|Home | About | Journals | Submit | Contact Us | Français|
BLBC represents a distinctive group of invasive breast carcinomas with specific genotype and immunopro-file. BLBC is usually defined by gene expression profiling and is currently associated with poor outcome. BLBCs are estrogen receptor (ER) negative, progesterone receptor (PgR) negative, HER2 negative, and usually show a variable expression of basal cytokeratins (CKs), EGFR and CD117. p16 INK4a is a tumor suppressor protein, encoded by the CDKN2A gene, which regulates cell cycle. The reported association of abnormalities in the p16/Rb pathway with increased risk of malignancy prompted us to determine the expression of p16INK4a in a group of BLBC; the results were compared with a group of high-grade invasive carcinoma (HG-IC) of breast. Tissue microarrays (TMA) were constructed in triplicate including 18 BLBC and 18 HG-IC. All BLBC cases were ER-/PgR-/HER2-. Seventeen (94%) BLBC were CK 5/6+/CK 14+; 14 (78%) BLCB showed EGFR expression and 13 (72%) were CD117 positive. BLBCs showed a strong positive reaction with p16 INK4a antibody in 16 of 18 (89%) cases. Although the significance of p16 INK4a expression in breast cancer is not fully understood, we have shown that p16INK4a is strongly expressed in breast cancers with basal-like phenotype. Since it is known that p16INK4a is associated with aggressive behavior in human carcinomas, these data suggest that p16INK4a play a role in the poor prognosis of BLBC.
Microarray studies have allowed the identification of molecular breast cancer subtypes, distinguished by differences in their gene expression profile, which usually correlate with prognosis and response to the therapy . Triple-negative breast cancer (TNBC) is defined by the lack of expression of estrogen receptor (ER), progesterone receptor (PgR), and human epidermal growth factor receptor 2 (HER2) . The basal-like breast carcinoma (BLBC) class is a molecular subtype of breast cancer that usually shows a triple-negative phenotype  and constitute about 15% of invasive cancers . No current targeted therapies are available for BLBC. Although BLBC and TNBC are not synonymous, their morphological features overlap and are not enough to provide a distinction between the tumors. BLBCs usually show a variable expression of basal cytokeratins (CKs) (CK5/6, CK14, CK17) and/or overexpression of human epidermal growth factor receptor 1 (HER-1/EGFR) and CD117 (c-Kit) [4-5]. The purpose of identifying BLBCs is to provide insight in terms of prognosis, treatment and response to therapy.
p16/Retinoblastoma (Rb) pathway abnormalities have been widely reported in human neoplasm and associated with increased risk of malignancy [6-10]. Rb protein is a tumor suppressor that arrests cell cycle at G1. p16INK4a is a tumor suppressor protein, located at chromosome 9p21, encoded by the CDKN2A gene, which regulates the transition from the G1 to the S phase of the cell cycle . p16 INK4a protein is a cyclin-dependent kinase (cdk) inhibitor that decelerates the cell cycle, inactivating cdks that phosphorylate Rb. Inactivation of p16 INK4a results in loss of inhibition of Rb phosphorylation, facilitating loss of control of cell cycle arrest . Interestingly, p16INK4a overexpression has been found in high-grade carcinomas of the oropharynx, genital and genitourinary tract, in association with human papilloma virus (HPV). The reported association of abnormalities in the p16/Rb pathway with increased risk of malignancy prompted us to determine the expression of p16INK4a in a group of BLBC; the results were compared with a group of high-grade invasive carcinoma (HG-IC) of breast. We examined the significance of p16INK4a positivity in carcinomas of the breast, and its utility for identification of BLBC.
Cases were selected based on the availability of paraffin blocks. In total, 18 cases of high-grade invasive carcinoma (HG-IC) and 18 cases of BLBC were retrieved from the files of the Department of Pathology of UPAEP University Hospital, Mexico, after approval from the IRB. The patients were diagnosed at the hospital between 2006 and 2008. All tumor slides were reviewed, and morphological parameters including mitotic activity, necrosis and Nottingham Grade were recorded. Donor blocks were prepared after evaluation of Hematoxylin-eosin (H&E) slides. Tissue microarrays (TMA) were constructed in triplicate for both HG-IC and BLBC by extracting cores of invasive breast carcinoma from the original paraffin blocks and re-embedding the cores into a receptor block. Tissue controls (smooth muscle and liver) were included in the blocks. In addition, unstained whole tissue sections were also prepared for BLBC.
