The aim of this study was to determine alterations in the expression of proteins involved in proliferation and cell cycle in HUT cells of patients with and without invasive breast cancer. Our analysis showed no difference in the cell cycle related proteins immunoexpression in HUT from the three different groups, in spite of age and menopausal status. Similar results were obtained by Gobbi et al. (2005) evaluating ER expression in usual hyperplasia without atypia of patients who developed breast cancer compared with patients who did not.
In our study, we found that ER and PR immunoexpression was significantly higher in HUT cells than in neoplastic cells of IMC specimens. Even in all 16 cases of ER-negative tumors the epithelial cells of HUT were positive for ER. Our results are in agreement with other investigators who found higher levels of ER expression in benign breast epithelium of patients who developed breast cancer compared to controls [7, 9–11]. The presence of positive ER staining in normal lobules increases the breast cancer risk and the likelihood of progression to cancer . It occurs through the increase of the rate of cell proliferation by both recruiting non-cycling cells into the cell cycle and by shortening the overall cell cycle time due to a reduction in the length of G1 phase .
Previous comparison between normal and precancerous breast biopsies has shown that ER expression is relatively low in normal epithelium and slightly increased in HUT . Recent studies indicate that HUT is a heterogeneous entity containing subgroups identified according to the criterion of ER-α (+) proliferating cells and this fact could explain the different biologic behavior of HUT . In our study, the ER positive cells were more often found at the periphery of hyperplastic lesions. Similar pattern of ER positivity was previously described by Gobbi et al(2005). It is possible that the ER+ epithelial cells at the periphery of HUT represent the most proliferative group of cells, in spite of low positivity for Ki-67 in sequential sections of the same lesion. The estrogen exposure may stimulate a clonal proliferation of some ER+ cells or may increase the chance of spontaneous mutations .
In normal breast, there is a negative association between expression of ER and Ki-67, indicating either that ER+ cells are non-dividing or that the receptor is down-regulated as cells enter cycle [12, 14]. This important correlation breaks down in many ER+ cancers, where the receptor is often detected in proliferating cells . However, cells co-expressing ER-α and Ki-67 have been found in precancerous lesions and correlate positively with the level of risk of developing breast cancer . In our series, we found higher ER and Ki-67 immunoexpression in HUT areas compared to the immunopositivity for these markers in adjacent normal lobules. Our results are similar to those described by Schmitt (1995)  who found the existence of a positive correlation between ER status and proliferation in hyperplasic epithelium and a progressive inversion of this relationship in lesions evolving towards malignancy. The observation of higher rates of proliferation in ER positive benign proliferative breast lesions fits with the concept of an initial hormone-dependent status in breast carcinogenesis . In addition, HUT with higher expression of ER and Ki-67 could represent a subset of hyperplastic lesions with increased risk of subsequent breast cancer development.
Some previous studies suggest that at least some HUTs are clonal [9, 15]. Nevertheless, it remains unclear whether the dysregulation of ER has arisen prior to clonal expansion of the HUT, since cells with apparently abnormal regulation of ER during cell division are scattered randomly throughout the HUT in a non-contiguous pattern in some cases. The variable number of ER+ cells might indicate that in HUT the dysregulation of ER expression is incomplete and may be absent in some lesions, or apparent under certain conditions .
We detected p53 and p21 positivity in neoplastic cells of IMC, especially in high grade carcinomas. Mutations in tumor suppressor gene TP53, which mediates G1 arrest and apoptosis leads to an increased half-life and accumulation of the p53 protein . A way to investigate the functional status of TP53 is to evaluate some of its downstream effectors such as p21 gene whose product acts by blocking cyclin-dependent kinases . In our study, we found a positive association between expression of p21 and p53 in neoplastic cells, which is in agreement with the studies of Bankfalvi et al (2000) and Pelikainen et al (2003).
High expression of p21 would result in decreased cell proliferation subsequent to inhibition of cyclin/CDK activity . However, in our study p21 expression in neoplastic cells was related to higher proliferative index. Our results are in agreement with the theory of p53 independent pathways of p21 regulation in breast cancer [21, 22]. High amounts of p21 in high proliferating cells may reflect an unsuccessful effort to halt proliferation. This may result from the presence of other cell cycle regulatory pathways, which bypass the p21 mediated cell cycle block, such as c-Myc or B-myb  or due to mutant non-functional forms of p21 which posses prolonged half lives . In addition, p21 expression can indeed be up-regulated by epidermal growth factor receptor and transforming growth factor β1  which are associated with higher tumor grade and disease progression in breast carcinoma [24, 25].
In our study, cytoplasmic expression of p21 was found in 17.6% of benign hyperplastic cells and in 23.5% of neoplastic cells of IMC. Previous studies have reported exclusive nuclear localization of p21 in neoplastic cells of breast carcinomas [19, 23], and in epithelial cells of HUT  However, other authors reported p21 immunoexpression in the cytoplasm of breast and ovarian tumors and it was considered critical for promoting cell transformation [26, 27]. There is no other data in current literature concerning cytoplasmatic p21 expression in HUT similar to our findings. It remains unclear how the elevated cytoplasmic p21 expression might contribute to tumorigenesis. One possibility is that p21 is sequestered away from the nucleus thereby preventing it from binding to nuclear cyclin/CDK complexes, thus allowing sufficient cyclin/CDK activity for cell cycle progression . Alternately, relocalization of p21 to the cytoplasm may target cytoplasmic molecules such as apoptosis signal-regulating kinase 1 (ASK1) thereby promoting cell survival .
In our study, p63 was exclusively expressed in myoepithelial cells of normal breast lobules and ducts, partially expressed around the HUT cells, and rarely expressed in invasive breast carcinoma. We observed that p63 staining was discontinuous in 38.1% in HUT-ARM, in 73.3% in HUT-MCC, and in 64.7% in HUT-IMC. The discontinuous p63 expression pattern in HUT was different from continuous expression in normal lobules and could suggest that there is loss of p63 expression in the progression to invasive carcinoma. Our data is similar to the findings of Wang et al (2002) that demonstrate non-continuous expression of p63 in usual ductal hyperplasia. P63 expression has been useful to differentiate DCIS from microinvasive and invasive carcinomas based on lack of myoepithelial cells in invasive tumors without continuous distribution [31–33]. Although p63 is the most specific marker for myoepithelial cells, limitations exist because discontinuous myoepithelial layer seen in benign lesions, such as in our study may potentially cause diagnostic problems in clinical practice .
Although our data and genetic studies failed to demonstrate molecular changes in HUT, that are present in columnar cell lesions, ADH cells and in neoplastic cells of DCIS and IMC  the argument that HUT may be an early precursor is still supported by consistent data from epidemiological studies [1–4, 37, 38].