Fenretinide Disrupts Actin-Cytoskeleton Interactions in Human Oral Keratinocytes

Patients with lesions of oral epithelial dysplasia (OED) that recur despite microscopically confirmed clear surgical margins are a reality of OMFS practice.  Such recalcitrant lesions are problematic for both surgeons and patients due to the need for multiple biopsies and/or laser ablations, as well as concerns regarding potential for malignant transformation. Oral cancer chemoprevention, which entails the use of natural or synthetic compounds to prevent cancer occurrence, is a focus of our lab. We have recently developed an interest in the vitamin A derivative, Fenretinide, due to its recognized abilities to induce maturation or programmed death in precancerous or malignant epithelial cells. While previous human Fenretinide oral cancer chemoprevention trials were not highly successful, these studies relied on systemic delivery and did not determine drug levels at the treatment site [Chiesa et al., Int. J of Cancer, 2005].  To eliminate the prospect of insufficient target levels, our labs developed and evaluated a mucoadhesive local-delivery Fenretinide patch in a rabbit model.  Our data confirmed that patch-mediated delivery (30 minutes q.d. for 10 days) provided therapeutic drug levels (as shown by protective enzyme modulation), without any local or systemic toxicities [Holpuch et al., Carcinogenesis, 2012].  Previous studies conducted in other human cancer cells have introduced the prospect that Fenretinide may also inhibit cell motility.  Confirmation of an additional chemopreventive mechanism i.e. perturbation of cytoskeletal reorganization which is essential for basement membrane invasion and OED transformation, in a human oral keratinocyte model, will further clarify Fenretinide's chemopreventive repertoire.  

A spectrum of human oral keratinocytes spanning the normal (HOK3437)-dysplastic-oral squamous cell carcinoma (OSCC) continuum was used to determine whether Fenretinide effects cell motility. Fenretinide significantly suppressed keratinocyte migration, with greatest inhibition observed in the chemoprevention-targeted dysplastic keratinocytes. These data show: 1) significant migration inhibition within the normal-dysplastic-OSCC continuum via Platypus Cell Migration Assay assessed at 5μM over a 48-hour period (normal HOK 3437 and dysplastic cells [p<0.0001, n=8], SCC15 [p<0.01, n=8], SCC2095 cell lines [p<0.0001, n=8]) 2) constant viabilities (hemocytometer counts/ trypan blue exclusion, post 24h-96h treatment) among all SCC cultures, despite the effect of 5µM Fenretinide reducing SCC-cell proliferation related to matched controls (SCC15 + SCC2095 p<0.05, 72h; SCC2095 p<0.01, 96h; n=4) 3) Fenretinide elicited a concentration-dependent effect on dysplastic cells, demonstrating a marked decrease in viability (control vs 5μM p≤0.05, over time-course; n=4) and 4) significant decrease in HOK3437 keratinocytes proliferation with concurrent reduction in cell viability. ANOVA (Bonferroni post hoc) and Kruskal-Wallis (Dunn's Multiple Comparison post hoc) statistics were used to analyze these normally-distributed data.  Companion fluorescent microscopy analyses provided mechanistic insights regarding how Fenretinide inhibits cell mobility. Dual fluorescent staining for F-actin (green, phalloidin) and nuclei (DAPI, blue), used to depict actin networks/cell nuclei, revealed Fenretinide perturbed actin-myofilament reorganization in a dose-dependent fashion (highest with 10 µM) in all cell strains and lines evaluated. Fenretinide application elicited loss of cellular polarity, deviating from normal oral keratinocyte polygonal shape.  The lack of histologic grade specificity observed in these fluorescent studies highlights the generalizable nature of fenretinide's anti-migration effects.  

Observed cytoskeletal reorganization refocused attention to the STAT3 signaling hub, a clinically-relevant pharmaceutical signaling target instrumental in head and neck SCC (HNSCC) growth and survival.  Our results also demonstrate Fenretinide inhibits STAT3 signaling (confirmed using Western Blotting analysis), as well modulates cytoskeletal-dependent proteins such as ERM (ezrin/radixin/moesin) and Cofilin towards immobility. These data may reflect a variety of mechanisms including: 1) binding of the lipophilic drug, Fenretinide, to phospholipid membranes, 2) Fenretinide's redox-cycling interference with signaling, 3) Fenretinide-protein interactions. These findings expand Fenretinide's scope from growth-modulatory to also include mobility-regulatory chemopreventive effects.