Abstract
In addition to the spectrum of conditions known collectively as the Metabolic Syndrome, obesity is now recognized to be associated with increased risk of several cancers including colon, endometrial, and breast cancer. Obesity and carcinogenesis share 2 characteristics in common. On the one hand, they involve inflammatory pathways, and on the other hand, they involve dysregulated metabolism. In this review we focus on postmenopausal breast cancer and discuss the metabolic and cellular mechanisms whereby obesity and breast cancer are related. Because a majority of postmenopausal breast tumors are estrogen responsive, we include a discussion of the action of obesity-related factors on estrogen formation within the breast.
It has become a truism that we are facing a global “pandemic” of obesity, which currently affects hundreds of millions of women and men worldwide, not only in the developed world but in many developing countries also. Obesity is commonly associated with the metabolic syndrome, which is a spectrum of conditions including insulin resistance, type 2 diabetes, hyperlipidemia, hypertension, as well as increased risk of cardiovascular disease, stroke, and kidney failure. In general, there is a positive correlation between body mass index and the risk of type 2 diabetes, as has been shown in a number of epidemiological investigations. However, added to these pathologic abnormalities noted above, the metabolic syndrome is now recognized to be an established risk factor for a number of cancers (1). A conclusion from several studies (reviewed in 2) is that both obesity and type 2 diabetes increase the risk of postmenopausal breast cancer. The results concerning women with premenopausal breast cancer were less clear as positive as well as inverse relationships have been described in women from different ethnic groups. On the other hand, when employing waist-to-hip ratio rather than body mass index, a positive correlation has been observed between elevated waist-to-hip ratio and risk of breast cancer in premenopausal women as well as postmenopausal women (reviewed in 3). Although most postmenopausal breast cancers are estrogen receptor (ER) -positive, many premenopausal breast cancers are ER-negative or triple negative, that is to say they lack ER, PR, and HER2-neu (4).
Obesity and carcinogenesis share 2 properties in common. On the one hand, they involve inflammatory pathways (5), and on the other, they are characterized by dysregulated metabolism (6). In this review we discuss the cellular and molecular mechanisms whereby obesity and cancer are linked, with particular reference to postmenopausal breast cancer. In this case it should be emphasized again that most postmenopausal breast cancers are ER-positive and that the ovaries cease to synthesize estrogens at the time of menopause. There is now widespread acceptance of the view that the increased risk of breast cancer in postmenopausal women is due to the production of estrogens by the adipose tissue (reviewed in 7). Estrogen synthesis in adipose tissue increases with obesity but also with aging and is associated with an increased expression of aromatase, the enzyme responsible for estrogen biosynthesis (8, 9). Thus the importance of estrogen formation in the adipose tissue and aromatase expression in the breast adipose, in particular, has been emphasized. Accordingly, in this review we devote space to discussing how inflammatory processes and dysregulated metabolism affect aromatase expression in the breast. A schematic of the gene encoding human aromatase (CYP19A1) is shown in Figure 1.
Diagram of the human CYP19A1 (aromatase) gene. The 9 coding exons numbered II through X are shown in yellow, and the untranslated first exons are shown in red together with associated promoters. Also indicated are stimulatory factors and coregulators associated with each promoter. Abbreviations: ERRγ, estrogen related γ; FSH, follicle stimulating hormone; HBR, heme binding region.
Inflammation and Adiposity
Obesity is associated with inflammation in the adipose tissue, which is classified as subclinical and is characterized by infiltration of macrophages into the adipose depots in both subcutaneous and visceral sites (10). These macrophages typically form what are known as crownlike structures around the lipid-filled adipocytes of obese individuals and characteristically stain for CD68. Hypoxia is another characteristic of adipose tissue of obese individuals and may be due in part to the physical limitation of the blood vessels to supply the large lipid-engorged adipocytes (11). This is associated in turn with an increase in the expression of hypoxia-inducible factor 1-α (HIF1α) by these adipocytes, which increases the levels of monocyte chemotactic protein 1. This in turn stimulates the recruitment of macrophages to these sites. Saturated fatty acids produced by the lipid-engorged adipocytes can act to stimulate the activity of inflammasomes present in both macrophages and adipocytes (12), which gives rise to a cascade featuring caspase 1, IL-1β, and ending with nuclear factor κB (NFκB). In addition, saturated fatty acids such as palmitate can increase the activity of toll-like receptors, such as toll-like receptor-4 (13), which can also stimulate the formation of NFκB. This inflammatory cofactor together with the macrophage recruitment factors can bring about an increase in the expression of inflammatory mediators such as tumour necrosis factor α (TNFα), IL-6, and prostaglandin E2 (PGE2). These results are summarized in Figure 2. This figure also indicates that these factors can in turn stimulate the expression of aromatase in the adipose tissue, which is discussed in the next section.
