Targeting of Ras-mediated FGF signaling suppresses Pten-deficient skin tumor.

Deficiency in PTEN (phosphatase and tensin homolog deleted on chromosome 10) is the underlying cause of PTEN hamartoma tumor syndrome and a wide variety of human cancers. In skin epidermis, we have previously identified an autocrine FGF signaling induced by loss of Pten in keratinocytes. In this study, we demonstrate that skin hyperplasia requires FGF receptor adaptor protein Frs2α and tyrosine phosphatase Shp2, two upstream regulators of Ras signaling. Although the PI3-kinase regulatory subunits p85α and p85ß are dispensable, the PI3-kinase catalytic subunit p110α requires interaction with Ras to promote hyperplasia in Pten-deficient skin, thus demonstrating an important cross-talk between Ras and PI3K pathways. Furthermore, genetic and pharmacological inhibition of Ras-MAPK pathway impeded epidermal hyperplasia in Pten animals. These results reveal a positive feedback loop connecting Pten and Ras pathways and suggest that FGF-activated Ras-MAPK pathway is an effective therapeutic target for preventing skin tumor induced by aberrant Pten signaling.

Pten loss induces autocrine FGF signaling to promote skin tumorigenesis.

Inactivation of the Pten tumor suppressor negatively regulates the PI3K-mTOR pathway. In a model of cutaneous squamous cell carcinoma (SCC), we demonstrate that deletion of Pten strongly elevates Fgf10 protein levels without increasing Fgf10 transcription in vitro and in vivo. The translational activation of Fgf10 by Pten deletion is reversed by genetic disruption of the mTORC1 complex, which also prevents skin tumorigenesis in Pten mutants. We further show that ectopic expression of Fgf10 causes skin papillomas, whereas Pten deletion-induced skin tumors are inhibited by epidermal deletion of Fgfr2. Collectively, our data identify autocrine activation of FGF signaling as an essential mechanism in promoting Pten-deficient skin tumors.

Frs2α and Shp2 signal independently of Gab to mediate FGF signaling in lens development.

Fibroblast growth factor (FGF) signaling requires a plethora of adaptor proteins to elicit downstream responses, but the functional significances of these docking proteins remain controversial. In this study, we used lens development as a model to investigate Frs2α and its structurally related scaffolding proteins, Gab1 and Gab2, in FGF signaling. We show that genetic ablation of Frs2α alone has a modest effect, but additional deletion of tyrosine phosphatase Shp2 causes a complete arrest of lens vesicle development. Biochemical evidence suggests that this Frs2α-Shp2 synergy reflects their epistatic relationship in the FGF signaling cascade, as opposed to compensatory or parallel functions of these two proteins. Genetic interaction experiments further demonstrate that direct binding of Shp2 to Frs2α is necessary for activation of ERK signaling, whereas constitutive activation of either Shp2 or Kras signaling can compensate for the absence of Frs2α in lens development. By contrast, knockout of Gab1 and Gab2 failed to disrupt FGF signaling in vitro and lens development in vivo. These results establish the Frs2α-Shp2 complex as the key mediator of FGF signaling in lens development.

The functional role of the Meis/Prep-binding elements in Pax6 locus during pancreas and eye development.

Pax6 is an essential transcription factor for lens, lacrimal gland and pancreas development. Previous transgenic analyses have identified several Pax6 regulatory elements, but their functional significance and binding factors remain largely unknown. In this study, we generated two genomic truncations to delete three elements that were previously shown to bind to the Meis/Prep family homeoproteins. One 3.1 kb deletion (Pax6(∆DP/∆DP)) removed two putative pancreatic enhancers and a previously identified ectodermal enhancer, while a 450 bp sub-deletion (Pax6(∆PE/∆PE)) eliminated only the promoter-proximal pancreatic enhancer. Immunohistochemistry and quantitative RT-PCR showed that the Pax6(∆PE/∆PE) pancreata had a significant decrease in Pax6, glucagon, and insulin expression, while no further reductions were observed in the Pax6(∆DP/∆DP) mice, indicating that only the 450 bp region is required for pancreatic development. In contrast, Pax6(∆DP/∆DP), but not Pax6(∆PE/∆PE) mice, developed stunted lacrimal gland and lens hypoplasia which was significantly more severe than that reported when only the ectodermal enhancer was deleted. This result suggested that the ectodermal enhancer must cooperate with its neighboring sequences to regulate the Pax6 ectodermal expression. Finally, we generated conditional knockouts of Prep1 in the lens and pancreas, but surprisingly, did not observe any developmental defects. Together, these results provide functional evidence for the independent and synergistic roles of the Pax6 upstream enhancers, and they suggest the potential redundancy of Meis/Prep protein in Pax6 regulation.