Patient sample acquisition. Patients with OS treated surgically at our institution between 2002 and 2016 were identified and FFPE archival samples from primary and metastatic tumors were obtained. Primary tumor samples included biopsies (B), primary tumor resections (R) and recurrent tumor resections (C). Chart reviews were performed to obtain the pertinent clinical characteristics of each patient. We identified samples from 28 unique patients. Four patients had samples available from both primary and LM, 19 patients had samples exclusively from the primary site, and five patients had samples exclusively from LMs. Among the 38 total samples, there were 14 samples from primary tumor sites in patients that never developed LM (L), 14 samples from primary tumor sites in patients that did develop LM (M), and 10 LM samples (LM). Sixteen of the 28 patients (57%) were male. Of the 12 female patients in this study, only 5 patients (< 42%) developed metastatic tumors (Table 1). Patient histories were obtained up to current follow-up or their time of death. Fourteen patients are alive at the time of this publication. Survival for the other 14 patients ranged from two months to seven years from the time of primary diagnosis.
FFPE sample processing, RNA extraction, and quality control. A board-certified pathologist reviewed hematoxylin and eosin slides from each sample and marked areas with high tumor cellularity and less than 40% necrosis. Subsequently, five to seven 10 uM sections from the paraffin blocks were cut, and the areas of interest were macrodissected with a sterile, disposable No. 15 scalpel (Fisher Scientific, catalog 50-822-460). RNA extraction was then performed with the AllPrep DNA/RNA FFPE kit (Qiagen, catalog 80234) according to the manufacturer’s instructions under sterile RNase/DNase-free conditions. RNA concentration was determined with the Qubit 4.0 Fluorometer (ThermoFisher Scientific, Q33227) using the Qubit RNA BR Assay Kit (ThermoFisher Scientific, Q10211). Quality RNA integrity number scores, and fragment sizes (DV200 metrics) were obtained utilizing the Agilent 4200 TapeStation system (Agilent Technologies, Inc).
ecRNA-seq, expression quantification, and normalization. For patient samples which passed quality control tests, we performed exome capture RNA-sequencing (ecRNA-seq) as previously described with minimal changes27. Briefly, sequencing library preparation was performed using a minimum of 40ng RNA according to Illumina’s TruSeq RNA Access Library Preparation protocol. Indexed, pooled libraries were then sequenced on the Illumina NextSeq 500 platform with high-output flow cell-producing stranded, paired-end reads (2 × 75 bp, paired end). A target count of 40 million reads per sample was used to plan indexing and sequencing runs. For expression quantification and normalization, the RNA transcripts from paired-end FASTQ files were mapped and quantified using k-mer–based quasi-mapping with seqBias and gcBias corrections (Salmon v1.1.0, 31-kmer index built from GRCh38 Ensembl v99 transcript annotations)28. Transcript-level abundance estimates were collapsed to gene-level estimates using tximport29. Gene-level counts or log2normCPM values were implemented for subsequent analyses30,31.
Differentially expressed genes analysis (unsupervised hierarchical clustering). Differentially expressed genes (DEGs) between patient lung metastases samples and primary tumor patient samples were calculated using a paired or unpaired DESeq2 v1.24.0 differential gene expression analysis controlling for method of collection (biopsy vs. resection); design = ~ method of collection + condition (primary/met). This enabled us to test for the effect of primary in contrast to LM while controlling for the effect of the different collection procedures, which included biopsy or resection. DEGs were defined as genes with an absolute value fold change > = 1.5; FDR < = 0.01; and minimum TPM of 1 in at least 10% of the samples. Gene ontology analysis was performed to determine biological processes represented in the upregulated and downregulated gene sets. We performed hierarchical clustering utilizing the heatmap.3 function (https://raw.githubusercontent.com/obigriffith/biostar-tutorials/master/Heatmaps/heatmap.3.R). AR expression across cell lines was examined using Cancer Cell Line Encyclopedia (CCLE) data. Gene expression levels from processed RNA-seq of 1379 cell lines were obtained from CCLE along with the sample information containing primary disease types and subtypes for all samples (version DepMap Public 21Q2). Prior to analysis, all bone cancer cell lines were sorted into their respective subtypes: Osteosarcoma, Chondrosarcoma, and Ewing Sarcoma. Median AR expression values and quartile ranges for primary diseases and bone subtypes were calculated in base R and plotted using ggplot2 v.3.3.5. Pathway analyses were performed using Ingenuity Pathway Analysis (IPA) (Qiagen, v68752261). Cut-offs for Exp. Log ratio= -0.58 and 0.58 and FDR = 0.01.
