Abstract
Background:
Food allergy (FA) often occurs in early childhood with and without atopic dermatitis (AD). FA can be severe and even fatal. For primary prevention, it is important to find early biomarkers to predict the future onset of FA before any clinical manifestations.
Objective:
To find early predictors of future onset of FA in the stratum corneum (SC).
Methods:
Skin tape strips (STS) were collected from the forearm of newborns (n=129) at the age of 2 months before any signs of clinical FA or AD. Children were clinically monitored until they reached 2 years of age to confirm the presence or absence of FA and AD. STS were subjected to lipidomic analyses by the LC-MS/MS and cytokine determination by Meso Scale Discovery U-Plex assay.
Results
Overall, 9/129 (7.0%) infants developed FA alone, and 9/129 (7.0%) infants developed FA concomitantly with AD (ADFA). In the SC of future FA and ADFA children, absolute amounts of unsaturated (N24:1)(C18S)Ceramide and (N26:1)(C18S)Ceramide and their relative percentages within the molecular group were increased in comparison to healthy children with p-values between <0.01 and p<0.05, ANOVA. Future AD children had normal levels of these molecules. Interleukin (IL)-33 was upregulated in future FA but not AD infants, whereas thymic stromal lymphopoietin (TSLP) was upregulated in AD but not FA. Logistic regression analysis revealed strong FA predicting power for the combination of dysregulated lipids and cytokines, with an odds ratio reaching 101.4 (95% CI 5.4 – 1910.6).
Conclusion
Non-invasive STS analysis at 2 months of age can identify infants at risk of future FA.
Keywords: Food allergy, Stratum corneum, Skin lipids, Protein-bound ceramides, Unsaturated ceramides
Capsule Summary
Lipid and cytokine levels in the stratum corneum at the age of 2 months selectively predict with high probability the future onset of food allergy up to the age of 2 years.
Graphical Abstract:
Introduction
In a recent issue of JACI, we demonstrated the ability to predict the future onset of atopic dermatitis (AD) through a comprehensive analysis of lipid and cytokine components in stratum corneum (SC) collected at the age of two months prior to the onset of symptomatic AD or food allergy (FA)1. During this prospective two-year clinical follow-up study, a limited number of newborns developed FA and FA associated with AD (ADFA). The analysis of SC lipids has revealed a unique signature specific to the future onset of FA irrespective of future AD status. Importantly, in the current study, lipid changes observed in future FA infants were uniquely different than those observed in future AD infants. This suggests fundamentally different underlying mechanisms involved in the regulation of lipid metabolism in FA versus AD. Future FA and AD infants also demonstrated divergent dysregulation of skin Type 2 cytokine levels that may be responsible for dissimilar changes in skin lipid metabolism. As such, we found unique lipid and cytokine signatures that predicted the future onset of FA and unique lipid profiles that differentiated FA from ADFA and AD endotypes.
Results and Discussions
This prospective birth cohort study enrolled 129 infants from South Korea who were followed for potential onset of FA and AD until 24 months of age. Major details of this clinical trial (WHO International Clinical Trials Registry Platform registration number KCT0007979) were described in1 and are further provided in Supplementary Materials. Out of 129 newborns, 9 children developed FA without AD, another 9 children developed ADFA, and 28 infants developed AD without FA during the first two years of life, with 83 infants staying healthy (Supplementary Table E1). The most common food allergen was egg white (16/18); one child each had an allergy to walnut and cow’s milk. Two children had an allergy to peanuts in addition to egg white, and one infant had an allergy to egg white, peanuts, and walnuts. At the age of two months, when the SC was collected by skin tape stripping, none of the infants had a history of clinical FA or AD.
