Breast cancer is the most frequently diagnosed cancer in women, with 1.4 million new cases annually worldwide, leading to more than 450 000 annual cancer deaths.1 However, there is large geographical variation in its incidence,2 and breast cancer was the most frequently diagnosed cancer among United Arab Emirate (UAE) nationals in 2002, accounting for 23% of all cancers in women.3
The UAE has a young and multiethnic population, and UAE women come from diverse ethnic backgrounds, such as the Arabian Gulf region, Mediterranean region, Middle East, Iran, the Indian subcontinent and Far East Asia.4 It has been established that in many Asian and African countries breast cancer tends to affect younger females, who present at an advanced stage with poor prognostic features, and that outcome is poor in these women compared with their counterparts in Western countries.5–7
Although morphological characterization remains a fundamental diagnostic tool, in most instances a diagnostic approach based solely on morphology lacks the ability to specifically address the prognosis and treatment options for the individual patient. Therefore, additional diagnostic characterizations are required to achieve the ultimate goal of a personalized approach to the management of breast carcinomas.8
Traditionally, pathological determinations of tumour size, lymph node status, endocrine receptor status and human epidermal growth factor receptor 2 (HER2) status have driven prognostic predictions and adjuvant therapy recommendations for patients with early-stage breast cancer. However, these prognostic and predictive factors are relatively crude measures, resulting in many patients being over- or undertreated.9
Candidate prognostic biomarkers in breast cancer include expression of oestrogen receptor (ER) and progesterone receptor (PR), amplification and overexpression of HER2, cyclin D1 and c-myc, as well as p53 nuclear protein accumulation, B-cell lymphoma 2 (Bcl-2) expression and alteration in angiogenesis proteins such as vascular endothelial growth factor.10
Numerous studies, mostly conducted in the West, have demonstrated variable marker expressions and different breast cancer characteristics among different racial groups.11 Alterations in hormone receptor expression in breast cancer have been extensively studied and the expression of ER is shown in 70–80% of cases of invasive ductal carcinoma and the expression of PR is shown in 60–70% of cases. These cases are associated with better outcome and good response to hormone therapy in both adjuvant and metastatic settings.12
Human epidermal growth factor receptor 2/neu, also called ErbB2, is a member of the human epidermal growth factor family and encodes a transmembrane tyrosine kinase receptor. It is a proto-oncogene that is located on chromosome 17 and is amplified, or the protein HER2/neu overexpressed, in 15–25% cases of breast cancer.13 It has been shown that HER2/neu is a significant prognostic marker in both node-positive and node-negative breast cancer patients.14 The positivity of HER2/neu in breast cancer is associated with high histological grade and tumour aggressiveness. It is therefore assessed for both the prognosis and prediction of response to trastuzumab treatment, chemotherapy and endocrine therapy.15
Apoptosis-related genes have been extensively studied in breast cancer as both prognostic and predictive markers.16 Bcl-2, a cytoplasmic protein belonging to the Bcl-2 family, is expressed in normal glandular epithelium but is overexpressed in 25–50% of breast cancers. Furthermore, Bcl-2 expression in breast cancer has been found to be associated with favourable prognostic factors such as smaller tumour size, ER positivity and low nuclear grade.17
Reduced cell adhesion brought about by altered surface expression of E-cadherin has been implicated in invasive and metastatic malignant growth.18 Recent studies have shown that inactivation of E-cadherin is important in the progression of sporadic breast cancer19 and, moreover, altered E-cadherin expression has been found to be associated with decreased overall survival in cases of breast cancer.20
The transcription factor p53 is expressed in most cell types. The activation of p53 results in induction or repression of a number of genes involved in cell cycle regulation, DNA repair and apoptosis. Thus, p53 allows DNA repair or induces apoptosis, protecting against the accumulation of genetic changes.21 Mutations in p53 suppress the regulatory functions of the protein and are the most common genetic change in breast cancer with a frequency of approximately 30% (range 15–71%).22
The recent development of gene expression profiling from cDNA microarrays has led to an advanced understanding of the heterogeneous biology of breast carcinoma, but the method is not yet relevant to clinical practice. In contrast, molecular classification based on immunohistochemistry provides several advantages in terms of costs, simplicity, tissue requirements and popularity over cDNA microarray analysis.23 Additionally, as reported by Park et al. in a study from 2012, molecular subtyping based on immunohistochemistry could be used as a standard to determine treatment and surveillance strategies in breast cancer patients.24
The primary molecular subdivision is by ER status, with further subdivision of ER-positive tumours into a luminal A category or a luminal B category, which is associated with a worse prognosis. ER-negative tumours are subdivided into HER2/neu-positive and a basal-like group, the latter being approximated by a triple negative (ER–, PR– and HER2–) phenotype.25
As for the existing data, the clinicopathological characteristics and the immunohistochemical profile of breast cancer in the northern emirates was not examined in previous studies. Hence, the aim of this study was to evaluate the immunohistochemical expression of ER, PR, HER2/neu, Bcl-2, E-cadherin and p53 in primary breast carcinomas in Ras Al Khaimah (RAK) and to analyse their expression to the established prognostic factors of breast cancer. We also attempt to classify breast cancers in the present cohort using immunohistochemical marker panels consisting of ER, PR and HER2 and study their correlation to the other clinicopathological characteristics.
