Re: the Wnt signaling pathway in non-small cell lung cancer.

We read with great interest the article by Stewart (1) in the December issue of the Journal. The author gives a useful update of genomic alterations affecting Wnt signaling components that are frequently observed in non-small cell lung cancer (NSCLC) and reminds us that Wnt component mutations seen in other cancers such as colon are uncommon in NSCLC. We believe it is necessary to bring further attention to the fact that Wnt pathway disruption in lung cancer differs greatly from what is observed in other malignancies. As described by Stewart, numerous Wnt pathway components are disrupted in lung adenocarcinoma (LUAD). The alterations summarized in Table 1 from Stewart (1) are also prominent in our own LUAD cohort (n = 77; Figure 1). Further investigation of our tumor set revealed that when considering gene expression signatures associated with canonical Wnt pathway transcription, up to 30% of tumors showed signatures consistent with active Wnt signaling (TCF1 13–30%, TCF4 4%, LEF1 2%; gene set enrichment analysis [GSEA], P < .01). These results agree with other studies assessing Wnt pathway activation in lung cancer, indicating Wnt/TCF is active in a significant fraction of lung tumors. For example, a recent study demonstrated that 45% of 309 NSCLCs stained positive for nuclear β-catenin, which is often associated with Wnt/TCF activity (2). However, despite the occurrence of component disruption and the apparent Wnt activity across lung tumor datasets, we found no correlation between Wnt/TCF activity (defined using GSEA) and disruption status of the Wnt pathway component genes (Table 1 from Stewart [1]). The frequencies of component alteration do not differ between tumors classified as Wnt/TCF active vs inactive for CoRReSPoNdeNCe