Estrogens play a central role in the growth and differentiation of normal breast epithelium, stimulating cell proliferation and regulating the expression of other genes, including the progesterone receptor (PgR) 454647. In the normal pre-menopausal breast, ER(+) cells comprise the 7% of the total epithelial cell population 48. ER(+) cells are luminal epithelial cells, evenly distributed, and seem to secrete factors which paracrinely influence the proliferation of the adjacent ER(-) cells 4849. It should be noted that ER positivity and proliferation (as depicted by ki-67 expression, see below) are almost mutually exclusive in normal epithelial breast tissue 49. ER+ cells increase with age, reaching a plateau after menopause 50.

The intensity of ER expression in the normal epithelium is a risk factor for breast cancer, conferring a 3-fold increase in risk 51. More specifically, regarding non-atypical epithelial hyperplasia, ER positivity together with ki-67 expression, seem to make an important distinction between lesions: the existence of ki-67(+)/ER(+) cells seems to correlate with progression to more severe lesions 52. Alternatively, it has been suggested that an increased percentage of ER(+) cells in the adjacent normal lobules seems to be associated with elevated risk for invasive breast cancer rather than ER-positivity within the lesion per se 53.

In other benign breast lesions, such as sclerosing adenosis, radial scars, papillomas, fibroadenomas, and phylloides tumours, the percentage of ER(+) cells is higher than normal breast tissue 54. Similarly, ER-alpha expression has been proven significantly elevated in the hyperplastic enlarged lobular unit (HELU), which represents the earliest histologically identifiable lesion with premalignant potential. On the contrary, as far as enlarged lobular units with columnar alteration (ELUCA) are concerned, an intriguing result emerged, since more intense ER-alpha staining has been associated with a lower risk of subsequent invasive carcinoma in these lesions 55.

In the context of ADH, LN and DCIS, and contrary to the normal breast, ER(+) cells are surrounded by contiguous cells characterized also by ER-positivity 49. Moreover, in DCIS, the above explained diptych ER(+)/Ki-67(+) is a hallmark 54. In general, non-comedo carcinomas exhibit more frequently ER positivity 5657.

In parallel, the less studied ER-beta, seems also to be of importance and to exhibit an inverse pattern to that of ER-alpha, declining during the progression from normal breast tissue to ADH, DCIS, and IDC 5859. Of notice, according to a recent article, a high ratio ER-alpha/ER-beta in non-atypical epithelial hyperplasia seems to predict progression to carcinoma 60. Finally, for the optimal envisagement of the ER network, it should be kept in mind that important regulators exist, such as hsp-27 61 or AIB1 62, exhibiting more intense expression in breast cancer.

Similarly to estrogen receptors, progesterone receptors (PR) have been found elevated very early in premalignant breast lesions, at the hyperplastic enlarged lobular unit (HELU) 63. On the other hand, a decreasing trend for PR expression has been documented along with progression to malignancy 64. In DCIS, PR positivity is associated with ER positivity and lack of comedo necrosis 656667. Concerning PR positivity, studies are contradictory with respect to tumor grade 65666869 and recurrence rate [707172, reviewed in (72)]. In IDC, progesterone receptors have been associated with histological grade, but not with lymph node involvement, tumor size, or prognosis of the patients 64. With respect to lobular neoplasia, there is scarcity of data; PR seem however to be expressed in the majority of cases 7475.

More interestingly however, it is the ratio progesterone receptor A (PRA)/progesterone receptor B (PRB) which seems to play a central role. In normal breast tissue and non-atypical hyperplasia, receptors are homogenously coexpressed, but early during progression, one receptor (especially PRA in advanced lesions) predominates at a heterogeneous manner 76. Indeed, the physiological role of PRA predominance has been supported by in vitro studies, demonstrating its modulating effects on cell morphology and adhesion 7778. Of notice, in the normal tissue of BRCA mutation carriers, PRB isoform is strikingly absent 79.

HER-2/neu, a gene located on 17q, encodes for c-erbB-2 oncoprotein, a tyrosine kinase receptor. Alterations of c-erbB2 (HER-2/neu) are suggested to be an important event in malignant transformation 808182. According to a variety of studies, c-erbB-2 has not been found overexpressed at the protein level in benign proliferative breast disease or ADH 83848586. However, amplification of Her-2/Neu has been documented with the use of FISH in ADH, supporting the notion that the degree of HER-2/neu amplification increases with progression to carcinoma 87. Accordingly, patients with benign breast lesions showing low levels of amplification of the HER-2 gene have a two-fold increased risk of breast cancer 88; however, according to another study, c-erbB-2 overexpression in benign lesions was not a significant risk factor 89.

With respect to lobular neoplasia, one fourth of LCIS cases have been found positive for c-erbB-2, irrespectively of the coexistence of an invasive component 90. Occasional positivity has been found also in pleomorphic lobular (ductal-lobular carcinomas in situ 91.

