Early‐onset inflammatory bowel disease as a model disease to identify key regulators of immune homeostasis mechanisms

Summary Rare, monogenetic diseases present unique models to dissect gene functions and biological pathways, concomitantly enhancing our understanding of the etiology of complex (and often more common) traits. Although inflammatory bowel disease (IBD) is a generally prototypic complex disease, it can also manifest in an early‐onset, monogenic fashion, often following Mendelian modes of inheritance. Recent advances in genomic technologies have spurred the identification of genetic defects underlying rare, very early‐onset IBD (VEO‐IBD) as a disease subgroup driven by strong genetic influence, pinpointing key players in the delicate homeostasis of the immune system in the gut and illustrating the intimate relationships between bowel inflammation, systemic immune dysregulation, and primary immunodeficiency with increased susceptibility to infections. As for other human diseases, it is likely that adult‐onset diseases may represent complex diseases integrating the effects of host genetic susceptibility and environmental triggers. Comparison of adult‐onset IBD and VEO‐IBD thus provides beautiful models to investigate the relationship between monogenic and multifactorial/polygenic diseases. This review discusses the present and novel findings regarding monogenic IBD as well as key questions and future directions of IBD research.

on identifying genetic risk factors using genome-wide association studies (GWAS). In the last decade, intensive research using GWAS has identified over 230 IBD-associated loci comprising approximately 300 potentially associated genes, [8][9][10][11] including NOD2, ATG16L1, IRGM, IL23R, CARD9, RNF186, and PRDM1. Although there are only a few GWAS SNPs with evidence of biological involvement in IBD, such as missense SNPs in NOD2 12 and ATG16L1, 13 identification of such associations pinpointed crucial mechanisms such as autophagy, pattern-recognition, Th17 involvement, and maintenance of the epithelial barrier in IBD pathogenesis. 11 Recent efforts have focused on meta-analysis and fine mapping of existing GWAS datasets using innovative approaches such as Bayesian analysis, 14  Despite current advances in data collection and analysis, our understanding of SNPs outside the coding regions is still elusive. It has been shown that SNPs for autoimmune disease tend to be enriched in regulatory regions, and in differentially expressed genes, and that risk variants for autoimmune diseases show particular enrichment in active chromatin regions of immune cells. [18][19][20] In addition, several efforts have been made to unravel how SNPs at a locus affect mRNA expression of genes. These efforts combining GWAS with transcriptome analysis have revealed that pinpointing the causal SNP in a haplotype block is a non-trivial task and that many of the SNPs have a detectable effect only in a cell type-dependent or stimulus dependent context. [21][22][23] As our knowledge for non-coding regions of the genome is growing relating SNPs to regulatory regions, as well as assaying the cell type specificity of loci, will be important goals for the future. Notably, it is unclear how the identified susceptibility loci and associated genes identified in these GWAS studies relate to the early-onset, Mendelian form of IBD. Identification of high impact SNPs in NOD2 that are associated with adult Crohn's disease with clear involvement in IBD pathogenesis has illustrated a genetic continuum between adult and early-onset IBD, in contrast to the classical view of two genetically independent diseases. 24 In this context, we can hypothesize that adult and Mendelian IBD arise as a result of a spectrum of varyingly pathogenic genetic lesions that impact common key pathways in IBD. Despite these advances, the exact relationship between adult and Mendelian IBD is still poorly understood. The lack of understanding of (adult) IBD is also reflected in the fact that there are currently only a few stratified/ personalized treatment strategies despite the recent expansion of therapies based on immune modulation, mostly using monoclonal antibodies. 25 Given these challenges, the precise mechanisms of IBD disease pathogenesis, the relationship between adult and early-onset IBD, and the complex interplay between host genetics and environmental factors have remained partially elusive with major gaps in our understanding in the genetic processes governing IBD pathology. 8

