Karyotypic abnormalities in precursor lesions of human cancer of the breast

Aneuploidy and structurally aberrant chromsomes were identified in most cultures of ductal epithelial cells from proliferative lesions of human fibrocystic disease. These were qualitatively similar to those observed in cells comprising overt invasive mammary cancers. This represents some evidence in support of views indicating that proliferative fibrocystic disease may represent a precursor lesion of cancer of the breast, if not indeed its earliest morphologic expression.     

Microsomal prostaglandin E2 synthase-1 in breast cancer: a potential target for therapy

The anti-tumour actions of cyclooxygenases (COX) are thought to be mediated by inhibition of prostaglandin E(2) (PGE(2)) synthesis. However, COX-2 inhibition also alters cellular production of other prostaglandins such as prostacyclin (PGI(2)). The latter action is believed to be important for the development of adverse cardio-vascular events. Microsomal PGES (mPGES-1) is an enzyme downstream to COX-2 and affects PGE(2) production only. It is possible that targeting mPGES-1 could decrease PGE(2) production without affecting PGI(2) production. In order to assess the potential of mPGES-1 as a target for therapy, we analysed its expression in breast cell lines and normal and malignant breast tissues. The expression of mPGES-1 and COX-2 was correlated in tumour cells and vascular endothelium, and with prognostic parameters in breast cancer. Although not detectable in normal epithelial cells, expression was noted in areas of fibrocystic change and in situ carcinoma. mPGES-1 expression was noted in 79% of breast cancer tissues. Its expression did not correlate with COX-2 overexpression or with prognostic markers of breast cancer. Endothelial cells did not show mPGES-1 expression. Upregulation of mPGES-1 is therefore frequent in pre-malignant and malignant breast disease. In this study, coordinate over-expression of COX-2 and mPGES-1 was not observed, particularly in the endothelial cells of blood vessels. Targeting mPGES-1 might prove to be an alternative therapeutic strategy to inhibit PGE2 production.     

Detection of basement membrane components and basal cell keratin 14 in noninvasive and invasive carcinomas of the breast.

Using immunohistochemistry, the distribution patterns of basement membrane components type VII collagen (monoclonal antibody LH7.2), type IV collagen, and laminin were investigated in normal and malignant human breast tissue and compared with that of keratin 14 (monoclonal antibody LL002), which is expressed only by the basal (myoepithelial) cells in the secretory epithelia of the mammary gland. In normal breast tissue as well as in intraductal carcinomas, a more or less continuous basement membrane was observed at the epithelial stromal interface. Unlike laminin and type IV collagen, type VII collagen was not detected in the basement membrane of blood vessels. The keratin 14 antibody stained the basal cell layer of normal ducts and ducts with in situ cancer. In 85% of the invasive carcinomas no basement membrane or basal cells were detected. In 13 cases, however, laminin, type IV collagen, and/or type VII collagen were detected around tumor nests and individual tumor cells. Five of these tumors also showed a positive reaction with the keratin 14 antibody. In five cases keratin 14 expression was found without detectable basement membrane components. It is concluded that 18 of 103 invasive ductal breast carcinomas examined in this study exhibit a basal cell phenotype as determined from the expression of keratin and the deposition of basement membrane components.     

Immunohistological detection of tissue factor in normal and abnormal human mammary glands using monoclonal antibodies

Summary Tissue factor (TF) is the primary cell-bound initiator of the coagulation protease cascade. The cytological distribution of TF in various tissues may be described on the basis of immunohistochemistry with epitope-defined monoclonal antibodies and the extravascular distribution of TF apparently represents a haemostatic envelope ready to activate coagulation when vascular integrity is disrupted. The present study localized TF in human breast cancer tissues when compared with normal breast gland tissues and benign disorders of the mammary gland. By use of a cocktail of three epitopedefined monoclonal antibodies, TF was detected only in the myoepithelia of the resting breast gland. In proliferating disorders like fibrocystic disease or in fibroadenomas, both myoepithelia and luminal epithelia showed TF expression. Of 115 breast cancers 93 reacted with anti-TF, in an inhomogeneous manner in terms of intensity and number of positive cells. There was a tendency for more positive and intensely stained cells to be found in well-differentiated structures such as tubules. Invasive ductal carcinomas exhibiting more positive and more strongly stained cells were less commonly metastatic to lymph nodes when compared with the tumours with no detectable or very low TF immunostaining. A semiquantitatively recorded score of TF immunostaining correlated with the procoagulatory activity measured (7 fibroadenomas and 24 carcinomas). The results of this study suggest that proliferation and differentiation of the mammary gland is associated with enhanced TF expression in the epithelia which are negative for TF staining in the resting gland. Malignant growth is characterized by randomly expressed epithelial TF, which expression is enhanced and more frequent in well-differentiated tumour cells.     

