Tumor fragments were collected from 30 female dogs with mammary neoplasia that were brought to the veterinary clinics (São José do Rio Preto and region) during the years of 2009 and 2010. After tumor excision, the animals were followed-up from 1–18 months, with a median of 540 days. During follow-up time, the vets evaluated tumor metastasis and recurrence, as well as the cause of death of the animal.
For histopathologic diagnostics, the tumor biopsies collected were classified according to Misdorp et al.  by the AFIP (Armed Forces Institute of Pathology). The parameters employed for the classification of clinical tumor staging were in accordance with the TNM system (size, lymph node involvement, metastasis) established by WHO for canine mammary gland tumors modified , which recommended tumor mass size (T) – T1: < 3 cm - T2: between 3 and 5 cm - T3: > 5 cm; lymph node involvement (N) - N0: no apparent involvement - N1: unilateral involvement - N2: bilateral involvement; and distant metastasis (M) - M0: no evident metastasis - M1: distant metastasis including non-regional lymph nodes. Clinical staging was assigned as I, II, III or IV according to the tumor extension and established prognosis.
The presence of local tumor recurrence, metastasis and death were described and the overall survival was determined from the date of diagnosis until last follow-up or death. The cause of death was evaluated by the attending veterinarian and only female dogs that died of the illness were included in the group for the study. The dogs that had died of respiratory failure were diagnosed with lung metastasis as shown by X-ray. This study was approved by the Ethical Committee of the Faculdade de Medicina de São José do Rio Preto (Protocol number 3945/2009).
For immunochemistry, tumor samples were embedded into paraffin blocks and cut to give 3 μm sections. The samples were prepared on silanized glass slides before the paraffin was removed, the sections rehydrated in an ascending range of alcohol concentrations and incubated with 3% hydrogen peroxide for 30 min to block endogenous peroxidase activities. Antigenic recovery was made in a recipient at 95°C in buffer for 35 min for each specific antibody. The GSH primary antibody was diluted in the proportion of 1:100, GSTpi 1:4000 and GSH-Px 1:1200 in bovine serum albumin solution (BSA). After cooling, the slides were covered with BSA for 30 min and incubated at 4°C overnight with the antibodies.
After they had been washed with phosphate buffer (PBS) for 15 min and incubated with Easy Path kit (Erviegas, São Paulo, Brazil), composed of the secondary antibody “anti-mouse, rabbit and goat immunoglobulin with Biotin” for one h and “streptoavidin complex with peroxidase” for 30 min, followed by washes with PBS for 15 min, 0.5% of 3,3′-diaminobenzidin-tetra-hydrochloride (DAB, Signet® Laboratories, Dedham, MA, USA) was applied to the slides for 2–5 min at 20-22°C. The slides were counterstained with Harris hematoxilin for 40 min. Negative controls were obtained by the omission of the primary antibody, and one sample of prostate was used as an internal control in each assay.
The slides were photographed and the proteins quantified by the AxioVision software at 40× magnification under an AXIOSKOP2 Zeiss microscope. For each sample, three regions of the tumor tissue were selected and 20 points of the tumor cells were marked in each region. In this way, 60 different points were analyzed in each sample to obtain an average relative intensity of immunoreactivity. The values were given in arbitrary units (a.u.) and the mean optical density (M.O.D.) showing the specific immunostaining intensity at immunoreactive areas.
In vitro study
The in vitro study was performed with tumor biopsies of 10 dogs out of the 30 with mammary tumors from the immunohistochemistry study. The tumor biopsies were sliced into microfragments and incubated at 37°C in 5% CO2 in RPMI1640 (Cultilab) supplemented with 20% BSA, 1% penicillin/streptomycin and 1% L-glutamine. The cells were cultured until there was 80% confluence and then immunocytochemistry was carried out with the primary antibodies anti-cytokeratin, anti-vimentin and anti-calponin for epithelial origin confirmation.