Materials and methods
Patients and methods
This study included 28 patients with primary invasive carcinomas of the breast diagnosed at the Department of Surgery, Saqr Hospital, RAK, over the period 2001–10. The patients’ medical records were reviewed for demographic data and clinical history.
The archival formalin-fixed, paraffin-embedded blocks of the tumour specimens and the pathology reports were retrieved from the Department of Pathology, Al Baraha Hospital, Dubai. The original haematoxylin and eosin-stained microscopic slides from all cases were reviewed by the three contributing pathologists for confirmation of diagnosis and standardization of grading according to the Elston–Ellis modification of the Bloom–Richardson system and clinical stage was determined according to the American Joint Committee on Cancer (AJCC) Cancer Staging Manual, 6th edition.26
Lymph node status was determined according to TNM staging system as follows: 0, negative nodes; 1, one to three positive nodes; 2, four to nine positive nodes; and 3, ≥ 10 positive nodes.
The research protocol was approved by the Research and Ethics Committee of RAK Medical and Health Sciences University and the Ministry of Health in UAE.
The method of immunohistochemical staining for ER, PR, HER2/neu, Bcl-2, E-cadherin and p53 was carried out according to standard techniques: 4-μm sections had been prepared from the blocks, deparaffinized in xylene, rehydrated in a series of graded alcohols and placed in a Tris buffer bath (pH 7.6). Manual immunohistochemical staining was carried out according to the manufacturer’s instructions.
The antibodies used were as follows:
oestrogen receptor: mouse monoclonal antihuman ER (DAKO, Glostrup, Denmark) – Clone 1D5;
progesterone receptor: mouse monoclonal PR (DAKO) – Clone PgR 636;
human epidermal growth factor receptor 2/neu: polyclonal rabbit antihuman c-erbB-2 oncoprotein (DAKO);
B-cell lymphoma 2: mouse monoclonal antihuman Bcl-2 oncoprotein (DAKO) – Clone 124;
E-cadherin: monoclonal MxH E-cadherin (DAKO) – Clone NCH-38;
p53: mouse monoclonal antihuman p53 (DAKO) – Clone DO-7.
Quantification of immunohistochemistry
The ER and PR antibody reactions were graded using the Allred scoring system, which includes scores for staining intensity (0, none; 1, weak; 2, moderate; and 3, strong) and scores for the percentage of positive tumour cell nuclei (0, none; 1, < 1%; 2, 1–10%; 3, 11–33%; 4, 34–66%; and 5, 67–100%). Scores ≥ 3 were considered positive.
Human epidermal growth factor receptor 2 was graded as follows: 0, faint incomplete staining in ≤ 10% of cells; 1, faint incomplete staining in > 10% of cells; 2, weak to moderate complete staining in > 10% of cells; 3, strong complete staining in > 10% of cells. Fluorescence in situ hybridization (FISH) for HER2 amplification was not performed. Bcl-2 staining was considered positive if > 30% of the tumour cells showed cytoplasmic immunoreactivity. Only cytoplasmic staining was scored as positive for Bcl-2, regardless of the intensity of the stained cells.
For E-cadherin, immunoreactivity was interpreted as normal (strong) or aberrant (weak or absent). Aberrant expression was reported when the immunohistochemical reaction was negative or < 70% of the tumour cells showed a positive reaction with membrane pattern. Normal expression was stated when ≥ 70% of the tumour cells showed positive reaction with membranous pattern. p53 staining was interpreted as positive when > 10% of the tumour cells showed distinct nuclear staining. The adjacent normal glandular structures served as internal positive control for ER, PR and E-cadherin.
The studied carcinomas were classified according to ER, PR and HER2/neu expression into four subtypes: luminal A (ER+ and/or PR+, HER2–), luminal B (ER+ and/or PR+, HER2+), HER2/neu-enriched (ER–, PR– and HER2+) and triple-negative breast cancer (TNBC) (ER–, PR– and HER2–).