Each set of tissue sections (5 μm) were deparaf-finized and rehydrated through a series of graded ethanols. The slides were stained with commercially available antibodies for ER, PgR, HER2, CK5/6, CK14, EGFR, CD117, p53, Ki67 and p16 INK4a. Immunohistochemistry for HER2 was performed using A0485 antibody (prediluted, DakoCytomation, Carpinteria, CA, USA), according to manufacturer's instructions. HER2 expression was scored according to membranous staining of tumor cells: negative (0/1+), equivocal (2+) and positive (3+). ER (clone SP1, 1:100, NeoMarkers, Fremont, CA, USA) and PgR (clone SP2, 1:150, NeoMarkers, Fremont, CA, USA) antibodies were used. Tumors were considered positive if there are at least 1% positive tumor nuclei. Internal and external controls were used. Tumors were also stained with CK5/6 (clone D5/16B4, 1:100, DakoCytomation, Carpinteria, CA, USA), CK 14 (clone LL002, 1:20, NeoMarkers, Fremont, CA, USA), EGFR (clone H11, prediluted, DakoCytomation, Carpinteria, CA, USA), CD117 (polyclonal, 1:100, DakoCytomation, Carpinteria, CA, USA), Ki67 (Clone MIB-1, 1:150, DakoCytomation, Carpinteria, CA, USA) and p53 (clone Rabbit SP3, 1:100, NeoMarkers, Fremont, CA, USA). CK5/6 and CK14 were scored positive if any cytoplasmic and/or membrane staining was observed and a percentage was given. EGFR was considered positive if any strong membranous staining was seen, and CD117 was scored positive if any cytoplasmic staining was present. p16 INK4a antibody was used (1:20, DakoCytomation, Carpinteria, CA, USA); cases were scored based on percentage of positive invasive tumor cells showing nuclear and/or cytoplasmic staining. Positive and negative controls were applied. To detect p16INK4a, CK5/6, CK14, ER and PgR, tissue sections were pretreated with Trilogy retrieval buffer in sequence in steam heat; for HER2 and EGFR, Citrate buffer and proteinase K were used, respectively. The sections were stained using DAKO autostainer. The slides were reviewed by two pathologists (OB, SS) and consensus was achieved.
Statistical analyses were performed using the Fisher exact test. A 2-sided P value of <0.05 was considered as significant.
The majority of BLBC tumors were high grade as expected, with Nottingham grade 3 in 14 cases (78%) (Figure 1A) and grade 2 in the remainder cases as shown in Table 1. Necrosis was seen in 10 of 18 cases (55.5%) (Figure 1B). p53 expression was seen in 14 of 14 cases (100%), ranging from 5 to 100% (Figure 1C), and Ki67 showed a high proliferative index in 13 of 15 cases (86.6%), ranging from 10 to 90% (Figure 1D). IHC results are summarized in Table 2 and Table 3. All eighteen (100%) BLCB cases lacked of expression for ER, PgR and HER2. Seventeen (94%) BLBC demonstrated membranous and cytoplasmic staining co-expression for CK 5/6 and CK 14 (94%) (Figure 2A and and2B);2B); fourteen (78%) BLBC showed EGFR expression and thirteen (72%) were CD117 positive. Sixteen BLBC showed diffuse nuclear and cytoplasmic expression of p16INK4a (89%), in 80% to 100% of the tumor cells (mean 80.5%) (Figure 2C and and2D).2D). Either basal CKs, EGFR or both were expressed in all p16 INK4a cases. In order to exclude the possibility of immunophenotypic heterogeneity that have been previous described with regard to basal CKs, c-Kit and EGFR, we decided to use whole tissue sections in BLBC. The results using whole tissue were consistent to those found in the TMA.
The results are presented in Table 2 and Table 3. Among eighteen invasive carcinomas, two were considered metaplastic-type and the remainder of cases, invasive ductal carcinoma Grade III, NOS type. Thirteen cases (72.2%) were TNBC (ER-/PR-/HER2-), including the two metaplastic carcinomas. Twelve of 18 cases (66.6%) lacked of immunoreactivity for CK5/6 and CK14. Tumors expressing CK 5/6 or CK14 were ER negative. Sixteen of 18 (88.8%) lacked of CD117 expression. p53 expression was seen in 16 of 16 cases (100%), ranging from 15 to 90%, and Ki67 showed a high proliferative index in 16 of 16 cases (100%), ranging from 20 to 90%. Diffuse nuclear and cytoplasmic p16INK4a staining was present in 10 of 18 (55.5%) HG-IC, in 5% to 100% of the tumor cells (mean 26.3%). Only one of the three HER2 positive cases (3+) showed reactivity for p16 INK4a. Eight of 10 cases expressing p16 INK4a were negative for EGFR, CK5/6, CK14 and CD117.
