Epidermal Immunity and Function: Origin in Neonatal Skin

Abstract

The fascinating story of epidermal immunity begins in utero where the epidermal barrier derives from the ectoderm and evolves through carefully orchestrated biological processes, including periderm formation, keratinocyte differentiation, proliferation, cornification, and maturation, to generate a functional epidermis. Vernix caseosa derives from epidermal cells that mix with sebaceous lipids and coat the fetus during late gestation, likely to provide conditions for cornification. At birth, infants dramatically transition from aqueous conditions to a dry gaseous environment. The epidermal barrier begins to change within hours, exhibiting decreased hydration and low stratum corneum (SC) cohesion. The SC varied by gestational age (GA), transformed over the next 2-3 months, and differed considerably versus stable adult skin, as indicated by analysis of specific protein biomarkers. Regardless of gestational age, the increased infant SC proteins at 2-3 months after birth were involved in late differentiation, cornification, and filaggrin processing compared to adult skin. Additionally, the natural moisturizing factor (NMF), the product of filaggrin processing, was higher for infants than adults. This suggests that neonatal skin provides innate immunity and protection from environmental effects and promotes rapid, continued barrier development after birth. Functional genomic analysis showed abundant differences across biological processes for infant skin compared to adult skin. Gene expression for extracellular matrix, development, and fatty acid metabolism was higher for infant skin, while adult skin had increased expression of genes for the maintenance of epidermal homeostasis, antigen processing/presentation of immune function, and others. These findings provide descriptive information about infant epidermal immunity and its ability to support the newborn's survival and growth, despite an environment laden with microbes, high oxygen tension, and irritants.

Keywords: epidermal barrier; genomics; immunity; neonatal; proteomics; skin; stratum corneum; vernix caseosa.

FIGURE 1

(A) A comparison of the ceramide profile of vernix caseosa and adult stratum corneum. The ceramide profile of vernix was compared to that of adult stratum corneum. Samples of vernix caseosa were collected at birth from full-term infants and adult tissues were obtained during cosmetic surgery (Rissmann et al., 2006). The lipids were extracted, separated by high-performance thin layer chromatography, and quantified. Values are given as percent weight as mean and ± standard deviations. Ceramide AH (AH contains α-hydroxy acids and sphingosines) was the most abundant, followed by NS (NS contains non-hydroxy fatty acids and sphingosines), AS/NH (AS contains α-hydroxy fatty acids and sphingosines and NH contains non-hydroxy fatty acids and 6-hydroxy sphingosine and EOS (EOS contains ester-linked fatty acids, ω-hydroxy fatty acids, and sphingosine). The relative ceramide levels were higher in vernix compared to adult SC except for ceramides AP (AP contains α-hydroxy fatty acids and phytosphingosine) and NP (NP contains non-hydroxy fatty acids and phytosphingosine) that were lower in vernix. The ceramide distributions were similar in vernix and adult SC. Statistical comparisons were not reported. (B) A comparison of the ceramide profiles in vernix, fetal stratum corneum (16–20 weeks GA), mid-gestational SC (23–25 weeks GA), infant SC (1–11 months), and child SC (1–6 years). Vernix caseosa from healthy full-term infants, tissue samples that required surgery, and fetal tissue from spontaneous abortions were quantified by high-performance thin layer chromatography (Hoeger et al., 2002). Ceramide (AH) was the highest fraction, followed by AS, NS, EOS, and EOH, with AP and NP being the lowest species. Vernix ceramide AH was significantly higher and vernix NP and NS were significantly lower than for infants of 1–11 months (p < 0.05). Ceramide levels in vernix and premature infant stratum corneum at 23–25 GA were comparable, except for ceramide AP that was higher in vernix (p < 0.05).

