Animals, tumor detection and surgery
Tammar wallabies used in this study were part of a captive breeding colony maintained in an open enclosure at The University of Melbourne Macropod Research Facility (Melbourne, Victoria), with stock originally from Kangaroo Island, South Australia. Care and treatment of animals conformed to the National Health and Medical Research Council of Australia and were approved by the Victorian Department of Sustainability and Environment Ethics Committee. The animal presenting with a mammary lesion was part of the breeding colony for two years (February 2002 – February 2004) and was checked monthly during this time for signs of mating and presence of pouch young. The animal did not reproduce in this time but presented with a mammary lesion. The animal was subsequently monitored weekly for three weeks then euthanized. The lesion was measured, excised and either immediately frozen for future RNA extraction, placed in Hanks' Balanced Salt Solution (HBSS) (Sigma Aldridge, Sydney, Australia) at 4°C in preparation for isolation of mammary cells or fixed in formalin for histology. The tissue from the mammary gland was divided into two portions, one comprising the main body of the mammary lesion and the other comprising normal mammary tissue. Mammary gland tissue was also collected from a wallaby of comparable age and reproductive status for preparation of mammary epithelial cells.
Preparation of WalBC cell line and wallaby primary mammary epithelial cells
Tissue was immediately transferred to 1× Hanks' Balanced Salt Solution (HBSS) (Gibco, USA) with 20 μL/mL penicillin/streptomycin (Gibco, USA) and 5 μg/mL Fungizone (Gibco, USA) on ice and transported back to the laboratory for enzymatic digestion to harvest mammary epithelial cells.
Tissue was dissected free from fat, weighed, sliced finely and digested with 1 μg/mL Collagenase Class2 (Worthington UK), 10 mL/L penicillin/streptomycin (Gibco, USA), 10 mL/L fungizone (Gibco, USA), 0.35% Bovine Serum Albumin (Sigma) and a final concentration of 0.5 mM glucose. 10 g of tissue per 100 mL of media was digested shaking at 200 rpm at 37°C for 4 h. Cells were harvested by filtration through 150 μm nylon mesh using a Nalgene filter unit. The suspension was centrifuged at 80 g for 5 min and pellets were washed twice with HBSS containing 0.02 mg/mL DNase 1 (Invitrogen) and 1 mg/mL Trypsin Inhibitor (Sigma). Cells were again suspended in wash media, then filtered through 53 μM nylon mesh, re-centrifuged and finally resuspended in FCS/10%DMSO (DMSO-Sigma, Sydney, Australia), and frozen at a density of ~2 × 107 cells/mL.
WalBC cells and wallaby primary mammary cells were cultured in either 25 or 80 cm2 culture flasks in 10 mL or 20 mL respectively of M199/Hams/Hepes media with 1 μg/mL cortisol, 10 ng/mL EGF, 1 μg/mL insulin supplemented with 10% fetal bovine serum. WalBC cells were passaged (more than 20 times) when confluent by scraping and use of versine solution in phosphate buffered saline (PBS) (Sigma-Aldrich, Sydney, Australia). Passaging of cells with 0.1% trypsin-versine in phosphate buffered saline (PBS) (Sigma-Aldrich, Sydney, Australia) for 2 minutes at 37°C disrupted cellular aggregates but cells failed to retain viability following this treatment.
Tissue was fixed in 10% formalin for 24 h. Samples were processed (Citadel; Shandon Scientific Ltd., Cheshire, England), and embedded in paraffin using routine procedures. Paraffin-embedded sections of 5 μm thickness were cut, mounted on 3-aminopropyltriethoxysilane-coated slides and submerged in histolene to remove the paraffin. After rehydration, sections were stained with haematoxylin and eosin. Finally, sections were coverslipped and examined using an Olympus BX40 microscope and Coolscope digital microscope (Nikon) for light microscopy.
