Clinicopathological and molecular characteristics of Ku 70/80 expression in Nigerian breast cancer and its potential therapeutic implications

Ku 70/80, a regulator of the Non-Homologous End Joining (NHEJ) has been shown to have clinicopathological and prognostic significance in breast cancer (BC) from Caucasian populations. However, its significance in the Nigerian BC population, which is characterized by a higher rate of triple-negativity, and basal phenotype with high p53 mutation rate and BRCA1 deficiency, is yet to be investigated. We hypothesized that Ku70/80 expression would show adverse clinicopathological and survival characteristics in Nigerian and also, likely to have therapeutic implication on Black BC management. p27. Ku 70/80 was predictive of breast cancer-specific survival (BCSS) survival in multivariate analysis, but not of disease-free interval (DFI). Ku 70/80 expression is associated with metastatic disease, down-regulation of the Homologous Recombination pathway of DNA repair, loss of the G1-S phase checkpoint, high EMT potential and poor prognosis.


INTRODUCTION
The Non-Homologous End Joining (NHEJ) is an important cellular DNA double-stranded break (DSB) repair mechanism which is undertaken by a complex of proteins including the Ku heterodimer, Ku 70/80 [1][2][3][4]. The specific functions of the Ku heterodimer include recognition of and binding to the DSBs, recruitment of DNA protein kinase c (DNA-PKc) to the ends of the DSB. Once bound to the DSB ends the DNA-PKcs undergoes auto-phosphorylation, disengages from the damaged ends and allows downstream repair factors, such as Artemis, polynucleotide kinase 3'-phosphatase (PNKP), DNA polymerases, the MRN complex, XLF and XRCC4/DNA Ligase IV, to access the damaged site [1][2][3][4]. Ku 70/80 therefore acts as a regulator of the NHEJ, and its expression can be used as a marker for the NHEJ [4].
The NHEJ is involved in the repair of DSBs induced by radiation and some chemotherapeutic agents such as cyclophosphamide and topoisomerase poisons [1,5]. In addition, the NHEJ process itself is error-prone and therefore induces additional mutations and genomic instability in tumour cells [6][7][8][9][10][11]. Recent studies show that Ku 70/80 is over expressed in tumour more than in normal cells [6,10]. Furthermore, Ku 70/80 protects cancer cells from apoptosis [11]. Ku 70/80 can therefore be regarded as an oncogene in cancer cells as it promotes tumour progression.
The prognostic significance of Ku 70/80 expression in relation to clinical, pathological and survival indices are organ-specific. In advanced head and neck carcinoma, patients whose tumours had higher Ku 70/80 expression show better response to chemotherapy with 5fluorouracil and cisplatin [12]. In gastrointestinal malignancies, high Ku 70/80 expression is associated with higher gastric cancer clinical stage; with depth of invasion, pathological stage, histopathological grade and prognosis in advance colorectal carcinoma. Furthermore, in rectal carcinoma its relationship with pathological stage, histopathological grade, radio-sensitivity and disease-free interval has been reported [6,13,14]. Conversely, in endometrial carcinoma, disease-free interval is longer in patients whose tumour show low expression of Ku 70/80. However, there was no association between clinicopathological features and other survival characteristics [15].
A study of a Caucasian BC cohort showed that nuclear expression of Ku 70/80 is associated with higher histological grade, lympho-vascular invasion, negative estrogen receptor (ER) expression, basal-like phenotype, p53 and checkpoint kinase1 (CHK1) positivity. Ku 70/80 expression also showed an association with disease-free interval in univariate but not in multivariate analysis [16].
In this study we sought to determine the expression of Ku 70/80 in a well-characterised Nigerian BC cohort using Tissue Microarray and immunohistochemistry; and compared its expression with the clinicopathological and survival characteristics of the patients.

Western blotting
The KU 70/80 antibody was first validated using Western blotting as described in [16]. Briefly, protein extraction was accomplished using RIPA buffer containing protease and phosphatase inhibitors. About 50µg of protein was used to perform a protein gel electrophoresis. The resolved proteins were blotted onto a nitrocellulose membrane. Membrane blocking was accomplished with a solution of 5% non-fat dried milk in PBS-Tween-20. Primary incubation was performed in 1:1000 dilution of KU 70/80 mouse monoclonal antibody for 1hour at room temperature. Betaactin was used as control. Detection of proteins was performed by chemiluminescence.

Immunohistochemistry (IHC) staining
TMAs were constructed as previously described [17]. The expression of markers was determined using IHC in 4um TMA sections. The standard strept Avidin-Biotin complex method as  Table 1.

Immunohistochemistry scoring
The scoring of immunoreacitivty for Ku 70/80 was performed by determining the percentage of invasive malignant cells within showing positive staining. All samples were scored by one observer (JA) and a further observer (T-AF) countered scored a proportion. The cases were scored without knowledge of the patient outcome. The whole tissue mounts and TMA samples were scored twice. The mean of the scores were calculated to reach a final score. The median score of 74% was used to dichotomise immunoreactivity into high/low groups for subsequent analysis. All the other markers used in this study were scored as shown in Table 1 Statistical Analyses Statistical analysis was performed using SPSS 16.0 statistical software. Chi-squared analyses were used for inter-relationships between the Ku 70/80 expression, clinicopathological parameters and other biomarkers. The Kaplan-Meier survival method and the log-rank test were used for survival curves. Multivariate analyses using Cox proportional hazard regression models were performed and from the model both the risk factor and 95% confidence intervals were generated. A two-sided p-value of <0.05 was considered significant.