BLBC had higher and stronger p16 INK4a expression (89%) compared with the HG-IC (55%). The difference was at borderline statistical significance (p=0.05). Similarly, BLBC cases showed higher expression of CK5/6, CK14, EGFR and CD117 than HG-IC, and it was statistically significant (p<0.0001).
Gene profiling studies have provided new insights into the classification of invasive breast cancer, prognosis and benefit to therapy [1, 13-15]. To the date, five molecularly distinct breast cancer subtypes have been identified, including Luminal A, Luminal B, HER2 group, normal breast-like and the basal-like group. Currently, BLBC is defined by gene profiling (which is considered the gold standard for identification of basal-like cancers  and/or immunohistochemistry . Basal-like cancer is a heterogeneous group that do not cluster with ER or HER2 tumors , associated with poor outcome and limited therapeutic tools [18-19]. In a study, Fulford etal  reported that the specific morphologic features including the presence of a central scar, tumor necrosis and high mitotic count are strongly associated with BLBC. Although these findings are in agreement with our observations (certain histological features may strongly suggest the basal-type), the phenotype is not pathognomonic of BLBC-subtype, and we did not identified specific histological features that distinguish basal-like cancers from high grade invasive carcinomas. Because of the poorly-differentiated morphology associated with BLBC, the use of morphological features alone may misclassify cases as such, affecting the results of many studies in relationship to prognosis and therapy. Interestingly, recent studies have shown that BLBC possess the worse overall survival and disease free survival when compared to the other molecular subtypes [2, 17].
We selected a group of basal CKs usually expressed in BLBC (CK5/6 and CK14), and CD117 and EGFR, because these biomarkers have been used in previous studies as part of the immunohistochemical definition of BLBC. [19, 21] These CKs have been strongly associated with poor prognosis [22-23]. Bryan etal  have reported immunophenotypic heterogeneity with basal CKs, CD117 (c-Kit) and EGFR. To avoid discrepancies in the results, we assessed whole tissue sections, in addition to the breast cancer TMA slides; the results were consistent, validating the higher expression of CKs in basal cancer subtype when compared to HG-IC. We have also found that the majority of BLBC showed high proliferative Ki67 index and demonstrated p53 overexpression, but the differences between BLBC and HG-IC were not statistically significant.
We evaluated p16 INK4a expression by immunohistochemistry on TMAs. This study has shown that strong p16 INK4a positivity coupled with co-expression of basal CKs (CK5/6 and CK14) is seen in a high number of BLBC cases. Interestingly, the ratio of p16-positive cases was higher in the BLBC group than in HG-IC group. The importance of p16 INK4a expression is not fully understood in breast cancer. p16 INK4a expression has been studied in human carcinomas. For instance, in a study, Hui etal  demonstrated an inverse relationship between p16INK4aand ER mRNA levels in cell lines and primary breast cancers, suggesting that p16INK4a inactivation by hypermethylation and overexpression is a marker of poor prognosis. Similary, Milde-Langosch et al  have described high p16 INK4a reactivity (both nuclear and cytoplasmic) as indicative of a more undifferentiated phenotype in mammary carcinomas. The reciprocal correlation between p16/Rb has been observed in several human HPV-associated malignancies. The functional loss of Rb1 has been found in basal-like tumors, playing a possible key role in their aggressive behavior and metastatic potential . Subhawong et al  demonstrated that some breast carcinomas lacking of Rb expression and positive immunoreactivity for p16 INK4a, showed histological features similar to HPV-related squamous cell carcinomas. However, we did not identify morphologic parameters that suggest HPV infection. Although our study lacks of HPV infection data, several studies have shown the utility of p16 INK4a to distinguish HPV-related lesions in different organs including the cervix and the urinary tract, and further studies regarding this issue are required.
Finally, we propose p16 INK4a as a biomarker for identification of truly basal-like cancers. We are aware that the limitations of this study are due to the number of cases, however, but it raises the possibility that TNBC with basal CKs and p16 INK4a co-expression may adequately identify BLBC, and serve as a potential diagnostic/ prognostic biomarker. p16 INK4a overexpression is easily detectable with immunohistochemical stains; therefore is a feasible and simple tool for routine practice.
In summary, although many phenotypic features of BLBC are commonly seen in other high grade breast cancers, BLBC represents a distinctive group of invasive breast carcinomas with specific genotype and immunoprofile. Although the significance of p16INK4a expression in breast cancer is not fully understood, we have shown that p16INK4a is strongly expressed in breast cancers with basal-like phenotype. Since it is known that p16INK4a is associated with aggressive behavior in human carcinomas, these data suggest that p16INK4a play a role in the poor prognosis of BLBC; therefore, the presence of this abnormality may be used in the development of a targeted therapy in such group.