Cytoskeletal protein Flightless I inhibits apoptosis, enhances tumor cell invasion and promotes cutaneous squamous cell carcinoma progression

Abstract

Flightless I (Flii) is an actin remodeling protein that affects cellular processes including adhesion, proliferation and migration. In order to determine the role of Flii during carcinogenesis, squamous cell carcinomas (SCCs) were induced in Flii heterozygous (Flii+/-), wild-type and Flii overexpressing (FliiTg/Tg) mice by intradermal injection of 3-methylcholanthrene (MCA). Flii levels were further assessed in biopsies from human SCCs and the human SCC cell line (MET-1) was used to determine the effect of Flii on cellular invasion. Flii was highly expressed in human SCC biopsies particularly by the invading cells at the tumor edge. FliiTg/Tg mice developed large, aggressive SCCs in response to MCA. In contrast Flii+/- mice had significantly smaller tumors that were less invasive. Intradermal injection of Flii neutralizing antibodies during SCC initiation and progression significantly reduced the size of the tumors and, in vitro, decreased cellular sphere formation and invasion. Analysis of the tumors from the Flii overexpressing mice showed reduced caspase I and annexin V expression suggesting Flii may negatively regulate apoptosis within these tumors. These studies therefore suggest that Flii enhances SCC tumor progression by decreasing apoptosis and enhancing tumor cell invasion. Targeting Flii may be a potential strategy for reducing the severity of SCCs.

Keywords: Flightless; cell invasion; skin cancer; squamous cell carcinoma.

Figure 1. Flightless protein levels are increased in invasive cutaneous SCC

A–B. Immunofluorescence analysis of Flii (greyscale in signal channel images and green in Merge) expression in normal margin skin and invasive cutaneous SCC of human (A, N = 21) or murine (B, N = 5) origin. Cytokeratin 14 (K14) immunofluorescence (greyscale in signal channel images and red in Merge) marks keratinocytes. Box and whisker plots show the median, range and quartiles of area coverage analysis carried out on the green channel to quantify Flii expression. Statistical significance was calculated using the Mann-Whitney test with Dunnet's post hoc test. *** P < 0.001. Magnification ×10 Insert ×20. Scale Bar = 100 μm C. Levels of secreted Flii were also examined in serum of three SCC patients. Flii is increased in serum of SCC patients (n = 3) compared to normal human serum (NHS) control. Albumin was used as a loading control. Graphical representation of Flii protein expression determined by densitometry. Figure is representative of three independent experiments. Mean +/− SD *p < 0.05.

Figure 2. Overexpression of Flii results in severe SCC development

A–B. Using the MCA induced model of SCC in Flii^+/−, WT and Flii^Tg/Tg mice we demonstrate that over-expression of Flii leads to quicker onset and more severe progression of necrotic and ulcerated SCC tumor lesions with significantly larger tumor area (n = 12/genotype) Mean +/− SD * and ° = P < 0.05. * denotes significance vs Flii^Tg/Tg and ° denotes significance vs Flii^+/−. C. Epithelial origin of invading cells and hyperkeratotic nodules (red arrows) present in the dermis of Flii^+/−, WT and Flii^Tg/Tg SCC induced mice was confirmed with positive cytokeratin staining in all samples. D–F. Representative images of tumor margins (dotted green line) in Flii^+/−, WT and Flii^Tg/Tg mice and macroscopic analysis of tumor volume and microscopic analysis of length of tumor epithelium revealing significantly larger tumors in Flii over-expressing mice. (n = 12/genotype) Mean +/− SD *P < 0.05. t = tumor n = normal skin G. Representative images of increased Flii expression in tumors of Flii^Tg/Tg mice compared to wild-type controls. (n = 12/genotype) Mean +/− SD *P < 0.05. Magnification ×20 and ×10. Scale Bar = 100 μm and 500 μm.

Figure 3. Flii does not influence the proliferation of cutaneous SCC tumors

A–B. Cutaneous SCC tumors in Flii^+/−, WT and Flii^Tg/Tg mice were stained for PCNA and number of positive proliferating cells analysed showing no effect of Flii on tumor proliferation. (n = 12/genotype). Magnification × 20. Scale Bar = 100 μm. Mean +/− SD *P < 0.05.

Figure 4. High levels of Flii lead to an evasion of apoptosis in cutaneous SCC tumors

A–C. Cutaneous SCC tumors in Flii^+/−, WT and Flii^Tg/Tg mice were stained for caspase-1 and Annexin-V markers of apoptosis and number of positive cells analysed showing significantly decreased levels of apoptosis in tumors of Flii overexpressing mice. (n = 12/genotype). Magnification ×20. Scale Bar = 100 μm. Mean +/− SD *P < 0.05.