Paraffin-embedded sections of 5 μm thickness were cut and mounted on 3-aminopropyltriethoxysilane-coated slides and submerged in histolene to remove the paraffin. After rehydration, tissue peroxidases were blocked for 30 min with a 1% hydrogen peroxide solution and washed with 1× PBS. For detection of Vimentin, sections underwent treatment with sodium citrate buffer. Sections were then blocked for 60 minutes with 10% goat serum (Sigma)/1% BSA/PBS prior to addition of Vimentin antibody (V9, 1:800; Dako, Australia) at 4°C overnight. An HRP-conjugated goat anti-mouse secondary antibody (1:250 dilution; Dako, Australia) was then applied for 1 hour to allow a brown precipitate to develop using AEC (Dako, Australia). Finally, sections were counterstained with eosin, dehydrated, coverslipped and examined using a Coolscope digital microscope (Nikon) for light microscopy.
Proliferation assay and cell morphology
WalBC or wallaby MEC's cells were plated (2000 cells/well) in 96 well plate formats with growth media containing either insulin (I; 1 μg/ml), cortisol (F; 1 μg/ml) and prolactin (P; 1 μg/ml) or I, F, Epidermal growth factor (EGF; 10 ng/ml). Cells were grown for a further 10 days before being fixed and stained with Sulforhadamine B as previously described . Each time point was performed in quadruplicate. Cells were visualized by phase contrast microscopy using an Olympus BX40 microscope and photographed using a DigitalSight DSL1 (Nikon) camera.
Matrigel outgrowth assay
Matrigel outgrowth assays were performed in 48 well plates as previously described (Price and Thompson, 1999). Cells (2 × 104) were dispersed in 75 μL of undiluted Matrigel (approx.10 mg/mL) and then overlaid onto 100 μL of polymerized undiluted Matrigel. Once the top layer had polymerized the cultures were incubated in MEM/10%FBS media for up to 10 days and photographed at 20 × magnification by phase contrast microscopy using an Olympus BX40 microscope and photographed using a DigitalSight DSL1 (Nikon) camera.
In vivo studies
Mice (3–4 week old intact female Balb/C nu/nu) were purchased from Australian Resource Center (Perth, Australia), housed in individually ventilated cages under filtered air (Techniplast, Milan, Italy) and acclimatized for one week prior to manipulation. Anesthesia was achieved by i.p injection of ketamine/xylazine (Provet; Australia; 40 μg/g mouse and 16 μg/g mouse, respectively). The mice were allowed to recover from the anesthesia before being returned to their cages and monitored daily. Animal studies were conducted with ethical approval of the St. Vincent's Hospital Animal Ethics Committee (Melbourne, Australia), and in accordance with the Australian National Health and Medical Research Council's Guidelines for the Care and Use of Laboratory Animals.
Mammary fat pad inoculation of WalBC cells
WalBC cells were harvested from near confluent conditions, aspirated into cell suspension (cell aggregates and single cells were present), washed thrice and resuspended in PBS before inoculation. Two groups of 8 mice received mammary fat pad inoculation of WalBC cells (5 × 105 cells/15 μL) as described by Price et al . Tumor growth was assessed by monitoring the fat pad for palpable tumors at weekly intervals. Mice were sacrificed after 6 months.
WalBC cells grown on plastic were fixed in 4% paraformaldehyde (30 min) and washed three times in PBS. Cells were permeabilized with 0.1% Triton X/PBS (5–10 min) and washed thrice with PBS before blocking in 1% BSA for 30 min. Either Vimentin9 (Dako; 1/800) or Rabbit anti-cytokeratin (Bectin Dickinson1/400) antibody was added in blocking buffer and incubated overnight at 4°C. Cells were then washed five times in PBS to remove non-specific binding of primary antibody. Cells were then incubated in goat anti-mouse (FITC) (DAKO; 1/400) or goat anti-rabbit (FITC) (DAKO; 1/400) in blocking buffer for 1 hour and washed thrice with PBS. Nuclei were visualized using Propidium Iodide (Invitrogen). Images were visualized for fluorescence using an Olympus BX40 microscope and photographed using a DigitalSight DSL1 (Nikon) camera.
RNA preparation for gene expression by microarray analysis
WalBC cells, primary mammary cells cultured from the same animal and primary wallaby mammary cells cultured from a virgin animal of similar age were scraped from tissue culture dishes using Tripure (Roche). Total RNA was isolated from the aqueous phase and further purified using the Qiagen RNeasy miniprep kit (Sydney Australia) following the manufacturer's instructions.