The detailed analysis of lipids in the SC of infants who later in life developed FA revealed unique lipid abnormalities that separated them from healthy infants as well as from infants who developed only AD later in life. In particular, the SC of future FA infants had increased levels (Figure 1A) and proportions (Fig. 1B) of N(C18S)-ceramides with unsaturated fatty acids (N24:1 and N26:1, here and thereafter, amide-linked fatty acids, carbon chain length:number of double bonds) and C18-sphingosine. Infants who developed both FA and AD later in their lives also had increased levels and proportions of these unsaturated NS-ceramides, while children with future AD only were not different from healthy children for this lipid parameter. This signifies that such an increase in NS-ceramide unsaturation is unique to FA and not to AD. Combining FA and ADFA infants into one group further strengthened the difference between FA, AD, and healthy groups (Figure 1C).
Figure 1. Stratum corneum unsaturated NS-ceramides as predictors of the future onset of FA and ADFA.
Increase in the absolute levels (A) and relative proportions (B) of unsaturated NS-ceramides N24:1(C18S)CER and N26:1(C18S)CER in the stratum corneum of infants with future development of FA or ADFA (A, B) or combined future FA and ADFA infants (C).
Protein-bound OS-ceramides are another group of SC lipids that may help to identify future FA children. In contrast to infants with future AD who had decreased levels of OS-ceramides, the SC of future FA infants had a tendency to have increased levels of OS-ceramides as compared to healthy skin (Figure 2). This increase, while found so far to be statistically significant only for one of the OS-CER molecular species (with the longest measured fatty acid (N32:0O)(C18S)CER)) (Figure 2B), contrasted to an overall decline in OS-ceramides in future AD infants as compared to SC of healthy infants. At two months of age, in the skin of future FA children, the levels of N32:0O-S-ceramides with all three sphingoid bases (C18-C22) and total levels of protein-bound OS-ceramides (the sum of twelve measured molecular species) were significantly increased as compared to the skin of children who later developed AD (Figure 2). This finding has to be confirmed in other studies with larger numbers of future FA infants, but already indicates that there is no impairment in biosynthesis of OS-ceramides in the SC of future FA infants. Such an increase in OS-ceramides, especially in the most hydrophobic ultra-long-chain OS-ceramide species, might be a compensatory mechanism to counteract the increase in unsaturated NS-ceramide species that are known to impair barrier properties of lipid ultrastructures by reducing the packing density and enhancing the permeation and water loss2. Interestingly, in future FA infants with concomitant AD, FA and AD metabolically counteract each other, leading to a net no change in OS-ceramides (Figure 2).
Figure 2. Stratum corneum of future FA infants has increased levels of protein-bound OS-ceramides with ultra-long-chain omega-hydroxy fatty acids.
The separation of the future FA group from the healthy and ADFA group is best pronounced for OS-ceramide with C18-sphingosine (N32:0O)(C18S)CER (B). Note the decrease in OS-CER levels in infants with future AD (A-D).
We have reported in our original JACI observation that the SC of infants with future AD onset demonstrates an increase of unsaturated sphingomyelin (SM) species with 24:1 and 26:1 fatty acids1. Remarkably, infants with future onset of FA or ADFA have no increase in such SM molecular species (Figure 3) in SC. Such a dichotomy in the distribution of 24:1 and 26:1 fatty acids between N(C18S)-ceramides and sphingomyelins between infants with future AD only and those with future onset of FA or ADFA suggests a unique, AD-specific mechanism of regulation of SM biosynthesis in the skin that is counteracted and abolished by the development of FA. Currently, there is no clear understanding of how ceramide redistribution between biosynthesis of SM and glycosphingolipids is regulated. Ceramides are the immediate precursors for both of these complex sphingolipids. However, the biosynthesis of SM and glycosphingolipids occurs in trans-Golgi, while ceramides are synthesized in the endoplasmic reticulum and need to be transported to cis-Golgi by a CERT carrier protein3. Therefore, multiple potential targets involved in ceramide transport and complex sphingolipid biosynthesis should be further explored as differentially controlled in FA versus AD skin.
Figure 3. Unsaturated sphingomyelin species are not increased in the stratum corneum of infants with future FA.
The increase in absolute levels (A) and relative proportions (B) of 24:1- and 26:1-sphingomyelin molecular species in the SC of infants with future AD but not FA or ADFA. Note that the development of FA cancels the increase in 26:1-SM in the ADFA group.