Statistical analysis was performed with Microsoft Excel 2010, version 14 (Microsoft Corporation, Redmond, WA, USA) and Statistical Package for the Social Sciences, version 13.0 (SPSS Inc., Chicago, IL, USA). The sociodemographic data of the patients were presented in the form of percentage, mean and standard deviation. Statistical significance was set at P < 0.05 and associations between immunohistochemical markers and clinicopathological features, including histological grade, tumour size and lymph node involvement, were assessed using a chi-squared test, analysis of variance and t-test.
This study included 28 patients with primary invasive carcinomas of the breast and, of these, 10 (35.7%) underwent radical mastectomy. The mean age of the patients at diagnosis was 48.75 years ± 10.86 years with a range of 25–74 years. UAE nationals constituted 35.7% of the studied cases and their mean age was 47.85 years ± 16.92 years. A total of 19 (67.9%) tumours were classified as invasive ductal carcinoma, no special type (NST), five (17.9%) as lobular and four (14.3%) as medullary carcinoma. The majority of the tumours were detected in the left breast (15 tumours, 53.6%) and were most frequently located in the central area (11 tumours, 39.3%), followed by the upper outer quadrant (nine tumours, 32.1%). Thirteen (46.4%) of the studied tumours were poorly differentiated and vascular invasion was detected in six tumours (21.4%). The majority (19 tumours, 67.9%) showed histological evidence of carcinoma in situ, found in 5% to 50% of the sections examined. The majority of the patients (15 patients, 53.6%) presented with tumours > 2 cm, whereas only six patients (21.4%) had lymph node metastasis at the time of diagnosis. Collectively, 8% of tumours were stage I, whereas 50% were stage II and 42% were stage III. Table 1 displays the clinicopathological characteristics of the patients and tumours and their relation with the immunohistochemical expression of the studied markers.
LOQ, lower outer quadrant; NA, not applicable; UIQ, upper inner quadrant; UOQ, upper outer quadrant.
Overall, expression of ER, PR, HER2, Bcl-2, E-cadherin and p53 was observed in 12 (42.9%), seven (25%), six (21.4%), 14 (50%), 12 (42.9%) and five (17.9%) tumours, respectively (Figures 1–6). Overall, steroid hormone receptor expression (ER and/or PR) was detected in 50% of the studied tumours. In patients < 50 years, tumours were more frequently ER (62.5%), PR (75%), HER2 (87.5%), E-cadherin (62.5%) and p53 (87.5%) negative. However, the relationship between age at presentation and the immunohistochemical expression of the markers was statistically not significant. Among UAE nationals, higher percentages of tumours showed negative expression of ER (60%), PR (70%), HER2 (60%), Bcl-2 (70%) and p53 (70%).
Lobular carcinoma showed significantly higher expression of ER (four out of five, 80%; P = 0.05) and loss of E-cadherin expression in all the cases (100%; P = 0.006) compared with invasive ductal carcinoma (NST). Conversely, HER2- and p53-overexpressing tumours were more frequently of the invasive ductal carcinoma type (four out of six, 66.7%, and four out of five, 80%, respectively). However, the correlation between histological subtype and the expression of HER2, P53, PR and Bcl-2 was not significant.
Tumours with histological evidence of in situ components were observed to significantly express ER and Bcl-2 (12 out of 19, 63.2%, P = 0.001, and 13 out of 19, 68.4%, P = 0.004, respectively).
Association between tumour grade and ER and PR expression was highly significant as the majority of poorly differentiated tumours were ER and PR negative (84.6%, P = 0.02, and 92.3%, P = 0.01, respectively). Similarly, there was a trend towards loss of E-cadherin expression with increased tumour grade, as 63.6% (7 out of 11) of grade II tumours were E-cadherin positive, whereas 61.5% (8 out of 13) of grade III tumours were positive, but this difference did not reach statistical significance (P = 0.08). Conversely, p53 was exclusively expressed in poorly differentiated tumours, whereas all well and moderately differentiated tumours were p53 negative and the relationship between tumour grade and p53 expression was statistically significant (P = 0.029).
For 16 patients who initially underwent excision and Trucut biopsies, information about the lymph node status was not available. It was observed that a higher proportion of tumours displaying negative expression of ER (four out of seven), PR (6 out of 10), Bcl-2 (three out of five), E-cadherin (three out of five) and p53 (6 out of 11) had axillary lymph node metastasis at the time of diagnosis. However, none of these differences reached statistical significance. Similarly, there were no significant associations between the immunohistochemical expression of the studied markers and the tumour size or AJCC stage.
The frequency of expression and coexpression of the studied markers is shown in Table 2. Coexpression of steroid hormone receptors (ER+, PR+) was detected in five of the tumours (17.9%), while 14 out of 28 (50%) were both ER and PR negative. The remaining nine tumours expressed either ER or PR (seven were ER+ and PR–, two were ER– and PR+). The relationship between ER and PR expression was not statistically significant (P = 0.07).
