A total of 102 biopsies were collected de novo (adjacent and tumour tissues) from patients with BC between September 2011 and December 2012 at the Instituto Estatal de Cancerología (IECan) Dr. Arturo Beltrán Ortega, a Tertiary-Level Hospital (Acapulco, Guerrero, Mexico). Inclusion criteria were the following: patients who voluntarily accepted to be part of the study and who signed an informed consent letter; patients residing in the state of Guerrero; patients of any age at the time of diagnosis of invasive ductal carcinoma (range 31–85 years); clinical stage I, II and III, patients not currently under neoadjuvant treatment, and diagnosis of BC confirmed by the hospital's pathology department. Medical records were checked thoroughly to obtain the patients' clinic-pathological and gynaecological characteristics. The exclusion criteria were as follows: presence of cancer antecedents, recurrence of cancer and/or pregnancy. The elimination criteria were as follows: insufficient biopsy material and incomplete clinical history. The project was evaluated and approved by the IECan Ethics and Research Committee.

We investigated the genetic alterations of p53 and TOP2α, which were classified as: small-scale mutations (point mutations and deletions) for p53 gene and large-scale mutations (amplifications or deletions) for TOP2α gene.

Genomic DNA was extracted from breast tissue biopsies (adjacent and tumour tissues) using the DNeasy Blood & Tissue kit (Qiagen GmbH, Hilden, Germany) according to the manufacturer's instructions. The genomic DNA obtained was quantified, and its integrity was verified by GAPDH end-point PCR.

Total protein was extracted from breast tissue biopsies (adjacent and tumour tissues) using an Ambion PARIS kit (Applied Biosystems, Carlsbad, CA, USA) according to the manufacturer's instructions. The proteins were quantified using a Pierce BCA protein assay kit (Thermo Fisher Scientific, Rockford, IL USA), and their integrity was verified by resolving the proteins in 8% sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and Coomassie staining.

To amplify the DNA binding domain of p53 (exons 5–8), two sets of primers were designed using Primer Express ver. 3.0 software (Applied Biosystems, Foster City, CA, USA). One set of primers amplifies a 459 bp fragment from exons 5–6 (F, 5′-TTCCTCTTCCTACAGTAC-3′ and R, 5′-AGTTGCAAACCAGACCTCA-3′); the second set amplifies a 510-bp fragment from exons 7–8 (F, 5′-GTGTTATCTCCTAGGTTG-3′ and R, 5′-TCCTCCACCGCTTCTTGT-3′). PCR products were purified using a QIAquick PCR Purification kit (Qiagen GmbH) according to the manufacturer's instructions. The amplicons obtained were quantified and sequenced in a Genetic Analyzer 3130 (Applied Biosystems). The sequences were analysed using MEGA6 software.

Copy number of the TOP2α gene were analysed by performing duplex reactions in triplicate using 20 ng of genomic DNA (obtained from adjacent or tumour tissues), a TOP2α TaqMan Copy Number Assay probe labelled with FAM, a RNase P TaqMan Copy Number Reference probe labelled with VIC, and TaqMan Genotyping Master Mix (all from Applied Biosystems) according to the manufacturer's instructions. The Ct values obtained were analysed using CopyCaller ver. 2.0 software (Applied Biosystems). The method was validated by performing standard curves generated by serial dilutions of the genomic DNA (1,000-0.1 ng) in triplicate. As a diploid copy number control, genomic DNA was extracted from pooled peripheral-blood samples of healthy women (n=5). As a positive amplification control, genomic DNA from the breast-cancer cell line SKBR3 was utilized.

To perform the proximity ligation assay (PLA), 200 nM aliquots of the probes 3′ Prox-Oligo and 5′ Prox-Oligo (Applied Biosystems) were separately incubated with 200 nM biotinylated polyclonal p53 antibody (R&D Systems, Minneapolis, MN, USA) for 60 min at room temperature. Afterward, Assay Probe Storage Buffer was added and incubated for 30 min at room temperature. One microgram of total protein extract (from adjacent or tumour tissue) was incubated overnight with 250 pmol of the probes 3′ and 5′ Prox-Oligo antibody conjugated at 4̊C. Next, the ligation mix was added to each tissue lysate containing total protein. The ligation reaction was incubated for 10 min at 37̊C and after this, at 4̊C for an additional 10 min. Finally, the reaction mix was supplemented with protease, incubated for 10 min at 37̊C, inactivated at 95̊C for 5 min and maintained overnight at 4̊C (Applied Biosystems). The qPCR assay was performed using the TaqMan^® Protein Assays Fast Master Mix (Applied Biosystems) and 9 µl of the protein lysate obtained from the PLA assay. PCR conditions were as follows: 2 min at 95̊C and 40 cycles of 15 sec at 95̊C and 1 min at 60̊C. Ct values were imported into ProteinAssist™ software v1.0 (Applied Biosystems), in which the relative quantification analysis was performed. Prior to analysis of the protein lysates, a dynamic range was elaborated, and the reaction efficiency was verified with a serial dilution standard curve in triplicate. The cut-off limit for the relative expression of the p53 protein was <1.0 for underexpression and >1.0 for overexpression.