Cytokines are well known to regulate lipid metabolism4, and Type 2 cytokines such as interleukin (IL)-4 and -13 are critical for IgE synthesis and lipid changes in the SC in AD1, 5. Keratinocyte-produced alarmin, thymic stromal lymphopoietin (TSLP), has been demonstrated to be a predictive biomarker in the SC of future AD infants6, and keratinocyte-generated IL-33 (but not TSLP) has been shown to play a role in IgE-mediated gut mast cell activation and FA initiation in mice7, 8. Therefore, we examined Type 2 SC cytokines in an attempt to differentiate skin cytokine dysregulation in FA versus AD. Several cytokines (TNF-α, IL-6, IL-13, macrophage-derived chemokine (MDC)) were equally elevated in future FA, AD, and ADFA groups (Figure 4A). IL-1α and interferon gamma-induced protein 10 (IP-10) were not increased compared to healthy infants (Figure 4A). TSLP was not elevated in future FA infants but was elevated in the SC of future AD infants (as we already demonstrated in1) and in future ADFA infants (Figure 4A). This correlates with the known role of TSLP in the development of AD9. IL-33 was not elevated in the skin of future AD infants, but demonstrated moderate elevation in the skin of future FA-only infants and was statistically significantly higher (P<0.05) in SC of the future ADFA group in comparison to either healthy or future AD infants (Figure 4A). Combining future FA and ADFA infants into one group led to the complete separation of future FA infants, irrespective of their AD status, for SC IL-33 levels as compared to healthy infants (Figure 4B). Models of allergic inflammation have shown distinct requirements for IL-33 in skin sensitization and intestinal manifestations of FA independent of TSLP7, 10. Notably, in future ADFA, the increase in both TSLP and IL-33 was detected in SC. This is supported by experimental studies in animals that have shown TSLP promoting the expression of IL-33 receptors on many inflammatory cell types, allowing IL-33 to amplify type 2 immune responses11. Thus, TSLP and IL-33 upregulation differentiates future FA, AD, and ADFA groups.
Figure 4. TSLP and IL-33 levels in the stratum corneum are unique for infants with future FA.
(A) TSLP is not increased in the skin of infants with future FA but increased in the skin of future AD and ADFA infants. IL-33 is increased in the skin of future ADFA infants but not in the skin of future AD infants. No selectivity in other cytokines measured was found. (B) IL-33 levels separate infants with future FA irrespective of AD status.
Finally, we evaluated the power of identified lipid and cytokine parameters to predict the future onset of FA irrespectively of the future AD status using a logistic regression approach. Figure 5 demonstrates that unsaturated NS-ceramide species as well as protein-bound OS-ceramides (the most distinctive OS-ceramide molecular species chosen), have strong predicting power on their own (odds ratios (OR)s from 7.2 to 11.3). Furthermore, altered lipid profiles in combination with IL-33 (FA-specifically upregulated cytokine) and TNFα (non-specifically upregulated cytokine) substantially increased the possibility of predicting FA, with OR reaching a value above 100 (Figure 5 and graphical abstract).
Figure 5. Logistic regression analysis reveals the strong combinatorial power of STS cytokines and lipids to predict the future onset of FA by 24 months of age, irrespective of AD status.
