Estrogen, by interacting with ER, plays an important role not only in the regulation and differentiation of the normal breast epithelium but also in breast tumorigenesis [7]. Estrogen deprivation has been shown to result in a marked reduction in angiogenesis, which returns to pretreatment levels following estrogen replacement [8,9].

There is considerable experimental evidence showing that tumor growth is dependent on angiogenesis; moreover, a tumor will not grow to more than 1–2 mm, around 10⁶ cells, nor metastasize if neovascularization from preexisting capillaries fails to occur [10,11].

The cases were then divided into two groups: Group A: ER-positive, n = 35; and Group B: ER-negative, n = 29. Patients ranged in age from 47 to 82 years (mean 52.40 years) in Group A and from 46 to 88 years (mean 58.38 years) in Group B. The size of the tumors in the two groups ranged from 2.5 to 5 cm, stage II, mean tumor size 3.36 and 3.26 in groups A and B, respectively. The groups were considered homogeneous with respect to age and to the size and histological grade of the tumor (Table ​(Table1).1). For CD34 antigen immunostaining, specific primary monoclonal antibody for CD34 was used (Anti-human Hematopoietic Progenitor Cell, CD34 Class II, Clone QBEnd 10, Code M 7165, DAKO Corporation, Carpinteria, CA, USA) at a dilution of 1:25 with bovine serum albumin (BSA). Antigen retrieval was carried out in a steamer containing sodium citrate buffer (pH = 6) at 90°C for 30 minutes. Sections were then incubated overnight with the primary specific antibody at 4°C. Color was developed using DAB (3-3' diaminobenzidine, SIGMA code 5637), a chromogenic substrate. Liver hemangiomas were used to test positive and negative controls for the immunohistochemical reaction.

Microvessel count was performed by two independent observers, who were blinded with respect to the patients' identities. A Nikon Eclipse E400 light microscope was used to count microvessels and was coupled to a color video camera that transmitted the image to a microcomputer and hence to a monitor, counting being performed with the aid of an image analysis software program (Imagelab). Initially, the areas with greater microvessel density were identified at a magnification of 40×. Microvessel count was performed in 10 fields in each of these areas of highest density (40× objective lens, 10× ocular lens), total magnification 400×. Units of vessel counts were identified according to the criteria established by Weidner et al [19], who described them as a group of brown-stained endothelial cells, clearly separated from the adjacent microvessels, tumoral cells and other conjunctive tissues. The vessel lumen, although usually present, was not a criterion used to define a microvessel, and red blood cells were not used to define vessel lumens. Partially identified vessels that were not completely contained within the fields under analysis were not considered in the vessel count. In each case, microvessel density consisted of the mean number of vessels counted in ten fields.

In the present study, microvessel count in estrogen receptor-negative breast carcinomas of postmenopausal women was significantly greater compared to estrogen receptor-positive breast carcinomas. The women in the two groups were homogenous with respect to age and to the size and histological grade of the tumor, which made comparison between the two groups feasible. The tumor specimens in the present study were obtained from wedge biopsy; therefore, it is unlikely that the results obtained simply reflect tumor heterogeneity, which is a recognized methodological problem [17]. Studies carried out to evaluate the effect of tumor heterogeneity in microvessel count in breast cancer specimens have focused on the use of core biopsies [17], which contain less tumor volume than the preoperative wedge biopsies used in this study. CD34 was selected from the available markers because it is a sensitive marker that stains the neoplastic endothelium more strongly than the normal endothelium [21].

The presence of a significantly lower number of vessels in the ER-positive breast tumors observed in the present study indicates a correlation between tumor angiogenesis and estrogen receptor status. Moreover, in an in vivo study using nude female mice, Ali et al. [15] showed that high levels of ER-alpha downregulate angiogenic factors VEGF and integrin alphavbeta3 (αvβ3), leading to inhibition of tumor angiogenesis.

On the other hand, Erdem et al. [16] failed to show any difference between mean MVD values in estrogen receptor-positive and negative tumors. The samples in that study consisted of archival specimens collected around 7–10 years previously. However, for some investigators, material kept in storage for a long time may result in loss of antigenicity for some markers [22]. Likewise, in samples obtained by core needle biopsy from breast tumors of 158 patients, Vamesu [7] showed that a high microvessel density was significantly more common in patients with ER-positive/PR-negative tumors. Nevertheless, the possible effect of tumor heterogeneity on microvessel counts in specimens originating from core biopsies should be emphasized [17] as well as the overexpression of variants of the subtypes of estrogen receptor that affect the regulation of tumor angiogenesis [23].