Epidermal barrier alterations in tape-strips compared to biopsies
We next evaluated DEGs between AD and controls in either tape-strips or biopsies among a previously published epidermal barrier gene-subse4,11,30,66 (Figure 4) . While terminal differentiation (FLG2, SCEL, PSORS1C2, LCE5A) and keratin (KRT77, KRT79) products were significantly decreased in both techniques, greater decreases were observed in tape-strips. AKR1C3, which is expressed by differentiated keratinocytes, was the most down-regulated gene unique to tape-strips (Figure E1 ). Significant decreases in gap junction genes (GJB3, GJB5) were only seen in tape-strips (Figure 4, Table E1 ). Tight junction components like claudins and cadherins (CLDN8, CDH12)67 were comparably suppressed in tape-strips and biopsies, while CLDN1 was uniquely significantly decreased in biopsies. Markers of epidermal hyperplasia/proliferation (KRT6A, KRT6B, KRT16, KRT17, MKI67, AKR1B10, AKR1B15) and cornification (SPRR1A, SPRR2C, SPRR3) showed higher or unique increases in biopsies.
There was considerable variability in lipid biosynthesis and metabolism genes. PNPLA3, GAL, FA2H, FABP7, DGAT2, ACER1, ORMDL3, and DHCR7 were significantly decreased in both, with slightly greater FCHs in tape-strips, and DEGS2 and SPTLC3 were significantly down-regulated only in tape-strips. However, several other lipid-related genes (ELOVL3, FAR2, FADS2, ACOX2) were uniquely down-regulated in biopsies. A few genes (AGPAT3, FADS1, ELOVL5, SOAT1, LPIN1, PPARG) were decreased in biopsies, but increased in lesional tape-strips.