FOXA2 suppresses endometrial carcinogenesis and epithelial-mesenchymal transition by regulating enhancer activity

FOXA2 encodes a transcription factor mutated in 10% of endometrial cancers (ECs), with a higher mutation rate in aggressive variants. FOXA2 has essential roles in embryonic and uterine development. However, FOXA2’s role in EC is incompletely understood. Functional investigations using human and mouse EC cell lines revealed that FOXA2 controls endometrial epithelial gene expression programs regulating cell proliferation, adhesion, and endometrial-epithelial transition. In live animals, conditional inactivation of Foxa2 or Pten alone in endometrial epithelium did not result in ECs, but simultaneous inactivation of both genes resulted in lethal ECs with complete penetrance, establishing potent synergism between Foxa2 and PI3K signaling. Studies in tumor-derived cell lines and organoids highlighted additional invasion and cell growth phenotypes associated with malignant transformation and identified key mediators, including Myc and Cdh1. Transcriptome and cistrome analyses revealed that FOXA2 broadly controls gene expression programs through modification of enhancer activity in addition to regulating specific target genes, rationalizing its tumor suppressor functions. By integrating results from our cell lines, organoids, animal models, and patient data, our findings demonstrated that FOXA2 is an endometrial tumor suppressor associated with aggressive disease and with shared commonalities among its roles in endometrial function and carcinogenesis.


b) Fixation and staining
For immunostaining, the day 12 organoids were harvested from Matrigel by scraping with a pipette tip, washed in PBS, and fixed in 4% PFA (Electron Microscopy Sciences) for 20 min at room temperature (RT). Following fixation, organoids were gently pelleted and overlaid with 200 µl of low melting agarose (Lonza) at 42 0 C and permitted to solidify. The agarose plugs containing organoids were embedded in paraffin and 5 µm sections were prepared. The sections were deparaffinized in xylene followed by rehydration through a series of graded ethanol. Antigen retrieval was performed using sodium citrate buffer (pH 6.0, Vector Laboratories). Slides were washed with TBS containing 0.1% Tween 20 and incubated with cytokeratin, E-cadherin and GM130 antibodies (details in the resources table) overnight at 4°C. Sections were washed and incubated with Alexa Fluor secondary antibodies for 1 hr at RT and cover slipped using VECTASHIELD antifade mounting media with DAPI (Vector Laboratories). Images were taken on a Nikon ECLIPSE Ni fluorescence microscope. For 3D z-stack imaging, whole-mount organoids were fixed, permeabilized and stained with antibodies followed by imaging on a Zeiss LSM880 inverted confocal microscope.

c) Proliferation assays
Organoids were harvested from Matrigel using cell recovery solution (Corning) and trypsinized in TrypLE express enzyme (Thermo Fisher Scientific). The cells were resuspended in a 15 µl Matrigel drop containing 500 cells in the wells of a 96-well plate in triplicate, overlaid with 150 µl of organoid expansion medium and incubated at 37°C containing 5% CO 2 . The luminescence signal was recorded with a Tecan Spark 10M microplate reader on day 3, 5, 7 and 10 with Cell Titer-Glo 3D reagent (Promega).

Tumor grafting
All procedures for tumor xenograft mice were approved by the University of Texas Southwestern Medical Center Institutional Animal Care and Use Committee. Six-to eight-weekold female NOD scid gamma mice (Jackson Laboratory) were used. Ishikawa (ISK) or FP cells (1 X 10 6 ) +/-transduced with FOXA2 lentiviral particles (OriGene, as described above) were injected subcutaneously into the flanks of the mice (n = 4/group) in 200 μL of sterile DPBS and Matrigel (1:1). Tumor dimensions were measured twice in a week by a digital caliper and volumes were calculated using the formula (length x width 2 )/2. Tumor weights were determined at the end of the experiment.

