Vitamin D and the mammary gland: a review on its role in normal development and breast cancer

Breast cancer is a heterogeneous disease associated with diverse biological behaviours and clinical outcome. Although some molecular subgroups of breast cancer have a targeted therapy, the most aggressive tumours still lack a molecular target. Despite vitamin D being classically associated with the physiological role of calcium regulation and phosphate transport in bone metabolism, several studies have demonstrated a wide range of functions for this hormone, which are particularly important in the field of cancer. The mechanisms underlying the protective actions of vitamin D in cancer development are only sparsely understood, but evidence shows that vitamin D participates in cell growth regulation, apoptosis and cell differentiation. In addition, it has been implicated in the suppression of cancer cell invasion, angiogenesis and metastasis. Most of vitamin D biological actions are mediated by the vitamin D receptor and the synthesis and catabolism of this hormone are regulated by the enzymes CYP27B1 and CYP24A1. In the present review we highlight research data concerning the function of this hormone in the mammary gland, with a special focus on breast carcinogenesis. Hence, and although the available data are controversial, we consider not only updated information on the epidemiology of vitamin D in breast cancer and its potential value as a therapeutic agent or prophylactic (with an emphasis on molecular mechanisms and effectors of vitamin D action), but include data on its role in other stages of breast cancer progression as well. Accordingly, we review data on the influence of vitamin D in the development of normal breast and the expression of vitamin D-related proteins (VDR, CYP27B1 and CYP24A21) in benign mammary lesions and ductal carcinomas in situ.


Introduction
Vitamin D was fi rst identifi ed in 1919 by Edward Mellanby as a lipid soluble substance with anti-rachitic properties and is classically associated with its physiological role of calcium and phosphate regulation in bone metabolism. Humans can obtain vitamin D from two main sources: from the diet and from sunlight exposure. Few natural foods contain vitamin D in signifi cant amounts; among these, fatty fi sh, eggs and sun-dried mushrooms can be highlighted. Still, the majority (90 to 95%) of the required vitamin D is produced by the skin when exposed to sunlight (ultraviolet B radiation) [1], which has caused vitamin D to be nicknamed 'the sunshine vitamin' . Th e biologically active form of vitamin D (1α,25-dihydroxyvitamin D 3 or 1α,25(OH) 2 D 3 ) is synthesised in the kidney by the mitochondrial enzyme CYP27B1 [2]. In order to maintain the homeostasis of the organism, especially regarding the levels of calcium and phosphate, the amount of circulating vitamin D has to be tightly regulated. Th e enzyme CYP24A1 plays a key role in this process, and is activated by 1α,25(OH) 2 D 3 whenever there is an increase in the levels of this hormone. Th is enzyme is responsible for converting 1α,25(OH) 2 D 3 to biologically inactive metabolites [2]. In this review, for simplicity purposes, whenever we state vitamin D, we are referring to the biologically active form (1α,25(OH) 2 D 3 ) unless stated otherwise.
Vitamin D is a steroid hormone that exerts most of its biological activities by binding to a specifi c high-affi nity receptor, the vitamin D receptor (VDR). VDR belongs to the superfamily of nuclear receptors for steroid hormones and regulates gene expression by acting as a ligandactivated transcription factor [3]. However, vitamin D can also induce VDR-independent eff ects, indicated by the fact that the anti-proliferative eff ects of vitamin D in MCF-7 cells are not exclusively dependent on the presence of the VDR [4]. Vitamin D is also known to exert rapid eff ects that are not dependent on gene transcription [5]. Presumably, these eff ects are mediated by cell surface membrane receptors. Two proteins have been implicated in expression-independent vitamin D action: membrane VDR and the membrane-associated rapid response steroid binding (1,25D 3 -MARRS) protein. Th e evidence for the existence of a vitamin D membrane receptor came from two observations: fi rst, the existence of vitamin D analogues that can cause rapid actions of vitamin D, but show low levels of affi nity to the VDR [6,7]; and second, the existence of a vitamin D binding protein that has been described in the basolateral membrane of rat and chick enterocytes [8]. Besides the role of 1,25D 3 -MARRS as a vitamin D binding protein, the VDR can mediate non-transcriptional eff ects [5]. Th e most striking evidence supporting this hypothesis was the demon stration that the vitamin D-induced rapid actions are lost in osteoblasts from Vdr knockout mice [9]. More over, VDR has been identifi ed within caveolae-enriched plasma mem brane fractions from various cell types [10], suggesting that it can also work as a membrane receptor.
Th e role of vitamin D in health and disease is far from understood and the literature is full of contradictory fi ndings. Th e role of vitamin D in cancer is particularly controversial. Still, evidence that vitamin D is capable of modulating several features of cancer exist. Anti-carcinogenic properties of this hormone include the inhibition of cell proliferation, invasion, metastasis and angiogenesis, and the induction of apoptosis and diff erentiation [5]. Th ese issues will be further elaborated on in the light of breast cancer development in the following sections and controversial issues discussed.