RESULTS
Western blot confirmed the specificity of Ku 70/80 antibody (Ab3108 Abcam) used in this study.
The nuclear Ku 70/80 biomarker expression was dichotomised according to the frequency histogram distributions using the median of the percentage of the nuclear staining. Zero to seventy-four percent (0-73%) staining was considered negative/low expression while 74% and above was regarded as positive/ high expression. A total of 113 (60.1%) cases were considered positive and 76 (39.9%) were negative/low for Ku 70/80. Table 2 shows the relationship between the clinicopathological parameters and Ku 70/80 protein expression, where the majority of the tumours that expressed Ku 70/80 showed positive vascular invasion (p=0.009). There were no other significant associations between Ku 70/80 and other clinicopathological characteristics.

Prognostic significance of Ku 70/80 expression in the Nigerian breast cancer cases
Multivariate analysis shows that tumours which were positive for Ku70/80 protein expression had a significantly poorer BCSS compared to tumours with negative/low Ku70/80 protein expression, independent of tumour grade, tumour size and lymph node involvement (Table 4).
There was no significant association with disease-free interval, DFI.

DISCUSSION
Ku 70/80 is one of the key determinants in the DNA damage response [1][2][3][4]. The roles of this biomarker in relation to clinicopathological parameters, biological behaviour and patients' outcome in BC, particularly in women from an indigenous Black African population, had hitherto not been studied. In line with its role in tumour progression, this study found a high expression of the NHEJ marker, Ku 70/80, in more than 60% of BC from a Nigerian population In concurrence with previous studies, the expression of Ku 70/80 was associated with adverse clinicopathological and survival characteristics: vascular invasion, triple-negative and basal-like subtypes of BC, as well as the molecular markers of poor prognosis [12][13][14][15]. In particular, these findings are similar to the results that were obtained from a UK study [16].
Ku 70/80 expression was a predictor of poor clinical outcome in Nigerian BC patients, independent of tumour grade, size and lymph node involvement. This is the first study to report Therefore, the high expression of cyclin B found in this study may be in keeping with the loss of p53 function, i.e. p53 overexpression, seen in Ku 70/80-positive tumours. Furthermore, the p27 loss, which denotes, in conjunction with p53 loss of function, the absence of the G1 to S phase checkpoint, in these Ku 70/80 tumours may be due to increased activity of the PIK3CA pathway in our study, since the PIK3CA pathway has been shown to down-regulate p27 [29], this might probably induced aneuploidy cells capable of enhancing tumours metastasis.
With respect to the therapeutic significance of Ku 70/80 expression in our BC cohort, it is noteworthy that cyclophosphamide-based chemotherapy, hormonal therapy and radiotherapy were offered to many patients in the Nigerian series, and yet the majority of them died within 5 years of diagnosis [30]. It is plausible to assume that the poor prognosis observed in our BC cohort is partly due to the adeptness of tumour cells at repairing therapy-induced DSBs using NHEJ pathway, and also, probably subjected them to high resistant rate to chemotherapy and radiotherapy [1,5]. Going further, it might be arguable that there is probably a need to change from the current "one-size-fits-all" approach to therapy, where all patients are treated with cyclophosphamide-based chemotherapy, to a biomarker-assisted selection of patients for chemotherapy and radiotherapy. For example, a study has shown that patients whose tumours with higher expression levels of Ku 70/80 respond better to cisplatin and 5-fluorouracil [12] Also, low expression of Ku 70/80 was shown to predict good response to radiotherapy in early BC [31]. More recently, patients with deficient BRCA1 expression have been shown to respond better to cisplatin than patients with proficient BRCA1 expression [32,33]. Biomarker-assisted chemo-and radiotherapy may therefore be a superior strategy for clinical cancer therapy than the current approach. A further potential therapeutic implication of the findings in this study is that the Ku 70/80 expression pattern in our cohort opens up a possibility of targeting the NHEJ pathway in black BC This is particularly important in the triple-negative BC subtype which currently has paucity of targeted therapy and commonly occurred among the black women. For example, inhibition of the NHEJ pathway such as with DNA-PK inhibitor, Nu7026 was reported to sensitize cells to topoisomerase II poisons [34]. Considering that most tumours which show high expression of Ku 70/80 in our study are also deficient in BRCA 1 and BRCA 2 (i.e. the HR pathway), it becomes even more attractive to target the NHEJ pathway in Nigerian BC cases. In many preclinical studies, inhibition of both the HR and NHEJ pathways has been demonstrated to enhance therapeutic sensitivity to radiotherapy and chemotherapy [35][36][37][38].
In conclusion, this study has for the first time shown that Ku 70/80 protein expression is high in Nigerian BC cases, has adverse clinicopathological significance, defective HR pathway of DNA repair, loss of G1-S checkpoint, high EMT potential and also confers poor prognosis. In addition, it has revealed a possibility for targeting the DNA repair pathways to improve patients' outcome among the Black BC.