As far as the role of c-erbB-2 in DCIS is concerned, C-erbB-2 immunoreactivity has been primarily associated with DCIS of higher grade, in the absence 63 or presence 92 of IDC, and with comedo type 93. Interestingly, given the association of higher grade with c-erb-B2 amplification, the latter has been regarded as an independent prognostic factor 94. Allred et al 95 documented that the percentage of c-erb-B2 immunoreactivity is significantly higher in DCIS than IDC: one of the possible explanations the authors gave was that c-erb-B2 may be more important for the initiation than the progression of breast cancer, or that c-erb-B2 may be downregulated during the progression of breast cancer.

P53 is a tumour suppressor gene located on 17p. p53 protein mediates its tumor suppressor functions via the transcriptional regulation or repression of a variety of genes 969798 and is an important component of breast cancer pathophysiology 99. Regarding the role of p53 as a risk factor in benign breast lesions, there is controversy of data: the immunohistochemical detection of p53 in benign breast lesions has been associated with elevated cancer risk 89, although there are studies with conflicting results 100.

Considering the various types of lesions in the continuum between benign lesions and breast cancer, various studies have assessed the role of p53. In epithelial hyperplasia without atypia, p53 mutations have not been detected 101. In ADH, the presence and role of p53 mutations is still an open field: p53 mutations were initially not documented 102; then, studies pointing to p53 mutations appeared 103, and, more recently, the presence of mutated p53 in ADH has been demonstrated with the use of laser capture microdissection microscope, single-stranded conformational polymorphism (SSCP) and sequencing 104. Regarding LN, there is scarcity of data: in two studies, no p53 immunoreactivity was demonstrated in LN lesions 105106, whereas a more recent one on LCIS reported p53 immunoreactivity in one fifth of cases 90.

p53 mutations/accumulation are present in a significant percentage of DCIS 107108109110111, especially in the comedo type 112. However, the clinical significance of p53 accumulation remains still elusive; although it has been found to influence the proliferation rate 113, a recent study showed that it does not affect the proliferation rate of the DCIS lesion per se 107. Of notice, the coexistence of DCIS with IDC is not associated with a different degree of p53 immunostaining 114.

Ki-67 is a cell cycle-associated nuclear protein, which is expressed in all cycle phases, with the exception of G0 and early G1, and reacts with MIB-1 antibody 115. Protein Ki-67 is extensively used as a proliferative index and is linked with malignancy, even in FNA specimens 116. Moreover, its intrinsic association with apoptosis (bcl-2 status, see below) and p53 expression (see above) seems to be of importance in the diagnosis and prognosis of precursors and preinvasive breast lesions: low Ki-67 expression/bcl-2 positivity and p53 negativity are a trait of ADH and, subsequently, well-differentiated carcinomas. On the contrary, high Ki-67 expression/bcl-2 negativity within the lobules implicate lesions with a potential of poorly differentiated carcinoma 117. As mentioned above, also in the context of non-atypical hyperplasia, high Ki-67 and ER-alpha expression seem to predict progression to cancer 51118.

Interestingly enough, a clinical application of Ki-67 expression intensity seems to emerge. In non-atypical ductal hyperplasia, lesions with high Ki-67 expression can be clinically detected scintimammographically, since high (99m)Tc-(V)DMSA uptake seems to be their feature. According to the authors, this could prove useful in identifying women with benign but high-risk breast pathologies 119.

The bcl-2 gene is located on 18q. Bcl-2 protein, belongs to a family of proteins playing a central role in the regulation of apoptosis [reviewed in (120), 121122] and other pathways [reviewed in (123)]. With respect to the overall role of apoptosis in breast cancer pathogenesis, there seems to be an intriguing pattern incorporating the proliferation of the lesion. Growth imbalance in favour of proliferation seems crucial in the transition from normal epithelium to hyperplasia and later, from preinvasive lesions to IDC. On the contrary, apoptosis becomes more important at an intermediate stage: in the transition from hyperplasia to preinvasive lesions, the imbalance is in favour of apoptosis 124. Bcl-2 is present in the whole spectrum of breast lesions: predominantly in benign lesions, ADH, LN, and well-differentiated DCIS 105125126127. More specifically, there is a gradual increase in the extent of apoptosis 124128 and a parallel decrease in bcl-2 expression in benign/precursors/preinvasive/invasive lesions as they become histologically more aggressive 128. Bcl-2 positivity tends to coincide with p53 negativity in normal breast tissue, non-atypical ductal hyperplasia, ADH, LN and in the majority of the DCIS 105. The role of Bcl-2 expression as a risk factor for breast cancer is described above, together with Ki-67 (see above).

VEGF is a potent angiogenic growth factor, commonly involved in tumor-induced angiogenesis, with a putative therapeutical significance in the context of breast cancer 129. Of notice, VEGF gene polymorphisms have been associated with modified breast cancer risk in various populations 130131.