Immunohistochemistry and quantification
Human EC TMA slides (US Biomax) were deparaffinized in xylene followed by rehydration in graded alcohols and water. Antigen retrieval was performed using sodium citrate buffer (pH 6.0, Vector Laboratories). Slides were washed with TBS containing 0.1% Tween 20 and incubated with 3% hydrogen peroxide (H 2 O 2 ) in water for 15 min to block endogenous peroxidase. Sections were blocked with 2% BSA in PBS containing 0.1% Triton X-100 for 1 hr at RT. After blocking, tissue sections were incubated with primary antibodies against FOXA2, ERα or PR (details in the resources table) at 4ᵒC overnight. Slides were then washed and incubated with peroxidaseconjugated secondary antibodies (Vector Laboratories) and DAB substrate (Agilent Dako) to detect bound antibodies. Tissue sections were counterstained with hematoxylin. Images were captured using an Aperio AT2 slide scanner (Leica Biosystems) and analyzed using Aperio ImageScope software. Immunohistochemistry intensity score (H-score) was calculated from pixel intensity values (the sum of 3 x % of pixels with strong staining + 2 x % of pixels with moderate staining + 1 x % pixels with weak staining).

Immunostaining
Ishikawa and HEC-1-B cells were cultured on glass slides, fixed in 4% paraformaldehyde (Electron Microscopy Sciences) for 20 min and processed for immunofluorescence. Cells were permeabilized in PBS containing 0.5% Triton X-100 for 10 min at 4 0  RT and cover slipped using VECTASHIELD antifade mounting media with DAPI (Vector Laboratories). Images were captured on a Nikon ECLIPSE Ni fluorescence microscope using NIS-Elements BR software.

Western blots
EC cell lines were lysed in ice-cold RIPA buffer (Thermo Fisher Scientific) containing protease and phosphatase inhibitors (Sigma). Following lysis, the protein extracts were centrifuged at 12,000 rpm for 5 min at 4 0 C and supernatant was collected. 20 µg of protein was mixed with 4X NuPAGE LDS Sample Buffer (Thermo Fisher Scientific), incubated at 95 0 C for 5 min and resolved by 4-12% NuPAGE gels (Thermo Fisher Scientific). Transfer to PVDF membrane (Sigma) was with the Trans-Blot Turbo Transfer System (Bio-Rad Labs). The membrane was blocked in TBS-T (0.1% Tween 20 in TBS) containing 5% skim milk (w/v) for 1 hr at RT and probed with primary antibody at 4 0 C overnight. The membrane was washed and incubated with horseradish peroxidase (HRP) conjugated secondary antibody (Amersham) for 1 hr at RT and developed using SuperSignal West Dura Extended Duration Substrate (Thermo Fisher Scientific) for detection of HRP and the protein bands were detected by chemiluminescence using a ChemiDoc Imaging System (Bio-Rad Labs). β-Actin or GAPDH were used as loading controls.

Total RNA isolation, cDNA synthesis and qRT-PCR
Total RNA was isolated from EC cell lines using the RNeasy Plus Mini Kit (Qiagen) following the manufacturer's instructions. cDNA was synthesized from 1 µg of total RNA using SuperScript VILO master mix (Thermo Fisher Scientific). The cDNA was amplified using sequence specific primers (see resources table below). Quantitative real-time PCR (qRT-PCR) was performed using RT² SYBR Green ROX qPCR Mastermix (Qiagen) on a QuantStudio Real-Time PCR System (Applied Biosystems) through a pre-incubation step, and 40 amplification cycles. The comparative Ct (2 -ΔΔCt ) method was used to compare the expression level of the target gene with the housekeeping gene (GAPDH) in different groups.

Library preparation, RNA seq and data analysis
Total RNA samples were quantified by Qubit fluorometer and submitted to the McDermott Next Generation Sequencing Core at UT Southwestern for library preparation. 1 μg of DNase treated high quality RNA (RIN score ≥ 8 per Agilent 2100 Bioanalyzer) was prepared with the TruSeq Stranded Total RNA LT Sample Prep Kit (Illumina). Total RNA was depleted of its rRNA and fragmented before strand-specific cDNA synthesis. cDNA was poly-A tailed and indexed adapters were ligated. After adapter ligation, samples were PCR amplified, purified with Ampure XP beads, and validated again on the Agilent 2100 Bioanalyzer. Samples were Qubit quantified before being normalized and run on the Illumina HiSeq 2500 using SBS v3 reagents with a read configuration of 75 bp, single end reads. 25-35 million reads were generated per sample.
Only genes with log 2 CPM >0 and log 2 FC >1 were considered. Different significance levels were used to define differentially expressed genes depending on the scenarios. Pathway analysis was performed using webgestalt (http://www.webgestalt.org/). The threshold for enriched biological processes in the gene ontology was set to FDR <0.05. The dot plots were generated using ggplot2 (https://ggplot2.tidyverse.org/reference/geom_dotplot.html).