Oestrogen receptor expression was significantly associated with Bcl-2 expression and 10 out of 12 (83.3%) ER-positive tumours coexpressed Bcl-2, while loss of Bcl-2 expression was evident in 12 out of 16 (75%) ER-negative tumours (P = 0.002). Similarly, there was a strong positive correlation between PR and Bcl-2 expression (P = 0.03).
Although HER2 overexpression was more frequently detected in ER- and PR-negative tumours (66.67% for both), the relationship between HER2 overexpression and hormone receptor expression failed to reach statistical significance.
The relationship between expression of hormone receptors and E-cadherin was not found to be significant; nevertheless, a positive trend was observed that was relatively stronger for PR, as five out of seven (71.4%) PR-positive tumours coexpressed E-cadherin and 14 out of 21 (66.7%) displayed loss of both markers (P = 0.07).
A total of 40% (two out of five) of p53-expressing tumours showed HER2 overexpression compared with 82.6% (19 out of 23) displaying loss of both markers and this difference was highly significant (P = 0.0001). The majority of p53-positive tumours were found to be ER, PR and Bcl-2 negative (80%, 100% and 80%, respectively). However, the correlation between p53 expression and ER, PR and Bcl-2 expression was not significant.
The studied tumours were classified into four breast cancer subtypes: luminal A (ER+ and/or PR+ and HER2–), luminal B (ER+ and/or PR+ and HER2+), HER2 enriched (ER–, PR– and HER2+) and TNBC (ER–, PR– and HER2–). The subtypes accounted for 39.29%, 10.71%, 10.71% and 39.29%, respectively, and UAE national patients presented with equal frequencies of luminal A, TNBC and HER2 enriched subtypes (30% each) while only one patient presented the luminal B subtype. Table 3 shows the clinicopathological characteristics of the molecular subtypes.
With respect to age at diagnosis, no significant differences were found between the subtypes; the luminal A and TNBC subtypes were equally distributed among the studied age groups.
The ductal (NST) histology was more frequently shown in the luminal A subtype followed by the TNBC subtype (42.1% and 36.8%, respectively), while the lobular histology was predominantly shown in the luminal A subtype (three out of five, 60%) and medullary carcinomas were shown in the TNBC subtype (three out of four, 75%). However, the relationship between the histological type and the immunohistochemical subtype was not significant. Likewise, 50% of tumours with histological evidence of ductal in situ component were of the luminal A subtype and the lobular in situ component was exclusively seen among the luminal A subtype (P = 0.04).
The histological grade was significantly associated with the immunohistochemical subtypes. Luminal A subtype constituted the majority of well and moderately differentiated tumours (9 out of 15, 60%) and poorly differentiated tumours appeared most frequently in the TNBC subtype (53.8%) followed by the HER2 enriched subtype (23%) (P = 0.03).
The proportion of tumours that were < 2 cm in size was highest among the TNBC subtype (38.4%), whereas the proportion of tumours > 5 cm was highest among the luminal A subtype (57.1%). Nevertheless, tumour size was not significantly associated with the subtypes. Association between axillary lymph node status and the immunohistochemical subtypes was also not significant; however, a greater percentage (four out of six, 66.7%) of the TNBC tumours were shown in patients with lymph node metastasis. Collectively, a significant number of the TNBC subtype patients (three out of five, 60%) presented with AJCC stage III diseases (P = 0.02).
An ER-positive and PR-negative expression was detected in 50% of luminal A subtype and 33.34% of luminal B subtype, whereas ER-negative and PR-positive expression was shown in 10% and 33.34% of luminal A and luminal B subtypes, respectively.
As shown in Table 4, the luminal subtypes presented a significantly higher percentage of Bcl-2 expression (64.3% in luminal A and 21.4% in luminal B). In contrast, Bcl-2 negativity was predominantly evident in the TNBC subtype (64.3%), followed by the HER2-enriched subtype (21.4%) (P = 0.001). There were no significant associations between E-cadherin and p53 expression and the studied immunohistochemical subtypes.
Overall, patients with luminal A subtype were more likely to be < 50 years of age (7 out of 11; 63.6%). These tumours were more likely to show ductal histology (72.7%) and a prominent in situ component (100%) and to be staged as grade II (63.6%) or AJCC stage II (three out of four; 75%) and more frequently expressed ER (90.0%), E-cadherin (54.5%) and Bcl-2 (81.8%). In addition, tumours of the luminal A subtype showed HER2 overexpression and loss of p53 expression in 100% and 90.9%, respectively. In contrast, luminal B tumours were exclusively found in patients > 50 years of age with equal frequency of ductal, lobular and medullary histogenesis. Two-thirds of these tumours were < 2 cm and the only case with known lymph node status was N3 and AJCC stage III. Negative Bcl-2 expression, negative E-cadherin expression and positive p53 expression were demonstrated in 100%, 66.7% and 0%, respectively.
