MDA-MB-231 and MDA-MB-468 human breast carcinoma cell lines were acquired from ATCC and cultured in RPMI 1640 medium. MVT1 mouse mammary tumor cells have been described previously  and were grown in DMEM medium. Media contained 10% fetal bovine serum (FBS), 2% pen strep, 1% L-glutamine. Cells were kept at 37˚C with 5% CO2.
Small interfering RNA (siRNA)-mediated knockdown of RALA and RALB was achieved through the transfection of MDA-MB-231 or MDA-MB-468 cells with 50 pMol of siRNA targeting human RALA (GCAGACAGCUAUCGGAAGA; Dharmacon, Lafayette, CO, USA), RALB (GAAAGAUGUUGCUUACUAU, Dharmacon) or simultaneously targeting both isoforms (GAGCUAAUGUUGACAAGGU; Dharmacon) or non-targeting control siRNA pool (D-001810-10-05, Dharmacon) using Lipofectamine RNAiMAX (Invitrogen, Carlsbad, CA, USA) for 72 hours.
Human RALA (TL309957V) and RALB (TL309956V) targeting shRNA lentiviral particles and a non-targeting control (TR30021V) were purchased from OriGene (Rockville, MD, USA). MDA-MB-231 cells were transduced with lentiviral particles and selected using 5 µg/mL puromycin for > 7 days. Similarly, mouse RALA targeting shRNA lentiviral particles and a non-targeting control (sc-41843; Santa Cruz Biotechnology, Inc., Dallas, TX, USA) were used to achieve knockdown of RalA in MVT1 cells. Again, cells underwent selection in puromycin-containing medium for > 7 days.
CRISPR/Cas9 Gene Editing
For CRISPR-mediated knockout of RALA or RALB in MDA-MB-231 cells, paired gRNAs were designed using ATUM CRISPR gRNA design tool (atum.bio; Newark, CA, USA). Sequences for used to target RALA are 5’-aaa gtc atc atg gtg ggc ag-3’ and 5’-gcc aaa gaa ttc tga ccc tt-3’ and for RALB the sequences: 5’-atg gtt ggc agc gga ggc gt-3’ and 5’-tgt gga ggg cca agg agc tc-3’ were used. Following gRNA design, a customized all-in-one vector (pSpCas9n-BB-2A-Puro (PX459) v2.0) was generated by GenScript (Piscataway, NJ, USA). Plasmids were transfected into MDA-MB-231 cells using FuGENE HD Transfection Reagent (Promega, Madison, WI, USA). Cells were selected in puromycin-containing media for > 7 days. Single-cell colonies were chosen following additional sub-culturing in puromycin-containing media and were evaluated for RalA knockout by western blot analysis. Three to five individual clones with confirmed knockout were combined as a pool for further study.
See Sizemore et al.  for western blotting and immunodetection protocol. Antibodies were obtained from Cell Signaling Technologies (RalA: 4799, RalB: 90879, β-actin: A1978, anti-mouse-IgG HRP: 7076, anti-rabbit IgG HRP: 7074) and Millipore Sigma (GAPDH: G8795 or MAB374).
Breast cancer cell line gene expression data was obtained from the Broad Institute Cancer Cell Line Encyclopedia (CCLE, ). METABRIC  patient mRNA data was obtained through Oncomine. TCGA patient mRNA and protein data either came from Oncomine (www.oncomine.org) or cBioPortal (Firehose Legacy, cbioportal.org, [27, 28]). ROC Plotter (rocplot.org, ) was used to determine gene expression correlation with treatment response.
Patient Tissues Samples and PDX Model
A commercially available tissue microarray (TMA) was obtained from US Biomax, Inc. (HBre-Duc150Sur-01) and consisted of 150 individual BC patient samples. TNBC Tissue microarrays (TMA) were constructed from archived breast tumor tissue isolated from 60 patients treated at the OSU James Comprehensive Cancer Center upon informed consent following approval from The OSU Institutional Review Board (IRB). The TMAs are maintained by the Columbus Breast Cancer Tissue Bank and were described previously . Use of the TMAs herein was approved under IRB protocol #2016C0025. The patient-derived xenograft model (PDX-TM00096) was derived from a TNBC lung metastasis and purchased from Jackson Laboratories and maintained by the Target Validation Shared Resource (OSUCCC).
Animal Treatments and Procedures
Animal use was in compliance with University Laboratory Animal Resources (ULAR) regulations under the OSU Institutional Animal Care and Use Committee (IACUC)–approved protocol 2007A0120-R4. Adult, female FVB/N mice were purchased from Jackson Laboratories (Bar Harbor, Maine, USA), and nod scid γ (NSG) mice were acquired through the Target Validation Shared Resource (OSUCCC).
