Antibodies and other reagents
The following primary monoclonal (mAb) and polyclonal (pAb) antibodies were used to detect cytoskeletal, membrane, and signaling proteins: anti-anillin (A301-405A and A301-406A) pAbs (Bethyl Laboratories, Montgomery, TX); Keratin 14 mAb and Keratin 17 pAb (Proteintech, Rosemont, IL), anti-E-cadherin, calponin-1, L-caldesmon, and vimentin mAbs (BD Biosciences, San Jose, CA); anti-P-cadherin mAb MAB 861 and E-cadherin pAb (R & D Systems, Minneapolis, MN); anti-cadherin 11 and GAPDH pAbs, anti N-cadherin, cleaved caspase 3, and Ki-67 mAbs (Cell Signaling Technology, Danvers, MA); anti-SM22-pAb (Abcam, Cambridge, MA). PE-conjugated anti-CD4 and APC-conjugated anti-CD24 mAbs were obtained from BioLegend (San Diego, CA). Alexa Fluor-488-conjugated donkey anti-rabbit and Alexa Fluor-555-conjugated donkey anti-mouse secondary antibodies and Alexa Fluor-488 and 555-labeled phalloidin were obtained from Thermo Fisher. Horseradish peroxidase-conjugated goat anti-rabbit and anti-mouse secondary antibodies were acquired from Bio-Rad Laboratories. pFULT-tdTomato-Luciferase virus particles were bought from the DNA/RNA Delivery Core facility of the Northwestern University School of Medicine. All other chemicals were of the highest purity and obtained from Millipore-Sigma or Thermo Fisher.
MDA-MB-231, MDA-MB-468, MCF7, BT474, Hs578T, BT549, and HEK293 cells were acquired from the American Type Culture Collection (Manassas, VA). MCF10A, MCF10AneoT, and MCF10A1d.cl1 cells were purchased from the Barbara Ann Karmanos Cancer Institute, Wayne State University (Detroit, MI). MDA-MB-231, MDA-MB-496, MCF7, BT474, Hs578T, and HEK293 cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum and penicillin/streptomycin. BT549, cells were grown in RPMI, supplemented with 10% fetal bovine serum, insulin, and penicillin/streptomycin. MCF10A cells were grown in DMEM/F12 containing 5% horse serum, human EGF, human insulin, hydrocortisone, penicillin/streptomycin, and cholera toxin. All cells were maintained at 37 °C in an atmosphere containing 5% CO2. The cells were grown in T75 flasks and were seeded on collagen-coated coverslips or six-well plastic plates for immunolabeling and biochemical experiments, respectively.
CRISPR/Cas9-mediated gene editing for anillin knockout and overexpression
A stable knockout of anillin in MDA-MB-231 and BT549 cells was carried out using CRISPR-Cas9 technology. Different anillin-targeting single-guide RNAs (sgRNAs) or non-targeting sgRNAs were selected from two genome-wide sgRNA libraries generated by Zhang and Lander laboratories . The selected sgRNA oligonucleotides (sequences are included in Additional file 1: Table S1A) were annealed and cloned into the BsmBI site of a lentiCRISPR v2 vector (Addgene, 52961) and the obtained constructs were verified by sequencing. Transfer plasmids possessing the annealed guide oligonucleotides were transformed into recombination-deficient One Shot™ Stbl3™ Chemically Competent E. coli (Thermo Fisher), and the amplified plasmids were isolated using a Qiagen midi prep plasmid isolation kit. Lentiviruses were produced by transfecting HEK-293T cells with a transfer lentiCRISPR v2 plasmid and packaging pLTR-G (Addgene, 17532) and pCD/NL-BH*DDD (Addgene, 17531) plasmids. Viral supernatants were collected 48 and 72 h after transfection and used to infect MDA-MB-231 and BT549 cells in the presence of polybrene. After 48 h of the infection, the lentivirus-containing medium was replaced with fresh cell culture medium containing puromycin (5 μg/ml for MDA-MB-231 and 1 μg/ml for BT549 cells) and puromycin-resistant cells were collected after 7-day selection.
For the anillin overexpression in MCF10AneoT cells, a CRISPR/Cas9-based transcriptional gene activation system  obtained from Santa Cruz Biotechnology (sc-403342-LAC) was used. Cells were transfected with either anillin-activating lentiviral particles or control lentiviral particles (sc-108084) in the presence of polybrene, according to the manufacturer’s protocol. After 48 h of viral transduction, stable cell lines were selected by co-treatment with three different antibiotics: blasticidin (10 μg/mL), puromycin (5 μg/mL), and hygromycin B (300 μg/mL).