Cell lines and culture conditions. Human primary OS SaOS-2 (Cat. # HTB-85) and dedifferentiated chondrosarcoma HT-1080 (Cat. # CCL-121) cell lines were purchased from American Type Culture Collection (Manassas, Virginia). While HT-1080 was characterized as a fibrosarcoma of bone, the cell line has since been reported to possess an IDH1 mutation which suggests that a diagnosis of dedifferentiated chondrosarcoma is more appropriate26. The metastatic LM2 OS cell line was gifted from Dr. Eugenie Kleinerman (MD Anderson Cancer Center). Cell lines were authenticated at the MD Anderson Cancer Center Cytogenetics and Cell Authentication Core. The AR-positive castrate-resistant, C4-2, and AR-negative, PC3, prostate cell lines were generously provided by Dr. Zhou Wang (UPMC Department of Urology). Cells were determined to be mycoplasma-free after testing with MycoAlert Plus mycoplasma kit (Fisher Scientific, NC0529908). SaOS-2 and LM2 OS cell lines were cultured in complete media (Dulbeco’s Modified Eagle Medium (DMEM, Cat. # 11-995-065) + 10% FBS + 1% MEM Non-Essential Amino Acids Solution (NEAA, Cat. # 11-140-050) + 1% MEM Vitamin Solution (Cat. # 11120052) + 1% Penicillin-Streptomycin Solution (Corning, Cat. # MT30001CI). HT-1080 cells were cultured in DMEM supplemented with 10% FBS and 1% Penicillin-Streptomycin. C4-2 and PC3 were cultured in RPMI-1640 medium (Catalog No. A1049101) supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin. All cells were maintained in a humidified atmosphere
Detecting AR with western blotting OS cell lines. Protein lysates of PC3, C4-2, SaOS-2, and LM2 cell lines were extracted with RIPA Lysis Buffer (Sigma-Aldrich, R0278) plus Halt Protease and Phosphatase Inhibitor (Thermo-Fisher, 78442). Protein concentrations were measured using the Pierce BCA Protein Assay Kit (Thermo-Fisher, 23225). Equal amounts of protein (20 µg) were loaded and run on a 4–20% SDS polyacrylamide gel (Bio-Rad, 4561094). Electrophoresis was performed at 115V for 70 minutes. Mini Trans-Blot Cell System (Bio-Rad,1703930) was used to transfer protein (100V for 60 minutes) from gel to nitrocellulose membrane (Bio-Rad, 1620112). The membrane was blocked with 5% milk at room temperature for 1 hour, then incubated with primary antibodies at 4°C overnight. Antibodies for AR (1:1000 dilution) and GAPDH (1:5000 dilution) were purchased from Abcam (ab133273) and Cell Signaling (5174S), respectively. Next, the membranes were incubated with secondary antibody (1:3000) (Bio-Rad, 1706515) in 2.5% milk for 1 hour at room temperature. Blots were washed 3 times for 6 minutes each after primary and secondary antibody incubations. Western blots were incubated with ECL (Thermo Scientific, 32106) and imaged for chemiluminescence with the Kodak X-Omat 2000 Processor on BioBlot BXR film (Laboratory Product Sales Inc., BX57).
Cell line RNA extraction and qRT-PCR. RNA was collected from SaOS-2, LM2, PC3, and C4-2 cell lines (1.25X106 cells) according to the manufacturer protocol (Qiagen RNeasy Mini Kit, 74106). The highly metastatic PC3 and C4-2 prostate cell lines were used as negative and positive AR controls, respectively, based on their androgen-sensitivity. In this study, 1 µg of total RNA per sample was used to synthesize the first strand cDNA using iScript reagent (Biorad) in a total volume of 20 µL. Amplification of triplicate cDNA template samples for the target genes were performed with denaturation for 15 minutes at 95°C, followed by 45 cycles of denaturation at 94°C for 15 seconds, annealing at 55°C for 30 seconds, and extension at 72°C for 30 seconds using the BioRad CFX 384 machine. Primers were designed for AR (F: AATCCCACATCCTGCTCAAG, R: AAGTCCACGCTCACCATG) and ALDH1A1 (F: AGCAGGAGTGTTTACCAAAGA, R: CCCAGTTCTCTTCCATTTCCAG). Cycle threshold (Ct) values were normalized to ANKRD28 (F: TTGGAGTGCCTAAACCTTCTG, R: AGGTCATTCACACTTGCTCC), GAPDH (F: ACATCGCTCAGACACCATG, R: TGTAGTTGAGGTCAATGAAGGG), and SYMPK (F: CTTCACCAAGGTTGTGCTGGAG, R: GCGCTTGAAGATCAGGTCTCGA). The changes in fluorescence of SYBR green dye in every cycle were monitored and calculated by the BioRad CFX384 system software and the Ct for each reaction. The relative amount of PCR products generated from each primer set was determined based on the threshold cycle or Ct value. PCR analysis was performed on each cDNA in triplicate. All primers were supplied by Integrated DNA Technologies.
