The Effect of In Utero Exposure to Maternal Inflammatory Bowel Disease and Immunomodulators on Infant Immune System Development and Function

Abstract

Autoimmune and inflammatory disorders, including inflammatory bowel disease (IBD), commonly affect women of childbearing age, warranting the use of immunomodulatory agents at a time where pregnancy may be desired. In utero exposure to pro-inflammatory mediators from maternal IBD, IBD-associated intestinal dysbiosis, and immunomodulatory drug use may impact neonatal immune system development during what is considered to be a critical period, with potential long-lasting impacts on susceptibility to disease.

Both the innate and adaptative immune systems of the neonatal differ to that of the adult in terms of both cellular composition and sensitivity to antigenic and innate stimulation. The infant immune system gradually develops to more closely resemble that of the adult. Exposure to maternal inflammation in utero may aberrantly impact this period of infant immune system development, with maternal autoimmune and inflammatory disorders shown to affect the physiologic changes in serum cytokine abundance observed during pregnancy.

The maternal and neonatal intestinal microbiome greatly influence infant mucosal and peripheral immune system development, and thereby impact the susceptibility to short-term inflammatory diseases, the adequacy of vaccine response, and later life risk of atopic and inflammatory disorders. Maternal disease, mode of delivery, method of feeding, time of weaning to include solid foods in the diet, and neonatal antibiotic exposure all influence the composition of the infant microbiome, and thereby infant immune system maturation.

How exposure to specific immunosuppressive medications in utero alters infant immune cell phenotype and response to stimulation has been explored, but with existing studies limited by the time at which samples are performed, heterogenicity in methods, and small sample size. Furthermore, the impact of more recently introduced biologic agents have not been explored. Evolving knowledge in this field may influence therapeutic preferences for individuals with IBD planning to conceive, particularly if substantive differences in the risk of infant infection and childhood immune disease are identified.

Summary.

Immunomodulatory agents are commonly required for the management of inflammatory bowel disease and other inflammatory disorders in pregnancy. The effects of such agents and maternal inflammation on infant immune system development following in utero exposure should factor into risk stratification and therapeutic decision-making.

Chronic inflammatory conditions, including autoimmune diseases, affect 8% of the adult population, with a significant female preponderance and a predilection for onset in the childbearing years.¹ Fifty percent of patients with inflammatory bowel disease (IBD) are diagnosed under the age of 35, and 25% of women in this cohort conceive their first child following their diagnosis of IBD.² IBD often necessitates immunomodulatory therapies during a stage in life when pregnancy may be desired: in an Australian tertiary IBD pregnancy service, less than 10% of pregnant patients were not receiving any medical therapy, with over 60% receiving a biologic agent, and almost 50% receiving a thiopurine.³ A similar Danish cohort reported rates of biologic use in pregnant patients of 14.5% for ulcerative colitis (UC) and 30% for Crohn’s disease (CD), with 40.5% of those with CD receiving a thiopurine, and 46% with UC exposed to a 5-acetylsalicylic acid.⁴

The first 3 months of life are hypothesized to be a critical window of time in which immunological demands shape longer-term immune function, including susceptibility to disease, risk of atopy, and response to vaccination.5, 6, 7 The transition from a tolerogenic fetal immune system with a dominance of T helper 2 (Th2)/type 2 immune polarisation and functionally immature neutrophils, monocytes, natural killer (NK) cells, and dendritic cells (DCs), to a neonatal immune system that efficiently protects from pathogens in the initial antigen-rich post-partum period, is essential to avoiding severe postnatal infection.

As the therapeutic armamentarium expands, so does the need to personalize IBD therapy with consideration for the reproductive preferences of the individual. In this review, we aim to describe the changes in immune system composition and function that physiologically occur in pregnancy and explore the known impacts of IBD and immunosuppressive medication exposures on these adaptations. We describe critical differences between the adult and neonatal immune system and summarize the functional immunological and clinical implications to the infant following in utero exposure to IBD relevant immunomodulators, maternal inflammation, and biologics. Such knowledge will assist in informing rational therapeutic decisions in individuals with IBD and a childbearing wish, while also identifying areas where further research is needed.

Differences in the Neonatal and Adult Immune Systems

Innate Mediators

Innate immune system activators, including pathogen-associated molecular patterns (PAMPs), danger associated molecular patterns (DAMPs), toll-like receptors (TLRs), and the intracellular nucleotide oligomerization domain-like (NOD) receptors, are examples of pattern recognition receptors (PRRs).⁸ PRRs, located on or within antigen-presenting cells, respond to innate immune system activators, and their activity needs to be carefully regulated; insufficient responses will enable pathogenic organisms to infiltrate relatively undetected, whereas excessively sensitive responses to PAMPS and DAMPs may lead to deleterious inflammation directed against commensal microbes or the host themselves.⁸ Expression of TLR1, TLR2, TLR3, TL4, TLR8, and TLR9 on monocytes and lymphocytes the cord blood of full-term neonates is comparable to that of adults, including following ex vivo stimulation^9,10; however, their function was reported to be different under some circumstances.¹¹ For example, TLR1-, TLR2-, TLR4-, and TLR6-mediated induction of type I interferons (IFNs), tumor necrosis factor (TNF) and interleukin (IL)-1α from stimulated cord blood-derived monocytes and conventional and plasmacytoid dendritic cells was significantly decreased compared with adults.^10,12 TLR4-mediated IL-1β secretion in response to bacterial lipopolysaccharide (LPS) was also lower in the cord blood of term neonates vs adults.¹²

Data specifically addressing how maternal immunosuppression or maternal inflammation affects the expression and activity of PRRs is presently lacking. However, variations in TLR responsiveness and expression have been observed in neonates.¹⁰ For example, in human neonatal enterocolitis (NEC) and murine NEC models, TLR4 expression was increased in the small intestinal mucosa.^13,14 Supporting a role for TLR dysregulation in the pathogenesis of NEC, augmenting IL-37, which protected mouse pups from NEC, decreased TLR4.¹⁴

Myeloid Cells and Neutrophils

Chemotaxis is reportedly reduced in neonatal neutrophils and monocytes, due to a decrease in L-selectin (CD11b), complement receptor 3, and membrane attack complex-1 (CD18).¹⁵ Neonatal neutrophils were also shown to be less responsive to chemokines as a result of lower pliancy from actin polymerization, leading to reduced endothelium penetration.¹⁶ In vitro studies showed that the chemotactic velocity of neonatal neutrophils is less than that of children and adolescents; however, over time a similar percentage of neutrophils successfully migrate towards a chemotactic gradient.¹⁷ This difference in neutrophil chemotaxis may attenuate localized inflammatory responses in neonates.¹⁷