| Monogenic and Mendelian IBD
Very early-onset IBD (VEO-IBD) denotes a subgroup of IBD patients with a disease onset before the age of 6 years. 27 In contrast to adult IBD, VEO-IBD is a rare disease where mutations in causal genes may be inherited in a Mendelian fashion, as illustrated by our discovery of IL10R deficiency. 26 VEO-IBD patients usually present with a severe clinical course including (often bloody) diarrhea and abdominal pain. 27 Most patients with VEO-IBD receive immunosuppressive treatment, and many patients require surgical intervention during the course of their disease. 28 To date, there are only a handful of monogenic defects that result in a predominant IBD phenotype, including ADAM17, IL10, IL10RA, IL10RB, GUCY2C, IL21, LRBA, TTC7A, and XIAP. 26,[29][30][31][32][33][34] Identification of these gene defects have provided proof of concept for genetic diagnosis and stratified therapeutic choices, shaping our understanding of the immune system and illustrating molecular mechanisms underlying the delicate balance in keeping the homeostasis in the gut.
Intriguingly, a spectrum of inborn errors of immunity (IEI) can present with an IBD-like phenotype, sometimes as the initial disease manifestation. IEIs are a heterogeneous group of more than 330 different disorders with around 300 genes currently identified to be associated with monogenic, Mendelian forms. 35 The main characteristics of IEIs are increased susceptibility to infections due to improper function or dysregulation of key players of the immune system. These observations highlight the interesting fact that a spectrum of different immunopathological processes can underlie GI inflammation and point to the GI tract as an exceptionally sensitive site to immune disturbances. Current consensus estimates that about 20% of genetic defects underlying IEIs can develop bowel inflammation ( Figure 1A). The International Union of Immunological Societies (IUIS) recognizes 9 phenotypic groups of IEIs. 35 Among the functional groups of IEIs, diseases of immune dysregulation present most often with an IBD-like phenotype in up to 40% of the different genetic defects. On the other hand, complement deficiencies tend to present without bowel inflammation (95% of known gene defects do not cause IBD, Figure 1A). To date, considerably accelerated by the advent of next-generation sequencing, >60 monogenic diseases that present with IBD have been described 27,36 ( Figure 1A). Between the year 2015 and 2018 alone, several new gene defects have been identified that underlie some type of bowel inflammation ( Figure 1B).
Interestingly, some gene defects in subgroups of IEI do present with bowel inflammation, while other gene defects in the same group do not. While currently, there is no comprehensive and satisfactory explanation for the varying frequency of the IBD phenotype in individual gene defects, one can speculate that (a) due to the few patients and therefore small sample size in rare diseases, it is possible that certain phenotypes of inborn errors of immunity have not yet been captured, especially when it comes to disease with only one patient described at present, (b) our knowledge of the explicit effects of genetic aberrations is incomplete; therefore, it is plausible that in some gene defects counter-mechanism are in place and can maintain a pseudo-homeostatic state in the gut, therefore not inducing an IBD-like phenotype, and (c) since our understanding of the influence of factors extrinsic to genetic triggers is incompletely studied and understood in EO-IBD, it is likely that (similarly to adult IBD) in some cases the EO-IBD phenotype only emerges as a result of strong non-genetic triggers on a genetically susceptibly host.
Investigating the consequences of genetic aberrations in patients with monogenic defects causing IBD allows for a precise dissection of genotype-phenotype relationship. Moreover, through understanding of the mechanistic effect of pathogenic mutations on gene regulation, we have widened our knowledge on principal immune processes. Therefore identification of monogenic defects underlying IBD has not only provided genetic diagnosis to patients, but also proven to yield invaluable insights into how the immune system works. We here review monogenic defects underlying IBD and how dissection of their molecular pathophysiology has contributed to our understanding of immune homeostasis in the gut in health and disease.