Testing the effects of enzalutamide on cell viability. C4-2, SaOS-2, and LM2 cells lines were harvested after reaching 80% confluence and 105 cells were plated in 6 well plates for 24 hours. C4-2 was used as a positive control for the expression of AR. Cells were then treated with enzalutamide (Fisher Scientific, 50-701-7) at fold dilutions from 100 to 3 µM. In addition to enzalutamide treatment, untreated and methanol-treated (positive control) conditions were also tested. Each condition was performed in triplicate. Cells were incubated in their respective conditions for 72 hours, after which they were harvested and counted for viability with Trypan blue (Bio-Rad, 1450022).
Determining ALDH activity in OS cell lines using fluorescence-activated cell sorting (FACS) analysis. SaOS-2 and LM2 cell lines (106) were cultured as described above and plated in 60 mm dishes (Fisher Scientific, 08-772B) for 24 hours before any drug intervention to ensure adhesion of cells. Plated cells were then treated with enzalutamide, at IC50 concentrations (SaOS-2: 109.7 uM and LM2: 98.46 uM) for 48 and 72 hours. At the end of the treatment periods, cells were harvested with Tryple Express (Fisher Scientific, 12604021). The ALDEFLUOR™ kit (Stemcell Technologies, Inc., 01700) was used to analyze the cells with ALDH enzymatic activity according to the manufacturer's protocol as described previously32. Briefly, cells were incubated in the ALDEFLUOR™ assay buffer containing the ALDH substrate BAAA at 37°C for 45 min. 7- Aminoactinomycin D (7-AAD) dye (Thermo Scientific, A1310) was added to stain dead cells. Stained cells were analyzed using BD LSR Fortessa (BD Biosciences) with Flowjo software (Version 10.5.2+, Flowjo LLC).
Testing the effects of enzalutamide on cell migration. The two OS cell lines with varying metastatic potentials, SaOS-2 and LM2, and HT-1080, a highly metastatic chondrosarcoma cell line, were plated in 60 mm dishes at 106 cells and serum starved for 24 hours prior to their harvest for cell migration assays. Treated cells had their respective IC50 concentrations of enzalutamide (Selleck Chemical LLC, Cat. # S1250) added to the serum-free media during this 24-hour incubation. After 24 hours of serum-starvation, with or without enzalutamide treatment, cells were plated on 6.5 mm transwell permeable supports at a density of 5 x 104 cells, in 24-well plates (Corning, Cat. # 07-200-174). The supports (upper chamber) held a final total volume of 150 µl serum-free, cell suspension. The (lower chamber) wells of the 24-well plate held 800 µl of the cell’s respective media either with or without FBS. Wells with serum-free media in the lower chambers served as discrete negative controls for each group of the study. Conditions were run in triplicate. Cells were incubated at 37oC (5% CO2) for 18 hours, at which point the inserts were stained with crystal violet to terminate the migration assay.
Quantifying migrated cells with crystal violet staining. Following the migration assay of SaOS-2, LM2, and HT-1080, transwell supports were washed twice with Dulbeco’s phosphate-buffered saline (DPBS, Gibco, Cat. # 14190144) and cells that did not migrate were removed from the upper chamber using a moistened cotton swab. Migrated cells that adhered to the lower surface of the support were stained with crystal violet (Fisher Scientific, Cat. # C581-25) for 10 minutes. Transwell supports were washed with DPBS three times and air-dried. Once dry, bright field images were taken using Olympus cellSens Dimension software (Version 2.1) on an Olympus SZX16-ILLT microscope (Olympus, Tokyo, Japan) with 5x objective lens and 2.9x zoom. The crystal violet bound to migrated cells was eluted from the supports by pipetting 400 µl of 33% acetic acid (Fisher Scientific, Cat. # A38-500) into each upper chamber and shaking the plates for 10 minutes. Half of the eluent for each sample, 200 µl, was transferred to a 96-well clear microplate (Corning, Cat. # 3595) and the absorbance at 590 nm was determined using the Tecan Infinite M200 plate reader (Tecan Group Ltd., Switzerland) and Tecan i-control software (Version 3.91.0). Standard curves were generated for each cell line used in the migration assay and were used to calculate the cell concentration from the experimental absorbance measurements.
Statistics. The statistical significance among the different treatment timepoints in the Aldefluor activity assay was calculated using two-tailed Student’s unpaired t test. Data represent mean ± SD: *p < 0.05; **p < 0.01; ***p < 0.001. Analyses for migration assay results were performed using two-way ANOVA with Tukey’s multiple comparisons test (Version 9.2.0, GraphPad Prism). Significance was defined as p < 0.05.
Study approval. The study was reviewed and approved by the University of Pittsburgh IRB (STUDY19060152). This study was determined exempt from written consent because the research presents no more than minimal risk of harm to subjects and involves no procedures for which written consent is normally required outside of research context.
Data availability. RNA sequencing data have been deposited in NCBI’s Gene Expression Omnibus and are accessible through GEO Series accession number GSE220538.