Epigenetic changes have been observed in innate immune cells of infants exposed in utero to maternal inflammation, resulting in an increased susceptibility to atopic disease.^18,19 A low maternal third trimester IFN-γ:IL-13 ratio has been associated with an increased risk of childhood asthma, with this ratio positively correlating with neonatal cord blood mononuclear cell IL-6 production in response to LPS stimulation.^18,19 This finding may be related to the hypermethylation of CpG sites of genes (like ZFYV E9/SARA) regulating the innate immune response to microbes.^18,19

Innate Lymphoid Cells

Innate lymphoid cells (ILCs) largely reside in tissues and lack diversified antigen receptors; instead, they respond rapidly to the presence of resident tissue cell-derived cytokines and shared microbial epitopes.²⁰ Similar to T cells, ILCs can polarize, sharing the T cell signature transcription factors and cytokines (Table 1).²¹ ILC3s predominate in the colon and serve a multitude of roles, including in the maintenance of mucosal immune integrity.²² ILC3s express RORγt and characteristically produce IL-17 and/or IL-22.23, 24, 25 Via the production of IL-22, ILC3s increased epithelial Fut2 expression, thus promoting intestinal epithelial surface fucosylation and providing an environmental niche for commensal bacteria.²⁶ Furthermore, IL-22 stimulated antimicrobial peptide and mucus production by epithelial and goblet cells, respectively, which further preserved intestinal barrier functions.²⁷ By failing to express necessary co-stimulatory molecules, mucosal ILCs also limit T cell responses to bacterial commensal antigens.³⁸ Aberrant polarisation and function of neonatal ILCs has been implicated in intestinal and systemic disease^14,39; for example, protective NKp46⁺RORγt⁺ ILC3s were reduced in number in murine necrotizing enterocolitis models, whereas pro-inflammatory NKp46⁻RORγt⁺Tbet⁺ ILC3 populations were increased compared with control pups.^14,39 Although direct evidence linking maternal inflammation and aberrant infant ILC function is lacking, murine models demonstrated that the maternal intestinal microbiome promoted ILC3 expansion in their offspring.⁴⁰ As discussed subsequently, intestinal dysbiosis occurs more commonly in pregnant women with IBD,⁴¹ and might influence abundance and function of ILCs in the neonate.

Table 1.

Major T Helper/ILC Subtypes

	Induced bya	Transcription factor	Signature cytokines	Role in immune defensea	Activity in neonatesa
Th1	IFN-y, IL-12, IL-18	T-bet	IFN-y	Intracellular pathogens (bacteria, viruses)	Subdued, but inducible in neonates28,29
ILC1	IL-12, IL-15				Expressed in neonatal liver, lung, thymus, and bone marrow.30, 31, 32 Minimal data regarding activity in humans.
Th2	IL-4	GATA-3	IL-4, IL-5, IL-13	Parasites and helminths	Predominant Th subtype in neonates, contributes to immune homeostasis28,33
ILC2	IL-25, IL-33, TSLP	GATA, RORγt	IL-5, IL-13	Parasites and nematodes, support epithelial barrier recovery following tissue damage	Located in the respiratory tract of infants, with a role in the pathogenesis of bronchiolitis32,34
Th17	TGF-p, IL-6	RORγt	IL-12, IL-21, IL-17A	Extracellular pathogens (fungi, bacteria)	Subdued, but inducible in neonates35,36
ILC3	IL-7, TSLP, IL-23, IL-15, IL-113		IL-22	Preserve intestinal mucosal barrier functions and promote an environmental niche for commensals	Implicated in the pathogenesis of NEC, neonatal pneumonia, and bronchopulmonary dysplasia.14,23, 24, 25,27,29,37
Treg	TGF-p, IL-10, IL-2	FOXP3	TGF-p, IL-10	Regulate T cell-mediated inflammation and facilitate tolerance	Increased tendency towards Treg phenotype following TCR engagement, but decreased capacity to suppress antigen presentation and T cell interaction28

IFN, Interferon; IL, interleukin; ILC, innate lymphoid cell; TCR, T-cell receptor; TGF, transforming growth factor; Th1, T helper 1; Th2, T helper 2; Th17, T helper 17; Treg, T regulatory cell.

Adapted from Leung S, Liu X, Fang L, et al. The cytokine milieu in the interplay of pathogenic Th1/Th17 cells and regulatory T cells in autoimmune disease. Cell Mol lmmunol 2010;7:182-189 and Zhu X, Zhu J. CD4 T helper cell subsets and related human immunological disorders. Int J Mol Sci 2020;21:8011.

^aIncluding, but not limited to.

T Cells

The majority of studies exploring neonatal peripheral blood cellular composition and function have relied on the use of cord blood, which may substantially vary from neonatal peripheral blood.³³ Polarization of T cells and ILCs is dictated by specific transcription factors and signature cytokines (for example, T-bet and IFN-γ for Th1)²⁸ (Table 1). The effects of in-utero exposure to maternal inflammation on the fetus/neonate, including in the setting of chorioamnionitis, systemic infection, and autoimmune disease, depend on how severely the fetus is affected.³³ For example, mild fetal inflammation caused by maternal chorioamnionitis was associated with enhanced CD4+ Th2 polarization, whereas in more advanced stages, Th1 polarization predominated.³³ Other studies reported that maternal inflammation promoted neonatal Th17 and Th1 responses, with such changes impacting the delicately balanced homeostasis of the fetal/neonatal immune system.^33,42, 43, 44 In cord blood of neonates born to women with autoimmune conditions such as systemic lupus erythematosus, mixed connective tissue disease, scleroderma, or rheumatoid arthritis, the serum abundance of TNF, IL-1β, IL-6, IL-17, IL-4, and IL-8 was more than 2 standard deviations above the control mean.⁴² Cord blood peripheral blood mononuclear cells (PBMCs) from neonates born to mothers positive for anti-Ro/SSA and anti-La/SSB antibodies were shown to have significantly higher expression of genes related to IFN signaling and responses and higher serum IFNα.⁴³

Reportedly, in the setting of chorioamnionitis, maternal neutrophil-derived inflammatory signals and mediators such as MPO and matrix metalloproteinase 9 increase Th17 relative to Treg responses in the fetus, thus possibly predisposing to the development of subsequent chronic inflammatory disorders.⁴⁵ In cord blood of infants exposed to histological chorioamnionitis, progenitor Th17 cells were observed at increased frequency relative to controls.⁴⁴ These same infants were also shown to have increased T cell CD69, CD35, and Ki67, signifying heightened T cell activation proliferation. FOXP3 expression was also lower in these infants, as was the frequency of CD25^hiCD127^lo Tregs, which play an important role in maintenance of peripheral self-tolerance, in a similar study.⁴⁴ Increased neonatal IL-33 abundance associated with chorioamnionitis may promote excessive Th2, eosinophil, basophil, and mast cell activation,^46,47 thereby increasing the risk of childhood asthma.⁴⁸ PPP4C (protein phosphatase 4 catalytic subunit) expression is increased in response to T cell damage, playing a role in T cell homeostasis as well as in promoting thymocyte development.^49,50 Chorioamnionitis results in the methylation, and hence reduced expression, of this gene,^51,52 and so may aberrantly affect neonatal immune system development. As such, avoidance of maternal inflammation in pregnancy by achieving remission preconception and maintain this remission through pregnancy, should be prioritized.