FIGURE 2

Functional classes of vernix proteins. Vernix was extracted and digested with trypsin, quantified by liquid chromatography–tandem mass spectrometry, and analyzed against the Swiss-Prot protein database (Holm et al., 2014). Proteins (n = 203) for p < 0.05 and belonging to 25 functional classes were identified (percent by weight). Hydrolases, proteases, and enzyme modulators encompassed 29, 22, and 22 proteins, respectively, with 11 proteins classified as immunity/defense.

FIGURE 3

Skin barrier repair following application of vernix or petrolatum versus untreated skin. Skin barrier damage was created by repeatedly tape stripping the SC in the hairless mouse model. The damaged skin sites were treated with 5 mg/cm² of vernix, 5 mg/cm² petrolatum, or left untreated as controls and barrier recovery monitored over time (one-way ANOVA, posthoc Bonferroni correction, and p < 0.05) (Oudshoorn et al., 2009b). Vernix treated skin demonstrated a significantly increased rate of SC barrier repair compared to untreated, damaged control skin. In the same study, treatment of damaged skin with petrolatum also significantly increased the SC barrier repair rate versus the untreated control, but the skin was more erythematous and thickened compared to the vernix treated skin (p < 0.05). *Indicates significant difference for untreated skin versus vernix and petrolatum treated sites (p < 0.05).

FIGURE 4

Differentially expressed biomarkers for infants 4–8 days after birth and 2–3 months later at corrected GAs of 46–48 weeks versus adult skin. Two sequential skin surface samples (stratum corneum) were collected from the lower legs of 61 infants at each time and from the volar forearms of 34 adults (parent) at one time. Samples were extracted, quantified with liquid chromatography tandem mass spectrometry, and analyzed with targeted proteomics (p < 0.05) (Visscher et al., 2020). The proteins classified by function were: filaggrin processing, protease inhibitor/enzyme regulators, antimicrobials, keratins/structural proteins, lipid processing, and cathepsins (p < 0.05). (A–C) show the log2 fold changes for the specific proteins in each class versus adults for PT, LPT, and FT infants at both times. The differentially expressed biomarkers were decidedly different for infant skin compared to stable adult skin. For PT infants, the differentially expressed proteins increased from 12 to 54 versus adults over 2–3 months, suggesting substantial adaptive changes over time.

FIGURE 5

Changes in SC proteins for PT infants compared to FT infants 4–8 days after birth. Two sequential skin surface samples (stratum corneum) were collected from the lower legs of 61 infants at each time and from the volar forearms of 34 adults (parent) at one time. Samples were extracted, quantified with liquid chromatography tandem mass spectrometry, and analyzed using targeted proteomics (p < 0.05) (Visscher et al., 2020). PT infant SC had decreased expression of filaggrin processing biomarkers FLG, FLG2, AGR1, and TGM3, antimicrobial S100A8, protease inhibitor CSTA, and protective protein CTSA (cathepsin A) soon after birth compared to FT infant SC.

FIGURE 6

Changes in SC proteins for LPT infants compared to FT infants 2–3 months after birth at comparable corrected GA. Two sequential skin surface samples (stratum corneum) were collected from the lower legs of 61 infants at each time and from the volar forearms of 34 adults (parent) at one time. Samples were extracted, quantified with liquid chromatography tandem mass spectrometry, and analyzed using targeted proteomics (p < 0.05) (Visscher et al., 2020; Visscher et al., 2021). LPT infants had increased expression of protease inhibitors PI3, SERPINB3, and SERPINB12, as well as FLG, CALML5, CTSC, and TF versus FT infants. Expression of S100A7, LY6D, SFN, MDH2, and DDAH2 was lower in LPT compared to FT 2–3 months later. These findings suggest that the rate of change of specific aspects of epidermal barrier development may vary with GA and/or time from birth.