| Epithelial barrier defects
The intestinal epithelium forms both a physical and biochemical barrier between gut microbiota and the immune cells within the mucosa. Therefore, dysregulation of the gut epithelium can result in immune overactivation that culminates in bowel inflammation. The onset of IBD can arise through the following mechanisms: (a) defects of epithelial organization, (b) defects leading to epithelial apoptosis and necroptosis, and (c) defects of epithelial-intrinsic cellular function.
TTC7A, a member of TPR domain-containing proteins is thought to have diverse functions in cell cycle control, protein transport, phosphate turnover, and protein trafficking and secretion. Patients with TTC7A deficiency typically present with features of severe combined immunodeficiency (SCID), along with severe exfoliative apoptotic enterocolitis. 30,37,38 In previous studies, mutations in TTC7A were reported to have multiple intestinal atresias (MIA) possibly due to the constant inflammation and apoptosis of the epithelium. It appears that patients with complete loss-of-function typically present with MIA-SCID phenotype, whereas milder (hypomorphic) mutations may present with EO-IBD as a predominant phenotype. 37,38 TTC7A-deficient patient-derived organoids show defective apicobasal polarity and have increased apoptosis that may cause a physical breach of the epithelium therefore aggravating the bowel inflammation. 37 However, the involvement thymic stromal-intrinsic Patients with Kindler syndrome have been reported to have ulcerative colitis, [39][40][41] and Fermt1 −/− mouse model shows gut epithelial detachment due to a lack of epithelial integrin activation. 42 This was F I G U R E 2 Cell types and molecular mechanisms involved in the pathogenesis of inflammatory bowel disease. The inner circle represents cell types and cell components involved in IBD pathogenesis, as detailed in the text. The middle circle depicts the molecular mechanisms affected by mutations in genes presenting with an IBD phenotype. The outer circle represents the molecular pathomechanisms leading to IBD. Treg IL10: T-cell immunodeficiencies with bowel inflammation and Defects in Tregs or IL10 signaling. Phagocytes: Congenital defects of phagocyte number or function. Complement: Complement deficiencies. Bacterial recognition: Defects in host-microbiota interactions, bacterial sensing. Epithelial barrier: Epithelial barrier defects. B cells and antibodies: Predominantly antibody deficiencies with IBD. Innate immune cells: Systemic autoinflammatory diseases and IBD. PRR: pattern-recognition receptor hypothesized to cause epithelial barrier breach, which culminated in bowel inflammation in this model. Mutations in the COL7A1 gene elicit an autoimmune response and autoantibodies to type VII collagen and cause epidermolysis bullosa dystrophica. 43 The mutated COL7A1 leads to a deficiency in anchoring fibrils, which in turn impairs the adherence between the epidermis and the underlying dermis similarly resulting in an impaired gut epithelial barrier.
Mutations in guanylate cyclase 2C (GUCY2C), an intestinal receptor for bacterial heat-stable enterotoxins cause relatively mild earlyonset chronic diarrhea and is associated with increased susceptibility to IBD, small-bowel obstruction, and esophagitis. 44 Although the exact molecular mechanism behind the familial diarrhea is yet to be determined, it has been shown that the expression of mutant GUCY2C results in increased production of cGMP, possibly underlying the hyperactivation of CFTR, leading to increased chloride and water secretion from enterocytes. Missense, splicing, and truncation mutations in SLC9A3, identified in nine patients from eight families lead to congenital sodium diarrhea (CSD). 45 Two of these nine patients developed IBD at 4 and 16 years of age. 45 SLC9A3 is an epithelial brush-border Na/H exchanger that uses an inward sodium ion gradient to expel acids from the cell. Several members of the SLC9A family of Na + /H + exchangers are expressed in the gut, with varying expression patterns and cellular localization. They participate in the regulation of basic epithelial cell functions, including control of transepithelial Na + absorption, intracellular pH, cell volume, and nutrient absorption, and also in cellular proliferation, migration, and apoptosis. In addition, these proteins modulate the extracellular milieu to facilitate other nutrient absorption and to regulate the intestinal microbial microenvironment. 46 The functional consequence of loss-of-function SLC9A3 gene variants (ie, reduced sodium uptake and proton exchange at the luminal surface) appears similar to that of gain-of-function (GOF) variants in the GUCY2C gene, showcasing a potential overlapping molecular mechanism. However, the underlying mechanism of bowel inflammation in these patients is unclear.
One potential hypothesis includes physical epithelial damage due to distended bowel resulting in microbiota-mediated immune activation and bowel inflammation.
Loss-of-function (LOF) mutations in the SLCO2A1 gene, encoding a prostaglandin transporter have been described to cause pediatric-onset chronic nonspecific multiple ulcers of the small intestine, accompanied with persistent blood and protein-losing enteropathy 47,48 in the Japanese population. Mutations in SLCO2A1 have been previously reported as the cause of primary hypertrophic osteoarthropathy (PHO). 49,50 Three out of five male patients with chronic enteropathy associated with SLCO2A1 had all of the major clinical features of PHO as well, such as digital clubbing, periostosis, and pachydermia. SLCO2A1, naturally expressed on the cellular membrane of vascular endothelial cells in the small intestinal mucosa, was absent from the patients' epithelium, pointing to a potential epithelial-intrinsic cell defect. Similarly, a LOF mutation in the PLA2G4A gene, encoding for cytosolic phospholipase 2α, has been identified in patients with cryptogenic multifocal ulcerating stenosing enteritis (CMUSE). 51 It was shown that these patients lack protein expression in their gut epithelium. Phospholipase 2α is an enzyme important in the formation of prostaglandin. Together, these gene defects point toward the role of prostaglandin in gut epithelial homeostasis, specifically in the context of epithelium-intrinsic defects. However, the exact molecular mechanism of prostaglandinassociated enteropathy is still unclear.
Familial hemophagocytic lymphohistiocytosis (FHL) is caused by recessive mutations that impair cytotoxic function and is characterized by fever, splenomegaly, bicytopenia, high triglycerides/low fibrinogen, hemophagocytosis, high ferritin, low natural killer (NK) cell cytotoxicity, and high soluble CD25. 52 FHL type 5 is initiated by mutations in the STXBP2 (Munc18-2) gene, encoding a protein involved in intracellular trafficking, the control of soluble NSF attachment protein receptor (SNARE) assembly, and the release of cytotoxic granules by NK cells. 53 Notably, MUNC18-2 deficiency (unlike other FHL) is often accompanied by colitis, 53 although GI symptoms are not a common feature of FHL type 2, FHL type 3, or Griscelli Syndrome type 2 patients, suggesting that the pathology of FHL does not necessarily lead to GI disease, even in the more severe FHL subtypes. Munc18-2 proteins have been described to have widespread expression in epithelial tissues, such as the kidney and intestines, with localization to the apical surface of the plasma membrane. 54,55 Thereby, Munc18-2 might be essential for maintaining epithelial integrity in GI epithelial cells, but more mechanistic studies are required to determine the how Munc18-2 deficiency lead to bowel inflammation. Mutations in the CYBB and CYBA, NCF1, NFC2, and NCF4 genes, encoding for the cytosolic subunits of NOX, abrogate its activity and compromise host immunity against certain bacteria and fungi. These defects cause chronic granulomatous disease which are characterized by immunodeficiency and can cause IBD-like intestinal inflammation. 57 Inflammatory reactions in CGD patients (namely colitis) might be a result of impaired anti-bacterial protection due to impaired NOX activity, resembling defects in ephithelial-specific NADPH Oxidase 1 (NOX1) and Dual Oxidase 2 (DUOX2) in patients with severe EO-IBD.