As an alternative approach, anillin was overexpressed in MCF10AneoT cells using a pTK88-GFP-Anillin retroviral plasmid (Addgene, 46354) with a pWZL-GFP plasmid (Addgene, 12269) as a control. Retrovirus packaging was performed by transfecting those plasmids together with packaging plasmids, ENV and GagPol, into 60% confluent Phoenix cells using a TransIT2020 transfection reagent. Retroviruses were collected at 48 and 72 h post transfection. MCF10AneoT cells plated at 30% confluency were infected by retroviruses with 5 μg/mL polybrene. After 72 h of viral transduction, stable GFP-anillin expressed cell lines were selected by flow cytometry.
siRNA-mediated knockdown of either E-cadherin or P-cadherin in control and anillin-depleted breast cancer cells was carried as previously described [38, 39]. E-cadherin was depleted by using Dharmacon siRNA duplexes (duplex 1, D003877-02; duplex 2, D003877-05), whereas P-cadherin was depleted using specific Dharmacon siRNA SmartPool (L003823-00). A non-targeting siRNA duplex 2 was used as a control. Cells were transfected using DharmaFECT 1 reagent in Opti-MEM I medium (Thermo Fisher) according to the manufacturer’s protocol, with a final siRNA concentration of 50 nM. Cells were used in the experiments on days 3 and 4 post transfection.
Scratch wound assay
Confluent breast cancer cell monolayers were mechanically wounded by making a thin scratch with a 200-μl pipette tip. The bottom of the well was marked in a cell-free area to define the position of the wound. Images at the marked region were acquired at the indicated times after wounding using an inverted bright-field microscope equipped with a camera. The percentage of wound closure was calculated using a TScratch software .
Matrigel invasion assay
A Matrigel invasion assay was performed using BD Biocoat invasion chambers (BD Biosciences). Cells were disassociated from the culture dish using a TrypLE Express reagent (Thermo Fisher), counted, resuspended into a serum-free medium, and added to the upper chamber at a concentration of 5 × 104 cells per chamber. Complete cell culture medium containing 10% FBS as a chemoattractant was added to the lower chamber, and cells were allowed to invade through Matrigel for 24 h at 37 °C. The Matrigel plugs were washed with phosphate-buffered saline (PBS) and fixed with methanol, and non-migrated cells were removed from the top of the gel using cotton swabs. The invaded cells were stained with DAPI, visualized by a fluorescence microscope, and counted by using an ImageJ program (National Institute of Health, Bethesda, MD).
Extracellular matrix adhesion assay
Cell-matrix adhesion assay was performed as previously described . Briefly, control, anillin-depleted, and anillin-overexpressing cells were dissociated by the TrypLE Express reagent, counted with a hemocytometer, and resuspended in the complete medium. 3 × 104 cells were seeded to each well of 24-well plates coated with either collagen I, fibronectin, collagen IV, or laminin and were allowed to adhere for 30 min at 37 °C. After incubation, unattached cells were removed and attached cells were fixed and stained with a DIFF stain kit (IMEB Inc., San Marcos, CA). Images of adherent cells were captured using the bright-field microscope, and the number of adhered cells was determined using the ImageJ software.
Cell proliferation and soft agar colony formation assays
To examine anchorage-dependent cell proliferation, MDA-MB-231 and MCF10AneoT cells were seeded on 60-mm cell culture dishes at the density of 4 × 104 and 1 × 105 cells per dish, respectively. Cells were allowed to proliferate for the indicated times and stained with 0.4% Trypan Blue Solution, and the number of live cells was counted using a hemocytometer.
The anchorage-independent cell growth was examined by using a soft agar colony formation assay according to a published protocol . Briefly, a 2× cell culture medium was mixed with an equal volume of 1.2% noble agar (BD Biosciences) to form a solid 0.6% agar layer at the bottom of a six-well plate. Cells were resuspended in the 2× culture medium and mixed with an equal volume of 0.8% agar to obtain an upper 0.4% agar layer containing 1 × 104 cells per well. These agar plates were cultured at 37 °C with the addition of fresh medium every other day. After 40 days of culturing, these agar plates were washed with PBS and stained with methylene blue (0.06 g/mL) dissolved in 1.5% (v/v) buffered glutaraldehyde. Cell colonies were imaged and counted using ImageJ software.
Mammosphere formation assay
A Mammosphere formation assay was performed using a MammoCult™ Human Medium Kit (STEMCELL Technologies, Cambridge, MA) according to the kit manual. Briefly, cells were grown to 80–90% confluency, rinsed with PBS, and scraped from the dish by using a sterile cell scraper. Harvested cells were transferred into the MammoCult medium and centrifuged, and the cell pellet was resuspended in the MamoCult medium. Cells (1 × 104) were plated in each well of a six-well ultra-low adherent plate (STEMCELL Technologies) and incubated at 37 °C for 7 days. Mammospheres were imaged with a bright-field microscope. ImageJ software was used to quantify the number and the size of spheres larger than 60 μm in diameter.