A mouse model of basal-like breast carcinoma with metaplastic elements

Breast cancers arising in carriers of germline BRCA1 mutations frequently have a basal-like phenotype. Basal-like cancers are characterized by high histological grade, central necrotic areas, foci with metaplastic differentiation, lack of hormone receptor and HER2 (ErbB2) expression, and consistent positivity for basal markers, including CK5/6, CK14, and EGFR. We have used germline manipulation to generate a conditional mouse model of Brca1 deficiency. Transgenic expression of Cre recombinase in the mammary gland of these mice results in deletion of exons encoding the C-terminus of Brca1 and leads to tumour formation when combined with heterozygosity for a p53 mutation. Histologically, these mammary gland tumours were characterized by high histological grade, central necrotic areas, and presence of homologous metaplastic elements. These metaplastic elements consisted of neoplastic spindle cells or squamous cell differentiation in the form of keratin pearls or individual cell keratinization. Immunohistochemical analysis revealed expression of basal-like markers in all cases. The tumour phenotype generated in our mouse model was compared with published data on human basal-like breast carcinomas and also with metaplastic breast cancers with a basal-like phenotype; the comparison showed that we have generated a mouse model of basal-like breast cancer, which should prove useful in testing new and targeted treatments for this type of breast cancer.     

Comparative analysis of peptidylarginine deiminase-2 expression in canine, feline and human mammary tumours

The peptidylarginine deiminase (PAD) enzyme family converts arginine residues in proteins to citrulline. In the canine mammary gland, PAD2 expression is first detected in epithelial cells in oestrus and becomes more widely expressed during dioestrus. PAD2 appears to modify nuclear histones, suggesting a role for the enzyme in chromatin remodelling and gene regulation. Recent evidence suggests that PAD2 plays a role in gene regulation in primary human breast epithelial cells. PAD2 may therefore be involved in gene regulation as it relates to mammary development, the oestrus cycle and potentially to neoplasia. The aim of the present study was to determine whether PAD2 expression was increased or decreased in mammary carcinoma compared with normal mammary tissue. A human mammary tissue microarray and archival surgical biopsy tissues from canine and feline mammary tumours were used to demonstrate differential expression of PAD2 in mammary carcinoma that appeared to be consistent across species. Normal human and canine mammary epithelium showed strong cytoplasmic and nuclear expression of PAD2, but there was reduced PAD2 expression in mammary carcinomas from both species. Feline mammary carcinomas had complete loss of nuclear PAD2 expression. Loss of nuclear PAD2 expression may therefore represent a marker of progression towards more aggressive neoplasia.     

Keratinocyte Growth Factor Causes Cystic Dilation of the Mammary Glands of Mice: Interactions of Keratinocyte Growth Factor, Estrogen, and Progesterone In Vivo

Keratinocyte growth factor (KGF) is a paracrine mediator of epithelial cell proliferation that has been reported to induce marked proliferation of mammary epithelium in rats. In this study, systemic administration of KGF into naive and oophorectomized mice causes mammary gland proliferation, as evidenced histologically by the appearance of cysts lined by a single layer of epithelium and by hyperplastic epithelium. Whole mount preparations of the mammary glands reveal that the histologically noted cysts are actually ducts that are dilated along much of their length. The histology of the mammary glands of KGF-treated mice is similar to the histology of fibrocystic disease in the buman female breast. The response in mice differs significantly from the appearance of the mammary glands in KGF-treated rats in which ductal epithelial proliferation is most prominent. Estrogen and progesterone when administered in combination but not alone cause the development of numerous endbuds in the mouse mammary gland. KGF in estrogen- and progesterone-pretreated mice causes the growth of dilated ducts, hyperplastic epithelium within ducts and endbuds, and a fibrous metamorphosis of periductal adipose tissue. The mammary epithelial hyperplasia caused by KGF is rapidly reversible in both mice and rats after cessation of KGF treatment. The spectrum of KGF-, estrogen-, and progesterone-induced mammary histopathology in mice provides a model for the study of fibrocystic and hyperplastic breast disease.     