Unsurprisingly, 70% to >90% of T cells in neonatal peripheral blood are naïve (CD45RA⁺CD45RO^-), as neonatal T cells are yet to be exposed to a wide variety of antigenic stimuli.⁵¹ Respiratory mucosal T cells, which are more commonly memory T cells in neonates and infants, were shown to be predisposed towards the formation of a specific effector rather than a resident memory (CD103⁺) phenotype. Neonatal mice infected with HSV-1 were shown to have a lower numerical peak in HSV-1 antigen-specific CD8⁺ T cells at the time of initial infection, but similar numbers of memory T cells 6 weeks post infection, and comparable CD8⁺ numbers following secondary infection. This secondary response, however, was characterized by proliferation of CD8⁺ T cells of limited TCRβ clonal diversity, thus failing to achieve an antigenic response comparable to that elicited by vaccination in adulthood. These findings may partially explain the frequent recurrences of respiratory tract infections observed in neonates and infants.^51,52

B Cells

B cells mature in secondary lymphoid tissues to CD10+CD38+CD27-transitional, CD24+CD38+IgM+CD27-marginal, and CD21+IgD+CD27-follicular B cells.⁵³ Depending upon the dominant cytokine milieu, including the presence of B-cell activating factor, follicular B cells may develop into plasma or memory B cells following antigenic exposure.⁵⁴ Transitional B cells express CD5 in both newborns and adults but comprise >50% of the circulating B cell pool in neonates, compared with only 5% in adults.⁵⁵ Prior to antibody exposure, mature B cells produce antibodies of the IgM or IgD subtype, with subsequent class switching to IgG, IgA, or IgE following CD40 or cytokine receptor engagement.⁵⁶ IgM+ memory B cells may be antigen-specific or may function in an extensively responsive, innate-like manner at mucosal surfaces. In the human gastrointestinal tract, such memory B cells secrete IgM that is reactive to a broad array of epitopes, thereby contributing to immune defense at the mucosal interface.

Neonatal IgM serum abundance increases in the first week of life, as does the production of the immunoglobulin joining (J) chain Ig.⁵⁷ The latter facilitates the polymerisation of IgM and IgA, engendering high valency of antigen binding sites, decreasing the tendency for these antibodies to activate complement and allowing for active transport of the antibody polymers through epithelial cells for subsequent secretion.⁵⁸ IgM abundance increases further over time, reaching 15% of physiological adult abundance within the first month of life, and 100% by 1 to 2 years of age.⁵⁹ Contrastingly, IgG antibodies are placentally transferred, with serum abundance in infants decreasing over the first 5 months of life, after which IgG synthesis increases sufficiently to maintain stable levels.^57,59 The maternal-fetal transfer of IgG antibodies results in fetal exposure to therapeutic antibody medications of the IgG1 subtype, including those utilized in IBD.60, 61, 62, 63, 64 The implications of these circumstances for the infant are discussed subsequently.

Neonatal Intestinal Microbiome and Mucosal Immunity

The composition of the neonatal intestinal microbiome is highly dynamic in the first 4 months of life.⁶⁵ Maternal vaginal skin, or externally acquired microbes originally predominate, before the infant’s intestinal microbiome eventually converges to one largely resembling maternal fecal sources.⁶⁵

The maternal intestinal microbiome influences neonatal intestinal ILC abundance and function in murine models. Transient inoculation via gavage of germ-free mice (GFM) with E. coli HA107 while pregnant leads to an increase in the number of intestinal ILCs identified in their pups at postnatal day 14 as compared with pups from GFM who are not transiently colonized, despite both groups of dams being germ-free at the time of delivery.⁶⁶ In particular, NKp46⁺RORγt⁺ ILC3s were shown to be more frequent, with an associated increase in serum IL-22 abundance. This ILC3 expansion is postulated to occur as a consequence of placental transfer of maternal intestinal microbiome-derived proteins and immunoglobulins,⁶⁶ which supposedly prevent translocation of E. coli to mesenteric lymph nodes in the pups. Microbial metabolic by-products, such as short-chain fatty acids (SCFAs) (eg, acetate, butyrate), are additionally postulated to affect the neonatal immune system, functioning by inhibiting histone deacetylases, thereby inactivating NF-κB and reducing innate immune cell TNF production.67, 68, 69 In murine models, SCFAs also promote differentiation of CD4+ T cells to Tregs in the colonic mucosa by increasing histone acetylation of the promoter region of the transcription factor Foxp3.69, 70, 71

Pregnant women with IBD exhibit lower bacterial species diversity as compared with healthy controls, with an increase in Gammaproteobacteria and a reduction in Bifidobacteria.⁴¹ Infants born to women with IBD were also shown to have reduced alpha diversity of the intestinal microbial communities in comparison to controls at postnatal days 7, 14, and 90.⁴¹ Supporting a role of the maternal intestinal microbiota in the development of the fetal enteric immune system, class-switched memory B cells were shown to be less abundant in the mesenteric lymph nodes and colonic lamina propria of mice inoculated with stool from pregnant women with IBD as compared with those inoculated with stool from non-IBD controls.⁴¹ Whether such vertically transmitted intestinal dysbiosis contributes to the heritability of IBD is under investigation.⁷² Supporting a potential association, antenatal maternal exposure to more than 3 courses of antibiotics (which may promote intestinal dysbiosis) was associated with an increased risk of IBD in the offspring.⁷³

Calprotectin is derived from neutrophils, monocytes, and macrophages. Potentially through interactions with the intestinal microbiome, calprotectin plays an important role in the homeostasis of the neonatal intestinal mucosal immune system and inflammatory responses.⁷⁴ For example, calprotectin was shown to restrict the expansion of potentially harmful Gammaproteobacteria and promote a microbiome rich in usually beneficial Bifidobacteria.⁷⁵ A lack of intestinal calprotectin in murine models was also shown to result in impaired Treg expansion and macrophage IL-10 production.⁷⁵ Reliable normal ranges for fecal calprotectin have not been identified in young children, and there is substantial inter-individual variation even among healthy infants.^76,77 Mean fecal calprotectin in the meconium of full-term neonates is up to 3 times the upper limit of normal for adults, and 6 times that in very low birth weight preterm infants.⁷⁴ Fecal calprotectin abundance remains elevated in comparison to adult normal ranges for 3 months after birth, before gradually waning in the subsequent 2 years of life, reaching adult normal ranges by 4 years of age.⁷⁷ The time at which these adult normal ranges are attained was shown to be delayed in infants born to mothers with IBD compared with those born to women not affected by IBD, with higher fecal calprotectin noted at 1 and 3 years of age.⁷⁸ These findings may reflect the effects of maternal intestinal, and resultant neonatal, microbial dysbiosis, but were also noted to occur in neonates born to women with acute histological chorioamnionitis, hence may be driven by in utero exposure to systemic inflammatory mediators more broadly.⁷⁹