Epidemiology of vitamin D in breast cancer
Th ere has been a great amount of information in the literature regarding a protective role of vitamin D in breast cancer. Two major types of epidemiological studies have been conducted: fi rst, those that focused on the association between solar radiation and breast cancer risk; and second, those that analysed the relationship between vitamin D intake and breast cancer risk.
Th e fi rst set of epidemiological studies demonstrated that decreased sunlight exposure, and consequent diminished vitamin D production by the skin, was correlated with higher breast cancer incidence and mortality [11]. One study described that women with breast cancer had, on average, lower vitamin D blood levels than women without breast cancer [12], hence proposing a potential protective eff ect of vitamin D. In line with this observation, early stage breast cancer patients show higher serum levels of vitamin D than those who have advanced bone metastatic disease and these changes in hormonal levels preceded clinical signs of progression and predicted disease response [13]. Additional evidence is provided by an inverse association between the circulating levels of 25-hydroxyvitamin D 3 (the inactive circulating form of vitamin D, which is used to measure the levels of vitamin D in the blood stream) and the risk for developing breast cancer, suggesting that high levels of 25-hydroxyvitamin D 3 may be associated with a reduced risk of breast cancer [14]. Furthermore, it has been described that patients with the most aggressive type of breast cancer (triple-negative and basal-like) display the lowest levels of 25-hydroxyvitamin D 3 [15,16]. Th ese results are supported by a recent study where it was observed that breast cancer patients with suboptimal vitamin D levels are more likely to have tumours with aggressive profi les and worse prognostic markers [17]. Hence, these reports indicate that vitamin D defi ciency may be linked with increased breast cancer risk, as well as with the development of more aggressive carcinomas. In contrast, other studies found no associations of vitamin D plasma levels or intake with reductions in breast cancer risk. For example, it has been reported that supplementation with calcium and vitamin D did not reduce invasive breast cancer incidence [18,19]. In line with this, it has been observed that circulating levels of vitamin D were not signifi cantly associated with breast cancer risk [19][20][21].
Although some of these studies demonstrate a potential protective role for vitamin D in breast cancer, suggesting that disruption of the vitamin D signalling pathway may be a predisposition to develop the disease, others have failed to support a causal relationship between vitamin D intake or supplementation and breast cancer risk. Th us, further studies are required in order to clarify this issue.

Role of vitamin D in normal breast development
Th e VDR is expressed in the normal mammary gland and vitamin D has been shown to play an important role in the development and function of the mammary gland. Most studies conducted to elucidate the role of vitamin D in breast development have been based on the use of Vdr knockout mice. Zinser and colleagues [22] performed an elegant study on the role of the vitamin D signalling pathway in the growth regulation of the mammary gland during pubertal development. Th ey showed that Vdr knockout female mice display more extensive ductal elongation and branching when compared with their wild-type counterparts. Furthermore, they observed that this enhanced morphogenesis was not associated with the deregulation of cell proliferation and apoptosis pathways in terminal end buds [22]. Moreover, they saw that VDR ablation leads to an increase of breast responsive ness to exogenous hormones (oestrogen and progesterone), represented by an increase in cell growth [23]. Collectively, these data indicate that VDR could have an important impact on breast development and suggest that the vitamin D signalling pathway participates in the negative growth regulation of the mammary gland. In another study published by the same group [24], it was shown that Vdr knockout mice had delayed mammary gland regression after weaning and that this eff ect was associated with reduced apoptosis in the epithelial cell compartment, hence highlighting the contribution of VDR to mammary gland biology during the reproductive cycle.
Immunohistochemical studies on human material have demonstrated that the VDR protein is expressed in samples from normal breast tissues [25,26]. However, there are some discrepancies concerning the expression of the main enzymes involved in vitamin D metabolism (CYP27B1 and CYP24A1) in normal breast, probably due to the use of diff erent methodologies. Studies assessing the mRNA of the two genes confi rmed their expression in the normal mammary gland [27][28][29][30] and demonstrated that both enzymes are functionally active in normal breast [30]. In contrast, our group showed, by immunohistochemistry, that both enzymes are detectable in normal breast, but not in all cases [26]. In fact, although the majority of the cases (over 60%) presented CYP27B1 expression, only a small percentage (about 30%) of these presented detectable CYP24A1 protein expression ( Figure 1). Th is work provided evidence for an imbalance in the enzymes that favours the presence of vitamin D in the normal mammary gland. Together with the fact that VDR has an important impact on breast development, it supports the role of the vitamin D signalling pathway in the growth and developmental control of the mammary gland.