Thirty micrograms of total protein (from adjacent and tumour tissues) were resolved in a 10% SDS-PAGE gel. The proteins were transferred onto a nitrocellulose membrane (GE Healthcare, Buckinghamshire, UK) and blocked with 5% non-fat milk for 1 h at room temperature. The membrane was then incubated overnight with polyclonal rabbit anti-human TOP2α antibody (Abcam, Cambridge, UK) or monoclonal mouse anti-human p53-HRP antibody (Santa Cruz Biotechnology, Inc., Dallas, TX, USA) at 4̊C. The membranes were washed three times with TBS/Tween-20 0.1%. Additionally, the membrane employed for TOP2α detection was incubated for an additional 1 h with a secondary horseradish peroxidase (HRP)-conjugated anti-rabbit antibody (Sigma Chemical Co., St. Louis, MO, USA). Protein detection was performed using the reagent SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific, Rockford, IL, USA), and the chemiluminescent signal was detected using a VersaDoc Imaging System (Bio-Rad, Hercules, CA, USA). The monoclonal mouse anti-human β-tubulin antibody (Sigma Chemical Co.) was employed as a loading control. The densitometric analysis was performed using ImageJ software. TOP2α protein expression was normalized against the loading control (β-tubulin); then, TOP2α expression in the tumour tissue was normalized against the adjacent tissue. P53 protein expression was analysed in all samples (adjacent and tumour tissues), but detection was only identified in the tumour samples. In this case, the p53 protein was normalized only against the loading control (β-tubulin), since the half-life of the p53 protein is short, and it was not possible to detect its expression in the adjacent tissue.

To determine the presence of the estrogen, progesterone and HER2 receptors, 5-µm-thick sections were obtained from paraffin-embedded tissue. The primary antibodies used were the monoclonal mouse anti-human estrogen receptor clone 1D5 (Biocare Medical, Concord, CA, USA), the monoclonal mouse anti-human progesterone receptor clone PgR 636, and the polyclonal rabbit anti-human c-erbB-2 oncoprotein (both from Dako, Carpinteria, CA, USA). The detection system utilized was the mouse/rabbit UnoVue HRP/DAB detection system (Diagnostic BioSystems, Pleasanton, CA, USA). For each biopsy, a previously selected positive control was also analysed by the pathologist as an internal quality control. Data interpretation was performed using the Allred system, and for HER2, cells were defined as 3+ positive, 2+ intermediate and 0–1+ negative. HER2 biopsies with intermediate values (2+) were outsourced to an immunohistochemistry and molecular pathology laboratory (INMUNO Q, México, D.F.) for validation by means of a Silver In Situ Hybridization (SISH) assay.

Patients were subdivided into two groups: non-TNBC patients, expressing at least one of the receptors (RE, RP or HER2) and TNBC, lacking all of these receptors. To compare clinical and pathological variables, point mutation or deletion in p53 gene, amplification or deletion in TOP2α gene, and SNPs in p53 gene by groups, we used an χ² test. Age at the time of diagnosis was compared with a non-linear Gaussian adjustment for accumulated relative frequencies. The relative expression of the TOP2α and p53 proteins determined through western blotting and/or PLA-qPCR, were analysed by the multiple Students t-test. Analysis of variance (ANOVA) was used to compare the relative expression of the p53 and TOP2α proteins vs. genetic alterations of these genes. In all cases, p<0.05 was considered to be significant.

