Introduction

In adults angiogenesis is primarily confined to the female reproductive system, including the uterus, the ovary, and the breast. More recently, it has become clear that hypoxia is a potent inducer of angiogenesis, and thus can represent a physiological stimulus within tumors for neovasculariza-tion, an essential feature of tumor growth as well as metastasis. The normal rodent mammary gland is an excellent model system for studying the process of angiogenesis and vascular remodeling because it exhibits controlled expansion and regression of blood vessels over the course of pregnancy, lactation, and involution.

As the epithelial cell population of the mammary gland expands during pregnancy to prepare for lactation, there is also a concomitant expansion of the mammary-associated vasculature. During the first half of pregnancy, the vasculature expands by sprouting angiogenesis, effectively doubling the number of blood vessels. During the second half of pregnancy, the vasculature expands through nonpro-liferative or intussusceptive angiogenesis, dilating at parturition to facilitate nutrient exchange. Upon pup weaning (involution), the vasculature regresses along with the mammary epithelium through an undefined process that involves capillary collapse. The mechanisms responsible for the angiogenic switch during breast tumorigenesis remain undefined.

Overview of Mammary Gland Development

In contrast to other organs, the majority of mammary gland development occurs postnatally, facilitating investigation of how the mammary vasculature develops in conjunction with the epithelium. The mammary gland is primarily composed of epithelial cells, which function to synthesize milk at lactation, myoepithelial cells, which contract to express milk at lactation, and the stroma (or adipose). The relative proportions of these cell types change over the course of the development. The nulliparous (virgin) gland is comprised of approximately 9 percent epithelial cells, whereas the lactating gland contains about 90 percent epithelial cells (Figure 1). To date, the function of the epithelial cell population has been the primary focus of mammary gland studies in transgenic and gene-deleted mice. Recently, however, several laboratories have begun to analyze the contribution of other cell types within the gland using genetic models; included in these studies have been the myoepithelium and immune cells. Thus far, the effects of endothelial cell-specific gene deletion have not been analyzed in the context of normal or tumor mammary epithelial cell biology. Furthermore, it remains unclear if unique molecular markers are expressed in either normal or tumor-associated mammary gland vasculature.

Postnatal development of the epithelium consists of four tightly regulated stages: ductal outgrowth into the stromal a

Figure 1 The gross histology of the mammary gland; hematoxylin-stained whole mount preparations from nulliparous (a) or lactating (b) glands observed at the same magnification (40x). In the mature nulliparous mammary gland, the mammary epithelium is organized into a ductal tree, which is often described as "naked" or void of alveolar development. The open arrow indicates a duct, whereas the black arrow points to the paler stained structure indicative of a large vessel. At lactation, the majority of the mammary fat pad is filled with epithelial cells, which are organized into lobules (white dotted circle) that contain individual alveoli, the polarized, secretory units of the gland that produce milk. Following involution, the mammary gland returns to a "nulliparous-like" state containing primarily ducts.

Figure 1 The gross histology of the mammary gland; hematoxylin-stained whole mount preparations from nulliparous (a) or lactating (b) glands observed at the same magnification (40x). In the mature nulliparous mammary gland, the mammary epithelium is organized into a ductal tree, which is often described as "naked" or void of alveolar development. The open arrow indicates a duct, whereas the black arrow points to the paler stained structure indicative of a large vessel. At lactation, the majority of the mammary fat pad is filled with epithelial cells, which are organized into lobules (white dotted circle) that contain individual alveoli, the polarized, secretory units of the gland that produce milk. Following involution, the mammary gland returns to a "nulliparous-like" state containing primarily ducts.

fat pad from 3 to 9 weeks of age, lobuloalveolar proliferation and differentiation during pregnancy, synthesis and secretion of milk at lactation, and involution of the secretory epithelium following weaning. Each stage depends on a critical balance among proliferation, differentiation, and apoptosis. Ductal morphogenesis progresses via a balance between proliferation and apoptosis within multilayered club-shaped structures known as terminal end buds (TEBs). As the ducts approach the edges of the fat pad, the TEBs disappear, signaling the end of ductal morphogenesis (Figure 1a). The virgin gland remains relatively quiescent until the onset of pregnancy, or the administration of exogenous hormones such as estrogen (E) and progesterone (P).

Pregnancy induces proliferation of the secretory units of the mammary gland, the alveoli, which originate from ductal progenitor cells and proliferate to fill the entire stromal fat pad at lactation (Figure 1b). During lactation, the secretory epithelium produces and secretes milk. At involution, which peaks at day four following pup removal, extensive tissue remodeling and apoptosis of the secretory epithelium occurs until the gland contains the simple ductal network observed in the virgin animal; relatively little epithelium is present by day 10 of involution. A concise review of the important genes that regulate mammary gland development at each of these stages can be found in Ref. [1].

