Skip to main content
. 2011 May 23;30(46):4666–4677. doi: 10.1038/onc.2011.180

Figure 2.

Figure 2

Expression profiling of quiescent cultures of early-passage keratinocytes separates cSCC from normal and identifies potential tumor drivers in vivo. (a) Growth rates of keratinocytes used in this study as assessed by a colorimetric assay of mitochondrial dehydrogenase activity when seeded at low density (upper panel) or at confluence (lower panel). Confluent keratinocytes are quiescent after 48 h of culture. RNA for subsequent array experiment was isolated between 48 and 56 h after confluence. (b) An un-supervised clustering dendrogram of in vitro gene expression data generated by BRB-ArrayTools v3.8.1. cSCC keratinocyte samples (red box) cluster independently of non-cSCC keratinocyte samples (blue box). (c) Average cSCC versus normal skin (NS) fold change for all 154 concordantly expressed in vivo cSCC genes plotted against average psoriatic lesional skin versus psoriatic non-lesional skin (NLS) fold change shows a strong overall correlation (r2=0.84) and identifies genes specifically differentially regulated in cSCC. cSCC, cutaneous squamous cell carcinoma; EBKs, primary non-SCC RDEB keratinocytes; NHKs, primary normal human keratinocytes. A full colour version of this figure is available at the Oncogene journal online