In addition, a multiple correspondence analysis (MCA) was used (16), which is a descriptive method to determine the relationship between specific groups. MCA uses an χ²-calculated distance to assess the association between clinicopathological variables (clinical groups, SBR histological grade), large-scale mutations in the TOP2α (amplification or deletion) and small-scale mutations in the p53 (point mutation or deletion) genes, and the relative expression of the TOP2α and p53 proteins. Based on these distances, the correspondence analysis registers two sets of combinations, generating a score calculated for each group. The first combination (dimension 1) defined a scale of variation that allowed the highest discrimination between groups. The second combination (dimension 2) defined the maximum dispersion between uncorrelated groups and the relative expression of the TOP2α and p53 proteins. To compare relative expression values of the TOP2α and p53 proteins, the results were divided in percentiles starting from 33 and 66%. Three sets of each variable were generated: underexpression <33% (<0.9), basal expression (normal) 33–66% (0.9–1.1) and overexpression >66% (>1.1) (16).

Finally, survival curves were determined using the Kaplan-Meier method and verified by a log-rank test (Mantel-Cox) with a follow-up period of 26 months, as patients continue in follow-up on the present study.

The analysed population consisted of 102 patients with BC, of whom 71.5% were diagnosed as non-TNBC and 28.5% were diagnosed as TNBC (Fig. 1A). The age of TNBC patients at time of diagnosis was estimated as 50.2 (±11.3) years of age on average, whereas the age of non-TNBC patients was estimated as 55.01 (±10.3) years of age on average. The accumulated relative frequencies of both groups were significant (p<0.0001) (Fig. 1B). TNBC patients exhibited 82.7% of SBR histological grade III (p=0.0089) (Fig. 1C) and 37.9% recurrence (p=0.017) (Fig. 1D) when compared with non-TNBC patients (50.6 and 15.1%, respectively).

Furthermore, TNBC patients presented only stages II and III (p=0.539) (Fig. 1E), larger tumour size (p=0.393) (Fig. 1F), and a higher frequency of positive lymph nodes in N2 status (p=0.935) (Fig. 1G); we did not observe significant differences between non-TNBC and TNBC patients.

Since the tumour suppressor gene p53 is mutated in ~50% of all types of cancer, we evaluated small-scale mutations (point mutations, deletions) in this gene through DNA sequencing of exons 5–8, which are frequently mutated. We found genetic alterations in the p53 gene in 27.5% of patients with BC, of which 9.8% were point mutations (missense) and 17.7% were specific deletions in exons 5–6 (Table I).

The majority of PCR amplicons spanning exons 5–6 exhibited a similar migration pattern by electrophoresis (459 bp, expected molecular weight); however, in some cases, we observed an amplicon of ~395 bp, suggesting that the latter represents an allele carrying a deletion (Fig. 2A). We were able to identify an ~64 bp deletion corresponding to codons 191–216 (exon 6) (Fig. 2B).

In addition, it was through DNA sequencing that we identified single nucleotide polymorphisms (SNPs) rs1294778 and rs12951053 localized inside intron 7 of p53, both of which had a frequency of 25.5% in the study population (Table I). Both SNPs (homozygotes or heterozygotes, according to the case) were identified in the same patients. No significant differences were found between TNBC and non-TNBC patients regarding either deletions or mutations (p=0.0655) (Fig. 3A), and SNPs (rs1294778 and rs12951053) (data not shown) identified for p53.

To determine whether point mutations or deletions found in the p53 gene compromise its protein expression, all of the BC samples were analysed by western blotting. A representative image of western blotting of the p53 protein is shown in Fig. 3B. When comparing p53 protein expression levels, we observed a statistically significant average of 0.78 (±1.63) in the TNBC group (p=0.005) compared with the non-TNBC group (mean, 0.10±0.43) (Fig. 3C). We normalized the relative expression of the p53 protein using the loading control β-tubulin since normal adjacent tissue lack or present low levels of p53 protein. In contrast, p53 mutant proteins have a characteristically long half-life (4,17).

Next, we validated the results of p53 protein expression using PLA-qPCR, a technique more sensitive than western blotting. We found an average p53 protein expression of 1.83 (±1.75) and 0.62 (±0.65) in the TNBC and non-TNBC groups, respectively. The difference among these results was statistically significant (p<0.0001) (Fig. 3C).

Furthermore, we analysed an association among p53 gene alterations with p53 protein expression. We observed that patients carrying deletions in the p53 gene present underexpression or null expression of the p53 protein, as shown by PLA-qPCR or western blotting analysis, respectively (Fig. 3D). Patients with different mutations in the p53 gene present either underexpression or overexpression of the p53 protein, with both situations observed through western blotting and PLA-qPCR (Fig. 3D). In some cases, patients showing no alterations (point mutations or deletions) in exons 5–8 also showed expression of the p53 protein. This may be due to mutations localized in the unexamined regions of the p53 gene (Fig. 3D).

