Stratum Corneum Interleukin-2 in Facial Eczema at 1-Month-Old Predicts Later Atopic Dermatitis

Abstract

Atopic dermatitis (AD) commonly begins in infancy, with about 45% of cases beginning before 6 months of age and 60% before 1 year [1]. In infants, facial eczema often appears within the first month, presenting as 1–2-mm nonpruritic erythematous papules. Clinically, facial eczema often either remits spontaneously or persists, eventually leading to a diagnosis of very early-onset AD within a few months. However, the immune abnormalities present in nonpruritic facial eczematous lesions preceding pruritic AD remain unclear.

This prospective cohort study aimed to identify biomarkers predictive of whether infantile facial eczema will remit spontaneously or diagnose later very early-onset AD. We enrolled 1-month-old infants without (n = 55) or with facial eczema (n = 98) and followed them up for potential AD until 5 months of age (Figure S1). At baseline, stratum corneum (SC) was collected from the cheek by tape stripping. Twenty-three cytokines/chemokines were analyzed using multiplex immunoassays. We also measured barrier function, sebum amount, and serum squamous cell carcinoma antigen 2 (SCCA2), a downstream molecule of IL-4/IL-13 [2], which can objectively assess the severity of pediatric AD [3].

The final analysis included 19 healthy infants (no infantile facial eczema group), 17 with spontaneous remission, and 46 diagnosed with AD at 5 months of age (infantile facial eczema group) (Table S1). In the infantile facial eczema group, potential risk factors for AD were not present in the patient backgrounds.

Unlike previously reported AD-predictive biomarkers identified in nonlesional skin of infants without eczema at 2 months of age [4, 5], among facial eczema cases, elevated SC IL-2, CCL26, and CCL20 levels at 1 month increased AD risk at 5 months of age (odds ratios [ORs]: 3.4, 1.7, and 1.8, respectively) (Table S2). IL-2, CCL26, and CCL20 levels were significantly higher in the AD group than in the healthy and spontaneous remission groups (Figure 1A–C). When IL-2 levels were dichotomized based on the receiver-operating characteristic curve cutoff value (7.4 pg/μg protein), the OR for AD by 5 months in the high-value group was 11.9. IL-2 at 1 month of age demonstrated a strong ability to predict AD at 5 months of age in the infantile facial eczema group, with the area under the curve (AUC) of 0.78 (Figure 1D). Among these three markers, IL-2 demonstrated a superior ability to predict AD across all severity levels (Figure 1D, Figure S2).

For other parameters, a progressive increase in TEWL and serum SCCA2 was observed across the spontaneous remission and AD groups, along with a decrease in sebum amount in the AD group (Figure 2). These findings indicate that more severe barrier dysfunction is associated with systemic immune activation, as indicated by elevated serum SCCA2, and contributes to AD progression, supporting the epithelial barrier theory. Thus, the severity of barrier dysfunction and elevated levels of SC IL-2, CCL26, and CCL20 may play critical roles in determining the prognosis of AD. Previous studies reported that IL-2 induces STAT5-mediated IL-4Rα expression and plays a crucial role in the polarization of naive CD4 T cells to the Th2 phenotype in vitro [6], supporting our results.

In conclusion, our study provides evidence that SC IL-2 levels in 1-month-old infants with facial eczema may serve as a predictive biomarker for later very early-onset AD, highlighting IL-2 as an early marker of immune abnormalities. Currently, there is no evidence linking IL-2 to skin barrier dysfunction, but it may interact with both immune abnormalities and barrier dysfunction. In the future, we plan to evaluate the long-term AD-predictive ability of IL-2 and assess the effectiveness of AD-predictive markers at various time points. These findings could be a valuable strategy to enable the identification of infants at high risk of AD and early intervention.

E.M. conceived the study, conducted the experiments, drafted the manuscript, and oversaw the entire project. E.M. designed the study with the help of Y.F., K.K., and S.K. E.H., Y.N., and E.U. contributed to the collection of infant samples. R.Y. and S.N. provided advice in the laboratory. G.T. and H.E. proposed experimental strategies and offered valuable feedback throughout the manuscript writing process. J.K. and H.W. reviewed the analysis results and contributed insights for the statistical evaluation. M.K. and T.N. guided the project, assisted with data analysis, and supported E.M. in refining the manuscript with significant feedback.

The authors declare no conflicts of interest.