Individuals with IBD have high rates (up to 49%) of caesarean section.^3,41 The impact of mode of delivery on the neonatal intestinal microbiome is well-established and may consequently impact infant immune system development, susceptibility to inflammatory disease, and vaccine response.^80,81 Although, without providing any mechanistic insight, a national registry study identified an association between caesarean section and an increased risk of childhood-onset IBD, including after stratifying by parental IBD status (incidence rate ratio, 1.29; 95% confidence interval [CI], 1.11‒1.49).⁸² In a study that quantified IgG antibodies to meningococcal C and pneumococcus serotype antibodies (anti-Ps6B IgG) in response to vaccination over the first 18 months of life, vaginal delivery was found to be independently associated with higher anti-Ps6B IgG and meningococcal C antibody concentrations, with median antibody titres also significantly higher in those infants subsequently breastfed as opposed to formula-fed.⁸¹ The gut microbial communities identified to be present in greater abundance in those born via caesarean section was additionally associated with the number of respiratory infections in the first year of life.^80,81 Thus, when examining the impact of maternal IBD on the infant microbiome and immune system, effects of caesarean section should be accounted for.

Animal studies have explored the influence of the intestinal microbiota on the peripheral immune system function of the neonate and its effect on trained immunity. Neonatal macaques were differentially treated with antibiotics or control saline from day 1 to day 7 of life.⁸³ The intestinal microbiome, peripheral immune cell frequencies, and cellular transcriptomes differed significantly between the 2 groups, with reduced microbial phylogenic diversity, reduced peripheral neutrophil, B cell, CD45RA+CD8+ cell abundance, increased CD45RA+CD4+ T cell abundance, and increased cellular stess and apoptosis gene network expression in those exposed to antibiotics.⁸³ When inoculated with S. pneumoniae 1 week after receiving antibiotics, the antibiotic exposed (and therefore dysbiotic) neonates had a more severe clinical course, with increased serum IL-1β, IL-6, IL-17, and IL-21 and reduced alveolar fluid IL-10, PDGF, and VEGF abundance. An increased frequency of pro-inflammatory alveolar neutrophils and macrophages was also observed. Thus, antibiotic induced dysbiosis may impact the relative expression of pro- and anti-inflammatory gene transcripts in both innate and adaptative immune cells following microbial stimuli. The aforementioned differences in gene expression and clinical phenotype in following S. pneumonia inoculation were partially, but not completely, mitigated by fecal transplantation.⁸³ This study supports an important role of the intestinal microbiome in education of the immune system in the very early neonatal period.⁸³

The neonatal microbiome may not only impact the severity of infant infections in the shorter term, but also influence later-life susceptibility to serious infection and inflammatory diseases. For example, a lower abundance of Veillonella, Lachnospira, Rothia, and Faecalibacterium in the neonatal intestine at 3 months was associated with the risk of atopy and wheeze at 1 year.⁸⁴ In atopic infants at 1 year of age, fecal concentrations of acetate were lower, and urinary concentrations of sulphated bile acids (glycolithocholate, glycocholenate, glycohyocholate, and urobilinogen) were higher. However, these differences did not correlate with an altered intestinal microbiome composition at this timepoint.⁸⁴ In a murine model, germ-free dams were colonized with feces from an infant with atopy and wheeze (and a later diagnosis of asthma) collected at 3 months of age, which contained a low abundance of Faecalibacterium sp., Lachnospira sp., Veillonella sp., and Rothia sp.⁸⁴ When pups were immunized with ovalbumin to induce an eosinophilic asthma phenotype, those born to control dams (ie, dams colonized with stool rich in Faecalibacterium sp., Lachnospira sp., Veillonella sp., and Rothia sp.) had lower total lung cellular inflammatory infiltrates, and lower lung tissue IFN-γ, TNF, IL-6, and IL17A abundance compared with those pups born to dams inoculated with the ‘atopic’ feces, signifying a reduction in the Th17/Th1-polarized inflammatory response.⁸⁴ A further murine study fed pregnant dams heat-killed Lentilactobacillus buchneri,⁸⁵ a bacterium that was shown to have anti-inflammatory effects in dextran sodium sulfate (DSS)-induced colitis models.⁸⁶ Accordingly, the DSS-induced colitis was significantly milder in the pups born to mothers supplemented with Lentilactobacillus buchneri antenatally. The number of IL10+CD103+ DCs in the mesenteric lymph nodes of Lentilactobacillus buchneri-exposed offspring was increased, with these dendritic cells implicated in the maintenance of intestinal homeostasis.⁸⁵ However, there was no difference in the alpha diversity of the intestinal microbiome between the experimental groups.⁸⁵ This observation suggests that, during pregnancy, the maternal intestinal microbiome influences the fetal and neonatal intestinal immune system via mechanisms beyond simple vertical transmission of the intestinal microbiome upon delivery.

The mucosal innate immune development in young infants may also be affected by maternal breast milk-derived proteins.^72,87 Thymic stromal lymphopoietin (TSLP) is an IL-7-like cytokine, produced by epithelial cells upon exposure to microbial-derived PAMPs and pro-inflammatory cytokines, including TNF.⁸⁸ TSLP was shown in vitro to activate the STAT3 and STAT5 pathways in CD11c+ DCs, leading to increased MCH class I and II and costimulatory molecule expression, and thereby enhancing the proliferation of naïve T cells.⁸⁹ TSLP is detectable in breast milk.^90,91 Compared with controls, the abundance of IgA and TSLP is significantly lower in breast milk of women with IBD, and breast milk TSLP was negatively correlated with infant fecal calprotectin at 1 year of age.⁶ This finding suggests that higher breast milk abundance of TSLP may be protective against intestinal inflammation in the infant, although the mechanism underlying this observation remains unclear.⁷² Furthermore, the composition of breast milk may change with maternal systemic inflammation, with increases in S100A8 and S100A9 (the components of calprotectin), pro-inflammatory neutrophil-derived proteins, and cytokines including IFN-γ, IL-6, IL-23, and IL-1α observed in women with IBD.⁶ The clinical implications of this finding have not been investigated, but an excess of pro-inflammatory stimuli could theoretically promote intestinal dysbiosis and aberrant development of the infant enteric immune system. Similarly, studies examining the effect of breastfeeding vs formula-feeding on the risk of developing IBD in children born to women with IBD have not been conducted. However, general population-based registry studies comparing rates of IBD in children who were breastfed vs formula-fed have identified an association between breastfeeding and a decreased risk of both UC and CD.⁹² As such, until further data are obtained, breastfeeding should be encouraged for women with IBD.