Inflammatory cascade associated with lipid-laden adipocytes characteristic of obesity. Adipocytes and macrophages synergize to increase the production of inflammatory mediators, including TNFα, IL-6, and PGE2. These inflammatory factors then act on the adjacent stroma to increase aromatase expression and estrogen biosynthesis. Abbreviations: MCP-1, monocyte chemotactic protein 1; MMIF, macrophage migration inhibitory factor.
Inflammation and Breast Cancer
These inflammatory factors have been suggested to contribute to the increased risk of breast cancer progression and mortality. For example, the NFκB signaling pathway has been shown to be important for growth of anti-estrogen-resistant breast cancer cells using an macrophage chemotactic factor (MCF)7-derived cell model (14). Cyclooxygenase-2 (COX2), responsible for the rate-limiting step in PGE2 biosynthesis, has also been associated with increased breast cancer risk and several studies have indicated that the use of nonsteroidal anti-inflammatory drugs (NSAIDs), inhibitors of COX2, can reduce the risk of breast cancer in preclinical studies (15). Upregulation of COX2 and resulting increased PGE2 synthesis are recognized as a marker for the progression of many cancers including breast cancer (16). PGE2 activates 4 receptors, epinephrine (EP) 1–4. EP2 and EP4 are coupled to adenylate cyclase and cAMP formation, whereas EP1 is linked to phospholipase C, DAG formation, and activation of protein kinase C. Downstream targets of PGE2 include vascular endothelial growth factor (17) as well as the CRE-binding protein (CREB), MAPK, Src, and Akt pathways, and also HIF1α. On the other hand, TNFα acts primarily through the IKKβ/NFκB and mammalian target of rapamycin (mTOR) pathways, whereas IL-6 and other class 1 cytokines act via the gp130/Jak1/STAT3 pathway as well as the Ras/MAPK pathway.
One way in which TNFα, class 1 cytokines, and PGE2 can increase the risk of breast cancer is by stimulating the expression of aromatase in adipose tissue, particularly that of the breast. Aromatase synthesizes estrogens in adipose tissue from circulating androgens. It is expressed in the fibroblasts or stromal cells, which surround the adipocytes and the mammary ducts. In these cells, aromatase expression is regulated primarily by 2 promoters, the proximal promoters II/I.3 and a distal promoter, I.4. (Promoters II and I.3 are splice-variants of each other and for the purposes of this discussion can be considered as one) (reviewed in 18). Expression driven by promoter I.4 is regulated by TNFα via an NFκB-linked mechanism. In the case of class 1 cytokines such as IL-6, expression of aromatase driven by promoter I.4 is via a JAK1/STAT3 pathway, with STAT3 binding to a GAS element upstream of the promoter. On the other hand, PGE2 regulates aromatase via promoters II/I.3 (Figure 3). Two signaling pathways are employed; one is via the EP2 receptor linked to adenylate cyclase and phosphorylation of CREB, which binds to 2 CREs in the aromatase promoter II. The second pathway via the EP1 receptor is linked to phospholipase C stimulation with the resulting increase in protein kinase C activity. This in turn stimulates the expression of LRH-1, a monomeric orphan member of the nuclear receptor superfamily, which binds to a nuclear receptor half-site on the aromatase promoter II, downstream of the proximal CRE. This binding is absolutely required for aromatase promoter II activity (19).
Schematic of action of PGE2 to stimulate aromatase promoter II activity in human breast adipose stromal cells.
Recent studies from Dannenberg's group have indicated that crownlike structures, namely macrophages surrounding the large lipid-filled adipocytes of obese individuals, are present in the breasts of obese women (20). These are associated with increased expression of NFκB and their numbers correlate with an increase in the expression and activity of aromatase in the adipose tissue of the breasts of these women. Further studies from this group have shown that the increase in aromatase expression is associated with an increase in activity of the proximal promoters II/I.3 and in turn is correlated with an increase in COX2 expression and PGE2 levels in the breasts of these women (21). COX2 is also expressed in many breast carcinomas where it correlates with tumor size, high-grade HER2 positivity, and a worse disease-free interval.