For orthotopic fat pad injections, 5 x 105 MVT1 cells, 2.5 x 106 MDA-MB-231 cells were injected into 6–8 week old female FVB/N, or 6–8 week old female NSG mice, respectively. For orthotopic tumor experiments utilizing MDA-MB-231 CRISRP lines, 3 x 106 cells were injected into 11 week old NSG mice. Once palpable tumors were detected, two-dimensional caliper measurements of tumor size were made three times per week. The following formula was used to calculate tumor volume: volume = ½ (length x width2). Tumor tissue and lung tissue were collected for histological evaluation at a predetermined experimental endpoint (additional details for each cohort are provided in the text and/or figure legend). Early removal criteria (ERC) was defined as tumor diameter in excess of 1.6cm, ulceration > 2mm, or body condition score < 2.
Luciferase-tagged MDA-MB-231 cells were generated by transduction with EF1a-luciferase lentiviral particles (GenTarget, Inc., San Diego, CA, USA; #LVP435). Cells were selected for > 10 days in Neomycin-containing media (1mg/ml) prior to use in any experiments. 5 x 106 cells were injected intravenously by tail-vein into 6–8 week old female NSG mice. Bioluminescence imaging was performed according to the protocol described by Yang et al. . Mice were imaged at day 0 (i.e. within 2 hours post-injection) and 2x weekly (M and Th) for long term tumor studies and at 24, 72 and 96 hours post-injection for short term analysis. All images were acquired on the IVIS Lumina II optical imaging system (PerkinElmer Inc.) available within the OSUCCC Small Animal Imaging Core.
BQU57 (50µM) in DMSO was administered 3x weekly (M-W-F) until early removal criteria were met. DMSO was given as a vehicle control. For both control and treatment arms, body condition score was noted, and individual mouse weights were measured 3x per week as a measure of drug-related toxicity. Treatment of MDA-MB-231 tumor-bearing mice commenced 21 days following orthotopic injection of cells. For the TNBC PDX model, BQU57 treatment began 24 days after implantation when the PDX tumors were on average 100 mm3.
Quantification of Lung Metastatic Tumor Burden
Hematoxylin and eosin (H&E)-stained lung sections were either scored for metastatic nodules, foci, and emboli visually by a veterinary pathologist (SEK) or sections were scanned on a high-resolution, Leica Aperio ScanScope XT for image analysis. The entire lung was sectioned and serial sections were taken 100µm apart. Either three sections (Fig. 3B and Supplemental Figure S1B), two sections (Fig. 6E), or one section (Fig. 1A and Supplemental Figure S1C) per mouse were used for analyses.
For image analyses of metastatic burden, whole-slide images were acquired with an Aperio XT slide scanner (Leica Biosystems, Inc., Buffalo Grove, IL, USA) using the 40x objective at The Ohio State University Comparative Pathology & Mouse Phenotyping Shared Resource. Images were imported into Visiopharm Image Analysis software (Visiopharm, Hørsholm, Denmark version 2017.27.0.3313). Images were segmented into areas of tumor metastases and normal lung tissue using different color labels for each tissue type. Tumor metastases (blue), normal tissue (green), and bronchiolar epithelium (yellow), area of red blood cells (red), and air spaces (pink) labels. A custom algorithm/App was written using a Visiopharm Decision Forest classification algorithm set at an accuracy of 50 (range 0-100). To account for size variance and metastases too small to detect, metastases measuring 8500µm2 and above were labeled and measured as metastases. Small misshaped areas and small metastatic areas under 8500µm2 were included in the normal tissue quantification. Tissue types were established, and mark-ups reviewed in consultation with a veterinary pathologist board-certified by the American College of Veterinary Pathologists (SEK) to ensure accurate measurements and differentiate between tissue types. These data were used to calculate average number of metastasis per section (total lung area), the number metastatic lesions per section (total lung area), and the percentage of total lung area occupied by metastasis as indicated.
Immunostaining and Quantification
For immunofluorescence of MDA-MB-231 CRISPR variants, 25,000 cells were seeded on an 8 well chamber slide (Corning, 354118). Cells were fixed in 4% Paraformaldehyde, permeabilized in 0.5% Triton X-100 and then blocked in a 5% BSA-0.3% Triton x-100 solution. RalA (Abcam, 1:200, ab126627) or RalB (Abcam, 1:200, ab223479) primary antibody were diluted in blocking buffer and applied overnight at 4°C. Alexafluor 594 anti-rabbit secondary antibody (Invitrogen, 1:250, A21207) was added for one hour at room temperature before removing chamber slides and coverslipping with Prolong-Gold Antifade Reagent with DAPI (Invitrogen, P36931). Images were taken on a Zeiss LSM 800 confocal microscope using the Zen software (Carl Zeiss Microscopy, LLC, White Plains, NY, USA).