During the weaning period, when solid foods are introduced to a diet that previously exclusively consisted of milk, the intestinal microbiome of the infant rapidly expands to more closely mirror that of the adult.⁹³ In mice, the following periods can be defined: a neonatal milk-feeding phase from birth to day 10, the weaning period from day 11 to the time of solid food consumption (usually day 21), and the post-weaning phase.⁹⁴ In the neonatal phase, the intestinal mucosa is predominantly exposed to Lactobacilli, whereas in the post-weaning phase, Clostridia and Bacteroidetes dominate.⁹⁵ Immunologically, the weaning period is characterized by an increased expression of genes for cytokine receptors, CXC and CC chemokines, and Tnf and Ifng by ileal innate immune cells. Mice treated with vancomycin or metronidazole (ie, antimicrobials directed against Clostridia) do not exhibit these changes in gene expression, nor do mice deficient in MyD88, an innate signaling molecule.⁹⁵

In murine models, the timing of exposure to microbial antigens relative to the 3 ‘weaning’ phases determined the development of peripheral Tregs recognizing commensal microbes, which were capable of preventing pathological imprinting against these antigens.⁹⁴ In GFM, CD4+ T cells specific for Lachnospiraceae or B. vulgatus were best induced by CD11c+ antigen-presenting cells on postnatal days 14 to 21 compared with earlier or later timepoints. These CD4+ T cells were only induced to express foxp3 following transfer on day 18, not if transferred earlier or later.⁹⁴ Moreover, blockade of Treg induction during weaning, achieved through treatment with CD4+CD25+ monoclonal antibodies, was associated with development of comparatively severe DSS colitis in adulthood.⁹⁵

Invariant NK T cells (iNK cells) are highly prevalent in the intestinal mucosa.⁹⁶ Following lipid antigen presentation via the major histocompatibility complex class-1 like protein CD1, intestinal iNK cells can be polarized to produce either Th2-like (IL-13 and IL-4), Th17-like (IL-22 and IL-17), or regulatory cytokines, providing protection against infection or promoting tolerance to commensal organisms.⁹⁷ GFM were found to be more susceptible to severe oxazolone-induced colitis and OVA-induced asthma, attributed to a higher intestinal iNK cell prevalence.⁹⁸ Fecal microbial transplant (FMT) at 1 day, but not 8 weeks, post birth re-established a normal density of iNK cells in the intestinal mucosa, with a reduced risk of severe oxazolone-induced colitis in the early FMT group.⁹⁸ A further study showed that germ-free pups who received FMT pre-weaning developed relatively mild DSS-induced colitis in adulthood as compared with those transplanted post weaning, further emphasizing the importance of microbial exposures in the pre-weaning phase.⁹⁵

The above studies speak to the concept of the ‘early life window of opportunity,’ during which ‘pathological imprinting’ can be prevented. The maternal intestinal microbiome during pregnancy, as well as the intestinal microbiome during the early neonatal period, seem to play a guiding role in the appropriate development of the infant’s immune system. In consequence, the risk of inflammatory, atopic, and metabolic diseases, responses to vaccinations, and prevention of excessive immune responses to commensal bacteria may be influenced.^95,99,100 Notably, these conclusions largely rely on data derived from animals, and only a few human studies.^72,80,84 However, such data overall support current practical recommendations to avoid maternal antibiotic therapy during pregnancy and in early infancy when possible, and to promote and support vaginal delivery and breastfeeding where safe and feasible.

Maternal Immune Changes in Pregnancy and the Differences Seen With Inflammatory Bowel Disease

Pregnancy reflects a unique immunological challenge, whereby the mother must tolerate the semi-allogeneic fetus.¹⁰¹ A detailed discussion of pregnancy-related changes is beyond the scope of this review; however, examples include a progesterone-induced reversible thymic involution results in a reduction in peripheral CD4+ and CD8+ T cells in pregnancy. Moreover, a baseline reduction in Th1 as well as Th2 cytokines in response to stimulation was observed.¹⁰² In the setting of decreased IL-2, IL-12, and IL-15, peripheral CD16+ NK cells numbers declined in pregnant women from 20 weeks gestation.¹⁰³ B cells were similarly less abundant and less readily activated, consequent in part to increasing estrogen levels.^104,105 IL-10 abundance also increased over the course of gestation, dampening Th1 responses and promoting tolerogenic DC and Treg activity.¹⁰¹

To maintain protection against infection, such reduction in the pro-inflammatory functionality of some parts of the immune system in pregnancy may partially be balanced by an increase in the number of circulating neutrophils, DCs and monocytes, and antimicrobial peptides.^101,106 Additionally, complement hemolytic activity increases over the course of pregnancy,⁷⁵ whereas basal neutrophil function as evidenced by reactive oxygen species production remained stable and neutrophil extracellular trap (designed to neutralize and kill extracellular organisms) activity increased.^107,108

Antenatal cytokine profiles of individuals with IBD may differ from those of healthy controls, with current studies overall inconclusive.^109,110 Currently, the best data relates to TNF levels. In the third trimester of pregnancy in healthy women, TNF increased.^109,110 These findings were not appreciable in patients with IBD, who rather exhibited a progressive decrease in TNF. Whether this decrease in TNF is attributable to IBD itself, or the immunosuppressive medications used to treat it, remains unclear. In individuals receiving biologics, azathioprine, or 5-aminosalycylates, TNF abundance was lower than in those not receiving any medical therapy,¹¹⁰ but this difference was not observed in a subsequent study.¹⁰⁹

Differences in antenatal cytokine abundance in women with IBD may contribute to the adverse pregnancy outcomes associated with active disease. IBD flares in pregnancy have been associated with an increase in serum TNF.¹¹⁰ Excessive circulating TNF was associated with recurrent pregnancy loss, with median TNF abundance derived from trophoblast antigen-stimulated PBMCs in the first trimester shown to be significantly higher in women who experience recurrent miscarriage.¹¹¹ A significant difference in median TNF was also observed in pregnancies with intrauterine growth restriction (IUGR) and placental insufficiency, as compared with pregnancies with IUGR without placental insufficiency, and pregnancies without IUGR.¹¹² In keeping with this observation, a randomized controlled trial including individuals with prior failure of in-vitro fertilization and an elevated TNF:IL-10 ratio found that those who received adalimumab, an TNF inhibitor, prior to implantation had higher implantation, clinical pregnancy, and live birth rates compared with those who did not receive this therapy.¹¹³ In IBD-specific studies, elevated TNF in patients with UC was additionally associated with an increased risk of caesarean section, although the indications for this (ie, elective vs emergency) were not provided, and infant birth weight was not impacted.¹¹⁰ Overall, further studies including preconception serum cytokine analysis, individuals with other inflammatory disorders, and individuals with both active and inactive disease are required to extricate the impacts of immunosuppressive medications from those of an active inflammatory disorder.

