In the skin model activating mutations in H-ras, K-ras and N-ras are sufficient to produce the clonal benign squamous papilloma expressing the phenotype of the initiated cell. Utilizing H-ras transduced isolated primary keratinocytes, we have demonstrated that oncogenic Ras upregulates EGFR ligands, establishing a positive feedback loop through EGFR, thus increasing proliferation and simultaneously activating src kinases to tyrosine phosphorylate and inactivate PKC delta, thereby inhibiting PKC-dependent keratinocyte terminal differentiation. Contemporaneous H-ras mediated activation of phospholipase C and PKCalpha establishes a second feedback loop through IL-1alpha and its receptor, activating MyD88 and NFkB to establish a unique gene expression profile characteristic of both initiated cells and papillomas. Interruption of any of these interacting pathways by drugs, knockdown or genetic deletion reduces or prevents tumor formation. An important downstream effector of these interacting pathways is the upregulation and release of chemokines such as CXCL1 and CXCL2 that act in a paracrine mode to attract inflammatory cells and stimulate angiogenesis and an autocrine mode to promote migration of initiated cells. Tissue array analysis of human skin tumors indicates that Met is frequently upregulated in squamous carcinomas. We induced tumors with DMBA and TPA on transgenic mice that overexpress HGF and noted a marked increase in papillomas and carcinomas relative to control groups. Further analysis revealed that activation of Met in keratinocytes is sufficient to initiate tumor formation after TPA promotion in the absence of DMBA. In isolated keratinoctyes, activation of Met mimics the H-ras initiated phenotype reproducing the EGFR and other required feedback loops. However, the application of DMBA to mice with activated Met produces more carcinomas with K-ras mutations than application of DMBA to control mice where H-ras mutations dominate, suggesting that tissue context influences the selection of ras mutations in tumors. RNA profiling of initiated keratinocytes produced by Met or H-ras activation has revealed signature gene expression changes common to both and not previously associated with tumor initiation. Current studies are extending the analysis to K-ras initiated keratinocytes and using knockdown or overexpression approaches to test for critical elements in the common expression profiles essential for initiation and tumor formation. Our interest in inflammation and cancer pathogenesis in skin prompted us to develop a model directly tied to a known human inflammatory disease. S100A7 (psoriasin) and S100A15 (Koebnerisin) are highly expressed in human psoriasis, and we demonstrated their proinflammatory functions as chemokines and ligands for RAGE and a G-protein coupled receptor. We cloned the single mouse homologue (mS100A7/A15) and created a conditional epidermal targeted transgenic mouse line that expresses enhanced amounts of mS100A7/A15 conditionally in skin, producing many biochemical and inflammatory features of psoriasis susceptibility and lesional flare. These mice develop papillomas at the same frequency as controls after DMBA/TPA, but induced expression of mS100A7/A15 protein in epidermis reduces malignant conversion. Since we have previously documented the components of the TPA-induced inflammatory milieu in the skin of mS100A7/15 mice, we are evaluating the tumor microenvironment for clues to molecules that may alter malignant conversion. We are also evaluating the influence of genetic deletion of MyD88 on the promotion component of carcinogenesis. The advantage of having a body surface squamous cancer model reflecting a major group of internal tumors and cutaneous cancers (the second most frequent human malignancy) provides a special opportunity to test cancer therapeutic agents both topically and systemically. Support for this approach comes from our analysis of ingenol-3-angelate now in the clinic as Picato, a topical treatment for actinic keratosis, basal cell and superficial squamous cell skin cancer. The mechanism of action of this PKC agonist was mysterious until we determined its activity as a substrate for epidermal ABCB1, the P-glycoprotein (P-gp) also known as MDR-1. Picato is transported across the epidermis by absorptive transport into the dermis where its PKC agonist activity disrupts endothelial cells leading to tumor necrosis. The absorptive transport model of topical drug therapy has been confirmed by studies with Cyclosporin A, a P-gp substrate which reaches therapeutic plasma levels after topical administration unless blocked by an ABCB1 inhibitor. We are currently engaged in a wider screen for potential topically active anticancer agents using a high throughput P-gp mediated efflux assay that employs ABCB1 overexpressing KB-V1 cells and calcein AM efflux. In a screen of 193 compounds with known therapeutic targets or from a kinase inhibitor library, we identified 36 positive hits and twenty compounds not previously known as substrates for ABCB1. Among these are a PI3K/mTOR dual kinase inhibitor BEZ235, the PLK1 inhibitor BI 2536 and the IKK1/2 inhibitor IKK16. Considering the major function of NFkB in our carcinogenesis and inflammation models, the current emphasis is directed to evaluating IKK16 as potential topical anti-cancer and anti-inflammation agent.
Showing the most recent 10 out of 21 publications