Immunohistochemistry (IHC) on BC patient TMAs was done using the Bond RX autostainer (Leica Biosystems, Inc.). Briefly, slides were baked at 65°C for 15 minutes and the automated system performed dewaxing, rehydration, antigen retrieval, blocking, primary antibody incubation with α-RalA (1:2000, #ab126627, Abcam), post primary antibody incubation, detection (DAB), and counterstaining. Samples were then removed from the machine, dehydrated and mounted.
Ki67 and Cleaved Caspase-3 [α-Ki67 (1:100; ab16667, Abcam); α-Cleaved Caspase-3 (Asp175); 9661, Cell Signaling Technologies] immunostaining was done as previously described . Sections were deparaffinized in xylenes, rehydrated, and antigen retrieval was performed using an EDTA Decloaker (Biocare Medical, LLC, Pacheco, CA, USA) for 40 minutes in a steamer (90°C). Hydrogen peroxide (3%) was used to quench endogenous peroxidase before tissues were blocked with 5% BSA + 0.5% Tween-20 in PBS. Sections were incubated overnight in primary antibody. The following day a biotinylated secondary antibody was added, and a Vectastain® ABC (HRP) kit along with DAB substrate (Vector Laboratories, Inc., Burlingame, CA, USA) were used to develop signal. Tissues were then dehydrated, counterstained with hematoxylin, and coverslipped.
IHC images were taken on the PerkinElmer’s Vectra® Automatic Quantitative Pathology Imaging System. The acquisition workflow has been described . Quantification and scoring of RalA immunostaining on BC patient TMAs was done using inForm® Advanced Image Analysis software (PerkinElmer). InForm® software was used to spectrally un-mix images and the DAB signal was scored based on a user-defined threshold into four categories (0+, 1+, 2 + and 3+). The percent of cells within each scoring category was determined based on cell segmentation with the hematoxylin counterstain. An H-Score was calculated by the following formula: [1x(%cells 1+) + 2x(%cells 2+) + 3x(%cells 3+)]. Quantification of Ki67 and Cleaved Caspase-3 (CC3) (all but CRISPR experiments) was performed similarly using inForm® software (PerkinElmer). Quantification of CC3 for MDA-MB-231 CRISPR Ctrl and KO cell lines was performed on ImageJ Fiji Software by a previously described method [33, 34]. Briefly, images were taken on an EVOS M7000 Imaging System (ThermoFisher Scientific, Waltham, MA, USA). Color deconvolution and threshold ranges for DAB and hematoxylin staining were set to the same range (user defined) for each image. DAB signal and cell count was then calculated. DAB intensity was normalized by the number of nuclei in each image.
Invasion, Migration and Wound Healing Assays
Cells were serum-starved overnight prior to invasion and migration assays. Invasion assays were performed using 24-well Corning Matrigel Matrix chambers with 8µm pore size (Fisher Scientific). Costar 24-well, 8µm polycarbonate pore membrane, 6.5mm inserts were utilized for migration assays (Fisher Scientific). Cells were trypsinized and 50,000 viable cells were seeded to the upper chamber. The bottom chambers held serum-containing medium to serve as a chemoattractant. Migration was allowed to proceed for 6 at 37˚C with 5% CO2, while invasion was assessed after 24 hrs. After the designated incubation period, all non-migrated or non-invaded cells were removed using a cotton swab. Cells were then fixed in methyl alcohol with 1.8 mg/L of Triarylmethane Dye and stained with 1.25 g/L Thiazine Dye Mixture (Dade Behring, Deerfield, IL, USA). Invaded or migrated cells were quantified using ImageJ (National Institute of Health, Bethesda, MD, USA) .
Confluent plates of cells were serum-starved overnight before wound healing assay. Plates were scratched with a p200 pipette tip then washed with PBS and replenished with serum-free media. Cell migration was imaged at 0 and 6 hours at 4X magnification on an EVOS XL Core Imager. Data was quantified by ImageJ software.
