Influence of the sebaceous gland density on the stratum corneum lipidome

Abstract

The skin surface lipids (SSL) result from the blending of sebaceous and epidermal lipids, which derive from the sebaceous gland (SG) secretion and the permeability barrier of the stratum corneum (SC), respectively. In humans, the composition of the SSL is distinctive of the anatomical distribution of the SG. Thus, the abundance of sebum biomarkers is consistent with the density of the SG. Limited evidence on the influence that the SG exerts on the SC lipidome is available. We explored the differential amounts of sebaceous and epidermal lipids in areas at different SG density with lipidomics approaches. SC was sampled with adhesive patches from forearm, chest, and forehead of 10 healthy adults (8F, 2M) after mechanical removal of sebum with absorbing paper. Lipid extracts of SC were analysed by HPLC/(-)ESI-TOF-MS. In the untargeted approach, the naïve molecular features extraction algorithm was used to extract meaningful entities. Aligned and normalized data were evaluated by univariate and multivariate statistics. Quantitative analysis of free fatty acids (FFA) and cholesterol sulfate (CHS) was performed by targeted HPLC/(-)ESI-TOF-MS, whereas cholesterol and squalene were quantified by GC-MS. Untargeted approaches demonstrated that the relative abundance of numerous lipid species was distinctive of SC depending upon the different SG density. The discriminating species included FFA, CHS, and ceramides. Targeted analyses confirmed that sebaceous FFA and epidermal FFA were increased and decreased, respectively, in areas at high SG density. CHS and squalene, which are biomarkers of epidermal and sebaceous lipid matrices, respectively, were both significantly higher in areas at elevated SG density. Overall, results indicated that the SG secretion intervenes in shaping the lipid composition of the epidermal permeability barrier.

Figure 1

(A) Principal component analysis (PCA), and (B) C-C plots of the entities found in 100% samples of at least one condition among lipid extracts of ARM, CHEST, and HEAD stratum corneum (SC). Within the first two principal components (PC) 50.77% of the total variance was explained, with 38.93% in the first dimension and an additional 11.84% in the second dimension. The C-C plots show the P-Cor and P-Cov values of entities assigned as ceramides (pink dots), free fatty acids (green dots) and cholesterol sulfate related compounds (red dots) among the entities detected in 100% samples in at least one condition (black dots).

Figure 2

Volcano plots of entities found differently regulated in ARM, CHEST, and HEAD stratum corneum (SC). Comparison of relative abundance of entities between ARM and CHEST (A), CHEST and HEAD (B), ARM and HEAD (C) were performed by applying univariate significance test analysis to 600 entities found in the 100% of samples belonging to at least one condition. The direction from low-to-high SG density is depicted at the top of each volcano plot. The log2 of the fold-change (FC) values were plotted on the x-axis, whereas the −log10 of the t-test p-values were plotted on the y-axis. The vertical and the horizontal green lines marked the threshold of the FC and of the p-values with Bonferroni’s correction set at 2 and 0.05, respectively. The dots corresponding to each entity were coloured according to the FC, with red and blue colour indicating the highest and lowest FC values, respectively, for each pair-wise comparison, as depicted in the colour grade scale. Annotation of species found differently expressed is provided elsewhere in the results and following figures. Volcano plots returned 159, 173, and 314 entities that were expressed at a significantly different extent in the comparisons ARM vs. CHEST (A), CHEST vs. HEAD (B), ARM vs. HEAD (C), respectively. (D) Eulero-Venn’s diagram depicting the number of entities that were commonly modified when comparing ARM to CHEST, CHEST to HEAD, and ARM to HEAD.

Figure 3

Hierarchical clustering of 52 entities that were commonly modified when comparing ARM to CHEST, CHEST to HEAD, and ARM to HEAD as reported in Fig. 2. The identity of annotated FFA and cholesterol sulfate was confirmed by comparison with RT and MS/MS spectrum of the respective authentic compounds. Ceramide annotations were confirmed upon generation of MS/MS spectra.

Figure 4

Results of the quantitative assessments of biomarkers of sebum secretion and lipids of the epidermal permeability barrier in ARM, CHEST, and HEAD SC lipid extracts. FFA and CHS were quantified by LC-MS, whereas amounts cholesterol and squalene were assessed by GC-MS. Box plots report concentration expressed as nmol/mg SC in ARM (green box), CHEST (orange box), and HEAD (grey box). The table summary included in the figure reports the mean concentration marked with a red cross in the respective box plot and the significance assessed with ANOVA.