Previous studies have examined the activity and expression of aromatase in the tissue of breast quadrants obtained at the time of mastectomy because of the presence of a tumor. These studies found that aromatase expression (22) and activity (23) were highest in the quadrant of the breast that contained the tumor, and that there was a gradient of aromatase expression that was highest in the tumor and in the tumor-bearing quadrant and decreased with increased distance from the tumor. This suggested that the tumor produced a factor or factors that stimulated aromatase expression in the surrounding adipose tissue and in particular the cancer-associated fibroblasts. When this aromatase expression was examined, it was found that the increase was due primarily to an increase in expression from promoters II/I.3, strongly suggesting that it was PGE2 produced by the tumor that was largely responsible for this increase in expression (24). Taking these data together, the results are indicative that inflammatory mediators and, in particular, PGE2 produced both by the adipose tissue in the breasts of obese women and by breast tumors can drive aromatase expression locally in the breast (Figure 2). This in turn leads to the local production of estrogens within the breast that stimulate the proliferation of the tumorous breast epithelium due to a positive feed-on mechanism resulting from epithelial-mesenchymal interactions.
Dysregulated Metabolism and Carcinogenesis
The concept of dysregulated metabolism and carcinogenesis was first enunciated by the great German biochemist/cell biologist Otto Warburg over 80 years ago. Otto Warburg received the Nobel Prize in 1931 for the discovery of cytochrome oxidase but he also showed that tumor cells exhibit high rates of aerobic glycolysis, a phenomenon that became known as the Warburg effect. Warburg then stated that “this was the cause of cancer” (25). This concept lay fallow for decades because people lost interest in metabolism following the development of molecular biology and recombinant DNA technology and it is only in the last 10 to 15 years or so that interest in metabolism has revived with the development of techniques to study its regulation. Therefore, interest in the Warburg effect has also revived and certainly it correlates with proliferative capacity and is the basis for positron emission tomography scanning.
However it has been generally assumed that these changes in metabolism follow the changes in gene expression associated with cellular proliferation. However, there is increasing evidence that the reverse is also true, thus at least partially vindicating Warburg. For example, oncogenes such as HIF1α and Myc are potent direct stimulators of glycolysis at multiple steps, including glucose uptake, hexokinase, phosphofructokinase, pyruvate kinase, and lactate dehydrogenase (26). In addition, HIF1α stimulates pyruvate dehydrogenase kinase, which is inhibitory of pyruvate dehydrogenase and thus inhibits the entry of pyruvate into the mitochondria and the formation of acetyl CoA. Thus these oncogenes recapitulate many aspects of the Warburg effect. Furthermore, the downstream signaling pathway of growth factors that activate mTOR, such as insulin, IGF-1, and epidermal growth factor, also regulates metabolism in the direction favoring cell proliferation. Thus mammalian target of rapamycin complex 1 (mTORC1), via posttranscriptional mechanisms, increases the levels of HIF1α (27), which as we have seen, results in a stimulation of glucose uptake and glycolysis. However, it also stimulates the processing and activation of SREBP1 and SREBP2, which lead to stimulation of lipid and sterol biosynthesis, respectively (28). Interestingly these factors also stimulate the oxidative limb of the pentose phosphate pathway, which is required for the production of reduced nicotinamide adenine dinucleotide (NAD) phosphate to drive lipid and sterol biosynthesis (Figure 4). These are examples of factors with oncogenic activity that are directly involved in dysregulated metabolism.
Action of growth factor signaling cascade to stimulate glycolysis and lipid and sterol biosynthesis. Abbreviations: PI3K, phosphoinositide 3-kinase; RTK, receptor tyrosine kinase; SREBP, sterol regulatory element-binding protein.