Tools to Assay Angiogenesis and Microvessel Density in the Mammary Gland

Historically, to look at the gross structure of the mammary vasculature, vascular corrosion casts were prepared from mammary tissues isolated over the course of development, a process that destroyed the epithelium and the stroma. These corrosion casts were then subjected to analysis by scanning electron microscopy (SEM). A more recent technique preserves the relationship of the epithelium and mammary-associated vasculature. In this procedure tomato lectin, which binds to the interior of blood vessels, is injected intravenously into live mice followed by perfusion with fixative. Tissues of interest may then be harvested, followed by preparation of thick frozen sections, and subsequent staining for any other marker of interest, such as phalloidin (which stains actin networks) or E-cadherin. Following confocal microscopy a highly refined three-dimensional reconstruction of the vascularization of the mammary epithelium may be visualized, as demonstrated for a late pregnant gland in Figure 2a.

A classic morphological technique to quantify the fine microvessels surrounding alveoli was to perfuse rodents with India ink, to section the mammary gland, and to count the dark spots corresponding to the ink. More recently, staining with anti-CD31 antibodies (Figures 2b-d) and performing Chalkley counts has been accepted as a reproducible, quantitative measure of microvessel density (MVD). Finally, to look at the fine ultrastructure of individual endothelial cells, such as pinocytotic vesicles, microvilli, and fenestrations, transmission electron microscopy (TEM) has been performed on thin sections of intact fixed glands.

Angiogenesis during Normal Mammary Gland Development Is Not HIF-1 Dependent

The architecture and patterning of the vasculature shift dramatically from nulliparous to pregnant to lactating to involuting mice. Relatively few vessels are present in the nulliparous gland, and these vessels either run parallel with

Figure 2 Tools to visualize the mammary-associated vasculature. (a) A three-dimensional composite image of a lectin and phalloidin-stained mammary gland harvested in the last third of pregnancy. The pregnant host was injected intravenously with lectin-FITC (green), the inguinal mammary gland was embedded in OCT and sectioned at 50 mm, and then poststained with phalloidin-Alexa Red 455 (image courtesy of Bryan Welm and Jeffrey Rosen, Baylor College of Medicine). Note the network of vessels enveloping the alveoli and the vessels that run in parallel with and encircle the ducts. To observe the fine microvessel structure, anti-CD31 immunohistochemistry (blue color) was performed on zinc-fixed paraffin sections counterstained with nuclear fast red (b-d) isolated from mice at day 15 of pregnancy (b), day 2 of lactation (c), or from nulliparous mice (d) (staining courtesy of Debbie Liao, UCSD). (see color insert)

Figure 2 Tools to visualize the mammary-associated vasculature. (a) A three-dimensional composite image of a lectin and phalloidin-stained mammary gland harvested in the last third of pregnancy. The pregnant host was injected intravenously with lectin-FITC (green), the inguinal mammary gland was embedded in OCT and sectioned at 50 mm, and then poststained with phalloidin-Alexa Red 455 (image courtesy of Bryan Welm and Jeffrey Rosen, Baylor College of Medicine). Note the network of vessels enveloping the alveoli and the vessels that run in parallel with and encircle the ducts. To observe the fine microvessel structure, anti-CD31 immunohistochemistry (blue color) was performed on zinc-fixed paraffin sections counterstained with nuclear fast red (b-d) isolated from mice at day 15 of pregnancy (b), day 2 of lactation (c), or from nulliparous mice (d) (staining courtesy of Debbie Liao, UCSD). (see color insert)

or encircle the ductal tree (Figure 2d). The vasculature rapidly expands during pregnancy. Yasugi et al. demonstrated by India-ink perfusion that the vascular density of the rat mammary gland increased twofold from early (day 5) to mid-pregnancy (day 10) [2]. Matsumoto and Djonov have demonstrated that in the first third of pregnancy, vessels expand by capillary sprouting, in which endothelial cells migrate, proliferate, and form tubes [3, 4]. These sprouts ramify and anastomose with each other and will organize into distinct basket-like uniformly sized networks surrounding the alveoli by the second third of pregnancy (Figures 2a and b). Toward the end of pregnancy, capillary sprouting decreases and the vessels begin to expand instead by trans-luminal pillar formation that is independent of endothelial cell proliferation, a process known as intussusceptive microvascular growth (IMG), or intussusception [4]. The highly organized, homogenous structure of the mammary-associated vasculature observed during pregnancy is altered during lactation, at which time the vessels begin to dilate and to become more tortuous, in a manner similar to that observed in tumor vessels. At an ultrastructural level, Matsumoto observed by TEM that the length of microvillus processes and the number of pinocytotic vessels, indicative of active secretion, increased over the course of pregnancy compared to the nulliparous gland. Finally, there is controlled regression/collapse of the capillary network during involution. Peak capillary regression occurs at day 6 of involution, slightly delayed compared to peak epithelial cell death at day four of involution [4]. The critical steps in mammary gland angiogenesis are summarized in Figure 3.