Vitamin D in benign lesions of the mammary gland
Benign lesions of the breast can be associated with distinct clinical behaviours and their accurate classifi cation is applicable to patient management in terms of surgical treatment and prophylaxis [31]. In contrast to the extensive studies addressing the expression of Vitamin D signalling and metabolic pathways in breast cancer, the studies in benign lesions of the mammary gland have been ignored. Recently, we have evaluated the expression of VDR, CYP27B1 and CYP24A1 in a series of breast benign lesions [26]. Th is series consisted of usual and atypical ductal hyperplasias, columnar cell lesions, papillomatosis and adenosis. Th e results we obtained indicate that the most benign lesions express VDR and CYP27B1 and less than 20% of these cases detectably express CYP24A1. Decreased levels of the three proteins was observed in benign lesions when compared with normal breast, although the results were not statistically signifi cant. Th is observation may indicate a disturbance in the levels of the proteins that regulate vitamin D metabolism and signalling in early stages of breast cancer development (Figure 1).

Vitamin D in breast cancer models
Suppression of cell growth by vitamin D paved the way for vitamin D to be considered as a potential therapy in cancer research. Since then, numerous studies have been conducted in various cancer models in order to identify the signalling pathways and molecular mediators associated with such eff ects and these have shown the ability of vitamin D to aff ect diff erent features of cancer ( Figure 2). It was convincingly demonstrated in many of these studies that vitamin D has a prominent role in tumour cell proliferation, and cell cycle genes have consequently become the centre of attention to decrypt the molecular mechanisms of vitamin D in cancer.
Treatment of breast cancer cell lines with vitamin D elicits a change in the expression of proteins involved in cell cycle regulation, such as cyclins, cyclin dependent kinases (CDKs) and CDK inhibitors [5]. Vitamin D and analogues induce increased expression of the CDK inhibitors p21 and p27, thus blocking cell cycle progression [32][33][34]. Concomitantly, treatment with this hormone leads to impaired expression and activity of CDK2, CDK4, cyclin D1, cyclin D3, cyclin A1 and cyclin E1, which provoked cell cycle arrest at the G 0 -G 1 transition [32][33][34], and this eff ect was correlated with hypophosphory lation of the retinoblastoma protein pRB. Together with these eff ects on cell cycle proteins, a downregulation of c-Myc oncoprotein was reported, further contributing to the inhibition of cell proliferation [35]. Altogether, these results demonstrate that vitamin D hampers prolifera tion by targeting several key masters of cell cycle progression. Additionally, it has been proposed that the transcription factor CCAAT enhancer binding protein alpha (C/EBPα) may mediate vitamin D growth inhibitory eff ects, since C/EBPα is induced by the treatment with vitamin D. A decrease in cell proliferation was observed in C/EBPα-negative MDA-MB-231 cells transfected with C/EBPα, while the knockdown of C/ EBPα suppressed the antiproliferative eff ects of vitamin D in MCF-7 cells, thus providing evidence for the involve ment of this transcription factor in breast cancer proliferation [36]. Th e breast cancer tumour suppressor TCF-4 is another molecular target of vitamin D action, as it has been reported that this transcriptional regulator is present at lower levels in Vdr knockout mice, which suggests a role for TCF-4 in the antiproliferative eff ects induced by vitamin D [37]. Furthermore, vitamin D induction of the BRCA1 gene (Breast cancer 1) has also been inversely correlated with cell proliferation [38], while it has been described that vitamin D decreases aromatase expression [39] and, thus, can modulate oestrogen receptor-positive breast cancer growth.
Together with growth arrest, vitamin D has a role in the induction of apoptosis in breast cancer cells (Figure 2), since cell shrinkage, chromatin condensation and DNA fragmentation are observed in MCF-7 cells treated with this hormone [40]. Th e most probable mechanism of such vitamin D-induced apoptosis is through the downregulation of the anti-apoptotic protein Bcl-2 [5]. Vitamin D is able to enhance tumour necrosis factor alpha through caspase-dependent (increased caspase 3 activity) and caspase-independent mechanisms [41]. In support of the role of caspase-independent cell death mediated by vitamin D, it was shown that induction of apoptosis in MCF-7 cells was correlated with disruption of mitochondrial function, which was associated with Bax translocation from the cytosol to the mitochondria, cytochrome C release and production of reactive oxygen species. Th ese mitochondrial eff ects did not require caspase activation, since these were