Other factors do the opposite, namely, those which have tumor-suppressing activity. Examples of these are p53 and AMP-activated protein kinase (AMPK), and indeed p53 is inhibitory of glycolysis at multiple steps, including the uptake of glucose via glucose transporter, GLUT1 and GLUT4; it stimulates TIGAR, which is inhibitory of phosphofructokinase-2. Thus it lowers the levels of fructose-2,6-bisphosphate and so inhibits phosphofructokinase-1 and thus glycolysis. p53 also inhibits phosphoglucomutase. On the other hand, it stimulates the expression of SCO2 (stimulator of cytochrome oxidase 2), which increases the expression of cytochrome oxidase in the mitochondria and thus stimulates oxidative metabolism via the mitochondrial respiratory chain (29). In the case of AMPK, the liver kinase B1 (LKB1)/AMPK pathway is now recognized to be a master regulator of energy homeostasis. AMPK in general stimulates pathways that are involved in the generation of energy, such as glucose uptake, glycolysis, fatty acid oxidation, and mitochondrial biogenesis, and is inhibitory of pathways that require energy, such as fatty acid and cholesterol biosynthesis, gluconeogenesis, mTOR activation, and protein biosynthesis (30).
Thus it appears that metabolism is a direct and not simply an indirect response to growth factor action. In what has been called the traditional demand model (31), the growth factor action is envisioned primarily to be to stimulate the transcription of genes involved in cell proliferation and their translation. This results in a decrease in the ATP to ADP ratio, which leads in turn to an increase in glycolysis and oxidative metabolism via the trichloroacetic acid (TCA) cycle and mitochondrial respiratory chain, to generate sufficient ATP to maintain the increased rate of cell proliferation. In this case, most of the carbon from glucose and other fuel metabolites ends up as carbon dioxide as a consequence of TCA cycle and mitochondrial respiratory chain activity. In the new so-called supply-based model of growth factor action (Figure 5) (31), the primary action of the growth factors is to stimulate glycolysis at multiple steps and also to redirect mitochondrial metabolism such that instead of the intermediates of the TCA cycle being oxidized to carbon dioxide, they and also intermediates of glycolysis are used to generate metabolites that serve as intermediates for the biosynthesis of proteins, purines, pyrimidines, and lipids. Thus, for example, 3-phosphoglycerate can be converted to serine and hence on to glycine and cysteine; Myc stimulates the uptake by the cell of glutamine, an important fuel metabolite, and its conversion in the mitochondrion to glutamate by the enzyme glutaminase. Glutamate in turn can be converted in the mitochondria to α-ketoglutarate and on to oxaloacetate and aspartate. Oxaloacetate together with acetyl CoA derived from pyruvate via the action of citrate synthase forms citrate, which can exit the mitochondria and be converted via ATP citrate lyase to acetyl CoA in the cytoplasm and hence on to fatty acids. ATP citrate lyase itself is stimulated by Akt, an important intermediate downstream of growth factor signaling.
Metabolism is a direct, and not an indirect, response to growth factor action; diagram of role of growth factors to reprogram metabolism to provide intermediates for the synthesis of proteins, lipids, and nucleic acids, rather than oxidation to CO2. Abbreviations: Asp,aspartate; α-KG, α-ketoglutarate; CS, citrate synthase; GLS, glutaminase; OAA, oxaloacetate; 3-PG, 3-phosphoglycerate; NADPH, reduced NAD phosphate; PDH, phosphate dehydrogenase.
In his original formulation, Warburg envisaged that the increase in glycolysis observed in tumor cells was a consequence of unspecified mitochondrial damage. Therefore, in this aspect he was incorrect because what in fact is taking place is that the mitochondria are reprogrammed. Therefore, instead of oxidizing TCA cycle intermediates to carbon dioxide, these intermediates are used to provide precursors for the synthesis of proteins, purines, pyrimidines, and lipids, all of which are essential for cell proliferation. Thus the supply of these key precursors comes under the control of metabolism and not the other way around (31).
However there is yet another even more direct way whereby metabolism can regulate the expression of genes, as was discussed in a recent review by Sassone-Corsi and his group (32). What they point out is that each cell contains 2 meters of DNA, that histone H3 tails play a critical role in gene regulation, and that there are approximately 3 × 109 of these per cell. The activity of these tails is in turn regulated by epigenetic modifications, mainly acetylation, methylation, and phosphorylation. The importance of epigenetic signals has been emphasized by the recent large-scale ENCODE study (33). The enzymes involved in these modifications, namely the markers, require as cosubstrates, acetyl CoA, S-adenosylmethionine, and ATP, respectively. These of course require metabolism for their formation and the epigenetic demand is such that it appears to be regulated by the metabolic activity of the cell. Thus gene regulation itself and not just the supply of precursors for proteins, lipids, and DNA would appear to come under the control of metabolism.
