TitleMaternal and infant serum carotenoids are associated with infantile atopic dermatitis developmentTo the Editor,Various epidemiological studies have shown that eczema/atopic dermatitis (AD) in infancy is a risk for skin sensitization1 and development of allergic diseases later in life2.Carotenoids are natural pigments biosynthesized by bacteria, fungi, and plants, but not by mammals3; thus, they need to be supplied via dietary intake of vegetables and fruits or certain animal products. Carotenoids positively impact human health and prevents allergic reactions via their provitamin A activity and high antioxidant potential. However, previous studies using food frequency questionnaires that evaluated the association between maternal vegetable intake during pregnancy and eczema in offspring have shown inconsistent results4–6. As these studies did not measure the levels of individual carotenoids in the serum of mothers or children and in breast milk, appropriate nutritional interventions in maternal and early infant intake of vegetables for the prevention of AD remain unclear.In this study, we have measured the levels of total carotenoids and some of their sub-types in the serum of mothers and children and in maternal breast milk, to evaluate the association between carotenoid levels and the presence of AD at 1 year of age in infants.We compared participants’ characteristics and exposures by 1 year of age (Table S1); carotenoid, retinol, and α-tocopherol levels (Table 1) in the serum of participants, with and without AD at 1 year of age, were compared to those in the serum and breast milk of their respective mothers. We found that both the presence of eczema (OR, 31.7; 95%CI[13.2–76.0]) and S. aureus carriage in the skin by 6 months of age (OR, 5.20; 95%CI[2.30–11.75]) were associated with higher odds of AD development at 1 year of age. On the contrary, certain carotenoid levels in the serum and breast milk, including total carotenoids in the maternal blood, were associated with lower odds of AD at 1 year.To avoid multicollinearity in the regression analysis, we selected seven relevant predictive variables among the carotenoid data using VIP scores in the PLS analysis (Table S2). Stepwise logistic regression analysis using explanatory baseline characteristics, exposure by 1 year of age, and the seven selected carotenoid levels revealed that the following variables were significantly related to AD at 1 year of age (Table 2): presence of eczema by 6 months of age (OR, 34.5; P < 0.0001), maternal blood lutein level (unit OR, 0.002; P = 0.002), and infant blood lycopene level at 1 year (unit OR, 0.01; P = 0.007).One strength of this study is that multiple biological sample types were used as proxies for carotenoid intake. The lutein concentration in the maternal blood at 36 weeks of gestation, which was associated with a reduced AD risk at 1 year of age in the multivariate analysis, was significantly correlated with the cord blood lutein level (Table S3). This suggests that lutein ingested during pregnancy is transferred to the fetus and may have an inhibitory effect on the development of AD in infancy. Another strength of this study is that multiple carotenoids were evaluated simultaneously; the concentrations of lutein, zeaxanthin, α-carotene, β-carotene, and lycopene were strongly correlated, suggesting that these nutrients are absorbed together (Table S4).In conclusion, the results of this study suggest that children of mothers with low carotenoid intake during pregnancy are at higher risk for developing infantile AD and are ideal targets for early intervention in allergy prevention. Further studies are needed to clarify whether carotenoid supplementation during pregnancy/in lactating mothers and infants after weaning has a preventive effect on AD development in infancy.Table 1 Levels of each carotenoid, total carotenoids, retinol, and α-tocopherol in each type of sample in participants with or without atopic dermatitis at 1 year of age

BACKGROUND: Chemokine (C-C motif) ligand 17 (CCL17; also known as thymus and activation-regulated chemokine or TARC) is a pro-allergic factor, and high CCL17 levels in cord blood (CB) precede the allergic predisposition later in life. Offspring of pregnant mice treated with short-chain fatty acid (SCFA) have been shown to be protected against allergic diseases. The maternal microbial metabolome during pregnancy may affect foetal allergic immune responses. To examine this, the associations between CB CCL17 and gut SCFA levels in pregnant Japanese women were investigated here. METHODS: This study was conducted as part of the Chiba Study of Mother and Child Health; 434 healthy pregnant women were recruited. The CB CCL17 and maternal non-specific IgE levels were measured using CB sera at birth and maternal sera at 32 weeks of gestation. Stool samples were collected from pregnant women at 12 (n = 59) and 32 (n = 58) weeks of gestation and used for gut microbiota analysis, based on barcoded 16S rRNA sequencing and metabolite levels. RESULTS: The CB CCL17 levels correlated negatively with butyrate concentrations at 12 weeks of gestation. In contrast, CB CCL17 levels correlated positively with isobutyrate levels at 12 weeks of gestation, and valerate and lactate concentrations at 32 weeks of gestation in maternal faeces. CONCLUSION: The metabolites in maternal faeces may alter the foetal immune responses. This study provides the first link between maternal metabolites during pregnancy and the risk of allergic diseases in human offspring, even before birth.