not blocked by a specifi c caspase inhibitor [42]. Another caspase-independent cell death mechanism induced by vitamin D relies upon cytosolic calcium accumulation associated with an increase in lysosomal protease activity [43]. Finally, vitamin D was still described as a prooxidant in breast cancer cells, causing an increase in the overall cellular redox potential [44], which may also be an important mechanism underlying the pro-apoptotic eff ects of this hormone. Vitamin D is able to potentiate and enhance the morphological eff ects of apoptosis when administered to MCF-7 cells in combination with tamoxifen [45]. Eff ects of other anticancer agents are improved by the administration of vitamin D: it has been reported that pre-treatment of breast cancer cells with vitamin D enhances the cytotoxicity of doxorubicin [46], lowers the threshold for taxol-induced cell death [47] and enhances the growth inhibition eff ect induced by cisplatin in MCF-7 cells [48]. Importantly, a recent study demonstrates that the interaction between p53 and VDR provides a mechanism for mutant p53 (the most common genetic alteration in human cancers) gain-of-function. Th is may have clinical implications and suggests that p53 status should be considered when studying vitamin D for cancer therapy [49]. Vitamin D plays an important role in the modulation of cancer invasion and metastasis as well (Figure 2). Hansen and collaborators [50] demonstrated that vitamin D has the ability to inhibit the invasive potential of human breast cancer cells in vitro. Th is reduced invasiveness was found to be associated with diminished activity of the metalloproteinase MMP-9 and downregulation of the plasminogen-activator, simultaneously with increased tissue inhibitor of MMP-1 activity and the induction of plasminogen-activator inhibitor [51]. Th is is indicative of the dual eff ect of vitamin D in the invasive process: on one hand, it decreases the activity and expression of metalloproteinases and serine proteases; at the same time, it induces their inhibitors. In vitro experiments using vitamin D analogues have demonstrated that they can inhibit the invasive potential of mammary cancer cells [52], as well as prevent skeletal metastasis and prolong survival time in nude mice transplanted with human breast cancer cells [53]. Th is is in line with the observation that vitamin D induces cell adhesion as well as impairs in vitro motility [54], hence contributing to a less aggressive phenotype in breast cancer cell lines. Further more, vitamin D is able to downregulate the expres sion of P-cadherin [54], an invasion promoter molecule in breast cancer cells [55]. We have addressed the eff ects of vitamin D in E-cadherin-negative breast cancer cell lines and have observed that treatment with vitamin D induces the de novo expression of E-cadherin in MDA-MB-231 cells by CDH1 promoter demethylation [56], providing further evidence for a vitamin D role in invasion and metastasis.
Angiogenesis is yet another cancer feature that vitamin D can modulate (Figure 2). An analogue of vitamin D was able to inhibit angiogenesis at low concentrations in vivo [57]. Using xenografted mice with vascular endothelial growth factor (VEGF)-overexpressing MCF-7 breast cancer cells, it was demonstrated that the administration  of vitamin D results in reduced vascularisation of tumours [58]. Th e tumours formed in the treated animals displayed smaller capillaries when compared with their littermates, suggesting that vitamin D may also inhibit vessel growth and maturation. Additional evidence for angiogenesis inhibition by vitamin D and analogues has shown that it occurs through a decrease in VEGF and tenascin-C expression [59,60]. Th e anti-proliferative, pro-apoptotic, anti-invasion and anti-angiogenic properties of vitamin D indicate that it could serve as a potential therapeutic agent. However, breast cancer is not one disease and is divided into many diff erent molecular subtypes [61]. Th is may partly explain the many diff erent functions described above for vitamin D. In order to transform vitamin D into a (targeted) therapy, a better understanding of the role and function of this hormone in solid tumours is required, accompanied by an upfront stratifi cation of the diff erent patient cohorts in the vitamin D research fi eld.

Conclusion
Th e data described provide evidence for an essential role of vitamin D in normal development of the mammary gland and breast cancer. Epidemiological studies are confl icting and their results not consistent, with the current view indicating no protective role for vitamin D in breast cancer development. Still, the diff erent functions and eff ects of vitamin D on cancer cell biology, such as the cell cycle, apoptosis, invasion and metastasis, as well as angiogenesis, virtually bring together the entire spectrum of tumour development under its purview.