RNA Amplification was done in 3 parts similarly to the Eberwine protocol . First strand synthesis utilized MMLV RNase H- (Promega M3681) and second strand synthesis was done with DNA Polymerase 1 (Promega, M2501). Lastly in vitro transcription was performed with the T7 Megascript Kit (Ambion 1334). The resulting amplified RNA was then further purified using the QIAGEN RNeasy miniprep kit.
The amplified RNA from each treatment group was labeled using amino allyl reverse transcription followed by Cy3 and Cy5 coupling. Samples of amplified RNA (10 mg) were reverse transcribed using 5 μg random hexamers (Geneworks), MMLV reverse transcriptase (Promega), RNAse H (Invitrogen) and 1× buffer at 42°C for 2.5 hours. The reaction mix was hydrolyzed by incubation at 65°C for 15 minutes in the presence of 55 mM NaOH, 55 mM EDTA followed by a subsequent addition of acetic acid to 50 mM. The cDNA was then adsorbed to a Qiagen QIAquick PCR Purification column. Coupling of either Cy3 or Cy5 dye was performed on the column by incubating the adsorbed cDNA with the appropriate dye in 0.1 M sodium bicarbonate pH 9.0 for 1 hour at room temperature in darkness. Each labeled cDNA was eluted in 80 μl of water and was then combined with its comparing sample during further purification on a second Qiagen QIAquick PCR Purification column. The joint Cy3 and Cy5 labeled probe in a final concentration of 0.4 mg/ml yeast tRNA, 1 mg/ml human Cot 1 DNA, 0.2 mg/mL Poly dA50, 1.25 × Denharts, 3.2 × SSC and 50% formamide was heated to 100°C for 3 minutes. SDS, to 0.1%, was added immediately after heating and just prior to application. Probes were hybridized to custom made tammar wallaby EST microarray slides overnight at 42°C in a HyPro20 (Integrated Science) humidified chamber. The slides were printed with 10,000 EST's from tammar mammary gland cDNA libraries generated from tissue collected across the lactation cycle (Lefevre, manuscript in preparation).
The tammar EST database was derived from several cDNA libraries comprising day 23 pregnant (n = 4), lactating at day 130 (n = 4), lactating at day 260 (n = 1), lactating at day 130 subtracted for all the major milk protein genes (n = 2), non-lactating (n = 2) and a normalized library (combined RNA from day 26 pregnant, lactating at day 55, day 87, day 130, day 180, day 220, day 260 and involuting at day 5).
Microarray's were washed in 0.5× SSC, 0.01% SDS for 1 minute, 0.5× SSC for 3 minutes then 0.006× SSC for 3 minutes at room temperature in the dark. Slides were centrifuged dry at 130 g for 5 minutes then scanned with a VersArray Scanner (BioRad). Images were analyzed using Versarray Software (Biorad).
Analysis of gene expression data
Gene expression data was normalized using the single channel normalization method in the Limma package of Bioconductor . These normalized expression values were analyzed using a two-stage process where all the expression values are considered simultaneously. The first stage involves fitting a linear mixed model of the formMadj = μ + Treatment + Probe + Treatment.Probe + ε
where Madj is the adjusted (loess-normalized) log intensity ratio for a probe on the cDNA array, Treatment is the fixed effect of the treatment (tumor cells, non-tumor cells, or virgin), Probe is the random effect of the probe on expression levels, regardless of the treatment, and Treatment.Probe is the random effect of a particular probe within a particular treatment, the effect of interest. Typically, the distribution of these random Treatment.Probe effects show a mixture of two distributions, one with small variance (non-differentially expressed genes, non-DE) and one with large variance (differentially expressed, DE). So the second stage involved fitting a two-component mixture model (DE vs non-DE) to these effects , and this will return the (posterior) probability that a particular gene is DE, given its Treatment.Probe effect, with a probability in excess of 0.5 indicating a gene is more likely DE, for that particular treatment. However, a stringent threshold has been used requiring a posterior probability in excess of 0.999 before a gene was classified as being DE, in order to reduce the false positive rate. The statistical package R was use for this two-stage process.
To determine the significance of EGF response within each cell type a paired students t-test was performed on quadruplicate samples at day 10. For Matrigel outgrowth, 20 outgrowths were measured for each cell type at day 4 in a given area and averaged. Unpaired t-test was used to determine significance differences in the length of outgrowths between the two cell types.