Chemopreventive effect of celecoxib and expression of cyclooxygenase-1 and cyclooxygenase-2 on chemically-induced rat mammary tumours

We investigated the chemopreventive effect of celecoxib on 7,12-dimethylbenz[a]anthracene (DMBA)-induced rat mammary tumours and also the expression and immunolocalization of cyclooxygenase-1 (COX-1) and COX-2 in the various stages of rat mammary carcinogenesis. Rats were divided into normal control group, DMBA-control group, 500 p.p.m. celecoxib-treated group, and 1500 p.p.m. celecoxib-treated group. Both incidence and multiplicity values of tumour for rats treated with celecoxib were less than those in rats of the DMBA-control group. The level of prostaglandin E2 was higher in tumours of the DMBA-control and both celecoxib-treated groups compared to normal mammary glands of each group. In Western blot analysis, all tumours of the DMBA-control group expressed COX-1, whereas normal mammary glands showed insignificant expression. COX-2 expression was observed in 67% of the DMBA-control group and 20% of both celecoxib-treated groups and was absent in normal mammary glands. COX-1 protein was localized in the nuclear membrane and cytoplasm of epithelial tumour cells abutting on glandular lumen, stromal cells, and endothelial cells. COX-2 protein was detected in the perinuclear cytoplasm of tumour cells bordering on glandular lumen and surrounding stroma, stromal cells, and vascular smooth muscle. In the DMBA-control group, invasive carcinoma cells showed higher positive immunoreactivity of COX-2 than carcinomas in situ and atypical tumours. Tumours displayed an increased number of mast-like cells with COX-2 expression in comparison to carcinomas in situ. Our results suggest that COX-1 and COX-2 expression in tumour cells and stromal cells play an important role in the various stages of DMBA-induced rat mammary carcinogenesis. In addition, we reconfirm that celecoxib reduces the growth of DMBA-induced rat mammary tumours.     

Immunohistochemical localisation of keratin and luminal epithelial antigen in myoepithelial and luminal epithelial cells of human mammary and salivary gland tumours

Rabbit antisera to human 40-63 000 MW epidermal keratin, one batch with restricted distribution of reactivity from an initial (aK1) and one with "broad spectrum" distribution of reactivity from a late bleeding (aK), and to "luminal epithelial antigen" (aLEA) were applied to formalin fixed paraffin embedded sections of human normal and neoplastic mammary and salivary glands using an indirect immunoperoxidase method. aK1 reacted with myoepithelial cells, aLEA with luminal epithelial cells and aK with both cell types in normal mammary and salivary gland. In breast carcinomas the majority of intraluminal and infiltrating carcinoma cells reacted with aLEA but not with aK1 which reacted only with surrounding myoepithelial cells. aK reacted with both myoepithelial cells and with intraluminal and infiltrating tumour cells. In the salivary gland adenomas the majority of cells reacted with aK, and those cells arranged in a tubular fashion reacted with aLEA.     

Hormone-responsive 3D multicellular culture model of human breast tissue

A hormone-responsive 3D human tissue-like culture system was developed in which human primary mammary epithelial cells (MECs) were co-cultured with two types of predominant mammary stromal cells on silk protein scaffolds. Silk porous scaffolds with incorporated extracellular matrix provided a compatible environment for epithelial structure morphogenesis and differentiation. The presence of stromal cells promoted MEC proliferation, induced both alveolar and ductal morphogenesis and enhanced casein expression. In contrast, only alveolar structures were observed in monocultures. The alveolar structures generated from the heterotypic cultures in vitro exhibited proper polarity similar to human breast tissue in vivo. Consistent with their phenotypic appearance, more functional differentiation of epithelial cells was also observed in the heterotypic cultures, where casein-α and -β mRNA expression were increased significantly. Additionally, this 3D multicellular culture model displayed an estrogen-responsive physiologically relevant response, evidenced by enhanced cell proliferation, aberrant morphology, changes in gene expression profile and few polarized lumen structures after estrogen treatment. This culture system offers an excellent opportunity to explore the role of cell-cell and cell-substrate interactions during mammary gland development, the consequences of hormone receptor activation on MEC behavior and morphogenesis, as well as their alteration during neoplastic transformation in human breast tissue.     

