Peroxisome proliferator-activated receptor subtypes in mammary gland development and breast cancer

Background: Peroxisome proliferator-activated receptors (PPARs) have a pivotal role in regulating genes that are involved in lipid


Introduction
Peroxisome proliferator-activated receptors (PPARs) are ligand activated transcription factors that are members of the nuclear hormone receptor superfamily [1]. There are three members of this subfamily, PPARα, PPARβ/δ and PPARγ with each subtype having overlapping but distinct tissue distribution [1]. PPARα is primarily expressed in tissues involved with fatty acid catabolism, e.g. liver, kidney, skeletal muscle, brown adipose tissue and heart, whereas PPARβ/δ is widely expressed and found in most tissues dependent on the stage of cell differentiation and proliferation. PPARγ is predominantly found in adipose, liver, kidney, gut, brain and vascular tissues [1,2]. The genes regulated by these transcription factors are involved in many different processes including cell cycle, differentiation and inflammation [1]. However, one of the main roles of these subtypes is their involvement in lipid homeostasis [3]. PPARα transcriptionally activates genes that control fatty acid catabolism [4], PPARβ/δ has a prominent role in fatty acid oxidation [5] and PPARγ activates genes that function in adipogenesis [6].
The mammary gland is an organ that undergoes distinct developmental stages post-birth under the influence of hormones [7]. The basic glandular network is established during puberty, but the onset of pregnancy marks the beginning of changes toward functional differentiation. The mammary epithelial cells proliferate and give rise to a branching network of ducts and alveoli. At the end of pregnancy and during lactation the alveolar epithelium becomes secretory and produces milk [7]. Finally after lactation ceases the mammary gland involutes, a stage that is characterized by a remodelling of the gland to produce a more rudimentary structure in a process involving apoptosis followed by remodeling of stroma and re-differentiation of adipocytes [8].
Breast milk is rich in lipids, the precise percentage is species dependent but ranges from 4% in humans to 50% in whales [9]. In mice, the lipid content of milk is approximately 30%, resulting in the secretion of approximately 30 g of lipid over a 20 day lactation period [10]. The major milk lipid is triacyclglyceride, which comprises 98% of the total milk lipid [9]. Enzymes to synthesize lipids de novo increase in expression within the mammary gland at lactation [11]. Given the role of the PPARs in cellular processes including differentiation and lipid homeostasis, their expression in the mammary gland during different stages of development may give some insight into the processes that control lipid biosynthesis in the mammary gland. Understanding the normal physiology of the gland is important in determining changes that may be deregulated in pathophysiological states. We have shown previously that a calcium transporter, known as SPCA2, is important during lactation [12] and is also upregulated in breast cancer cells and human breast tumors [13].
Many tumors are characterized by alterations in lipid homeostasis with concomitant changes in the expression and activity of many lipogenic and lipolytic enzymes. Many of these changes occur in enzymes regulated by PPARs [14,15]. PPAR subtypes are linked with tumorigenic processes in breast cancer cells including actions that are both pro-and anti-tumorigenic through differential effects on cellular proliferation, apoptosis and differentiation [14]. Given their role in lipid homeostasis, their effects are likely to extend to influencing the balance of lipids within a tumor and specifically their relationship with breast cancer could depend on the molecular profile of the tumor itself. Breast cancer is stratified into transcriptionally different subtypes based on gene expression patterns and includes subtypes described as basal-like, ERBB2 (HER2)enriched, luminal A and luminal B [16]. A potential association between specific PPAR subtypes and breast cancer subtypes is suggested by the regulation of PPARγ expression and cellular response by HER2 [17].
Our aim with this study was to explore the expression of the PPARs in physiological and pathophysiological states of the mammary gland, i.e. mammary gland development and breast cancer and assess possible relationships between PPAR subtype expression and breast cancer prognosis.

Microarray analysis
PPAR gene expression was assessed in breast cancer clinical samples in the NKI-295 dataset consisting of the microarray profiles of 295 primary breast tumors [20]. Raw data are available from the NCBI Gene Expression Omnibus (Accession number: GSE2845). Tumors were assigned to molecular subtypes based on their gene expression profiles and those matching to the basal-Like, HER2-overexpressing, luminal A and luminal B subtypes were analysed for PPAR gene expression. Samples matching to the "Normal-Like" subtype were excluded from the analysis as these are believed to be substantially contaminated by normal tissue. Kaplan-Meier analysis was used to determine correlation of PPAR gene expression with patient survival and statistical significance was assessed using the log-rank test.