The sole and combined effects of cytokines and lipids were evaluated by the estimates of odds ratios (ORs) from the logistic regression models. All study participants were included in the analysis. Dotted line – 2x fold change
Our findings are the first indication that changes in the SC lipidome precede future onset of FA. Importantly, we have identified lipid and cytokine signatures that are unique for the future onset of FA and AD and can be used as predictive biomarkers. Changes in lipid composition identified in the skin of future AD (decrease in protein-bound ceramides and increase in unsaturated SM species reported previously by our group1) and FA children (increase in monounsaturated NS ceramides reported here) suggest the critical contribution of the skin barrier in the initiation of allergic sensitization11–13. A detrimental role of fatty acid unsaturation for skin barrier function is supported by recent observations in an experimental SC substitution model demonstrating increased epidermal permeability and transepidermal water loss by ceramides with monounsaturated fatty acids2. Importantly, our SC cytokine analysis demonstrates selective activation of epidermal alarmins in the skin of future AD (TSLP), ADFA (TSLP, IL-33), and future FA children (IL-33). These epithelial cell-derived cytokines license dendritic cells to activate type 2 responses, but also act on basophils, eosinophils, mast cells and innate lymphoid cells to initiate and maintain allergic inflammation. Our study was not designed to investigate the causal relationship between AD and FA, and the factors leading to selective TSLP or IL-33 activation in the skin of future AD or FA children remain to be discovered. Similarly, our data do not answer the question about the interrelationship between IL-33 and IL-13 in the sequence of events leading to the initiation of AD and FA in ADFA infants. But, since targeted biologics are under development for IL-33 and are now available to TSLP and IL-4/IL-1314–16 (the latter is currently approved down to 6 months of age), early identification of infants at risk for future AD or FA development opens possibilities for targeted interventions to prevent and circumvent their clinical disease. Additional studies are needed in different ethnical groups and regions of the world to determine if the observed molecular signatures are universal.
Supplementary Material
Clinical Implications
Identifying children at risk of future FA development with high probability will allow the implementation of targeted preventive strategies to reduce sensitization and increase immune tolerance to foods.
Acknowledgments
We thank Samsung Medical Information and Media Services, Samsung Medical Center for the preparation of the graphical abstract for this manuscript. This work was funded in part by NIAID/NIH Grant 5UM1 AI 130780.
Funding:
The study was funded by Edelstein Family Chair of Pediatric Allergy-Immunology at National Jewish Health, NIH/NIAID (5UM1AI130780-5), by SMC Research and Development Grant (SMO1230141) and by the Korea Environment Industry and Technology Institute through Environmental Health Action Program funded by Korea Ministry of Environment (grant no. 2017001360006).
Abbreviations
- AD
Atopic dermatitis
- ADFA
Atopic dermatitis with food allergy
- C18S, C20S, C22S
C18-sphingosine, C20-sphingosine, C22-sphingosine
- FA
Food allergy
- IL
Interleukin
- IP-10
interferon gamma-induced protein 10
- LC-MS/MS
Liquid chromatography tandem mass spectrometry
- MDC
Macrophage-derived chemokine
- NC
Normal control
- NS-CER
Non-hydroxy fatty acid sphingosine ceramides
- (N24:1)(C18S)CER
NS-ceramide with N-linked 24:1 fatty acid and C18-sphingosine
- (N26:1)(C18S)CER
NS-ceramide with N-linked 26:1 fatty acid and C18-sphingosine
- (N32:0O)(C18S)CER
OS ceramide with C18-sphingosine and omega-hydroxy 32:0 fatty acid
- OR
Odds ratio
- OS-CER, OS-ceramides
omega-hydroxy fatty acid sphingosine ceramides
- SC
Stratum corneum
- SM
Sphingomyelin
- 24:1-, 26:1-SM
Sphingomyelin with C18-sphingosine and N-linked 24:1 or 26:1 fatty acid
- SPT
Skin prick test
- STS
Skin tape strips
- TARC
Thymus and activated-regulated chemokine
- TNFα
Tumor necrosis factor-alpha
- TSLP
Thymic stromal lymphopoietin
Footnotes
Conflict of Interest: EB reports research grants with LEO Pharma and Sanofi Genzyme. EG reports research grants with Sanofi Genzyme. DYML reports consulting with Aslan Pharma, Genentech, Jasper Therapeutics, Sanofi-Genzyme, LEO Pharma, and Incyte; reports research grant with Sanofi-Genzyme and NIH/NIAID. All other authors have nothing to report.