Together these observations suggest that expansion and regression of the vasculature is under the same hormonal control as alveolar proliferation. This hypothesis is supported by the observations of Soemarwoto and Bern in 1958, and of Matsumoto in 1992, that administration of E+P to ovariectomized mice promoted development of capillaries as well as the epithelium and that prolactin augmented this effect [5, 6]. Furthermore, epithelium-free or "cleared" fat pads fail to induce angiogenesis, indicating that there is a necessary paracrine relationship between the epithelium and the vasculature. Therefore, signals from the epithelium control both the expansion and regression of the mammary-associated vasculature.

But, the question remained does hypoxia, a classical inducer of angiogenesis, also play a role in normal mammary gland vasculature development? Several laboratories

Figure 3 Cartoon representation of the various stages of mammary-associated vasculature development in conjunction with epithelial cell development. In nulliparous mice, the ducts (black lines) do not contain any alveolar structures (indicated in yellow), and the vasculature (indicated in red) runs in parallel with the ductal tree. In the first half of pregnancy, the vasculature expands by rapid endothelial cell (EC) proliferation and capillary sprouting, doubling from day 5 to day 10 of pregnancy in the rat mammary gland. In the latter half of pregnancy, when the alveoli have organized into lobuloalveolar structures, the vasculature expands by nonproliferative intussusceptive angiogenesis. At lactation, the epithelial cells become flattened since the alveoli are expanded and engorged with milk and the vessels dilate and become more tortuous. Finally, the vasculature regresses along with the epithelium during involution. The entire process will repeat over the course of each successive pregnancy. (see color insert)

Figure 3 Cartoon representation of the various stages of mammary-associated vasculature development in conjunction with epithelial cell development. In nulliparous mice, the ducts (black lines) do not contain any alveolar structures (indicated in yellow), and the vasculature (indicated in red) runs in parallel with the ductal tree. In the first half of pregnancy, the vasculature expands by rapid endothelial cell (EC) proliferation and capillary sprouting, doubling from day 5 to day 10 of pregnancy in the rat mammary gland. In the latter half of pregnancy, when the alveoli have organized into lobuloalveolar structures, the vasculature expands by nonproliferative intussusceptive angiogenesis. At lactation, the epithelial cells become flattened since the alveoli are expanded and engorged with milk and the vessels dilate and become more tortuous. Finally, the vasculature regresses along with the epithelium during involution. The entire process will repeat over the course of each successive pregnancy. (see color insert)

have demonstrated that the transcription factor hypoxia-inducible factor-1 (HIF-1) is a master regulator of oxygen homeostasis. HIF-1 is a heterodimeric transcription factor complex induced under decreased oxygen tensions that is composed of two subunits: the aryl hydrocarbon receptor nuclear translocator (ARNT or HIF-1 b) and HIF-1 a, the oxygen-responsive subunit. One of HIF-1's direct tran-scriptional targets is vascular endothelial growth factor (VEGF). Although HIF-1a is a key component in the hypoxically induced regulation of VEGF, it should be noted that in both the mammary epithelium and stroma there is an increase in VEGF mRNA in response to estrogen and progesterone treatment. VEGF may be transcribed into three isoforms from a single promoter. This process produces three isoforms in the mouse: VEGF120, VEGF164, and VEGF188. These isoforms differ in their increasing affinities for the extracellular matrix and decreasing solubility, respectively. Recent studies have shown that mammary epithelial cells express the VEGF164 and VEGF120 isoforms, with expression increasing twofold during pregnancy and lactation, whereas the stroma constitutively expresses VEGF188 [7].

Given these relationships, we recently investigated the effect of mammary epithelial cell-specific conditional deletion of HIF-1a in pregnant mice using Cre/lox technology. Although deletion of HIF-1 a impaired mammary epithelial cell differentiation and milk metabolism, to our surprise, there was no difference in vasculature density, patterning, vessel diameter, MVD, or VEGF mRNA expression between wild type and HIF-1 a-/- glands [8]. These results suggest that either pregnancy hormones are sufficient to induce angiogenesis, or that hypoxia is not a key driver of the angiogenesis and remodeling that occurs during normal mammary gland development. It is possible that there is a switch to a hypoxia-dependent mechanism during breast tumorigenesis.