Impact of Immunomodulatory Therapy on the Immune System in Adults

A background understanding of the immunological consequences of immunomodulatory pharmacotherapies on the adult is integral to interpreting the potential implications of in utero exposure on neonatal immune system development. Table 2 provides an overview. Presently, time to functional immunological recovery following withdrawal of such therapeutic agents remains poorly understood, particularly for more recently introduced biologic agents.

Table 2.

Mechanism of Action, Immunological and Clinical Effects of Immunomodulatory and Biologic Therapeutics Commonly Used in Pregnant Women With IBD

	Mechanism of action (examples)	Immunological effects in the adult (examples)	Immunological effects on the infant following in-utero exposure	Associated infection risk in patients with IBD (examples)
Corticosteroids (eg, prednisolone)	Bind cytoplasmic glucocorticoid receptors, which translocate to the nucleus and modulate gene expression directly and indirectly. For example, corticosteroids prevent the nuclear translocation and function of pro-inflammatory transcription factors, including NF-κB and activator protein 1 (AP-1).114	Decreased production of pro-inflammatory cytokines and IL-2, thus suppressing T cell proliferation.114 Reduced expression of B cell activation genes.115	No difference in lymphocyte subpopulations, immunoglobulins, or in vivo cytokine production following in-utero exposure to dexamethasone.116, 117, 118, 119	Increased susceptibility to opportunistic infections; OR of 3.4 (95% CI, 1.8–6.2, comparison with subjects not receiving medical therapy for IBD).120 Increase in presentation with common infections to primary care physicians.121
Thiopurines (eg, azathioprine, mercaptopurine, thioguanine)	Disrupt DNA replication, particularly in highly proliferative cells by acting as “rogue” nucleic acids (ie, being inappropriately incorporated into the DNA double helix and leading to its termination).122 Inhibition of purine synthesis.122 Lymphocyte cell cycle arrest and apoptosis via Rac1 activation and prevention of CD28-mediated co-stimulation.123	Reduction in the absolute number of circulating lymphocytes, naïve and transitional B cells, and NK cells.124 Reduced number of peripheral CD27- (naïve) B cells, naïve CD8+ T cells, Th1, Th2 and Th17 subsets.125	No difference in lymphocyte subpopulations, immunoglobulins, or in vivo cytokines116, 117, 118, 119 with thiopurine monotherapy. In combination immunosuppression: Higher IgA in infants exposed to cyclosporin as opposed to tacrolimus. Lower IgG1 and IgG3 in cyclosporin-exposed infants as opposed to healthy controls.126,127	OR of 1.32 (95% CI, 1.2–1.4) for serious infections, 3.72 (95% CI, 3.0–4.6) for serious infections.120 HR of 0.57 (95% CI, 0.4–0.9) for viral opportunistic infections in comparison to subjects receiving anti-TNF.120
Calcineurin inhibitors (eg, tacrolimus, cyclosporin)	Bind to cyclophilin (cyclosporin) and FK-binding protein (tacrolimus), leading to decreased activation of NFAT and thereby decreased production of cytokines such as IL-2, TNF, IL-3, IL-4, CD40L, granulocyte-macrophage colony-stimulating factor, and IFN-γ.128	Inhibit cytokine production from memory CD4+ T cells.128 Prevent differentiation of naïve CD4+ T cells into memory CD4+ T cells.129	CD45RA+RO- naïve T cells increased in cyclosporin-exposed children CD45RA-RO+ memory T cells more numerous in azathioprine-exposed children.126,127,130,131 CD25 expression on T cells and CD5 expression on B cells fails to increment over time.127 Decreased Treg numbers at birth.127	High risk of serious (6.3%) and opportunistic (3.8%) infections in patients managed with cyclosporin for limited duration in acute severe UC respectively.132 No opportunistic infections in a cohort of 29 patients treated with tacrolimus for acute severe UC.133
Anti-TNF (eg, infliximab, adalimumab, golimumab)	Bind to and neutralize soluble TNF, a potent and pleotropic pro-inflammatory cytokine.134	Delayed CD25+ expression with delay in T cell activation following IL-2 stimulation.135 Delayed maturation from naïve (CD45RA+) to memory (CD45RO+) T cells.135 Increased peripheral CD4+ T cell production of IL-10.136 Decreased expression of costimulatory molecules CD28, CD27, and CD95.136 Decreased T cell proliferation and increased peripheral and lamina propria T cell apoptosis.135 Decreased expression of PTGS2, and PTGER4, encoding COX-2 and the prostaglandin E2 receptor E4.137 Downregulation of NF-κB transcription factor network genes.137	Lower memory & higher naïve T cells Decreased Tregs Decreased T cell activation of bCG exposure Fewer activated & class-switched B cells Higher naïve B cells Increased Bregs Reduced serum IL-6, IL-1β IL-1Ra, TNF in response to stimulation. No data available regarding impact on innate immune system138,139	Odds ratio 4.0 (95% CI, 3.1–5.3) for opportunistic infections vs those not receiving medical therapy for IBD.120
Vedolizumab	Binds to α4β7 integrin expressed on leukocytes and prevents interaction of α4β7 with mucosal addressin cell adhesion molecule-1 (MAdCAM-1) on the intestinal endothelium, thereby decreasing trafficking of these inflammatory cells to the intestinal mucosa.140	Reduced innate mucosal immune system activating gene expression (eg, RAC2 and TREM1).141 Increased intestinal macrophages polarized towards a reparative phenotype.142	No data available	No increased risk of infection (including gastrointestinal), opportunistic or severe infections.143
Ustekinumab	Blocks the binding of human IL-12 and IL-23 to their respective receptor complexes by binding to the shared p40 subunit.144	Decreased CD4+ T cell differentiation into T follicular helper cells, which are central to germinal center formation and memory B cell development.145 Decreased IL-12- and IL-23-mediated inflammatory responses, including reduction of STAT3 and STAT4 phosphorylation and production of IFN-γ, IL-17A, and IL-22.120,144	No changes in peripheral lymphocyte subsets or in the distribution of T or B cells in lymphoid tissues in macaques.146	Risk of infectious complications comparable to that of placebo.147 HR of 0.59; (95% CI, 0.4–0.9) for severe infection vs anti-TNF in rheumatologic patients.148

CI, Confidence interval; HR, hazard ratio; IBD, inflammatory bowel disease; IFN, interferon; IL, interleukin; NF-κB, nuclear factor-κB; NK, natural killer; OR, odds ratio; Th, T helper; TNF, tumor necrosis factor; UC, ulcerative colitis.