Figure 5. Reducing Flii expression during SCC initiation and development using FnAb results in decreased SCC progression

WT mice were treated with FnAb at time of SCC induction and every second week throughout the trial. A–B. Examining the effect of reducing Flii levels using FnAb at time of SCC induction and during development resulted in decreased tumor progression and delayed development with significantly smaller tumor volume. (n = 12/treatment) Mean +/− SD *P < 0.05. C. Representative images of H&E stained tumor sections treated with FnAb and IgG control show smaller tumor size and less advanced tumor histology with less invasion, outgrowth and hyperkeratotic nodule formation. Magnification ×4 Insert ×20. Scale Bar = 500 μm and 100 μm. D–F. Macroscopic analysis of tumor volume and microscopic histological analysis of tumors revealed significantly decreased tumor volume, length of tumor epithelium and cross sectional tumor width in FnAb treated tumors. (n = 12/treatment) Mean +/− SD *P < 0.05. G. Representative images and graphical representation of significantly reduced Flii expression in FnAb treated tumors compared to IgG treated controls. (n = 12/treatment) Mean +/− SD *P < 0.05. H. Western Blot analysis of Flii levels in serum of FnAb treated and IgG control SCC induced wild-type mice showing significant reduction in Flii serum levels following FnAb treatment. (n = 3/treatment) Mean +/− SD *P < 0.05. Figure representative of three experimental repeats.

Figure 6. Reducing Flii expression prior to SCC initiation and development using FnAb results in decreased SCC progression

As Flii is increased in human SCC samples, effect of preventative FnAb treatment prior to SCC induction was investigated in Flii^Tg/Tg mice. Mice were treated with FnAb two weeks prior to SCC induction and every second week throughout the trial and SCC development. A–B. Examining the effect of reducing Flii levels in mice skin using preventative FnAb treatment prior to SCC induction and during development resulted in decreased tumor progression and size in FnAb treated mice. IgG control mice have significantly larger and more developed necrotic ulcerated tumors. (n = 12/treatment) Mean +/− SD *P < 0.05. C. Representative images of H&E stained tumors treated with FnAb or IgG control show more severe ulcerated tumor pathology in IgG control group (black arrows). Magnification ×4 Insert ×20. Scale Bar = 500 μm and 100 μm. D–F. Macroscopic analysis of tumor volume and microscopic histological analysis of tumors revealed no significant difference in length of tumor epithelium however significantly smaller cross tumor volume and cross sectional tumor width was observed in FnAb treated tumors. G. Representative images and graphical representation of significantly reduced Flii expression in FnAb treated tumors compared to IgG treated controls. (n = 12/treatment) Mean +/− SD *P < 0.05.

Figure 7. Reducing Flii expression using FnAb reduces sphere formation and invasion properties of SCC keratinocytes in-vitro

Flii levels in MET-1 SCC cells were compared to both HaCaT immortalized keratinocyte cell line A. and MET-2 (cells from recurrent tumor of same patient) and MET-4 (cells from metastatic tumor of same patient) SCC cell line B. showing significant increase in Flii levels in MET-1 SCC cell line compared to HaCaT immortalized keratinocytes. C. The effect of reducing Flii using FnAb (50 μg/ml) or dose matched IgG antibody control on MET-1 SCC keratinocyte ability to form tumor spheres was analysed. MET-1 SCC keratinocytes treated with FnAb showed a significantly delayed ability and reduced sphere formation. Representative images of sphere size at day 10 are shown. N = 8. Mean +/− SD *p < 0.05. Scale bar = 500 μm refers to all images. Sporadic SCC cell line, MET-1, was used in a 3D organotypic model of cell invasion. Cells were treated with FnAb (50 μg/ml) at time of gel preparation and in media throughout the ten day period. D–E. Representative images of H&E stained gels illustrating cell invasion properties. Treatment of MET-1 SCC cells with FnAb significantly reduced the cell invasion properties (green arrow) into the gel matrix with 60% decreased depth of invasion. (n = 3/treatment) Magnification = 10×. Mean +/− SD *P < 0.05. F. Considering the function of Flii in actin remodelling, Flii association with invadopodia marker cortactin was examined using immunohistochemistry. Representative confocal images of dual immunofluorescence of MET-1 SCC cell lines shown Flii (green) was prominent throughout the epidermis highlighting all SCC cells, while cortactin (orange) was only prominent in invading cells highlighting the invadopodia. Flii co-localized with cortactin (yellow) suggesting a tumor promoting activity for Flii in SCC cancer cell invasion. Magnification = 400×. G. Anti-cortactin immunoprecipitate (IP) was prepared from confluent MET-1 SCC cell line and immunoblotted with Flii or Gelsolin (positive control) antibodies. Flii but not Gelsolin was found to associate with cortactin and specific bands were observed in MET-1 cell lysate (WCL) control. Data are representative of three independent experiments.