Flow cytometry analysis
Control, anillin-depleted, and overexpressing cells were lifted from plates using the TrypLE Express reagent and dual-labeled with a PE-conjugated anti-CD44 antibody and an APC-conjugated anti-CD24 antibody. Labeled cells were analyzed in the Lerner Research Institute Flow Cytometry Core using a BD LSRII flow cytometer and FACSDiva (Becton-Dickinson) software. APC fluorescence was excited by the red laser (639 nm) and detected with the 660/20 filter, whereas PE fluorescence was excited by the green laser (532 nm) and detected with the 575/26 filter.
Cells were homogenized in RIPA lysis buffer (20 mM Tris, 50 mM NaCl, 2 mM EDTA, 2 mM EGTA, 1% sodium deoxycholate, 1% Triton X-100 (TX-100), and 0.1% SDS, pH 7.4), containing a protease inhibitor cocktail (1:100, Sigma) and phosphatase inhibitor cocktails 1 and 3 (both at 1:200, Sigma). Lysates were cleared by centrifugation (20 min at 14,000×g), diluted with 2× SDS sample buffer and boiled. SDS-PAGE and immunoblotting were conducted by standard protocols with an equal amount of total protein (10 or 20 μg) per lane. Results shown are representative immunoblots of at least three independent experiments. Protein expression was quantified by densitometry, and signal intensities were calculated using ImageJ software. The densitometry data are presented as normalized values, with the expression level of control groups set at 1 arbitrary unit.
Immunofluorescence labeling and confocal microscopy
MDA-MB-231 cell monolayers plated on collagen-coated coverslips were fixed with 4% paraformaldehyde and permeabilized with 0.5% Triton X-100 at room temperature. Fixed cells were blocked for 60 min in PBS containing 1% bovine serum albumin. Cells were incubated with the appropriate concentrations of primary antibodies in blocking solution for 60 min, washed with blocking buffer, incubated with Alexa dye-conjugated secondary antibodies and Alexa-labeled phalloidin (to detect filamentous (F) actin) for 60 min, washed, and mounted on slides with a ProLong Antifade mounting reagent (Thermo Fisher). Labeled cell monolayers were observed using a Zeiss LSM 700 Laser Scanning Confocal Microscope (Carl Zeiss Microimaging Inc.; Thornwood; NY). The Alexa Fluor 488 and 555 signals were imaged sequentially in frame-interlace mode to eliminate crosstalk between channels. Image analysis was conducted using imaging software ZEN 2011 (Carl Zeiss Microscopy Inc.) and Adobe Photoshop. Images shown are representative of at least three experiments. Multiple images were captured from each slide.
Rho GTPase activation assays
RhoA activity in control and anillin-depleted MDA-MB-231 cells was determined by the enzyme-linked immunosorbent assay using a G-LISA activation assay kit (Cytoskeleton Inc., Denver, CO). The assay was performed according to the manufacturer’s protocol, with the internal active RhoA standard used as a positive control. The sample absorbance at 490 nm was measured using a microplate reader. Rac1 activity was examined using a pull-down assays kit from Cell Biolabs (San Diego, CA) according to the manufacturer’s instructions. The amount of precipitated Rac1, as well as total Rac1 in cell lysates, was determined by immunoblotting analysis.
RNA sequencing analysis
Total RNA was extracted from either control or anillin-deficient (two different CRISPR probes, referred hereafter as CRISPR1 and CRISPR2) MDA-MB-231 cell lines using mirVana miRNA Isolation Kit (Thermo Fisher Scientific, Cat #AM1560). The RNA quality was determined by Bioanalyzer (Agilent, Santa Clara, CA). To isolate the polyA RNA, NEBNext Poly(A) mRNA Magnetic Isolation Module (New England BioLabs, Ipswich, MA) combined with SMARTer Apollo NGS library preparation system (Takara, Mountain View, CA) was used with a total of 1 μg of good-quality total RNA as input. The NEBNext Ultra Directional RNA Library Prep Kit (New England BioLabs) was then used for polyA RNA library preparation, which is a dUTP-based stranded library. The library was indexed and amplified under the PCR cycle number of 11. After library Bioanalyzer QC analysis and quantification, individually indexed and compatible libraries were proportionally pooled and sequenced using an Illumina HiSeq 1000 sequencer. Under the sequencing setting of single read 1 × 51 bp, about 25 million pass filter reads per sample were generated.
Two biological replicates for each sample were sequenced. One set of replicates had an insufficient number of reads and was additionally sequenced. These additional sequences were concatenated with the under-sequenced samples.