In vitro cell treatment
The cells from each tumor biopsy were divided into two groups: control (untreated) and that treated with 0.2 mg of the drug (doxorubicin, Adriamycin) for 24 h. At the end of the treatment, cell viability was verified by cell counting with the Neubauer chamber after staining with trypan blue dye. The cells were submitted to immunocytochemistry for confirmation of the epithelium origin, with anti- cytokeratin antibody, resulting in a positive protein staining.
Quantitative RT-PCR (qPCR)
Total RNA was extracted from the cell cultures with TRIZOL reagent (Invitrogen). Each sample of total RNA was subjected to reverse transcription using a High Capacity cDNA kit (Applied Biosystems).
The quantitative PCR (qPCR) was performed according to Bustin et al. .
The polymerase chain reaction (qPCR) was made in triplicate using the equipment Step One Plus (Applied Biosystems). The reaction had a final volume of 20 μL, with 10 μL of Power SYBR Green PCR Master Mix (Applied Biosystems), 250 μL of each primer, and 10 ng of cDNA. The qPCR conditions were 50°C for 2 min, 95°C for 10 min, followed by 35 cycles of 95°C for 15 s and 60°C for 1 min. The dissociation curve was analyzed to confirm the desired genetic product: one cycle of 95°C for 15 s, 60°C for 1 min, and 95°C for 15 s.
Endogenous control of the genes RPS19 and RPL8 were used for normalization. The transcription level was calculated by the 2-ΔCt method . The Ct number (relative abundance) was calculated using the 2-ΔΔCT formula and was plotted as mean ± SD of the triplicates reaction.
The genes were selected from PUBMED. The GCLC and GSS genes were selected because they synthesize GSH, and the gene GST-Pi was not used in this study because its sequence was not available for the canine specie. Primers were designed with PRIMER3 software. The primers used for amplification were: GCLC sense (5′-CCAAGTCCCTCTTCTTTCCTG-3′) and antisense (5′-CGGAGACGGTGT ATTCTTGTC-3′) [GenBank: NG_012071.1]; GSS sense (5′-AGCCAATGCTCTGGTGCTAC-3′) and antisense (5′-ACCTTCGACGGATTACATGG-3′) [GenBank: NG_008848.1]; GSH-Px sense (5′-GGCATCAGGAAAACGCTA AG-3′) and antisense (5′-CCTCGCACTTCTCAAAAAGC-3′) [GenBank: NG_012264.1]; RPS19 (internal control), sense (5′-CCTTCCTCAAAAAGTCTGGG-3′) and antisense (5′-GTTCTCATCGTAGGGAGCAAG-3′) [GenBank: NG_007080.2]; and RPL8 (internal control), sense (5′- CCATGAATCCTGTGGAGC-3′) and antisense (5′-GTAGAGGGTTTGCCGATG-3′) [GenBank: NC_000008.10]. They were classified as subexpressed [samples with values lower than -1 (Log3)] and superexpressed [samples that presented values higher than -1 (Log3)].
All the analyses were done with the assistance the GraphPad Prism4 and StatsDirect software. The female dogs were grouped according to their clinical-pathological variables: age, time course, histological type, clinical staging, number of nodules, tumor ulceration, tumor vascularization, metastasis and death. The average referred to the densitometry, and quantification for the different mammary tumor groups was compared with Student’s t test or ANOVA, followed by Bonferroni test. The values were expressed as mean ± S.E.M.
The cutoff for the death risk was determined by the ROC curve. This analysis graphically represents the comparison between the sensitivity distribution specifically for each factor, or the true positive index or the false positive index. Whenever the ROC curve is closer to the left superior quadrant, the test variable is more precise, since the positive value (sensitivity) would be closer to 1 and false positive (specificity) closer to 0 . Survival curves were plotted using the Kaplan-Meier method and the differences between the curves were evaluated by a log-rank test and hazard function. All the 30 female dogs of the study were included in this analysis, and each death case is represented by the decrease of the survival percent in the graph in the correspondent day of the death. The overall survival was determined from the date of diagnosis until last follow-up or death.
To determine the possible association between the relative differences of the genetic expression of the groups, the Man-Whitney U test was used. P < 0.05 was adopted as significant.