Chromatin Immunoprecipitation (ChIP) and data analysis
The chromatin that was bound by acetylated Histone in ISK-EV and ISK-FOXA2 cells was analyzed. Crosslinking was carried out using 1% formaldehyde at RT. Formaldehyde was quenched with 1.25 M glycine. Cells were washed with PBS, lysed, sonicated, and the chromatin was used in a ChIP reaction using Histone H3 acetyl Lys27 antibody and HighCell# ChIP kit protein G (Diagenode) per the manufacturer's recommendation (4). The ChIP DNA was reverse crosslinked, eluted and purified (5), and used to prepare next generation sequencing libraries (KAPA Hyper Prep, KAPA Biosystems, MA). The ChIP sequencing (ChIP-Seq) libraries, after appropriate size selection using AMPure XP Beads (Beckman Coulter Sciences, IN) were sequenced with a read-length of 75bp at the UTSW McDermott Sequencing Core. The quality of the raw data was checked using fastqc (v0.11.8) (6). Raw reads were mapped to hg19 using BWA (v 0.7.17-r1188) (7). Low quality and duplicated reads were removed using Picard (2.20 Genomics Viewer (IGV, v2.8.0) (10) was used to visualize the ChIP-seq peak signals.

IncuCyte cell proliferation and scratch assay
The IncuCyte live-cell imaging and analysis system was used for cell proliferation and wound closure assays. FOXA2 +/-Ishikawa and FP cells (5,000 cells/each well of flat bottom 96well plate) were grown to 90-100% confluency with images acquired every 6 hrs. Cell proliferation curves were determined with IncuCyte software by plotting normalized confluency (ratio of confluency at each 6 hr interval to confluency at time 0).
For wound closure assays, Ishikawa cells were cultured to reach 90-100% confluency and a single homogenous scratch per well of 96-well plate was made with the WoundMaker tool (Essen Bioscience). Cells were washed twice and replaced with complete medium. Images of migrating cells near the wound were taken every 4 hrs and wound width (distance between leading edges of each cell front) was calculated with the IncuCyte software.

Cell adhesion assays
FOXA2 +/-Ishikawa cells (150,000) were seeded onto MicroMatrix ECM array slide (Advanced BioMatrix) and allowed to adhere for 16 hrs. Following cell attachment, the slide was washed in DPBS to remove unbound cells. The cells adhered to micromatrix spots after the wash were captured and counted per the manufacturer's instructions.

Transwell migration and invasion assays
HEC-1-B cells (control and FOXA2 knockdown) were grown till 70-80% confluency and serum starved for 4 hrs. For migration assays, 1 x 10 5 cells in serum-free medium were placed in the upper chamber of 24-well Transwell inserts (8-µm pore size, Fisher Scientific) and complete medium containing 10% FBS was added to the lower chamber. For invasion assays, inserts were coated with Matrigel and DPBS (20 µL Matrigel + 100 µL DPBS), incubated at 37 0 C for 2 hrs and then dried at RT overnight. Cells in serum-free medium were added to the upper chamber of the insert and medium containing 10% FBS was added to the lower chamber. After 16 hrs, migrated or invaded cells on the insert membrane were fixed in 70% alcohol and stained with crystal violet (0.5%) for 10 min. Cells in the inner chamber were removed with cotton swabs. Images of migrated or invaded cells attached to the membrane were quantified with ImageJ (NIH, USA).   (B) Heatmap shows differentially expressed EMT genes in FOXA2 knockdown HEC-1-B cells.