Patients with the HER2 enriched subtype were more frequently < 50 years of age (two out of three; 66.7%) and presented with poorly differentiated ductal carcinoma. Paradoxically, these tumours were < 2 cm in size with no axillary nodal involvement and stage I disease. Negative Bcl-2 expression, negative E-cadherin expression and positive p53 expression were demonstrated in 100%, 66.7% and 66.7%, respectively.
Patients with the TNBC subtype were more likely to have tumours ≥ 2 cm in size (54.5%) that were grade III (7 out of 11; 63.6%) and AJCC stage III (50%). TBNC tumours were also more likely than the other subtypes to be of the medullary histological type (75%). Negative Bcl-2 expression, negative E-cadherin expression and positive p53 expression were demonstrated in 81.8%, 63.6% and 18.2%, respectively.
Breast cancer is a clinically heterogeneous disease, and existing histological classifications do not fully capture the varied clinical course of this disease. Histological type, grade, tumour size, lymph node involvement and ER and HER2 receptor status all influence prognosis and the probability of response to systemic therapies.27 Immunohistochemistry performed on formalin-fixed, paraffin-embedded cancer tissue is currently the gold standard to assess ER status, which represents an essential part of the routine diagnostic workup for all breast cancer patients and predicts the response to adjuvant endocrine therapy. PR expression is usually reported together with ER; however, its prognostic value is independent of ER status.28
The present study showed immunohistochemical expression of ER and PR in 42.9% and 25% of cases, respectively. This is relatively consistent with the lower ranges reported in some Asian and African countries compared with Western countries. A study carried out in Mafraq Hospital, UAE, reported ER and PR expression in 46.3% and 44.1% of patients, respectively.29
In a study carried out in Yemen, ER and PR immunoexpression was found in 43.8% and 27% of tumours, respectively.30 In addition, a prevalence of 32.6% for ER-positive and 46.1% for PR-positive breast cancers was documented in a study in India.31 A study from Sri Lanka reported the prevalence of of tumours expressing ER and PR to be 45.7% and 48.3%, respectively.32 However, in a Jordanian study, 50.8% of tumours were ER positive and 57.5% were PR positive,33 and, in a Saudi cohort,15 64.6% and 57.3% of invasive ductal carcinomas showed ER and PR expression, respectively.
In contrast to other studies,15,29,30,34 we reported a lower proportion of ER- and PR-positive (17.8%) tumours and a higher proportion (50%) of ER- and PR-negative cases. This discrepancy could be attributed to the different methodologies used to determine ER and PR positivity. Although the threshold for determining ER positivity has been widely debated, it is currently defined by the American Society of Clinical Oncology/College of American Pathologists guidelines as positive nuclear staining in at least 1% of carcinoma cells.35
Another point to be considered is that our study population was heterogeneous and younger than in other series (mean age 48.75 years ± 10.86 years) and, subsequently, the frequency of hormone receptor-negative tumours was higher. Tumours from the UAE national patients included in our study were more frequently ER (60%) and PR (70%) negative. ER-positive carcinomas, in general, are more likely to be of lower nuclear grade and show histological subtypes such as classic lobular, tubular or colloid carcinomas.36 In our series, a significant difference was noted in ER expression among the histological subtypes, as the predominant histological type was the ductal carcinoma (NST) and the majority of these demonstrated loss of ER expression (57.9%), while 80% of tumours of the lobular histological type showed positive ER expression.
Determination of HER2 status is important because overexpression or amplification of HER2 is associated with more aggressive disease, poor clinical outcomes,37,38 lack of response to hormonal therapy and increased response to anthracycline-based chemotherapy.39,40 Perhaps more importantly, HER2 status determines eligibility for targeted therapy with trastuzumab, a humanized monoclonal antibody against HER2.8
The current study observed HER2/neu overexpression in 21.4% of patients, which falls within the range of 20–30% reported in the literature.32,41,42 Among the current cohort, tumours from the UAE nationals demonstrated more frequent HER2 overexpression than the other ethnicities studied. A review of the previous studies in the region showed a somewhat lower percentage (19%) of HER2 overexpression in Oman;43 however, HER2 overexpression was observed in 30.9% and 35.3% of the Yemeni30 and Saudi15 patients, respectively, and even higher frequencies of HER2 overexpression were reported in Iran44 and UAE29 (43% and 49%, respectively). Immunohistochemistry-based detection of HER2 has the advantages of being a routinely performed technique that can be executed rapidly, uses standard light microscopy for interpretation and offers the pathologist simultaneous correlation with morphology. However, many studies have reported inherent potential problems with immunohistochemistry-based detection of HER2 in addition to the multitude of commercially available anti-HER2 antibodies, with evidence suggesting variability in the ability to detect HER2 overexpression.8 In the current study, we used the polyclonal HER2 (A0485) antibody (DAKO), which is reported to have similar concordance with results obtained by FISH.45
When the entire data set of breast tumour samples were analysed for coexpression of ER, PR and HER2, we found a trend towards an inverse correlation between hormone receptor expression and HER2 overexpression. This difference was not significant, although Arafah,15 Ayadi et al.46 and Almasri and Al-Hamad47 reported a significant inverse correlation between hormone receptor expression and HER2 overexpression.