| Congenital defects of phagocyte number or function
Both NOX1 and DOUX2 are epithelial NADPH oxidases involved in the generation of ROS in the gut epithelium. 58 Mutations in NOX1 and DUOX2 result in reduced ROS production and cause a 10-fold increase in bacterial invasion. 59 Impaired mucosal defense may represent a key pathomechanism that results in intestinal inflammation and development of IBD. Another possible pathomechanism leading to colitis in CGD patients is inflammmasome hyperactivation.
Intriguingly, NOX-deficient mice exhibited a skewed Th17 phenotype suggesting a possible role of pathogenic Th17 cells in development of inflammatory reactions. 60 These data indicate that while reactive oxygen species are used by the immune system to eliminate infections they may also serve as signaling intermediates to coordinate the efforts of the innate and adaptive immune systems resulting in a complex etiology underlying phagocyte defects.
Although the exact molecular link has not been established yet, it has been shown that mononuclear phagocytes from CGD pa-

| Defects in host-microbiota interactions, bacterial sensing
Nucleotide-binding and oligomerization domain (NOD)-like receptors act as a first line of defense against invading bacteria. Within the NOD family, NOD2 functions as an intracellular sensor for peptidoglycans from the bacterial cell wall. NOD2 has long been studied and is recognized as a critical player in Crohn's disease pathogenesis, where it was shown to regulate innate immunity through NF-κB-induced proinflammatory responses. 12 Intriguingly, single gene defects involving NOD2 cause Blau syndrome, an inflammatory disorder phenotypically characterized by the triad of granulomatous polyarthritis, dermatitis and uveitis, however without bowel inflammation. 69 In this context, it is postulated, that gene defects that do not directly disrupt NOD2 function, but rather de-regulate proper NOD2 signaling, do present with IBD, whereas at least the Blau syndrome-associated mutations in NOD2 do not. The very first of discovery relating IBD to defective NOD2 signaling without NOD2 mutations was XIAP deficiency. 29,70 X-linked lymphoproliferative (XLP) disease is a rare immunodeficiency caused by mutations in the SH2D1A/SAP or XIAP genes, respectively. XLP is characterized by severe immune dysregulation that presents with susceptibility to EBV-triggered lymphoproliferative disease (EBV-LPD) or hemophagocytic lymphohistiocytosis (HLH), lymphoma, and dysgammaglobulinemia. 71 71,76 The IBD phenotype in XLP2 is hypothesized to be brought on by abrogated NOD2-mediated signalling and result in innate and adaptive immune defects including granulomatous colitis and perianal disease. Therefore, it is postulated hat colitis may be clinically and pathologically different between XLP1 and XLP2. 77 Two additional novel gene defects that influence NOD2 signaling and present with bowel inflammation have been described recently. Mutations in the NPC1 gene, encoding a protein that mediates intracellular cholesterol trafficking of endosomes and lysosomes, cause a neurodegenerative lysosomal storage disease, coupled with fistuling colitis with granuloma formation. 70 The pathogenic mutations in NPC1 is thought to elicit impaired autophagy due to defective autophagosome function. Similar to XIAP deficiency, mutations in NPC1 abolishes NOD2-mediated bacterial handling. However, NPC1 mutations do not impair RIPK2-XIAP dependent cytokine production.
Identification of patients with homozygous TRIM22 mutations provided additional links of NOD2 to VEO-IBD. TRIM22 is a ubiquitin ligase that influences NOD2 activity by ubiquitination. 78 Mutations in TRIM22 disrupt the ability of TRIM22 to regulate NOD2-dependent activation of IFNβ signaling and NFkβ.
Intriguingly, LOF variants in NOD2 have been shown to result in the loss of NF-κB-induced proinflammatory cytokine response to muramyl dipeptide (MDP), 12