FIGURE 7

Hierarchical clustering analysis of differentially expressed genes in newborn infant and adult skin samples. Full-thickness tissue samples (body site, non-foreskin) from 27 infants were collected at non-elective surgery and buttocks tissue (protected from ultraviolet radiation exposure) from 43 adults was processed to collect total RNA (Visscher et al., 2021). Gene expression was determined from mRNA using Affymetrix GeneTitan U219 array plates. The lowest 30% of the 49,386 gene transcripts were removed, assayed for quality, data was normalized and Log2 transformed, analyzed using linear models and differential expression analyses and analyzed and analysis of variance with a term for the combination was conducted, as previously described (Merico et al., 2010; Supek et al., 2011; Yu et al., 2012; Gu et al., 2016; Szklarczyk et al., 2019; Visscher et al., 2021). Rigorous quality control was applied and all data were MIAME compliant. The Empirical Bayes method (limma R-package) was used to test comparisons. Test statistics were moderated with the Empirical Bayes method (limma R-package). The Benjamini–Hochberg correction was used to control for false discovery rates. The complete linkage method using the R hclust function was used to perform hierarchical clustering. Genes that were significantly expressed were analyzed for enrichment of biologic themes (Gene Ontology) using the clusterProfiler package (Yu et al., 2012), EnrichmentMap (Merico et al., 2010), g:profiler (Raudvere et al., 2019), Revigo (Supek et al., 2011) and String database (Szklarczyk et al., 2019). The transcriptomics is in the NCBI Gene Expression Omnibus (GEO) repository with the dataset accession number GSE181022. Panel (A) is a heatmap of the normalized expression values (based on z-score) of the 1,086 differentially regulated genes with adjusted p value <0.05 and absolute fold change ≥1.5 for the two groups, infants (pink) and adults (blue). Euclidean distances between each sample were determined and cluster analysis was performed with an unsupervised hclust algorithm. Samples formed two clusters. From a hierarchical cluster, analysis genes were grouped for similarity where each column is an individual sample and each row is a single gene. Extracellular matrix (ECM, orange), immune-related (black), and epithelial (yellow) genes are indicated on the annotation bar under the gene type. Panel (B) is a heatmap of group average values for infants (pink) and adults (blue). Panel (C) are the values of the negative log₁₀ of the adjusted p values from the Limma testing for adults versus infants for each gene. The z-scores are shown in the blue-white-red gradient where 2 is the darkest red color, 0 is white and -2 is the darkest blue color shown on the right and labeled as “expression”. Many negative log₁₀ adjusted p values were greater than 10, indicating large differences.

FIGURE 8

Gene ontology themes with enriched gene expression in infant skin. Significantly expressed genes were analyzed for enrichment of biologic themes (Gene Ontology) using the clusterProfiler package (Yu et al., 2012), EnrichmentMap (Merico et al., 2010), and g:profiler (Raudvere et al., 2019). Significant pathways were selected using the false discovery rate (FDR)adjusted p values. NegLog₁₀Qvalue indicates -Log₁₀FDR adjusted p value. The lowest adjusted p values (highest NegLog₁₀Qvalue) were biological processes (BP) extracellular matrix (ECM) organization and ECM structure organization. Others included system development (e.g., blood vessel, cardiovascular, tube), and response (e.g., to lipid). The most significant molecular functions (MF) were ECM matrix structural constituent and structural molecule activity. ECM and collagen-containing ECM were the most significant cell components (CC).

FIGURE 9

Gene ontology themes with enriched gene expression in adult skin. Significantly expressed genes were analyzed for enrichment of biologic themes (Gene Ontology) using the clusterProfiler package (Yu et al., 2012), EnrichmentMap (Merico et al., 2010), and g:profiler (Raudvere et al., 2019). Significant pathways were selected using the false discovery rate (FDR)adjusted p values. NegLog₁₀Qvalue indicates -Log₁₀FDR adjusted p value. The lowest adjusted p values for BP were skin development, epidermis development, keratinocyte differentiation, keratinization, and cornification. Immune BPs included antigen processing and presentation of exogenous antigen, major histocompatibility protein complex, and antigen-binding were also prominent. Highly significant MF was peptide antigen binding and structural molecule activity and CC was a cornified envelope and major histocompatibility complex (MHC) protein complex.