot spots” are sequenced	• Same weaknesses as WES for technical approach although coverage is typically higher	• Usually takes the longest to run and analyze • Some exons are missed • Not sensitive (at this time) for deletions and duplications • Regulatory (non-coding) mutations missed • Off target findings common
Insurance coverage/cost	• The easiest of the three • Cost is highly variable but single genes are typically less than WES	• Mid-range difficulty • Less expensive than WES	• Most difficult • Most expensive generally
Notes		• Panel content highly variable • Some panels use WES data and targeted analysis, others use specific gene capture which leads to higher coverage	• Bioinformatic analysis variable • Strength of variant association can be difficult to discern

Interpretation of sequencing results

There are three conceptual categories of findings for each patient who has sequencing. There may be a clear variant identified that explains the condition because it has been previously reported in VEO IBD. This type of result is gratifying but well less than half the sequenced patients have this type of result. There may be a variant in a gene that is known to be associated with the condition but the specific variant has not been previously reported. It is tempting to call the new variant causal and report the result to the patient and indeed there are cases where that may be appropriate, however, this type of result generally requires some type of validation. A standard lab study may reveal findings compatible with that genetic diagnosis. An example would be a new mutation in the CYBB, associated with X-linked CGD. The variant may be easily validated using a dihydrorhodamine (DHR) assay. For some genes, there are no clear laboratory or clinical features that would allow validation, and in those cases, a research lab must be sought to perform analyses demonstrating that the variant affects protein function. Finally, in many cases either no gene variant is identified or a gene variant labelled as variant of uncertain significance (VUS) is found. For genes not previously reported to be disease-associated, GeneMatcher (https://genematcher.org) or ClinVar (https://www.ncbi.nlm.nih.gov/clinvar/) may be used to see if other patients have been identified in other settings. Lacking a finding of other similar patients, the road in this circumstance is long and arduous. A research lab must be identified with time, interest and funding to support a functional validation and possibly an animal model of the gene variant.

How can the patients be managed?

At our center, we strongly feel that identification of a variant can drive therapeutic decisions. For patients with VEO IBD, the gene often suggests a specific therapy (Table 4). In some cases, knowing the pathway can suggest a treatment. In the cases, where there is no variant identified, we individually determine whether additional studies should be pursued. When there is no guidance from sequencing, experience can be useful. While we have strategies that seem more successful to us, in the end, patients often need a unique therapeutic approach. While the goal is to prevent surgery, diversion and colectomy represent options with clear expectations set [73*,74*].

Table 4:

Genetic stratification for therapy

Gene defect	Therapeutic considerations
CYBB and other etiologies of CGD	IL-1 blockade. TNF inhibitors should not be used due to risk of fungal infections
CTLA4	Abatacept, sirolimus, HSCT
FOXP3	Sirolimus, HSCT
IL-10, IL-10RA, IL10RB	HSCT
LRBA	Abatacept, sirolimus, colchicine, HSCT
MEFV	IL-1 blockade, colchicine
MVK	IL-1, TNF blockade
NLRC4	IL-18 binding protein
TTC7A	HSCT can be considered, however, success has been limited
XIAP	HSCT, IL-18 binding protein study underway

HSCT: Hematopoietic stem cell transplant

CONCLUSIONS

Multidisciplinary approaches to VEO IBD can provide coordinated and comprehensive care for patients. The genetic underpinnings of VEO IBD represent an opportunity to advance therapeutic decision making beyond trial and error. Strategies that are successful in cases with causal genetic variants can also be applied for those without an identified genetic cause.

• A minority of patients with VEO IBD have a single gene cause, although these patients often benefit from specific targeted therapeutic approaches

• A multidisciplinary clinic can bring value to the diagnosis and care of patients with VEO IBD

• Traditional immunologic studies have limited value but sequencing can often identify patients with monogenic causality