| Predominantly antibody deficiencies with IBD
The molecular mechanisms of IBD in patients with defects in humoral immunity are not yet completely understood. Impaired antibody production, especially low IgA, may contribute to the development of gut dysbiosis, but the defects in antibody deficiency alone do not result in intestinal disease. The pathomechanism in this case is most likely due to combined T-and B-cell defects. 80  during these episodes. 88 Although the IBD phenotype is currently attributed to intrinsic defects of the gut tissues, whether defects in humoral immunity may contribute to GI inflammation is still to be investigated. Bone marrow transplantation in two patients led to encouraging results, although more long-term data are needed to clarify the disease etiology.
Inducible T-cell costimulatory (ICOS) is an activation-induced member of the CD28 family on T cells. Mutations in the ICOS gene cause ICOS deficiency, presenting with common variable immunodeficiency (CVID) including splenomegaly, autoimmune manifestations, recurrent bacterial infections, and IBD. 89 Absence of ICOS results in abrogation of germinal center formation leading to severe reduction of class-switched memory B cells, as well as reduction in naïve B cells. The presumed cause of the IBD phenotype in ICOS deficiency is insufficient IL-10 production by ICOS-deficient T cells. 89

| T-cell immunodeficiencies with bowel inflammation
Gene defects that disturb adaptive immune cell selection, ac- goclonal T-cell repertoire, which indicates that the thymic selection in patients with Omenn syndrome is restricted to the T cell in which recombinase activity is sufficient to generate a functional TCR. 97 Adenosine deaminase (ADA) deficiency leads to an accumulation of toxic purine degradation by-products, most potently affecting lymphocytes, but other manifestations include skeletal abnormalities, neurodevelopmental affects, and pulmonary manifestations associated with pulmonary-alveolar proteinosis. 98 The major consequences of ADA mutations are severe depletion of T and B lymphocytes and NK cells. The underlying mechanisms of this deleterious effect are the increased apoptosis due to the buildup of dATP in cells especially in developing thymocytes and T cells. 99 Although patients present with severe B-lymphocytopaenia and hypogammaglobulinaemia, B-cell development seems to be unaffected. 100 Interleukin receptor common gamma chain (IL2RG), is a cytokine receptor subunit that is common to the receptor complexes of at least six different interleukin receptors: IL-2, IL-7, IL-9, IL-15, and IL-21. 101 110 To date, both these report lack direct conclusion about the molecular mechanism, but highlighted chromosomal stability as one of the influential factors of IBD pathogenesis. One could hypothesize that as DNA repair is important in the context of T-and B-cell maturation through V(D)J recombination, the development of IBD may be pinpointed toward lack of immune regulation.
Defects of telomere maintenance, exemplified by mutation in DKC1 and RTEL1 underlie dyskeratosis congenital myelodysplasia which can present with IBD. [111][112][113][114] In these cases, manifestation of GI inflammation can be one of the first presenting symptoms as reviewed by Jonassaint et al. 115 They proposed that the onset of GI inflammation is due to defective epithelial barrier function as they found that these patients present with extensive apoptosis in the intestinal mucosa, potentially resulting in the breach of the epithelium and unprecedented activation of the gut immune system. However, it is likely that the T-cell deficiency has an additional pathogenic role in the onset of bowel inflammation in these diseases.