Angiogenesis in Breast Cancer

Breast cancer is the most frequently diagnosed noncuta-neous cancer of women in the United States, projected to affect 1 in 8 women over their lifetimes. All solid tumors must recruit a blood supply from neighboring vessels through angiogenesis in order to grow beyond a small size, known as the angiogenic switch. However, in contrast to normal development, which produces regular, controlled expansion of functional vessels, solid tumors develop tortuous, leaky vessels with disrupted endothelial linings, creating a hypoxic microenvironment. It has been demonstrated that the normal breast has relatively low angiogenic activity. For example, normal breast tissue implanted into nude mouse hosts induces a mild angiogenic response, whereas breast tumor tissue will induce a potent angiogenic response [9]. Therefore, understanding the factors controlling the angiogenic switch in breast cancer is clinically relevant.

Several studies have reported that increased microvessel density is an independent prognostic factor in several tumor types, including the breast. Given HIF-1's relationship to angiogenesis, it is not surprising that HIF-1 a protein was found in 1999 to be upregulated in a variety of human tumors and their metastases [10]. A subsequent study by Bos et al. followed in 2001 [11], which investigated the correlation between the levels of HIF-1 a overexpression and other prognostic factors of breast tumors, including proliferation rates, VEGF expression, MVD, expression of estrogen receptor (ER), and p53 expression. As in the previous study, it was found that HIF-1 a overexpression was detected in the majority of ductal carcinoma in situ (DCIS) lesions, as well as in all poorly differentiated invasive carcinoma samples.

In contrast, HIF-1a expression was below detection levels in normal breast tissue or ductal hyperplasias. Moreover, in DCIS lesions, HIF-1a overexpression was statistically significantly associated with increased tumor MVD, high levels of proliferation, strong expression of VEGF, and ER positivity, but not p53 expression.

Several rodent mammary tumor models have been well characterized during the stages of mammary tumorigenesis. The most commonly used transgenic mouse models utilize the mouse mammary tumor virus (MMTV) long terminal repeat (LTR) to drive transgene expression in the mammary epithelium. Angiogenesis in transgenic mammary tumors has been investigated by corrosion casting using the MMTV-neuT model, which produces constitutive expression of activated rat ErbB2-2/Neu, producing mammary tumors with almost 100 percent penetrance [4]. Evidence of sprouting angiogenesis and intussusception was observed in these tumors, with angiogenesis occurring preferentially at the tumor border and intussusception occurring in the tumor center [4]. Because intussusception is not dependent on endothelial cell proliferation, it is likely that the centers of large breast tumors would be resistant to standard cyto-toxic treatments.

Therapeutic Implications

Several lines of evidence indicate that hypoxia induces growth arrest of a variety of normal as well as transformed cell types both in vivo and in vitro. In addition, the percentage of proliferating cells in a tumor decreases as the distance from blood vessels increases. These observations are relevant to cancer therapy since sensitivity to radiation damage depends on the presence of oxygen and since most chemotherapeutic drugs are targeted against proliferating cells. Therefore, hypoxic areas of tumors in which HIF-1a is expressed should be the most resistant to cancer therapy. In support of this hypothesis, recent studies of the response of HIF-1a null transformed mouse embryonic fibroblasts to cytotoxic agents found that the HIF-1a null cells are more susceptible to treatment with carboplatin, etoposide, and ionizing radiation than HIF-1a wild-type cells both in vitro and in vivo [12].

Given the widespread nature of hypoxia in tumors, there has been impetus to search for tumor-specific and hypoxia-responsive cytotoxins to augment or replace current therapies. One possibility currently in testing is the creation of a vector in which a prodrug enzyme is placed under control of a hypoxic response element (HRE), the HIF-1 binding site. Upon subsequent delivery of the vector to the tumor site, the hypoxic environment would result in preferential expression of the construct under control of HIF-1. With respect to breast cancer treatment, other gene therapy-based approaches are being developed in which killing agents are induced under control of both an HRE and an estrogen response element (ERE), potentially increasing the specificity of the response.

10 Ways To Fight Off Cancer

10 Ways To Fight Off Cancer

Learning About 10 Ways Fight Off Cancer Can Have Amazing Benefits For Your Life The Best Tips On How To Keep This Killer At Bay Discovering that you or a loved one has cancer can be utterly terrifying. All the same, once you comprehend the causes of cancer and learn how to reverse those causes, you or your loved one may have more than a fighting chance of beating out cancer.

Get My Free Ebook


Post a comment