Effects of In Utero Exposure to Maternal Immunosuppression in Neonates and Infants

The majority of immunosuppressive medications cross the placenta and enter the fetal circulation (Figure 1). The effects of such exposure on the development of the neonatal immune system were evaluated in the setting of transplantation and immunosuppression for autoimmune disease, but in a small number of children at heterogenous timepoints, with variable outcomes. There is little data pertaining to biologic therapeutics specifically, particularly the more recently introduced agents, ustekinumab and vedolizumab.

Figure 1.

Relative maternal and infant concentration of immunosuppressive medication at the time of delivery following antenatal exposure.

Multimodal Immunosuppression

Unlike in the management of many autoimmune or inflammatory disorders, transplantation immunosuppression is almost universally multimodal. Consequently, the specific effects of an individual medical therapy on the immune system of a neonate or infant exposed in-utero to transplantation-related immunosuppression is difficult to delineate.

In neonates, infants, and children older than 1 year, in utero exposure to tacrolimus, cyclosporin, glucocorticoids, and/or azathioprine for maternal transplantation was not shown to result in significant changes in the serum abundance of immunoglobulins, barring higher IgA in mothers treated with cyclosporin as opposed to tacrolimus, and lower IgG1 and IgG3 in cyclosporin-exposed infants as opposed to healthy controls.^126,127 T cell proportions and B cell subtype populations were comparable to controls; however, in relevant studies, CD45RA+RO- naïve T cells were increased in cyclosporin-exposed children aged from newborn to 10 years of age, whereas CD45RA-RO+ memory T cells were more numerous in azathioprine-exposed children, suggesting delayed T cell maturation for cyclosporin, and accelerated T cell maturation for azathioprine.^126,127,131 In another study assessing cyclosporin-exposed infants over the first year of life, CD25 expression on T cells and CD5 expression on B cells did not increment over time, and B (including Bregs) and Treg numbers were decreased at birth.¹²⁷ The clinical significance of these abnormalities appeared to be minimal, with no increases in immunodeficiencies or autoimmune disorders reported in the included infants. However, the ability to draw firm conclusions from these reports is limited by the small number of infants (<15 participants in each study).^{126,127,130,131,149}

Immunomodulators

No significant differences in lymphocyte subpopulations, immunoglobulins, or in vivo cytokine production were reported between infants exposed in utero or not to cyclosporin, azathioprine, and/or dexamethasone at the ages of 11 months and between 2 and 12 years, with protective antibody titres following hepatitis B vaccination also observed.116, 117, 118, 119 These findings suggest that effects on infant immune function induced by in-utero exposure to immunosuppression may be minimal by 1 year of age; however, once again, the ability to draw conclusions is substantially impeded by the limited numbers of participants, divergent protocols, and heterogenous, single-sampling timepoints.

Biological Therapeutics

As IgG1 antibodies, the TNF antagonists adalimumab and infliximab are incrementally transferred to the fetus over the course of pregnancy, resulting in serum concentrations at delivery that exceed maternal concentrations^62,150 (Figure 1). With in-utero exposure to adalimumab and infliximab, complete clearance from the infant serum occurs at a mean of 4.0 and 7.3 months, respectively.^62,150 Seventeen percent and 27% of infants exposed in utero to adalimumab and infliximab, respectively, have detectable (although low) drug levels at 6 months of age.⁶² In comparison, median clearance time of ustekinumab in infants is 2.3 months, and 2.5 months for vedolizumab.^63,150 The variable time to drug clearance is critical to the design of scientific studies investigating the effects of these therapeutics on the infant. The biological activity of a medication is expected to persist while the drug remains present in the serum, and subsequent functional immunological recovery needs to be assessed longitudinally following complete clearance.

In comparison to controls, infants with detectable serum concentrations of adalimumab or infliximab at the time of delivery were shown to have increased proportions of naïve B and T cells as well as decreased proportions of memory T cells, IgD-CD27+ memory B cells, Tregs, and activated B cells. These abnormalities in the T and B cell compartments were shown to resolve by 12 and 18 months of age.^138,139 Following culturing with bacille Calmette-Guérin (BCG) and recombinant IL-12p70 or human recombinant IFN-γ, the abundance of IL-6, IL-1β, IL-1Ra, TNF, and lymphocyte activation markers such as CD69 and human leukocyte antigen-DR in whole cord blood neonates exposed to anti-TNF in utero was reduced, but recovered when the drug was removed from the cultures. These findings suggest that the presence of anti-TNF in the serum decreases lymphocyte activation and inflammatory response to mycobacterial antigens. Upon complete clearance at 3 to 12 months of age, IFN-γ, IL-12p70, IL-1β, and TNF increased when measured in unstimulated whole blood and whole blood cultured with BCG.¹³⁸ Thus, anti-TNF exposure in utero may result in a transient delay in T and B cell maturation, and a reduction in pro-inflammatory cytokines in response to stimulation.¹³⁸

There are no published human studies evaluating the effects of ustekinumab in utero exposure on the neonatal immune system. A study of macaques did not demonstrate any changes in peripheral lymphocyte subsets in ustekinumab-exposed fetuses, nor in the distribution of T or B cells in lymphoid tissues.¹⁴⁶

To our knowledge, there are also no published data exploring the immunological impact of in utero effects by vedolizumab in infants. However, in adult murine models of colonic injury, vedolizumab treatment was associated with an increase in the number of alternatively activated monocytes in blood and a decrease in the number of macrophages in the gastrointestinal tract.¹⁴² These changes might contribute to the delay in wound healing observed in vedolizumab-exposed mice as opposed to controls.¹⁴² Whether such changes also occur in infants following exposure to vedolizumab in-utero remains to be evaluated. Theoretical implications include impaired gastrointestinal healing following an inflammatory insult, as occurs for example in the setting of necrotising enterocolitis. Notably, a study showed that 95% of neonatal CD3+ T cells expressed α4β7, compared with 50% of adult T cells.¹⁵¹ Consequently, investigations into the potential effects of in utero vedolizumab exposure on neonatal T-cell homing to the intestinal tract and subsequent tolerogenic education against commensal microbial populations are of importance.

Outcomes in Infants Exposed to Immunomodulatory Therapies In Utero

Biologics

Exposure to anti-TNF monotherapy in utero does not appear to result in an increased risk of severe infection in infants in the majority of published literature. Prospective studies including almost 2000 infants reported similar rates of infection comparing anti-TNF-exposed and non-exposed infants at a median follow-up of 47 months and in comparison to infants at 12 months of age exposed to thiopurine or no medical therapy,^61,152 no matter whether the child attended childcare or not.¹⁵³ However, the detrimental impact of in utero biologic exposure on infant immunological function may have resolved by 12 months of age – and studies on adverse events occurring earlier (ie, at 3‒6 months of age) are limited.