| Defects in Tregs or IL10 signaling
The discovery of biallelic LOF mutations in the IL-10 receptor genes presenting with bowel inflammation as the main phenotype have highlighted the pivotal role of IL-10, and IL-10 in T reg cell function especially in the gut. Defects in the IL-10 receptor genes IL10RA and IL10RB and IL-10 itself lead to early-onset enterocolitis involving hyperinflammatory immune responses in the intestine due to abrogated interleukin-10-induced signaling and therefore improper function of regulatory T cells. 26 Similarly, immune defects abrogating proper T reg function can lead to bowel inflammation as well.
Immune dysregulation, polyendocrinopathy, and enteropathy (IPEX) is caused by mutations in the FOXP3 gene, a master regulator of the development and function of Tregs. In IPEX, the lack of or mutant Patients with mutations in ADAM17 present with early-onset pustular dermatitis, short and broken hair, paronychia, frequent cutaneous bacterial infections, cardiomyopathy, and early-onset diarrhea. 162 In a study of two related patients, patient-derived PBMCs showed high levels of lipopolysaccharide-induced production of interleukin-1β and interleukin-6 but impaired release of TNFα. 162 ADAM17 plays a role in the processing of other cell surface proteins, including a TNF receptor, the L-selectin adhesion molecule, and transforming growth factor-alpha (TGF-α). 163 Although direct links between the patient's phenotype and ADAM17 defects is still elusive, lack of TNFα is considered partly responsible for the increased susceptibility to infection and development of cardiomyopathy, and as Adam17 knockout mice present with impaired epithelial cell maturation in multiple organs, the lack of proper epithelial barrier could be postulated to stem the IBD phenotype. Numerous polymorphisms in MASP2 that causes lack of MBL pathway activation have been identified, 166   Patients may have splenomegaly, and in some cases, hepatosplenomegaly. Diarrhea, vomiting, and dehydration may also occur. 174,175 Pathogenic mutations in PEPD lead to reduction or loss of prolidase activity which may contribute to the multifactorial clinical presentation. Since phenotype, age of onset, and clinical course of prolidase deficiency are very variable even within the same family, and the number of molecularly characterized patients is very small, it is still difficult to define a genotype-phenotype relationship for this disease. 173 Complex dysregulation of transforming growth factor beta as a result of autosomal dominant mutations in TGFBR1 and TGFBR2

| Other gene defects
(Loeys-Dietz syndrome) cause a syndrome with a variety of phenotypes including skeletal involvement, arterial abnormalities and immunological abnormalities, IBD, and encelopathy. 176 Recently, biallelic LOF mutations in the TGFB1 gene encoding TGF-β1 have been described in patients with central nervous system disease including epilepsy, brain atrophy, and posterior leukoencephalopathy, and severe VEO-IBD. 177 The mutations in TGFB1 seemingly impaired the bioavailability of TGF-β1. Although the exact mechanisms of how impaired TGFβ signaling leads to IBD is yet to be determined, these findings suggest a pivotal role in of TGFβ immune function, especially in intestinal immune homeostasis.  Abatacept (CTLA-4-Ig) has proved to be successful in alleviating the infiltrative and autoimmune disease. 125,181 In the case of a genetic mutation in a gene that affects both the immune and epithelial barrier (for example TTC7A deficiency),

| G ENOMI C S AND ITS INFLUEN CE ON THER APEUTI C G U IDELINE S FOR VEO -IB D PATIENTS
HSCT did not correct for the epithelial-intrinsic defect and enteral tolerance. 182 This further highlights the importance of identifying underlying genetic cause of VEO-IBD to reduce treatment-related mortality. More research needs to be performed in order to elucidate the roles of gene defects in cell types in which they were not implicated before.

| BOWEL INFL AMMATI ON AND THE MICROB IOME
While the link between gut inflammation and gut dysbiosis is not a novel concept, the development of culture-independent techniques like next-generation sequencing and metagenomics exploded the field of microbiome-related studies. These techniques enabled the global assessment of the gut microbiota more accurately and in a more sophisticated manner. 183,184 The largest and perhaps the most ambitious initiative that has emerged in the last decade to study the changes of the human microbiome in health and disease is the NIH sponsored Human Microbiome Project (HMP). 185 It has resulted in the publication of 5177 microbial taxonomic profiles from a population of 242 healthy adults and serves as a comprehensive database for research in this field. 186 This project was followed up by the second phase that, in addition to phylogenic composition, aimed to analyze functional omic data including transcriptome, proteome, and metabolome. Such multi-omic approaches with simultaneous analysis of host and microbiome proteins and metabolites aimed to better our understanding of the biology of the microbiome and sophisticated molecular mechanisms of host-microbiota interaction. 187 Such integrative analysis is the key feature of the future microbiome research. 188,189 While microbiota from some body sites (for example skin) is easily accessible, the GI tract is much more challenging to sample and describe. The complex structural and functional features of the human GI tract is reflected by the differences in abundance and composition of bacteria and their dynamic variations along the intestine make human microbiome studies complex. 190 The excitement in studying the gut microbiome is not only driven by the fact that it is perhaps the most abundant and complex microbial community of the human body, but that it has also been associated with the development of wide spectrum of diseases. Indeed, numerous studies, including those that use integrative analysis of human gut microbiome and metabolome, have associated the gut microbiota with the promotion of health and development of IBD, obesity-related inflammatory disorders, allergic diseases, and infectious diseases. 191 Although the correlation between gut dysbiosis and IBD is well appreciated, the role of microbiome perturbations in disease development is not yet clearly defined. 192 The role of in two most abundant fila-Firmicutes and Bacteroidetes. 187,[193][194][195][196] While microbiome perturbations in IBD can have a complex etiology, dysbiosis in patients with VEO-IBD or IEI is driven primarily by the gene defects. A study of the gut microbiome in CVID patients showed significant differences in bacterial composition with dysbiosis and low alpha diversity characteristic of the patients with IBD. 197 Interestingly, while elevated dysbiosis index was a characteristic of patients "with infections only" and "with complications" subgroups, the latter had also reduced alpha diversity of the gut microbiota. Patients with enteropathy within the "with complications" subgroup did not show any significant differences in gut microbiota.
The lack of obvious differences in microbiomes between patients with or without gut pathologies in this study is difficult to explain.
No systematic studies to date involving genetically characterized The role of vitamins in maintaining T regs , as well as a number of lymphocytes and NK-cell activity, has also been established. 198 In this case, the effect is mediated through specific receptors broadly expressed on various subsets of immune cells.
The modulation of inflammasome signaling by bacteria-derived metabolites is another distinct mechanism involved in modulation of host immunity. Recent studies implicated several low-molecularweight compounds associated with metabolism, not immunity, in regulation of NLRP3 and NLRP6 activation. 201 In a recent study, it has been shown that microbial metabolites taurine, histamine, and spermine modulate NLRP6 inflammasome signaling, secretion of IL-18, and production of anti-microbial peptides shaping the hostmicrobiome interface. 202 The discovery of bacteria-specific vitamin B metabolites rec- on host-related metabolic pathways has been suggested. 208 The nature and efficacy of such potentially bioactive metabolites to be used for the therapy require further exploration.