In our study, HER2 overexpression identified a subgroup of breast carcinomas more likely to present in older patients with tumours sized < 5 cm and correlating with ductal histology and poor differentiation; however, none of these differences reached statistical significance. Previous studies reported contradictory results of the correlation between HER2 overexpression and the other prognostic factors. One study48 found that the youngest patients had a two- to threefold increased probability of developing a HER2-positive breast cancer than the oldest breast cancer patients. Some authors reported significant association of HER2 overexpression with tumour size,46,47 grade,41,49 histological type50 and lymph node involvement,51 whereas other studies did not show such significant associations.30,43
Most anticancer agents, independently of their mechanisms of action, kill cancer cells by inducing apoptosis in response to drug-induced damage. Alterations in the regulatory mechanisms of apoptosis are, therefore, responsible not only for the progression of breast cancer, but also for different response to treatment.52 In agreement with previous studies,16,17,43 the present study detected Bcl-2 in 50% of the tumours. Because Bcl-2 blocks apoptosis in vitro and, thus, contributes to malignant cell accumulation, its overexpression is expected to be associated with more aggressive tumour biology. However, our results observed more frequent expression of Bcl-2 in tumours > 5 cm in size as well as grade I and II tumours and those with no axillary lymph node metastasis or with metastasis to 1–3 lymph nodes, and these differences were not significant. When coexpression of Bcl-2 and the other markers was assessed, we demonstrated a highly significant positive correlation with the hormone receptor expression, while HER2, E-cadherin and p53 expression were independent of Bcl-2 expression. Many researchers have reported that Bcl-2 is associated with favourable prognostic features in breast cancer, such as smaller tumour size, well-differentiated tumours, less nodal metastasis, hormone receptor expression, low proliferation status and inverse correlation with p53.16,17,43,53
One possible explanation for this correlation is that the acquisition of Bcl-2 expression creates a restrictive environment for the expansion of genetically unstable and potentially malignant p53-deficient cells, causing a delay in tumour progression and explaining the different prognostic value of Bcl-2 and p53.54 In addition, Bcl-2 is known to be upregulated by oestrogen and to be downregulated by p53.55,56 The role of Bcl-2 antagonists, which negate its cytoprotective function, has also been documented.57
On the clinical front, a wide range of abnormalities of E-cadherin and its complex components have been recognized and characterized in breast cancer. These abnormalities vary from abnormal level of cellular redistribution to mutation of different types, and have strong prognostic values.58 Loss of E-cadherin function contributes to breast cancer progression by promoting cell proliferation, invasiveness, metastasis and recurrence.59 Furthermore, dysfunction of genes involved in cell-to-cell contact is a major common mechanism of drug resistance in breast cancer.60 The present study attempted to investigate the pattern of immunohistochemical expression of E-cadherin. In our cohort, E-cadherin expression was detected in 42.8% of tumours, compared with 57.2% exhibiting aberrant expression, which is consistent with the findings of Suciu et al.,61 who reported 45.5% cases had normal expression and 54.5% aberrant expression. Moreover, we demonstrated complete loss of membranous E-cadherin expression in 100% of infiltrating lobular carcinomas studied, a finding that is also in accordance with previous studies19,58,61,62 reporting complete E-cadherin loss in 86–100% of infiltrating lobular carcinomas. This further validates the role of E-cadherin in identifying infiltrating lobular carcinoma as a distinct entity and explains its histological appearance and distinctive growth patterns during metastases.