Results and Discussion
The requirement of the mammary gland for lipid metabolism alters with the developmental stage of the gland, particularly during the transition from a pregnant to a lactating gland in preparation for delivering nutrients to the new born [21]. Examining the profile of PPAR subtype expression with the morphological and physiological changes of the gland may promote a better understanding of lipid homeostasis.
As the structure and function of the mammary gland changed with pregnancy, lactation and finally early involution, the profile of each PPAR also altered, although each subtype was detectable at all mammary gland stages (Figure 1) as described earlier in other studies [21,22,23,24,25]. While our work did not focus on protein levels, the relative levels of PPARα mRNA transcripts were significantly (P<0.05) highest in the nulliparous gland (Figure 1A), whereas PPARβ/δ's expression was increased significantly (P<0.05) upon pregnancy before falling again once lactation had started ( Figure 1B). PPARγ expression significantly (P<0.05) dropped with pregnancy, but then continued to significantly decrease (P<0.05) beginning with early lactation (Figure 1C). Comparing the relative levels of each subtype suggested that PPARγ was the principal subtype expressed in the nulliparous mammary gland (Figure 2A), whereas in the pregnant gland the relative level of PPARβ/δ had increased so that the gland had equivalent levels of PPARβ/δ and PPARγ ( Figure 2B). The relative levels of PPARβ/δ and PPARγ remained high compared with PPARα during both lactation and involution (Figure 2C & 2D).
Since PPARα is pivotal in controlling genes important in the β-oxidation of fatty acids [4], it is not surprising that its expression level would decrease during the transition of the animal from pregnancy to lactation. The secretory phase of the animal is the time during which the production of milk for the neonate becomes foremost [21] and so the gland moves towards supporting lipogenesis and away from β-oxidation [21,26]. Studies with a transgenic mouse harboring a constitutively active PPARα show that PPARα activation during pregnancy results in a severe defect in lobuloalveolar development via a reduction in proliferation and an increase in apoptosis [27]. This effect of activated PPARα suggests a physiological reason why PPARα decreases in expression between the nulliparous and pregnant animal. There is a paucity of data on the role of PPARβ/δ in the mammary gland. Other studies have reported no change in PPARβ/δ with differentiation of the gland, however, these studies did not use real time PCR quantitative methods [23,28]. Our study suggests that the relative level of PPARβ/δ becomes more important once the gland begins to develop during pregnancy, highlighting a potentially important role for this transcription factor in regulating cell proliferation, differentiation and survival at this time in mammary gland development. PPARγ has a well-defined role in adipose tissue involving differentiation and promoting lipid storage [6].
One suggestion is that a decrease in PPARγ during mammary gland differentiation is fundamental to a switch of the gland's activities from β-oxidation to lipogenesis [21]. Interestingly, however, the role of PPARγ in the lactating mammary gland has been proposed to be important in maintaining the quality of milk by ensuring inflammatory lipids in the milk are not accumulating [29]. In our study using whole mammary glands, the alterations in PPARs may be in part related to a decreasing adipocyte contribution and increasing epithelial cell contribution to the gland architecture. Our study looking at the profile of PPAR subtype expression in a panel of breast cancer cell lines showed a lack of correlation between tumorigenicity and expression of individual subtypes (Figure 3), which is consistent with a previous study involving a smaller cohort of breast cancer cell lines [30]. Expression of both PPARβ/δ and PPARγ was highest in the MDA-MB-231 breast cell line compared with the tested non-tumorigenic and breast cancer derived cell lines (Figure 3B & 3C). MDA-MB-231 cells also had the highest expression of PPARα compared with the other tumorigenic breast cancer cell lines examined (Figure 3A) [31]. Further work examining knockdown of PPAR subtypes in MDA-MB-231 cells with a focus on functional outcomes such as cell proliferation and apoptosis will provide more information on the role of the PPARs in breast cancer cells. MDA-MB-231 cells are estrogen receptor negative and are used experimentally as a model of basal-like breast cancer [32]. PPARγ expression has been linked to a stellate morphology, which occurs when MDA-MB-231 cells are grown in three-dimensions, and is indicative of a subset of breast cancers referred to as having a triple-negative status [33]. Basal-like breast cancers (identified by gene expression profiling) and triple-negative breast cancers (identified by immunohistochemical staining for a lack of ER, PR and HER2) while not being identical do have some overlapping features including a poor prognosis [34]. Our results suggested a possible delineation of PPAR subtype expression with certain subtypes of breast cancers leading us to examine the expression of PPARs in a large cohort of human breast cancer tumors. Using the transcriptional subtypes described by Perou and colleagues [16], PPARα relative expression was significantly (P<0.05) higher in breast cancers of the basal subtype compared to those of the ERBB2+, luminal A or luminal B subtypes, and the expression of PPARα was associated with reduced human breast cancer patient survival (P<0.05, upper quartile versus others, Figure 4A). PPARβ/δ expression was significantly (P<0.05) higher in both basal and ERBB2+ populations of breast tumors and consistent with this, intermediate and high levels of PPARβ/δ expression were associated with poor outcomes (P=0.0007) (Figure 4B). In contrast, PPARγ expression did not differ between tumor types and levels did not correlate with patient survival (Figure 4C).
Women with basal tumors have a poorer outcome than those with tumors of other subtypes [35] [33]. Many of these tumors are classified as triple negative because they lack the estrogen receptor, progesterone receptor and the HER2 receptor, predictors of successful treatment with specific molecular therapies [34]. Studies using cancer cell lines and animal models have overwhelmingly looked at PPARγ and conclude that the evidence suggests an anti-tumorigenic role for PPARγ in human breast cancer [36]. Despite this, a clinical trial with the PPARγ agonist troglitazone failed to show a response in women with metastatic treatment-refractory breast cancer [37]. The relationship between PPARβ/δ and breast cancer is not as well researched as that of PPARβ/δ and colon cancer [38]. Recent work with animal models has shown a pro-tumorigenic role for PPARβ/δ in mammary gland tumorigenicity [28,39]. Since the expression level of PPARβ/δ correlates with patient survival and there is an upregulation of PPARβ/δ in the basal and ERBB2 breast tumor subtypes, further work should explore tumorigenic pathways that are enriched in these subtypes. Similarly there is a paucity of data on PPARα and breast cancer, although PPARα agonists do increase proliferation in the estrogen positive and estrogen negative cells lines MCF-7 and MDA-MB-231, respectively [31].

Conclusion
In summary, the profile of PPAR subtype expression changes with mammary gland morphological and functional development suggesting differential regulation. This may help the gland fulfil its role in providing a balanced lipid profile for milk production. PPARα and PPARβ/δ expression also differ between breast cancer subtypes, associating with the basal-phenotype, while high expression of these PPAR subtypes correlates with poor outcome. Together this evidence indicates a potentially important role, particularly for PPARα and PPARβ/δ, in breast epithelial homeostasis and tumorigenesis.