Co-expression of tenascin-C and vimentin in human breast cancer cells indicates phenotypic transdifferentiation during tumour progression: correlation with histopathological parameters, hormone receptors, and oncoproteins

Loss of epithelial morphology and the acquisition of mesenchymal characteristics are typical for carcinoma cells in tumour progression. In human breast carcinomas, up-regulation of tenascin-C (TN-C) and vimentin (Vim) is frequently observed in cancer cells and correlates with increased malignancy. Thus, it is possible that TN-C is co-expressed with Vim, representing cancer cells that have undergone epithelial-mesenchymal transition (EMT). This study examined 128 breast carcinomas using immunohistochemical techniques to demonstrate that mammary cancer cells are a prominent source of both TN-C and Vim. Statistical analysis revealed a significant association between TN-C and Vim expression in cancer cells. TN-C expression also correlated positively with overexpression of c-erbB-2 oncoprotein and down-regulation of oestrogen receptors (ERs). Eleven human mammary cancer cell lines and two 'normal' cell lines were examined by western blotting and immunohistochemistry. Co-expression of TN-C and Vim was detected in the carcinosarcoma cell line HS 578T, SK-BR-3 (B), fibroblast-like MDA-MB-231 cells, and the myoepithelial cell line HBL 100. These findings suggest that TN-C and Vim, when co-expressed in mammary carcinoma cells, represent regulator genes likely to be involved in EMT during mammary carcinogenesis.     

Proliferation of estrogen receptor-alpha-positive mammary epithelial cells is restrained by transforming growth factor-beta1 in adult mice

Transforming growth factor (TGF)-beta1 is a potent inhibitor of mammary epithelial proliferation. In human breast, estrogen receptor (ER)-alpha cells rarely co-localize with markers of proliferation, but their increased frequency correlates with breast cancer risk. To determine whether TGF-beta1 is necessary for the quiescence of ER-alpha-positive populations, we examined mouse mammary epithelial glands at estrus. Approximately 35% of epithelial cells showed TGF-beta1 activation, which co-localized with nuclear receptor-phosphorylated Smad 2/3, indicating that TGF-beta signaling is autocrine. Nuclear Smad co-localized with nuclear ER-alpha. To test whether TGF-beta inhibits proliferation, we examined genetically engineered mice with different levels of TGF-beta1. ER-alpha co-localization with markers of proliferation (ie, Ki-67 or bromodeoxyuridine) at estrus was significantly increased in the mammary glands of Tgf beta1 C57/bl/129SV heterozygote mice. This relationship was maintained after pregnancy but was absent at puberty. Conversely, mammary epithelial expression of constitutively active TGF-beta1 via the MMTV promoter suppressed proliferation of ER-alpha-positive cells. Thus, TGF-beta1 activation functionally restrains ER-alpha-positive cells from proliferating in adult mammary gland. Accordingly, we propose that TGF-beta1 dysregulation may promote proliferation of ER-alpha-positive cells associated with breast cancer risk in humans.     

Functional consequences of cyclin D1 overexpression in human mammary luminal epithelial cells

The proliferation of eukaryotic cells is primarily regulated by a decision made during the G1 phase of the cell cycle as to remain in the cycle and divide, or to withdraw from the cycle and adopt a different cell fate. During this time, environmental signals, which regulate the synthesis of the G1 cyclins, are coupled to cell division. In this context, mammalian D-type cyclins have been shown to control progression through the G1 phase of the mammalian cell cycle. Specifically, cyclin D1 has been reported frequently to be amplified, over-transcribed and overexpressed in human breast carcinomas. Although the effects of cyclin D1 overexpression have been examined in human breast carcinoma cell lines, the biological consequences of cyclin D1 expression in normal human mammary epithelial cells remain to be elucidated. In this study we have stably over expressed cyclin D1 in human mammary luminal epithelial cells in order to more directly address the role of cyclin D1 in cell cycle control and tumorigenesis of the human breast. Here, we demonstrate that the effect of cyclin D1 overexpression in these cells is to reduce their growth factor dependency, as well as shorten the duration of G1 and correspondingly reduce the mean generation time. Collectively, our data indicate that deregulation of cyclin D1 expression in human mammary epithelial cells can provide a growth advantage and hence contribute to the oncogenic potential of these cells.     