Dysregulated Metabolism and Breast Cancer
As indicated above, obesity must be considered the most common state of dysregulated metabolism in the human population. In addition to the role of inflammation in obesity-linked cancers, which has already been discussed in the context of breast cancer, examples of dysregulated metabolism in obesity include insulin resistance, increased synthesis of leptin by the adipose tissue (34), and decreased synthesis of adiponectin (35). Insulin as well as IGF-1 have both been associated with stimulation of growth and increased risk of a number of cancers. In particular, hyperinsulinemia, caused by insulin resistance in the liver, skeletal muscle, and adipose tissue, often predates the diagnosis of type 2 diabetes and has been shown to link obesity and type 2 diabetes to cancer (reviewed in 36). Higher levels of IGF-1 have also been correlated with elevated risk of cancer. In general, the mitogenic actions of insulin and IGF-1 are mediated by the activation of Ras and the MAPK pathway. However, both insulin and IGF-1 also signal via pathways involving phosphoinositide 3-kinase (PI3K), Akt, and mTORC1. mTORC1 activates the p70S6 kinase and inhibits 4E-BP1, in both cases leading to a stimulation of protein synthesis and in turn to increased cellular proliferation (37). Although leptin is generally viewed as signaling via a JAK-STAT pathway, it also signals through the PI3K, Akt, and mTORC1 pathway, leading to activation of p70S6 kinase. Recently it has been shown that p70S6 kinase phosphorylates AMPK on serine 491 of the α-2 subunit. This is an inhibitory site and so signaling through this pathway leads to inhibition of AMPK activity. Furthermore it was shown that leptin inhibits the activity of AMPK in the hypothalamus via phosphorylation of this site by p70S6 kinase (38).
AMPK is generally considered to be antitumorigenic and its role to inhibit lipid and cholesterol biosynthesis is well recognized (reviewed in 39). However, it also inhibits the activity of mTORC1. The tumor suppressor proteins TSC1 and TSC2, which are components of the mTOR complex, are known substrates of AMPK and many sporadic cancers are associated with mutations in TSC1 and TSC2, including those of breast, prostate, endometrium, colon, and lung (40). By inhibiting mTORC1 in this fashion, AMPK is inhibitory of protein synthesis. However, AMPK also plays an important role in the regulation of the tumor suppressor p53 whereby activation of AMPK leads to the upregulation of p53 and its phosphorylation at serine 15 (41). The apoptotic actions of p53 are well known but, in addition, p53 activates the cell-cycle inhibitors p21Sip and p27Kip and thus has an inhibitory effect on the cell cycle (42). Thus AMPK can inhibit cell proliferation at multiple sites. In addition, the emerging role of p53 as an inhibitor of aerobic glycolysis and stimulator of mitochondrial respiration provides an additional mechanism whereby AMPK regulates metabolism.
As stated above, leptin via its action to stimulate mTORC1 and thus p70S6 kinase would be expected to inhibit the activity of AMPK; this has been shown in the case of the hypothalamus and in breast adipose stromal cells (38). On the other hand, adiponectin, which although synthesized in the adipose tissue, is reduced in obesity, has been shown to activate AMPK in a number of cell types, including hepatocytes, myocytes, preadipocytes, and adipocytes, and it is known to inhibit the proliferation and metastases of breast cancer cells. In the ER-positive MCF7 and T47D cells, adiponectin treatment leads to an increase in AMPK activity and inhibition of p70S6 kinase, which is dependent on LKB1 (43). This has also been shown in MDA-MB-231 breast cancer cells. Moreover the adiponectin receptor peptide agonist ADP-355 has been shown to increase AMPK activity and inhibit the growth of orthotopic human breast cancer xenografts (44). Thus, although there are a few reports to the contrary, most studies suggest an antiproliferative role for AMPK in breast cancer.