| ORG ANOIDS
Although single gene defects affecting major immune pathways have been investigated in detail, little is known about how mutations of VEO-IBD-associated genes are involved in epithelial barrier function or the homeostasis of the host and microbiome. As the interplay between immune cells, gut epithelial barriers, and the gut microbiota represents a central axis in the onset of VEO-IBD, it is necessary to study the diverse genetic influences of EO-IBD-associated genes in these three players. Gut-derived organoid technology has been at the forefront of advancing our understanding of gut homeostasis, in particular in studying the elusive biology of the gut epithelium.
This technology has allowed us to take a reductionist approach in studying specifically gut epithelial derived from any area of the gut and has been shown to be crucial to further our understanding of the biology of some VEO-IBD genes. For example, the role of TTC7A in controlling the polarity of the gut epithelium has been shown in organoids derived from patients with TTC7A deficiency. 37 Patient-derived organoids provide a potential wealth of resources  209 In the future, generation of biobanks with patient-derived gut organoids with genetically defined backgrounds will allow us to advance personalized medicine for this heterogenous group of rare diseases. However, despite all efforts and recent advances, the genotypemolecular mechanism-phenotype link is still missing for many of the significant GWAS loci. It has been observed that although one expects GWAS signals to cluster in disease relevant pathways and genes, association signals for complex traits tend to be spread across most of the genome including in the vicinity of numerous genes without a clear connection to disease. An "omnigenic" view of diseases proposes that in contrast to Mendelian diseases, which are often caused by high impact mutations in protein-coding regions in a few genes, complex traits and diseases are mainly driven by lower impact variants that affect a multitude of genes and pathways often outside of the primary pathways and genes involved in the respective diseases. Given the interconnected nature of cellular systems, these lower impact variations in diverse pathways converge and create the disease phenotype. 210 The assessment of such combinatorial effect requires complex models of molecular networks and integration of multiple datasets.

| WE S , PANEL S , AND WG S
High-coverage exome or targeted exome (also known as panel) sequencing are routinely used to identify genetic aberrations causing early-onset bowel inflammation. Due to its high accuracy and moderate costs, panel sequencing is preferentially used as a screening method in many institutes. A combinatorial method is to use exome or genome sequencing, but prioritize "virtual panels"-a selected list of genes-for an initial screening and later extend the scope of the analysis to novel genes. When relying on a targeted panel approach, the design of the targeted panel is a crucial step toward appropriate gene discovery and diagnosis. On one hand, by building the panel by using only genes whose link to the disease phenotype is clear, one might miss important new genes and diagnosis.
Conversely, by including candidate genes (whose association to the phenotype is yet unknown), there is an increased chance of identifying variants of unknown significance. While WES conveys an advantage of looking at all known coding sequence and can be combined with a panel approach, incidental findings and variants of unknown significance are much higher.
The percentage of patients with genetically diagnosed VEO-IBD varies between centers and cohorts and ranges from 5%-31% depending the composition of the cohort. 211