Immunohistochemical COX-2 overexpression correlates with HER-2/neu overexpression in invasive breast carcinomas: a pilot study

Cyclooxygenase-2 (COX-2) is a prostaglandin synthase that catalyzes the synthesis of prostaglandin G2 and H2. It has been shown that COX-2 plays an important role in tumorigenesis of different tumor types and it is thought to take part in breast carcinogenesis. In the present study, we aimed to investigate the relationship of immunohistochemical COX-2 expression with clinicopathological parameters, including HER-2/neu overexpression in invasive breast carcinoma (IBC).Our study population comprised 10 normal breasts, 25 ductal carcinomas in situ (DCIS), and 51 invasive breast carcinomas. Immunohistochemical overexpressions of COX-2 and HER-2/neu were investigated in sections of formalin-fixed, paraffin-embedded blocks by 3 observers.In normal breast, DCIS and IBC, the COX-2 overexpression rate was 0%, 84%, and 58.8%, respectively. In IBC, COX-2 overexpression had a significant relationship with HER-2/neu overexpression (p = 0.026) and a high histological grade (p = 0.026). COX-2 expression in both DCIS (n = 25) and IBC (n = 51) was significantly higher than in normal breast tissue (p < 0.0001). In addition, the COX-2 expression rate was significantly higher in DCIS than in IBC (p = 0.042).Our results indicated that COX-2 overexpression correlates with aggressive phenotypic features, such as HER-2/neu overexpression and high histological grade in IBC. Increased expression of COX-2 in both DCIS and IBC in comparison to normal breast could indicate a role in breast carcinogenesis. COX-2 overexpression may provide a clinically useful biomarker for estimating tumor aggressiveness.     

Expression of parathyroidlike protein in normal, proliferative, and neoplastic human breast tissues.

Parathyroidlike peptide (PLP), or parathyroid hormone-related protein, is a protein of uncertain biological function that is structurally homologous to parathyroid hormone. Immunohistochemical studies have identified aminoterminal epitopes of PLP in breast carcinomas, but not in normal breast. In the present studies, immunohistochemistry and in situ hybridization were performed to evaluate further expression of PLP in normal, proliferative, and neoplastic breast tissues. Using a polyclonal antibody that recognizes epitopes within the middle and carboxyl-terminal domains of PLP, immunoreactive protein was detected within the cytoplasm of lobular and ductal epithelial cells in all normal and fibrocystic breast tissues from 74 patients. The intensity of cytoplasmic staining was increased in association with lactation, adenosis, and simple or atypical ductal hyperplasia and decreased in atrophic lobules. Cytoplasmic reactivity was also observed in 69% (56 of 81) of breast adenocarcinomas. Expression of immunoreactive PLP was inversely correlated with tumor stage and extent of nodal involvement at the time of diagnosis. However, there was no significant correlation with tumor grade, patient age, or hormone receptor status. In situ hybridization studies confirmed the epithelial expression of PLP messenger RNA in PLP-positive normal and neoplastic breast tissue. Interestingly, tumor-associated calcifications were identified in 43% of PLP-positive carcinomas, but in only 12% of PLP-negative carcinomas (P < 0.007). Our results suggest that PLP plays some role in the normal differentiated function of mammary epithelial cells and are consistent with the hypothesis that expression of this protein influences local calcium metabolism.     

A comparison of the pattern of cathepsin-D expression in fibroadenoma, fibrocystic disease, preinvasive and invasive ductal breast carcinoma.

In the metastatic process, proteolytic enzymes play an important role in mediating the passage of cancer cells through the basement membrane and extracellular matrix. We have compared cathepsin-D (CD) expression in a range of benign and malignant breast lesions so as to investigate its role in breast cancer progression. One hundred and sixty-two breast samples, comprising 18 fibroadenomas, 22 fibrocystic disease, 96 invasive ductal carcinoma and 26 lesions with intraductal carcinoma components, were evaluated for CD expression by the standard avidin-biotin-immunoperoxidase complex method on formalin-fixed, paraffin-embedded histological sections using a commercial antibody against human cathepsin-D. Of the invasive ductal carcinomas, 61.5% showed stromal cell CD positivity, whereas 48.9% expressed CD positivity in neoplastic cells. There was significant correlation between neoplastic cell and stromal CD positivity. The prevalences of CD positivity in both neoplastic and stromal cell components were significantly higher (P < 0.05 and P < 0.01, respectively) in histological grade III tumors compared to grades I and II carcinomas. CD expression by either neoplastic or stromal cells did not show significant correlation with patient age and tumor size. Only 15% of intraductal carcinomas were CD positive and expression was limited to neoplastic cells. Neither epithelial nor stromal cells in fibrocystic lesions and fibroadenomas were CD positive, but a weak to moderate positivity was observed within myoepithelial cells in mammary ducts. These findings provide insights into the mechanism whereby tumors with high histological grade mediate invasion into tissue. The role of stromal cells in tumor progression and the means of their recruitment deserve further study.