Figure 1.
Cellular effects in human reconstructed skin exposed to UVA1.
Sham-exposed (control) and UV-exposed samples were taken for classical histology and for vimentin staining (vimentin: green labeling, nuclei counterstaining: red labeling) at 48 h post UVA1 exposure. Arrows indicate fibroblast disappearance in human dermal equivalent.
Figure 2.
ROS and lipid peroxidation detection in reconstructed skin exposed to UVA1.
ROS assay was performed using sections of reconstructed skin after DCFH-DA probe incorporation and UVA1. Bracket and arrows indicated the fluorescent keratinocytes and fibroblasts, respectively, in UVA1-exposed samples. None of them were detected in unexposed reconstructed skin. The dotted line indicates dermal epidermal junction (A). Levels of DCFH-DA probe fluorescence in reconstructed skin after UVA1 exposure. AU, arbitrary units (B). Distance between dermal epidermal junction and the deepest positive DCFH-DA cells. Indicated values correspond to the mean of 6 measurements in each experimental condition (C). 8-isoprostane amount in culture medium of reconstructed skin 24 hours after UVA1 exposure (D). *, mean value significantly different from mean value at 0 J/cm2; a, mean value significantly different from mean value at 10 J/cm2; §, mean value significantly different from mean value at 20 J/cm2; #, mean value significantly different from mean value at 30 J/cm2 (p<0.05, Student's t test).
Figure 3.
Overall analysis of gene expression in reconstructed skin exposed to UVA1, using microarray.
Triplicates of reconstructed skins were unexposed (control) or exposed to 40 J/cm2 UVA1. Six hours later, a full genome transcriptomic study was conducted using Affymetrix microarray in fibroblasts (F) and keratinocytes (K), separately, for the 3 control reconstructed skins (samples F1-F3 and K1-K3) and for the 3 UVA1 exposed reconstructed skins (samples F4-F6 and K4-K6). A: Hierarchical clustering based on all probe set normalized expression data, using Ward's method and correlation distance. Y-axis of dendrogram represents the linkage distance that separates singletons or clusters. Height at which two clusters are merged in dendrogram reflects distance of the two clusters. B: Fold change comparison plot between fibroblasts and keratinocytes depicting number of significantly modulated probe sets 6 hours after exposure to 40J/cm2 UVA1. Each probe set is plotted. On y-axis: log of fold change value in keratinocytes; on x-axis: log of fold change value in fibroblasts. In keratinocytes and in fibroblasts, 502 and 494 probe sets were differentially expressed between UVA1 exposed and control reconstructed skins (fold-change >1.5 or <0.67, Adjp<0.001), respectively. Blue circles (n = 107) represent probe sets modulated in both keratinocytes and fibroblasts. Green (n = 395) and red (n = 387) circles represent probe sets modulated only in keratinocytes and in fibroblasts, respectively. Grey circles (n = 2787): un-modulated probe sets. C: Heat map showing relative expression levels of the probe sets differentially expressed between UVA1-exposed and control samples (fold-change threshold >1.5 or <0.67 and Adjp value<0.001). Two-dimensional hierarchical clustering was carried out with the 502 and 494 probe sets differentially expressed between UVA1 exposed and control samples in keratinocytes and fibroblasts, respectively. Euclidean distance and Ward's method, based on normalized log2-transformed gene expression value relative to median value of each row were used [85]. Each row represents a probe set, each column represents one sample. Red, high expression. Black, median expression. Green, low expression.
Table 1.
Summary of most significant enriched GO Biological Process terms in reconstructed skins exposed to UVA1.
Figure 4.
Distribution of UVA1 modulated genes in functional families.
In order to perform an exhaustive bibliographic analysis including literature related to skin and dermatology, the list of UVA1 modulated genes was reduced by using a fold change threshold>2 or <0.5, and an Adjp<0.001. Under these criteria, 134 and 141 genes were found modulated in fibroblasts and keratinocytes respectively of UVA1 exposed reconstructed skins. In fibroblasts 24 genes out of 134 could not be classified because their functions were poorly described; 110 genes were distributed in functional families (A). In keratinocytes 30/141 genes could not be classified; 111 genes were distributed in functional families (B). Some genes could be classified in several functional families. Lists of gene names associated with gene bank accession number, fold change values and their distribution in functional families are given in Tables S6 and S7, for fibroblasts and keratinocytes respectively.
Table 2.
Gene expression modulation assessed by quantitative PCR in fibroblasts of reconstructed skins exposed to UVA1.
Table 3.
Gene expression modulation assessed by quantitative PCR in keratinocytes of reconstructed skins exposed to UVA1.
Figure 5.
Levels of secreted proteins in culture medium of reconstructed skin exposed to UVA1.
Culture media were taken at 48 hours post UVA1 exposure and used to measure the amount of extracellular matrix remodeling proteins (matrix metalloproteinases, MMPs and GDF15), pro-inflammatory proteins (IL-6, GM-CSF/CSF2, CCL20 and GDF15), the HGF growth factor and the CXCL10 ( = IP10) interferon inducible protein. Values of control samples were adjusted to the 1 value. Asterisks indicate a significant difference between mean protein amount of control samples and mean protein amount of UVA1 exposed samples (p<0.05, Student's t test). AU, arbitrary units.