| Increasing need of high-quality metadata
In recent years, large-scale computation methods have been initiated to investigate the etiology of IBD. In these projects, researches often rely on public databases to provide them with data. Access to accurate genetic data is facilitated by resources such as Decipher, 239 HGMD, 240 OMIM 241 and ClinVar 242 that aim to aggregate clinical data. These resources and public data repositories are still incomplete and need to be queried manually or with specifically set-up local bioinformatics pipeline. These efforts are welcome steps toward efficient data access, but some issues with redundancy remain and variable quality and quantity of data that is still missing from these resources. A robust, unified database could be an approach worth considering.
It is becoming increasingly clear that beyond efficient access to genomic and variant information, there is a need for accurate metadata to describe clinical information not only in a genetic manner, but also phenotypically. Annotation of patients with accurate phenotype data, as well as the annotation of genes with pathways and molecular mechanisms requires standardized and objective language. Therefore, it is crucial to have a unified nomenclature and resource of disease-causing genes annotated with the corresponding physical, molecular, and cellular phenotypes. Along with annotating patients and genes with the correct disease and ontology, intrainstitute and laboratory collaborations would benefit immensely from precise and objective descriptions of phenotypic, molecular, and genetic abnormalities.
Human Phenotype Ontology (HPO) is a phenotype vocabulary initially published in 2008. 243,244 It is a tool that enables accurate phenotyping which further facilitates efficient data and patient exchange. HPO is being increasingly adapted into everyday use as the standard to describe phenotypic abnormalities. Gene ontology (GO) on the other hand, is a computational representation of the function and localization of genes and gene products on the molecular level.
Currently, the GO project has developed and constantly revised over 40 000 biological concepts and annotations. 245

| B E YOND G ENE TIC S: IB DOMI C S-SYS TEMS B I OLOGY AND INTEG R ATIVE ME THODS
One of the limitations of current approaches toward uncovering the key players in IBD pathology is that they are looking at individual contributors separately. In a case of IBD, which arises as a result of the combinatorial effect in multiple key players such as genetics, environmental factors, microbial perturbations, epigenetics, and lifestyle factors, understanding the disease cannot be tackled by studying each pathogenic component in isolation, without considering the interaction among the different "omes." Integration of the different "omes" requires intelligently designed data integration and processing. This integrative approach, recently introduced as the IBD interactome or IBDome, calls for new concepts and tools to implement a systems biology approach toward unraveling the processes behind IBD. 252 These tools should rely on unbiased data-driven investigation and include strategies to reveal key drivers and pinpoint central players of inflammation and enable development of targeted therapies.
Querying and integrating multiple "omes" with bioinformatics methods enables the integration of genomic, epigenomic, transcriptomic, proteomic, metabolomics, and microbiomic data to construct a comprehensive molecular map of IBD. Although seamless data integration is yet to be available, methods to integrate various types of information and use it toward identifying key players of IBD are already underway.
Recently, Peters et al 253 used individual networks constructed from molecular data generated from intestinal samples isolated from three populations of patients with IBD at different stages of disease, including two adult and pediatric cohorts of IBD. As a result, they developed a predictive model of the immune component of IBD that informs causal relationships among loci previously linked to IBD through GWAS using functional and regulatory annotations that relate to the cells, tissues, and pathophysiology of IBD. This network revealed potential key drivers of IBD pathogenesis. Among the key drivers were numerous known VEO-IBD genes, but also inborn errors of immunity and new candidates that were showed to have a role in inflammatory immune response. 253

| CON CLUS I ON AND FUTURE PER S PEC TIVE S
IBD is a complex, multifactorial condition with an onset brought on by a multitude of factors that can also present in rare, Mendelian fashion. While the direct link between the more common, adult version of IBD and rare, Mendelian EO-IBD is still elusive, we have gained tremendous understanding of different key players in immune regulation and cellular mechanisms required for immune homeostasis in the gut over the last years. The considerable fraction of inborn errors of immunity presenting with an IBD-like phenotype highlight that all patients presenting with VEO-IBD should be subjected to detailed genetic and immunological examination and investigation, as IBD can be an early sign of an underlying immunodeficiency. Recent advances in genomic technologies, organoid systems, as well as our increasing understanding and modeling of the interplay between the gut microbiota, immune cells, epithelial cells, and the environment promise to shed new light on to the complex molecular network behind IBD pathology. This novel understanding could allow for more efficiently identification of patient subgroups, and therefore increasingly direct treatment strategies toward personalized medicine.

This work was supported by a DOC Fellowship of the Austrian
Academy of Sciences (24486) to R.C.A. and Austrian Science Fund (FWF) grant I2250-B28 to K. B.

CO N FLI C T O F I NTE R E S T
The authors declare that there are no conflict of interests.