Evaluating the relation of E-cadherin expression with the clinicopathological variables studied, we observed that aberrant E-cadherin expression was more frequent in poorly differentiated tumours. In addition, a trend towards a positive coexpression of hormone receptors and E-cadherin was noted; however, except for the histological type, none of the studied variables showed a significant relation with E-cadherin expression. The association of E-cadherin alterations with the well-known prognostic factors has been controversial. Qureshi et al.62 reported that E-cadherin is retained in the majority of non-lobular invasive carcinomas, including poorly differentiated tumours, and is lost in the majority of lobular breast carcinomas irrespective of stage, grade, hormone receptor status, HER2 expression and nodal status. Some authors have found a strong correlation of aberrant E-cadherin expression with increased tumour grade,20,61,63,64 lymph node status20,65 and hormone receptors,66 although others have demonstrated no correlation of E-cadherin expression with tumour size, grade, tubule formation, nuclear pleomorphism, mitotic activity, ER and PR status or HER2 overexpression in invasive carcinomas.67 Interestingly, one study found no association between coexpression of E-cadherin and ER, PR, HER2 or p53 and Ki67 status or tumour grade, whereas coexpressions of E-cadherin and HER2, and of E-cadherin and p53 was significantly associated with differences in lymph node groups to which the tumour had metastasized.20
The tumour-suppressor gene p53 is considered one of the important predictive markers in breast cancer as it gives good information about the resistance to some chemotherapeutic agents and can be used as specific prognostic factor in breast cancer.68 The Catalogue of Somatic Mutations in Cancer reported p53 mutations in approximately 23% of breast cancer samples and it is the second most frequently mutated gene, after the PI3KCA proto-oncogene.69 According to the current release of the International Agency for Research on Cancer TP53 database,70 nearly 70% of breast cancer alterations in p53 are missense mutations.70
The current study detected p53 expression in 17.8% of tumours, compared with 25% reported in UAE,29 41.7% in Oman,43 48.2% in Yemen,30 57.3% in Kuwait71 and 57.4% in Iran.20 However, most of these earlier studies included only infiltrating ductal carcinomas (NST) and used different cut-off values for p53 positivity. Of the five tumours that were proven to be p53 positive in our study, three were diagnosed in patients > 50 years of age, four were of the ductal histology and four were < 2 cm in size; however, these findings were not significant.
The p53 mutations were found to be more frequent in high-grade, large, node-positive tumours and in ER- and PR-negative tumours.72 The present study demonstrated a significant correlation between p53 expression and poor differentiation in the studied tumours, a finding consistent with the results of studies by Ahmed et al.30 and Al-Moundhri et al.,43 whereas other authors have not observed such a significant association with tumour grade.71
The body of literature suggests that coexpression of p53 and HER2 occurs in > 42% of cases73,74 and may have more prognostic significance than the traditional prognostic factors such as T-stage and nodal status. Some authors have documented poor prognosis in patients with coexpression of p53/HER2 while other studies have shown no effect on prognosis or even a better prognosis for breast cancers coexpressing p53 and HER2.75 We found coexpression of p53 and HER2 in 40% of the studied tumours and a highly significant association between p53 and HER2 expression, whereas such coexpression was found in only 28.7% of another cohort and showed no significant association with Ki67 status, lymph node status or tumour grade.20 With regard to the association between p53 expression and the other markers studied in the current cohort, a tendency to an inverse correlation was observed as the majority of p53-positive tumours were ER (80%), PR (100%) and Bcl-2 negative (80%). However, previous studies have shown statistically significant correlations between p53 immunopositivity and lack of ER and PR30,43,71 and a strong inverse correlation with Bcl-2.43
As a result of gene expression assays, there is growing recognition that breast cancer is a molecularly heterogeneous disease.9 Gene expression profiling has led to the discovery of four distinct molecular subtypes of breast cancer: luminal A, luminal B, basal-like and HER2 enriched. Investigation of these subtypes in women with breast cancer has given insight into the heterogeneous biology and outcomes in patients with locally advanced disease. Moreover, these subtypes have been found to be predictors for survival, response to systemic therapy and locoregional recurrence.76
We were able to classify the study cohort into four breast cancer subtypes – luminal A, luminal B, HER2 enriched and TNBC – based on the expression of ER, PR and HER2, as proposed by earlier studies.77,78 The luminal A subtype accounted for 39.29% of the study population, luminal B accounted for 10.71%, the HER2-enriched subtype for 10.71% and TNBC for 39.29% of the study population. This distribution does not concur with the results of a study by Park et al.,24 among others,78–80 who reported luminal A as the predominant subtype and a lower frequency of the TNBC subtype. Racial differences, relatively younger age at presentation and higher frequency of negative hormone receptor expression among our study population are plausible explanations for this discrepancy. Interestingly, the UAE national patients included in our study showed equal frequencies of the luminal A, HER2-enriched and TNBC subtypes (30% each), and a Yemeni cohort30 showed a similar percentage of the TNBC subtype (28.5%).