Viacava et al have thoroughly examined the angiogenesis in precursor and preinvasive lesions. Increased vascularization is present in all preinvasive lesions and increases with lesion severity. In ductal lesions, angiogenesis is more intense in poorly/intermediately differentiated intraductal carcinomas than in non-atypical ductal hyperplasia and ADH. Similarly, LCIS, showing microvascular density similar to that of poorly/intermediately differentiated intraductal carcinoma, is more vascularized than ALH. In the same study, VEGF expression in normal glandular structures was lower than in lesions, with the highest levels found in ductal lesions. Interestingly, no correlation was found between VEGF expression and the degree of vascularization in that study 132. On the other hand, Hieken TJ et al. suggested that VEGF expression may help predict the biologic aggressiveness of DCIS 133. Additionally, in the context of DCIS, Vogl et al support VEGF expression is not regulated by the HER2 pathway 134.

E-cadherin, a tumor suppressor gene located on 17q, has been implicated especially in lobular breast cancer molecular pathogenesis 135. In clinical practice, immunohistochemistry for E-cadherin is a helpful marker for differential diagnosis, since most cases of low-grade DCIS exhibit E-cadherin positivity, whereas LN is almost always E-cadherin negative [136, reviewed in (137), (138)]. This implies that E-cadherin disruption is an early event, prior to progression, in lobular carcinogenesis 139140; more specifically, DNA alterations accompanying the loss of protein expression pertain to LCIS but not to ALH 140. As expected according to the above, only few studies have focused on E-cadherin in ductal lesions. In the context of DCIS, hypermethylation of E-cadherin 5' CpG islands has been demonstrated 141, and, at the protein level, E-cadherin has been linked to better differentiation 142. Moreover, mutational analysis of E-cadherin provided evidence to support that DCIS is the precursor of invasive ductal carcinoma in cases where LCIS coexists 143.

The transforming growth factor-beta (TGF-β) pathway has ambivalent importance in the pathogenesis of breast cancer [reviewed in (144)]. Serum TGF-beta levels do not differ between patients with breast cancer, DCIS and benign lesions 145; however, TGF-beta expression becomes more accentuated in IDC, compared with DCIS 146. Surprisingly enough, an interesting study recently showed that loss of TGF-beta-RII expression in epithelial cells of hyperplasia without atypia is associated with increased risk of IDC 147. No reports exist on ADH and LN, to our knowledge.

p16 is an inhibitor of cyclin-dependent kinases 4 and 6 [reviewed in (148)]. With respect to the role of p16, controversial results exist. According to some authors, aberrant methylation of p16 is not demonstrated in benign conditions, epithelial hyperplasia and intraductal papillomas, but is restricted in cancerous epithelium 149. On the contrary, another study showed that IDC demonstrate hypomethylation of p16 and hyperactivity of the p16 gene (enhanced expression of p16 mRNA), contrary to the hypermethylated, inactive state in the normal epithelium. 150. Independently, Di Vinci et al. distinguish between p16 hypermethylation and p16 protein overexpression; the former seems not to be specifically associated with malignancy and to occur both in benign and malignant lesions, whereas the latter, together with cytoplasmic sequestration, is a feature of breast carcinoma. 151. In the context of such controversy, no studies exist with respect to p16 as a risk factor, with the exception of a study in Poland envisaging p16 as a low penetrance breast cancer susceptibility gene 152.

The p27 gene encodes for an inhibitor of the cyclin – CDK (cyclin-dependent kinase) active complex, Although numerous studies exist with respect to the role of p27 in breast cancer [reviewed in (153), (154), (155)], there is lack of data regarding precursors, preinvasive lesions and other predisposing conditions. p27 expression has been documented in DCIS, but its clinicopathological significance is still uncertain 156.

p21 is a cell cycle regulator, implicated in a variety of pathways 157. p21 immunoreactivity has been detected both in benign and malignant epithelium, and thus its role is hard to interpret. 126. Studies focusing especially on ADH, or LN do not exist. As far as DCIS is concerned, p21 positivity has been independently associated with clinical recurrence 158. On the other hand, Oh YL et al. found a significant correlation between positive p21 immunoreactivity (67.3% of the cases) and well-differentiated histologic grade, non-comedo type, ER-positive and p53-negativity. According to them, DCIS with p21+/p53- is likely to be the non-comedo type 156.

Umbricht et al identified 14-3-3 sigma as a gene whose expression is lost in breast carcinomas, primarily by methylation-mediated silencing. Importantly, the hypermethylation of the locus was absent in hyperplasia without atypia, but was detectable with increasing frequency as the breast lesions progressed from atypical hyperplasia to DCIS, and finally to invasive carcinoma 159; of notice, methylated alleles existed in the periductal stromal breast tissue. Afterwards, a parallel, stepwise reduction at the 14-3-3 sigma protein level has been documented 160.

Despite the emerging role of 14-3-3 sigma in breast carcinogenesis, to date no studies exist assessing its role as a risk factor for breast cancer development.