Quality control at each processing step was performed using the FastQC tool v.0.11.2 (quality base calls, CG content distribution, duplicate levels, complexity level) https://www.bioinformatics.babraham.ac.uk/projects/fastqc/. Single-end reads were adapter-trimmed using Trimmomatic v.0.33  and aligned using the subread v.1.6.2 aligner and the latest assembly of the human genome (GRCh38/hg38). Gene counts were obtained using the feature-Count v1.6.2 software and analyzed for differential expression using the edgeR v3.24.3 R package. The sequences from two replicates were merged together. Variability (batch effect) due to the “merged” status of one set of replicates was accounted for in the detection of differentially expressed genes. P values were corrected for multiple testing using a false discovery rate (FDR) multiple testing correction method [https://www.jstor.org/stable/2346101]. Genes at FDR < 0.1 were selected as differentially expressed. Gene set enrichment analysis  was performed on a list of 275 genes differentially expressed in both CRISPR experiments using the clusterProfiler v3.10.1 R package. All analyses were performed in the R/Bioconductor v3.5.3 statistical environment.
Total RNA was isolated using the RNeasy mini kit (QIAGEN, Valencia CA), followed by DNase treatment to remove genomic DNA. Total RNA (1 μg) was reverse transcribed using the iScript cDNA synthesis kit (Bio-Rad Laboratories). Quantitative real-time PCR was performed using iTaq Universal SYBR Green Supermix (Bio-Rad Laboratories). Gene amplification was performed using a CFX96 Real-Time PCR System (Bio-Rad Laboratories) with the following reaction conditions: 1 cycle, 95 °C for 120 s; and 40 cycles, 95 °C for 1 s, and 60 °C for 20 s. A threshold cycle number for the gene of interest was calculated based on the amplification curve representing a plot of the fluorescence signal intensity versus the cycle number. The delta threshold cycle number was calculated as the difference between the threshold cycle number of the genes of interest between control and anillin knockout cells. Each value was normalized by the difference in the threshold cycle number for the housekeeping gene (GAPDH) amplification in the same samples. Primer sequences are included in Additional file 1: Table S1B.
In vivo studies of tumor growth and metastasis
NOD-SCID-IL2 gamma-receptor null (NSG) mice were purchased from Jackson Laboratory (#005557), bred at Virginia Commonwealth University (VCU) Cancer Mouse Models Core Laboratory (generation F2), and housed under standard pathogen-free conditions with food and water available ad libitum. To monitor tumor growth and metastasis, control and anillin-depleted MDA-MB-231 cells were stably transduced with a pFULT-tdTomato-Luciferase expressing lentivirus. The cells were sorted by flow cytometry to obtain populations with similar tdTomato expression before injection. Similar level of luciferase was also determined before injection by a serial dilution of cells in 96-well plates, addition of luciferin, imaging with In Vivo Imaging System (IVIS) Spectrum, and analysis with Living Image software (PerkinElmer). To assess the effects of anillin on tumor growth and spontaneous metastases from the primary site, MDA-MB-231 cells (2.5 × 105 cells/injection) were injected into the left and right sides of the lactiferous duct of the fourth mammary gland of the NSG mice (n = 12 animals per each experimental group) with the assistance of the VCU Cancer Mouse Models Core Laboratory. Calipers were used to measure tumor size, starting 1 week post-injection and continuing twice-weekly for the next 6 weeks. On day 42 after cell implantation, mice were subcutaneously injected with luciferin (150 mg/kg). Tumors, lungs, livers, kidneys, ovaries, and lymph nodes were dissected immediately after luciferin injection, and their luminescence intensity was measured using the IVIS Spectrum (PerkinElmer, Waltham, MA). Additionally, the weight of all dissected tumors was measured.
For the intracardiac injection, 2 × 105 MDA-MB-231 cells were injected into the left ventricle of the NSG mouse heart (n = 16 mice per each experimental group). Bioluminescence images of the dorsal and ventral parts of the mouse body were taken weekly. Twenty-eight days after injection, mice were sacrificed, their lungs, livers, kidneys, ovaries, bones, and brains were dissected separately, and bioluminescence images were taken for each tissue. Image analysis and quantification were performed by a Living Image Software-IVIS Spectrum Series.
Immunohistochemical (IHC) staining for cleaved caspase-3, E-Cadherin, and Ki-67 was performed in the Cancer Mouse Models Core Laboratory (CMMCL) with the Leica Bond RX autostainer using heat-induced epitope retrieval buffer 2 (Leica, EDTA pH 8.0). Stained slides were then imaged on the Vectra Polaris (Akova Biosciences) and quantified using InForm software (Akova Biosciences).
All numerical values from individual in vitro experiments were pooled and expressed as mean ± standard error of the mean (S.E) throughout. Obtained numbers were compared by two-tailed Student’s t test, with statistical significance assumed at P < 0.05. The numerical values from in vivo experiments were expressed as mean ± standard error of the mean (S.E) or scatter dot plot with a mean line. One-way ANOVA and Dunnett’s multiple comparisons test were performed for statistical significance assumed at P < 0.05.