Although there are substantially fewer reports on the infection risk associated with vedolizumab exposure in utero, the available data are reassuring. The published studies have not demonstrated an increased risk of infection in infants exposed to vedolizumab in utero when compared with those exposed to anti-TNF.^154,155 Studies specifically pertaining to infant infection risk with in utero exposure to ustekinumab are similarly limited^63,156,157; hence, more research is required prior to drawing conclusions with respect to the safety of in utero exposure to both vedolizumab and ustekinumab.

Immunomodulators

Results regarding the impact of in utero exposure to immunomodulators on the offspring are contradictory. A small case-control study reported that infants exposed to azathioprine in utero experienced an increase in episodes of common colds and conjunctivitis when compared with infants born to women with IBD not receiving immunomodulatory agents during their pregnancy.¹⁵⁸ However, analyses of French National Health registry data, multicenter retrospective cohort studies, and other small cohort studies did not support this finding, rather reporting similar rates of infection with or without maternal thiopurine treatment.^61,159,160

Data regarding the risk of serious infections following in utero exposure to combination therapy with anti-TNF and thiopurine are largely reassuring⁶¹; however, the risk of non-serious infections may be marginally increased. Using a propensity score-weighted model, a large registry study observed an adjusted hazard ratio (aHR) for infection of 1.36 (95% CI, 1.04‒1.79) of non-serious infection in first 12 months of life compared with infants born to healthy control women.¹⁶⁰ Similar findings were described in infants born to women receiving anti-TNF and thiopurine combination therapy in a smaller cohort (relative risk, 2.7; 95% CI, 1.09–6.78; P = .02).⁶²

Vaccine Efficacy

Serological vaccine responses in infants born to women with IBD not exposed to immunosuppression in utero were shown to be comparable to those exposed to biologics in utero,^161,162 including serological responses to vaccination with tetanus, diphtheria, pneumococcus, rubella, and measles.^138,163 Antibody responses to at least one hepatitis B vaccination were found to be preserved in infants exposed to azathioprine or cyclosporin in utero in one study,¹¹⁹ and in another cohort, maternal immunosuppression for renal or liver transplantation also did not result in a significant reduction in the rubella, varicella zoster, or hepatitis B antibody titres induced by vaccination when evaluated between 1 and 28 days, 28 days to 1 year, and older than 1 year.¹²⁶ Whether the timing of vaccination relative to the expected time of immunomodulator clearance from the infant influences vaccination response remains unclear.

Presently, live vaccinations are not recommended in infants exposed to biologics in utero until at least 12 months of age, as per the European Medicines Agency.¹⁶⁴ However, studies suggest that the incidence of severe adverse reactions is low in infants who inadvertently receive the rotavirus vaccine following exposure to biologics or triple immunosuppression in utero.¹⁶⁵ Serious adverse events are rare following BCG vaccination prior to 6 months of age in anti-TNF-exposed infants, but a single fatal case of disseminated BCG has been reported.¹⁶⁶ To inform the optimal timing of live vaccination in infants exposed in utero to biologic agents, further granularity, and a more sophisticated understanding of the effect of low residual serum levels of biologic agents on infant immune function and mucosal immune response to oral live vaccinations is required.

Autoimmune Disease

Exposure to maternal immunomodulators in utero was suggested to be a risk factor for development of autoimmune disease later in life, postulated to be a consequence of defective deletion of autoreactive T cells during induction of thymic self-tolerance. Rat studies, in which cyclosporine was administered during pregnancy, identified antibodies to gastric autoantigens in 11 of 53 rats, which was associated with an autoimmune gastritis-like inflammatory process in one rat pup.¹⁶⁷ Contrasting with this finding, a Danish registry-based nationwide cohort study identified 1047 children born between 1995 and 2015 exposed to thiopurine in utero with a median follow-up of 8.9 years and compared rates of autoimmune disorders in these children to those of 1,308,778 non-exposed children, adjusted according to underlying maternal disease. No increased risk of IBD (aHR, 1.45; 95% CI, 0.6‒3.3) or rheumatoid arthritis (aHR, 0.78; 95% CI, 0.4‒1.7) was identified.¹⁶⁸ Thus, such exposures in-utero may be less significant than previously postulated with regards to risk of autoimmunity.

Conclusion

Pregnancy is characterized by changes in both immune cell frequencies and functionality to facilitate maternal tolerance of the semi-allogenic fetus while maintaining protection against infection. Pregnant individuals with IBD may differ with regards to their antenatal cytokine abundance, in accordance with the medical therapies they receive and the activity of their disease. Such differences, best documented for TNF, may partially explain the adverse obstetric consequences of active disease.

Exposure to maternal inflammation in utero was also shown to impact the development and functionality of the neonatal innate and adaptive immune system. The impact of the maternal and neonatal intestinal microbiomes on the infant’s immune homeostasis is also increasingly recognized, particularly in animal models (albeit with human confirmatory data largely pending). These findings are important in the setting of maternal IBD, where a ‘dysbiotic’ intestinal microbiome is commonly observed.

Both directly and indirectly (eg, via interplay with the maternal and neonatal microbiome), a variety of external factors alter polarization and function of the neonatal immune system, in turn affecting the infant’s susceptibility to inflammatory and atopic diseases, and the robustness of vaccine responses. Such factors include mode of delivery, method of feeding and time to weaning, and maternal and neonatal antibiotic exposures. How exposure to immunosuppressive agents ties into this developmental process has been explored, with current data indicating only minor clinical implications overall. However, a detailed understanding of the functional immunological changes observed in the neonatal period following in-utero exposure to immunosuppressive therapies is still lacking. Similarly, studies investigating the therapeutic role of maternal or neonatal intestinal microbial manipulation in reducing adverse longer-term disease susceptibility have not been conducted. Such studies would present significant logistical challenges. Maternal participants would need to be meticulously characterized prior to, during, and following pregnancy from an immunologic, clinical, intestinal microbial, and demographic standpoint. Their offspring would similarly need to be frequently examined, including with temporal assessment of drug levels, immune phenotyping with assessment of response to stimulation ex vivo, and intestinal microbiome analysis. Appropriate validated surrogate endpoints for long-term outcomes, such as risk of atopy or IBD, would ideally be employed.

Despite this lack of a detailed understanding of immunological effects of the pharmacotherapies employed in pregnant women with IBD on the infant, the importance of controlling IBD activity during pregnancy is well-recognized and should be prioritized. While further data emerge, minimizing the impact of modifiable risk factors for aberrant neonatal immune system development is also advisable; breastfeeding and vaginal delivery where safe should be encouraged, and antibiotics should be prescribed judiciously during pregnancy and in the growing infant.