Similar to p53, TOP2α is altered in cancer. Thus, we evaluated large-scale mutations (amplifications or deletions) for the TOP2α gene by qPCR. We found that 32.4% of all patients analysed showed genetic alterations in this gene, among whom 20.6% presented deletions and 11.8% showed amplifications of the TOP2α gene (Table I). However, there were no significant differences in either TOP2α gene deletions or amplifications between the TNBC and non-TNBC patient groups (p=0.4508) (Fig. 4A). To determine whether such amplifications or deletions exert any effect on TOP2α protein expression, the BC samples were evaluated by western blot analysis (adjacent and tumour tissues). We observed that the average TOP2α protein expression was 1.36 (±2.95) in TNBC and 1.40 (±3.79) in the non-TNBC group. The patients showed similar expression levels of TOP2α protein between the groups without a significant difference (p=0.960) (Fig. 4B). We show a representative image of the western blotting of proteins obtained of two patients with BC for TOP2α protein. β-tubulin was used as the loading control (Fig. 4C). When comparing the amplifications or deletions and protein expression of TOP2α, we observed a higher expression of the TOP2α protein in patients lacking any TOP2α genetic alterations and decreased TOP2α protein levels in patients carrying TOP2α deletions. Individual group evaluations revealed higher TOP2α protein expression in TNBC patients carrying gene amplifications when compared with non-TNBC patients (p=0.023) (Fig. 4D).

To determine the association between the clinical and pathological characteristics (clinical group, histological grade) of patients with genetic alterations (previously described) in the TOP2α and p53 genes and the expression level of proteins, we carried out an MCA. We identified three groups by means of an MCA as follows: the first group of TNBC patients was associated with a point mutation in the p53 gene and p53 protein overexpression, due to a increase in the p53 protein's half-life. Additionally, we detected the amplification of the TOP2α gene in this group despite the fact that TOP2α protein expression was not associated with any clinical group. The second group of non-TNBC patients was associated with a p53 gene deletion, underexpression of the p53 protein, a non-altered TOP2α gene (although overexpressed protein), and SBR histological grade II. The third group represented by patients of any group (TNBC or non-TNBC) presented with a TOP2α gene deletion and a baseline expression level of both the p53 and TOP2α proteins (Fig. 5).

The follow-up period for patients studied with BC was set at 26 months, and the patients are undergoing continued follow-up. We observed a general survival of 91.2% in the study population. In the groups, we observed a 79.3% survival for the TNBC group and 95.8% for the non-TNBC group (p=0.015). In relation to time of survival for the study population, the mean was 25.36 months, with 24.1 months for the TNBC group and 25.8 months for the non-TNBC group (p=0.002) (data not shown).

After that, we analysed the prognostic factor of point mutations or deletions in the p53 gene. In patients with BC that presented a deletion in the p53 gene, the survival was 88.2%; in patients who had a point mutation, the survival was 80%. In neither case was the survival statistically significant (p=0.4396) (Fig. 6A). In contrast, when we analysed p53 protein expression through PLA-qPCR (Fig. 6B) in the same group of patients, we found that p53 overexpression was associated with a lower survival (76%) compared with p53 underexpression (96%), suggesting a poor prognosis factor (p=0.0014). Similarly, using western blotting we found an association between p53 overexpression and a lower survival (76.9%, p=0.0344) (Fig. 6C). In addition, when we analysed p53 protein expression and its association with a clinical group, we observed that TNBC patients with p53 overexpression presented a lower survival (p=0.0030) (Fig. 6D).

Finally, we analysed whether amplifications or deletions in the TOP2α gene are a prognostic factor. To perform this analysis, two groups were generated as follows: i) a group without genetic alterations of the TOP2α gene (normal) presented with a survival of 94.2%; ii) a group with amplification presented with a survival of 91.6%; and iii) a group with deletion presented with a survival of 82.6% (p=0.2340) (Fig. 7A). Although patients with TOP2α underexpression had a survival of 88.8%, this result was not statistically significant (p=0.1167) (Fig. 7B). However, neither TOP2α amplifications or deletions nor protein expression was a factor for poor prognosis in patients with BC.