Clinical Trial Registration
The study protocol was registered in the WHO International Clinical Trials Registry Platform with the registration number KCT0007979.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Berdyshev E, Kim J, Kim BE, Goleva E, Lyubchenko T, Bronova I, et al. Stratum corneum lipid and cytokine biomarkers at age 2 months predict the future onset of atopic dermatitis. J Allergy Clin Immunol 2023; 151:1307–16. [DOI] [PubMed] [Google Scholar]
- 2.Mojumdar EH, Helder RW, Gooris GS, Bouwstra JA. Monounsaturated fatty acids reduce the barrier of stratum corneum lipid membranes by enhancing the formation of a hexagonal lateral packing. Langmuir 2014; 30:6534–43. [DOI] [PubMed] [Google Scholar]
- 3.Kumagai K, Hanada K. Structure, functions and regulation of CERT, a lipid-transfer protein for the delivery of ceramide at the ER-Golgi membrane contact sites. FEBS Lett 2019; 593:2366–77. [DOI] [PubMed] [Google Scholar]
- 4.Kitaura J, Murakami M. Positive and negative roles of lipids in mast cells and allergic responses. Curr Opin Immunol 2021; 72:186–95. [DOI] [PubMed] [Google Scholar]
- 5.Danso MO, van Drongelen V, Mulder A, van Esch J, Scott H, van Smeden J, et al. TNF-alpha and Th2 cytokines induce atopic dermatitis-like features on epidermal differentiation proteins and stratum corneum lipids in human skin equivalents. J Invest Dermatol 2014; 134:1941–50. [DOI] [PubMed] [Google Scholar]
- 6.Kim J, Kim BE, Lee J, Han Y, Jun HY, Kim H, et al. Epidermal thymic stromal lymphopoietin predicts the development of atopic dermatitis during infancy. J Allergy Clin Immunol 2016; 137:1282–5 e4. [DOI] [PubMed] [Google Scholar]
- 7.Leyva-Castillo JM, Galand C, Kam C, Burton O, Gurish M, Musser MA, et al. Mechanical Skin Injury Promotes Food Anaphylaxis by Driving Intestinal Mast Cell Expansion. Immunity 2019; 50:1262–75 e4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Han H, Roan F, Johnston LK, Smith DE, Bryce PJ, Ziegler SF. IL-33 promotes gastrointestinal allergy in a TSLP-independent manner. Mucosal Immunol 2018; 11:394–403. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Luo J, Zhu Z, Zhai Y, Zeng J, Li L, Wang D, et al. The Role of TSLP in Atopic Dermatitis: From Pathogenetic Molecule to Therapeutical Target. Mediators Inflamm 2023; 2023:7697699. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Wang YH. Developing food allergy: a potential immunologic pathway linking skin barrier to gut. F1000Res 2016; 5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Han H, Roan F, Ziegler SF. The atopic march: current insights into skin barrier dysfunction and epithelial cell-derived cytokines. Immunol Rev 2017; 278:116–30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Goleva E, Berdyshev E, Leung DY. Epithelial barrier repair and prevention of allergy. J Clin Invest 2019; 129:1463–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Akdis CA, Akdis M, Boyd SD, Sampath V, Galli SJ, Nadeau KC. Allergy: Mechanistic insights into new methods of prevention and therapy. Sci Transl Med 2023; 15:eadd2563. [DOI] [PubMed] [Google Scholar]
- 14.Beck LA, Cork MJ, Amagai M, De Benedetto A, Kabashima K, Hamilton JD, et al. Type 2 Inflammation Contributes to Skin Barrier Dysfunction in Atopic Dermatitis. JID Innov 2022; 2:100131. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Bieber T, Paller AS, Kabashima K, Feely M, Rueda MJ, Ross Terres JA, et al. Atopic dermatitis: pathomechanisms and lessons learned from novel systemic therapeutic options. J Eur Acad Dermatol Venereol 2022; 36:1432–49. [DOI] [PubMed] [Google Scholar]
- 16.Wang Z, Tang N. Unpacking the complexity of nuclear IL-33 (nIL-33): a crucial regulator of transcription and signal transduction. J Cell Commun Signal 2023; 17:1131–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.