To the Editor In allergic diseases, not only ‘allergic inflammation’, but also ‘epithelial barrier impairment’ play important roles in their development1. Allergen penetration through epithelium to the body is essential for allergen sensitizations, which are the most critical risk factors for the development of allergic diseases. In children with atopic dermatitis (AD), in which epithelial barrier impairment is one of its famous features, the prevalences of allergen sensitization and food allergy (FA) are known to be very high1. Concerning about the assessment of epithelial barrier function, while transepidermal water loss (TEWL) is a clinically useful marker in AD skin, it is quite difficult to evaluate barrier function in gastrointestinal tracts.Zonulin (pre-haptoglobin 2), an epithelial tight-junction regulator, plays an important role in the regulation of epithelial barrier function2,3. Recently, zonulin was reported to play a pathogenic role in celiac disease and other chronic inflammatory diseases23. Sturgeon et al reported that zonulin transgenic mice, as a result of zonulin-dependent small intestinal barier impairment and altered gut permeability, increased morbidity and mortality in the DSS colitis model4.Because allergen sensitization and chronic inflammation are important in the pathogenesis of allergic diseases, zonulin has potential to play significant roles in allergic diseases. However, there are few reports about zonulin levels in pediatric allergic patients. So, the objectives of this study are to assess whether i) serum zonulin levels in allergic children are higher than those in children without allergic diseases and ii) serum zonulin levels are different among allergic diseases.To evaluate them, we measured serum zonulin levels in infants and school-age children (with or without allergic diseases) using zonulin enzyme-linked immunosorbent assay kit. All infants (9 months old) were selected from Katsushika cohort study5. Allergic infants (‘AD infants’) were defined as infants who had doctor-diagnosed AD, and whose total IgE levels were above normal levels and Eczema area and severity index (EASI) were above zero. ‘Healthy infants’ were defined as infants who did not have AD, FA, bronchial asthma (BA), nor allergic rhinitis (AR), and whose serum total IgE levels were below measureable limits and EASI were zero. There were significant differences in total IgE levels and EASI score between groups (AD infants vs. healthy infants) (Table S1). Allergic school-age children were recruited in our outpatient clinic in Chiba University hospital and were defined as children who had doctor-diagnosed FA or BA or both, and whose total IgE levels were above normal levels. ‘healthy school-age children’ 6 were defined as school-age children who did not have AD, FA, nor BA and their serum total IgE levels were below measureable limits. The details of study subjects are shown in online supplementary methods and table S1-S3.Our first question was whether or not serum zonulin levels in allergic children are higher than those in children without allergic diseases. In infants, median zonulin levels in AD infants and healthy infants were 28.1 ng/ml (interquartile range (IQR): 23.8-32.3) and 15.3 ng/ml (IQR: 10.0-23.7), so serum zonulin levels were significantly higher in allergic infants than those in healthy infants (p<0.01) (Figure 1). To confirm this result in other age group, we evaluated zonulin levels in school-age children. Same as in infants, serum zonulin levels in school-age children were significantly higher in allergic children with FA and BA than those in healthy children without FA, BA, nor AD (Median (IQR): 29.5 ng/ml (22.6-41.9) and 10.6 ng/ml (9.4-12.2), p<0.001) (Figure 2a). Those results suggest that zonulin levels are higher in allergic children than in non-allergic children, regardless of their age.To assess our second question, we compared serum zonulin levels in BA patients with those in FA patients. In this analysis, we selected asthmatic patients as children who had BA but not FA, and FA patients as children who had FA but not BA, and there were no significant differences in total IgE and age between groups (Table S3). Serum zonulin in FA patients was significantly higher than those in BA patients (Median (IQR): 38.5 ng/ml (35.0-46.7) and 31.0 ng/ml (13.6-34.0), p<0.05) (Figure 2b). Those results suggest that zonulin levels are higher in FA patients than those in BA patients.Sheen et al reported that Serum zonulin levels were elevated in children with AD7, and this result was confirmed in our present paper. On the other hand, zonulin primarily reflects epithelial permeability in the gastrointestinal tract, it is important to examine Zonulin levels in other age groups and in patients with other allergic diseases. As we showed in this paper, in school-age children, zonulin levels are higher in allergic children than in healthy children, and higher in food allergy children than in asthmatic children. In general, many AD infants also have FA and intestinal permeability is repoted to be increased in younger children with AD but not in older children with AD 8. Those facts might influence the serum zonulin levels in infants with AD.Zonulin regulates intestinal barrier function through the regulation of tight junctional 2. This suggests that zonulin may affect allergen permeability in epithelia of the intestinal tract. Zonulin may contribute to the development of allergic diseases through allergen sensitization and affect the appearance of allergic symptoms via their effect on allergen exposure.There were several limitations in this study. First, the number of study subjects was small. And secondly, we did not compare zonulin levels in allergic children to those in children with other non-allergic diseases. So it is not clear whether serum zonulin is an “allergic marker”, and zonulin plays a role in the “development” of allergic diseases.In conclusion, serum zonulin levels are significantly higher in allergic children than in healthy children. In addition, those levels are significantly higher in food allergy patients than in asthmatic patients. Based on the knowledge about the role of zonulin in the intestinal epithelial barrier regulation, zonulin may play a role in the development of allergic diseases, especially in FA.