Profiling of epidermal lipids in a mouse model of dermatitis: Identification of potential biomarkers

Abstract

Lipids are important structural and functional components of the skin. Alterations in the lipid composition of the epidermis are associated with inflammation and can affect the barrier function of the skin. SHARPIN-deficient cpdm mice develop a chronic dermatitis with similarities to atopic dermatitis in humans. Here, we used a recently-developed approach named multiple reaction monitoring (MRM)-profiling and single ion monitoring to rapidly identify discriminative lipid ions. Shorter fatty acyl residues and increased relative amounts of sphingosine ceramides were observed in cpdm epidermis compared to wild type mice. These changes were accompanied by downregulation of the Fasn gene which encodes fatty acid synthase. A profile of diverse lipids was generated by fast screening of over 300 transitions (ion pairs). Tentative attribution of the most significant transitions was confirmed by product ion scan (MS/MS), and the MRM-profiling linear intensity response was validated with a C17-ceramide lipid standard. Relative quantification of sphingosine ceramides CerAS(d18:1/24:0)2OH, CerAS(d18:1/16:0)2OH and CerNS(d18:1/16:0) discriminated between the two groups with 100% accuracy, while the free fatty acids cerotic acid, 16-hydroxy palmitic acid, and docosahexaenoic acid (DHA) had 96.4% of accuracy. Validation by liquid chromatography tandem mass spectrometry (LC-MS/MS) of the above-mentioned ceramides was in agreement with MRM-profiling results. Identification and rapid monitoring of these lipids represent a tool to assess therapeutic outcomes in SHARPIN-deficient mice and other mouse models of dermatitis and may have diagnostic utility in atopic dermatitis.

Fig 1. Monitored lipid ions in cpdm and WT epidermis by MRM scans in positive ion mode.

Clear discrimination of the phenotypes of WT and cpdm mouse strains was observed by PCA and cluster analysis. (A) Score plot of principal component analysis (PCA). PC1 explained 47.1% of the variability of the data. When PC2 was included, the explained variance increased to 65.7%. (B) Heat map with the distribution of lipids monitored individually in 36 samples. Lipids not identified are shown with their m/z and corresponding lipid class. Color of each cell corresponds to the relative abundance of the lipid feature monitored in the sample.

Fig 2. Ceramide profile in cpdm and WT epidermis by MRM-profiling.

(A) There was an increase of Cer[S] and a decrease of Cer[P] and Cer[DS] in the cpdm epidermis. (B) The relative amount of ceramides with α-hydroxy-fatty acid residues was larger in cpdm compared to WT. This finding was independent of the sphingoid base as it was observed for both Cer[S] and Cer[P]. (C) Cer[S] carrying fatty acid residues of 16–18 and 22–24 carbons were increased and those with 26 carbons were reduced in cpdm samples compared to WT. The vertical axis represents the relative amounts of ceramides detected in the epidermis of cpdm and WT mice (horizontal axis). Bars represent the mean +SE of 7 (cpdm) or 8 (WT) mice. * p < 0.05; *** p < 0.001, based on unpaired t-test with Holm-Sidak correction for multiple comparisons.

Fig 3. Monitored lipid ions in cpdm and WT epidermis by MRM scans in negative ion mode.

Clear discrimination of the phenotypes of WT and cpdm mouse strains was observed by PCA and cluster analysis. (A) Score plot of principal component analysis (PCA). PC1 explained 57.3% of the variability of the data. When PC2 was included, the explained variance increased to 70.6% (B) Heat map with the distribution of lipids monitored individually in 36 samples. Lipids no identified are shown with their m/z. Color of each cell corresponds to the relative abundance of the lipid feature monitored in the samples.

Fig 4. FFA profile in cpdm and WT epidermis by MRM-profiling.

(A) Relative amounts of polyunsaturated FFAs were increased and saturated and monounsaturated FFAs were decreased in the cpdm epidermis. (B) The relative amounts of FFAs with chain length of 12–18 and longer than 26 carbons were reduced in cpdm samples compared to WT, instead FFAs with 22–24 carbons were increased. Lipid ions were detected in negative ion mode with m/z 199–600 range and normalized by the total ion count. Values are means of 7 (cpdm) or 8 (WT) mice. ** p < 0.01; *** p < 0.001, based on unpaired t-test with Holm-Sidak correction for multiple comparisons.

Fig 5. Discriminative value of a set of three ceramides.

ROC curve analysis of sphingosine ceramides CerAS(d18:1/24:0)2OH, CerAS(d18:1/16:0)2OH and CerNS(d18:1/16:0) in cpdm and WT epidermis. The threshold (red dotted line) set to differentiate between the two groups is not crossed by any of the samples analyzed for any of the three ceramides.

Fig 6. Discriminative value of a set of three free fatty acids.

ROC curve analysis of free fatty acids (FFAs) ω-hydroxyl palmitic acid (16OH-16:0), cerotic acid (26:0), and DHA (22:6)in cpdm and WT epidermis. The threshold (red dotted line) set to differentiate between the two groups is crossed by one sample of the WT group analyzed for cerotic acid (26:0), and DHA (22:6). There was no overlap between the groups for ω-hydroxyl palmitic acid (16OH-16:0).