Assessment of the psoriatic transcriptome in a large sample: additional regulated genes and comparisons with in vitro models

Abstract

To further elucidate molecular alterations in psoriasis, we performed a gene expression study of 58 paired lesional and uninvolved psoriatic and 64 control skin samples. Comparison of involved psoriatic (PP) and normal (NN) skin identified 1,326 differentially regulated transcripts encoding 918 unique genes (549 up- and 369 downregulated), of which over 600 are to our knowledge previously unreported, including S100A7A, THRSP, and ELOVL3. Strongly upregulated genes included SERPINB4, PI3, DEFB4, and several S100-family members. Strongly downregulated genes included Wnt-inhibitory factor-1 (WIF1), beta-cellulin (BTC), and CCL27. Enriched gene ontology categories included immune response, defense response, and keratinocyte differentiation. Biological processes regulating fatty acid and lipid metabolism were enriched in the down-regulated gene set. Comparison of the psoriatic transcriptome to the transcriptomes of cytokine-stimulated cultured keratinocytes (IL-17, IL-22, IL-1alpha, IFN-gamma, TNF-alpha, and OSM) showed surprisingly little overlap, with the cytokine-stimulated keratinocyte expression representing only 2.5, 0.7, 1.5, 5.6, 5.0, and 1.9% of the lesional psoriatic dysregulated transcriptome, respectively. This comprehensive analysis of differentially regulated transcripts in psoriasis provides additional insight into the pathogenic mechanisms involved and emphasizes the need for more complex yet tractable experimental models of psoriasis.

Figure 1

Hierarchical clustering of the entire data sample. This figure has been presented previously (Gudjonsson et al., 2009b)and is presented here for comparison. Unsupervised hierarchical clustering showed near complete separation of the PP (n=58) samples from the PN and NN samples, whereas there was some overlap between PN (n=58) and NN (n=64) samples (Gudjonsson et al., 2009b).

Figure 2. Heatmap of differentially regulated transcripts between PP and NN skin

Differentially regulated transcripts between PP and NN skin (n=1,326) are shown in a heatmap image. Blue indicates low expression levels whereas red indicates high expression levels. Black bar above heatmap indicates NN skin, purple bar above heatmap indicates PP skin. Note that the number and intensity of up-regulated transcripts exceeds the number and intensity of up-regulated transcripts. Clustering was done only on rows and based on row means, for column samples were grouped into NN or PP groups without any clustering.

Figure 3. Venn diagram of up- and down-regulated genes in psoriasis skin

This diagram shows comparison of the overlap between PP vs. NN and PP vs. PN and PN vs. NN genes for all up- and down-regulated genes. Not unexpectedly, there is a large overlap between the PP vs. NN and PP vs. PN datasets. There was a larger overlap of PN vs. NN with PP vs. NN (27 up-regulated genes and 52 down-regulated genes) compared to the PP vs. PN (24 up-regulted genes and 19 down-regulated genes) dataset suggesting that PN skin is more similar to PP skin.

Figure 4. QRT-PCR confirmation of several of the differentially expressed genes

PI3 (SKALP), DEFB4, S100A12, SPRR2C and S100A7L1(S100A7A) were up-regulated 1,300-fold, 2,200-fold, 63-fold, 15-fold, and 4500-fold respectively in PP vs. NN skin, whereas ELOVL3, THRSP and BTC were down-regulated 30-fold, 8-fold and 36-fold respectively in PP skin. The results are shown relative to the expression of the housekeeping gene RPLP0 (36B4) and as mean + SEM, n=30. Statistical significance tested with Student's two-tailed t-test assuming equal variances and indicated by ** p<0.01 and *** p<0.001, for PP versus PN skin.

Figure 5. Enriched gene ontology (GO) categories for up- and down-regulated transcripts

Figure 6. Comparison of the psoriatic transcriptome to cytokine stimulated keratinocytes

We compared the PP vs. combined PN and NN transcriptome to published genomic maps obtained from cytokine stimulated keratinocytes using same criteria as were used for the PP transcriptome (≥ 2-fold change). This analysis was focused on cytokine previously implicated in the pathogenesis of psoriasis, such as IL-1α (Mee et al., 2007), OSM (Finelt et al., 2005), TNF-α (Banno et al., 2004), IFN-γ (Mee et al., 2007), IL-22 and IL-17 (Nograles et al., 2008). Note that there is no overlap between down-regulated genes in PP vs. NN/PN and IL-1α down-regulated genes. Several of the cytokines showed moderate overlap with the psoriatic transcriptome although this never represented more than a very small fraction of the entire PP transcriptome (maximum 53/649 = 5.6% for IFN-γ).