A dynamic model of inorganic arsenic-induced carcinogenesis reveals an epigenetic mechanism for epithelial-mesenchymal plasticity

Abstract

Inorganic arsenic (iAs) causes cancer by initiating dynamic transitions between epithelial and mesenchymal cell phenotypes. These transitions transform normal cells into cancerous cells, and cancerous cells into metastatic cells. Most in vitro models assume that transitions between states are binary and complete, and do not consider the possibility that intermediate, stable cellular states might exist. In this paper, we describe a new, two-hit in vitro model of iAs-induced carcinogenesis that extends to 28 weeks of iAs exposure. Through week 17, the model faithfully recapitulates known and expected phenotypic, genetic, and epigenetic characteristics of iAs-induced carcinogenesis. By 28 weeks, however, exposed cells exhibit stable, intermediate phenotypes and epigenetic properties, and key transcription factor promoters (SNAI1, ZEB1) enter an epigenetically poised or bivalent state. These data suggest that key epigenetic transitions and cellular states exist during iAs-induced epithelial-to-mesenchymal transition (EMT), and that it is important for our in vitro models to encapsulate all aspects of EMT and the mesenchymal-to-epithelial transition (MET). In so doing, and by understanding the epigenetic systems controlling these transitions, we might find new, unexpected opportunities for developing targeted, cell state-specific therapeutics.

Keywords: Epigenetics; Epithelial-to-mesenchymal transition; Heavy metal carcinogenesis; Inorganic arsenic.

Figure 1.. Chronic, low-dose iAs-transformed cells exhibit mesenchymal features and gain migratory and invasive characteristics.

Data shown are the mean of three independent experiments ± SEM. **** p<0.05. Cells were fixed and stained with crystal violet. (A) Diagram of the iAs-treatment model used in this study. BEAS2B cells were grown and exposed to 0.5 μM iAs for up to 17 weeks, then 2.0 μM iAs was added for another 10 weeks (28 weeks total iAs exposure). Cells were harvested at weeks NT, 3, 8, 17 and at the end of the 2.0T exposure. Cells were flash frozen and stored at −80 °C until pellet was used for analyses. (B) Representative images of cells at each week of the iAs-induced EMT. Cells gain a more mesenchymal phenotype (more elongated) through week 17 and begin to revert to a more epithelial state in the 2.0T. Scale bar is 100 μm. (C) Representative images of transwell migration assays from iAs-exposed cells. Scale bar is 100 μm (D) Quantification of transwell migration assays shown normalized to NT cells. (E) Representative images of wound healing assays of cells at 0 (Scale bar is 500 μm) and 24 hours after wounding (Scale bar is 200 μm). (F) Quantification of wound healed after 24 hours. Data is mean of n=3 ± s.e.m, p < 0.05 by one-way ANOVA as we compared all changes to NT.

Figure 2.. Late stage iAs-transformed cells gain tumorigenic properties.

(A) Representative images of anchorage-independent growth assays of cells throughout iAs exposure. Scale bar is 200 μm. (B) Quantification of anchorage-independent growth assays. (C) Representative images of spheroid growth by iAs-treated cells (Top). Zoomed in image showing filopodia formed in the 2.0T cells (Bottom). Arrows highlight characteristic filopodia protruding from cells’ edges. Scale bar is 200 μm. (D) Quantification of spheroid filopodia projections. (E) Representative images of xenograft tumors formed in mice (Left W17, Right 2.0T) seven weeks post subcutaneous injection of cells. Tumors only formed from W17 and 2.0T. Scale bar is 1 cm. (F) Xenograft volumes measured at each week post injection. Data shown are the average of four tumors ± SEM. * denotes p<0.05. (G) Final xenograft weights when harvested at seven weeks post subcutaneous injection. Data is mean of n=3 ± s.e.m, p < 0.05 by one-way ANOVA as we compared all changes to NT.

Figure 3.. EMT-related gene expression changes during iAs-induced transformation.

Shown are qRT-PCR fold change in gene expression of EMT markers through iAs-induced carcinogenesis. Data shown are the mean of three independent experiments ± SEM. (A) Epithelial marker genes (CLDN1, CDH1, TJP1). (B) Mesenchymal marker genes (CDH2, FOXC2, VIM). (C) EMT-associated transcription factors (SNAI1, SNAI2, ZEB1). (D) Western blot of EMT-associated Marker genes. (E) EMT markers in cells grown in 3D culture (CTNNB1, CDH1, CDH2, CDH3, SNAI1). For clarity, statistical notations (*) are not shown. Data is mean of n=3 ± s.e.m, p < 0.05 by one-way ANOVA as all weekly changes were compared to time-matched NT control. (F) Western blots of EMT-associated Marker genes in the 3D cell culture.

Figure 4.. SNAI1 bivalent chromatin state driven by iAs-induced carcinogenesis.

Data from ChIP-qPCR for the occupancy of histone modifications at EMT-related promoters throughout iAs-induced carcinogenesis. Data are mean from three independent experiments ± SEM. Inset graph show gene expression changes. Epithelial-related gene promoters: (A) CDH1, (B) CTNNB1, (C) TJP1; Mesenchymal-related gene promoters: (D) CDH2, (E) VIM; EMT-related transcription factor promoters: (F) SNAI1, (G) ZEB1, (H) SNAI2. For clarity, statistical notations (*) are not shown. Data is mean of n=3 ± s.e.m, p < 0.05 by one-way ANOVA as all weekly changes were compared to time-matched NT control.

Figure 5.. Epigenetic Landscape Changes of EMT-Associated genes.

Cartoons of EMT-related gene promoter regions showing relative levels of histone post-translational modifications throughout the iAs exposure. (A) CDH1 Proximal Promoter. (B) CDH2 Proximal Promoter. (C) SNAI1 Proximal Promoter