de Ronde et al.81 recommended that molecular subtyping should not be substituted for traditional prognostic factors, but rather should be used to increase clinical relevance and robustness combined with these classical approaches. Furthermore, a large population-based study using the immunohistochemistry surrogate markers for molecular subtype examined the association of molecular subtype with clinical characteristics and found that the luminal A, luminal B, basal-like and HER2-enriched subtypes differed significantly by age, race, menopausal status, lymph node involvement, histology group, tumour grade and mitotic index.82 In this study, we evaluated the clinicopathological characteristics according to the immunohistochemically defined subtype to enhance our knowledge of breast cancer biology and clinical behaviour in our population. In relation to age, the present study did not identify any significant difference among the subtypes, in particular, the luminal A and TNBC subtypes were equally distributed among the studied age groups. On the other hand, many authors documented older age at onset of the luminal A subtype than TNBC.24,79,80
This study found a significant relation between the histological type and the immunohistochemistry subtype and observed that the majority of the lobular tumours belonged to the luminal A subtype, which is similar to the results of Park et al.24 and Hugh et al.,25 but contrasts the results reported by Onitilo et al.78 We also observed that, among the TNBC subtype, ductal histogenesis was the predominant type, followed by the medullary type; this finding was not surprising, as the TNBC mainly belongs to the molecular basal-like subtype.83 The basal-like subtype itself is heterogeneous and encompasses the majority of breast cancer 1 gene (BRCA1)-related carcinomas, medullary carcinomas and metaplastic carcinomas.84
In keeping with their aggressive nature, the proportion of poorly differentiated tumours in our cohort was highest in TNBC and HER2-enriched subtypes, compared with luminal A tumours, which were predominantly grade I and II, and this difference was found to be statistically significant. These results are consistent with other Asian and Western series.24,25,77,78,85
In our study, we did not identify a significant relation between tumour size and the immunohistochemistry subtypes, which was in contrast with previous studies24,25,79 reporting that TNBC often presents as a large tumour. Additionally, patients with TNBC in our study showed the greatest likelihood of having axillary lymph node metastasis at the time of diagnosis, whereas patients with luminal A tumours presented predominantly with a node-negative form of the disease. However, these differences failed to reach statistical significance, which could be the result of the small number of patients in whom data regarding axillary lymph node metastasis were available.
Nevertheless, evaluating the differences in the distribution of the subtypes with respect to AJCC stage, based on the tumour size and lymph node status, we demonstrated a significant correlation: 60% of stage III tumours were of the TNBC subtype and stage II tumours were equally distributed among the TNBC and luminal A subtypes. In their series, Vallejos et al.79 also reported a significant association between AJCC stage and the molecular subtypes. The associations between the molecular subtypes and axillary lymph node metastasis have been controversial, as some studies24,79,82 have reported that the HER2-enriched subtype is significantly associated with lymph node metastasis. Conversely, some authors have reported that the basal-like subtype shows the highest percentage of node involvement,85 whereas others have demonstrated that the basal-like subtype predicts a lower incidence of axillary nodal involvement and have found a disconnect between tumour size and positive lymph nodes in this subtype.86,87
Consistent with its significant association with favourable prognostic factors, Bcl-2 expression was predominantly demonstrated in the luminal subtypes, with a significantly higher frequency among the luminal A subtypes.
Our observations support the results of earlier studies that report significantly higher p53 expression in TNBC and HER2 enriched subtypes compared with low expressions in well-differentiated and hormone receptor-positive luminal subtype A.25,88,89
In a recent study carried out in 2012 involving a detailed genome-scale analysis of nearly 2000 breast cancer cases, p53 mutations were found in 34% of basal-like, 22% of HER2-enriched, 13% of luminal B and 5% of luminal A molecular subtypes.90 In addition, p53 alterations were found to correlate with poor clinical outcome in HER2-positive cancers.91,92
Our study shed some light on the profile of breast cancer and its associated biomarkers in UAE. Steroid hormone receptor expression and p53 expression were relatively lower than in other series, whereas the expressions of HER2, Bcl-2 and E-cadherin were consistent with the range reported in the literature. Hormone receptor expression correlated significantly with low tumour grade and Bcl-2 expression; however, p53 was significantly associated with poor differentiation and HER2 overexpression, and E-cadherin validated infiltrating lobular carcinoma as a distinct entity. In our cohort, an equal frequency of breast cancers of the luminal A and TNBC subtypes was observed, with a proportion that was significantly different from the other evaluated groups. The molecular subtypes showed significant associations with tumour grade and stage, two of the well-established prognostic factors. Further studies with larger cohorts are warranted to investigate the significance of the studied markers as independent prognostic factors of breast cancer in UAE populations.