GILA and Soft Agar Assays
Growth in low attachment (GILA) was measured using the CellTiter-Glo luminescent assay (Promega, G7570). 2,000 MDA-MB-231 and MDA-MB-468, or 10,000 MVT1 cell variants were plated in Ultra Low Attachment polystyrene 96-well plates (7007; Corning, Corning, NY, USA) for 5 days. 100 µL CellTiter-Glo was added and the cells were shaken for 5 minutes. Following a 25 minute incubation the cells were moved to a white polystyrene 96 well plate (3610; Corning) to read luminescence signal using a Promega Glomax Discover microplate reader. For DMSO and BQU57 treatment studies, drug was added to growth media at day = 0.
For soft agar assays, 0.6% agar was layered in a 96-well plate. Cells were suspended in 0.4% agar at 10,000 cells/well. Each well was topped with 100µL of media. For DMSO or BQU57 treatment, media contained drugs at a 50µM final concentration. After incubation for 7 days at 37˚C with 5% CO2, anchorage-independent growth was determined using CytoSelect 96-well Cell Transformation Assay, Soft Agar Colony Formation Kit (Cell Biolabs, Inc., San Diego, CA, USA).
Spheroids were generated using a protocol adapted from Andersen, et. al., 2016 and Trevigen’s protocol outlined in their Spheroid Formation Kit (Trevigen; Cat#: 3511-096-K). In brief, 10,000 MDA-231 cells in 50uL of DMEM with Trevigen Spheroid Formation Matrix (Trevigen Cat#: 3500-096-SP) were seeded in 96-well, Ultra-low attachment (ULA) round bottom plates (Corning Cat# CLS7007-24EA). ULA-round bottom plates were centrifuged at 750 g’s for 10 mins at 4*C, and incubated at 37*C at 5% CO2 until ready to use. Spheroids used for invasion assays were used between 7 and 12 days old.
For spheroid invasion assays, p10 tips were cut with a sterile #10 blade under tissue culture conditions. 10uL of media with a single spheroid were transferred into a new 96-well, ULA round bottom plate. 50uL of thawed Cultrex Basement Membrane Extract (BME) (Trevigen; Cat# 3432-001-01) was carefully added with ice-cold tips to each well. Plates were centrifuged at 300 g’s for 5 mins at 4*C to center the spheroids. An additional 300 g’s for 5 mins centrifugation step was preformed if the spheroids were not positioned correctly. Plates were incubated at 37*C at 5% CO2 for 1 hours prior to the careful, dropwise addition of 100uL of pre-warmed DMEM as to not dislodge the spheroids. Spheroids were cultured for five days and then imaged using the EVOS Core XL microscope at the 4X objective daily and were analyzed using the ImageJ/Fiji software (Schindelin et. al., 2012).
Viability, Proliferation, and Apoptosis Assays
For determining cell viability, 50,000 MDA-MB-231 cell variants were plated into a 6-well plate. At 72 hours, cells were trypsinized, harvested, stained with trypan blue, and counted using a hemocytometer. Cells were deemed viable upon trypan blue exclusion. The Cell Proliferation Kit I MTT Assay (Millipore Sigma, St. Louis, MO, USA) was also used to evaluate tumor cell viability over time. 2000 MDA-MB-231, MDA-MB-468, or MVT1 variants were plated and measurements taken at 72 hours using a Promega Glomax Discover microplate reader. For DMSO, BQU57, and Taxol treatments, the media contained drugs at the indicated concentration and measurements were taken after 72 hours.
The BrdU Cell Proliferation Assay (Cell Signaling, Danvers, MA, USA) was used to evaluate tumor cell proliferation. 10,000 MDA-MB-231 or 5000 MVT1 cell variants were plated and the assay was performed following manufacturer’s instructions. At 72 hours, absorbance was read at 450 nm using a Promega Glomax Discover microplate reader.
The Dead Cell Apoptosis Kit with Annexin V Alexa Fluor™ 488 (Invitrogen, Carlsbad, CA, USA) was used to evaluate apoptotic populations. MDA-MB-231 or MVT1 cell variants were plated and allowed to reach ~ 90% confluency. Floating and harvested cells were collected and treated according to manufacturer's instructions. Cells were acquired using a LSR II.
Statistical analyses were conducted with GraphPad Prism 7. Sample size was not predetermined statistically. Mice were randomly allocated for treatment studies. Investigators were not blinded to group allocation but were blinded to downstream analyses. For all data, normality was checked by Kolmogorov-Smirnov normality testing. Comparison between two groups of normally distributed data was done by homoscedastic or heteroscedastic unpaired two-tailed Student’s t-test as appropriate. For data not normally distributed, statistical comparisons were done by the Mann-Whitney U test. Statistical significance of Kaplan-Meier survival curves was determined using log-rank. Statistical significance was established at P ≤ 0.05.