As indicated above, most postmenopausal breast cancers are estrogen-dependent and evidence suggests that it is the estrogen produced locally within the breast that is responsible for the increased proliferation of cancer cells. Recently AMPK has been shown to be a negative regulator of aromatase, the enzyme responsible for the biosynthesis of estrogens in human breast adipose stromal cells (45, 46). Results demonstrate that AMPK prevents nuclear translocation of the CREB coactivator CREB regulated transcription coactivator 2 (CRTC2) in these cells. This was shown initially by Montminy's group in terms of inhibition of PEPCK gene expression in the liver (47). This is due to phosphorylation of CRTC2 by AMPK, resulting in its sequestration in the cytoplasm by 14-3-3. However a similar result has been shown in human breast adipose stromal cells (Figure 6). CRTC2, a potent inducer of aromatase, is sequestered in the cytoplasm in situations where AMPK is activated (45). This appears to be predominantly due to an increase in expression of LKB1 as seen in cells stimulated by adiponectin, resulting in an increased phosphorylation of AMPK at threonine 172, which results in activation of AMPK.
Action of PGE2 and leptin to stimulate, and adiponectin and metformin to inhibit, aromatase promoter II expression in human breast adipose stromal cells. Abbreviations: CRE, cAMP response element; CRTC, CREB regulated transcription coactivator; LKB, liver kinase B; PKA, protein kinase A.
On the other hand, the inflammatory mediator PGE2 causes a downregulation of LKB1 expression and a decrease in phosphorylation of AMPK at threonine 172. In addition, PGE2 via PKA causes an increase in phosphorylation of AMPK at the inhibitory serine 485/491 sites on the α-1/2 subunits. The net effect of this inhibition of AMPK is to permit the translocation of CRTC2 to the nucleus where it activates CREB and hence aromatase via promoter II. Similarly, leptin also causes a decrease in LKB1 expression and a decrease in the phosphorylation of AMPK at T172 (Figure 6) (45). However, the fact that leptin stimulates p70S6 kinase in the hypothalamus resulting in an increased phosphorylation of AMPK at the inhibitory serine 491 site (38) suggests that this might also pertain in the human breast adipose stromal cells, although this is still to be established. Thus yet another way whereby obesity leads to an increase in breast cancer risk is likely due to the increased formation of leptin and the decreased formation of adiponectin in the breast adipose tissue of obese women, resulting in an increase in the expression of aromatase.
Compounds That Target Breast Cancer Therapeutically via Inflammatory and Metabolic Pathways
The fact that inflammatory mediators as well as factors involved in dysregulated metabolism are involved in breast cancer proliferation, invasion, and metastasis has led to a growing number of studies involving the use of these agents to determine their therapeutic use in breast cancer prevention and treatment. In terms of endocrine therapy for breast cancer, this field is currently dominated by the use of aromatase inhibitors, which are generally proven to be superior to tamoxifen in both neoadjuvant and adjuvant settings. The 3 compounds in clinical use are letrozole, arimidex, and exemestane. However there are a number of contraindications associated with the use of these compounds, such as bone loss, joint pain or arthralgia, hot flashes, and possibly cognitive defects, as some studies indicate. The reason for this is that these compounds inhibit the catalytic activity of aromatase and hence inhibit its activity globally, not just in the breast but in other body sites where estrogens have important roles to play such as bone, brain, and the cardiovascular system. Ideally, one would like to target the breast specifically and leave other body sites unaffected. This is possible to do postmenopausally in the case of aromatase because the aromatase promoters II/I.3 appear to be employed specifically in the breast in postmenopausal women in whom the ovaries no longer synthesize estrogens, and bone and brain employ different aromatase promoters.
Factors that target inflammation
As indicated above, COX2 is believed to be a key factor during tumor initiation in tissues subject to chronic inflammation (48). Furthermore COX2 expression in human breast cancer is correlated with reduced survival, increased tumor size, high tumor grade, Her-2 overexpression, as well as metastases to lymph nodes and other organs. Moreover COX2 is overexpressed in roughly 50% of breast cancer specimens inclusive of ductal carcinoma in situ and invasive carcinomas. As a consequence, several epidemiological studies including prospective, case-control studies as well as meta-analyses have sought to determine the efficacy of NSAIDs in terms of breast cancer prevention and treatment. The results from these studies have proven to be somewhat inconclusive (49). Although most studies have found a small benefit from the use of aspirin, results on the use of ibuprofen have been mixed with some studies showing as much as a 40% reduction in breast cancer risk (50) and others showing no benefit. At least one study was able to obtain data in a case-controlled study of the use of selective COX2 inhibitors for 2 years or more, namely, celecoxib and rofecoxib, before these compounds were withdrawn from the market (51). In this study 2 years' use of aspirin led to a benefit odds ratio (OR) of 0.5, whereas similar use of ibuprofen led to an OR of 0.4. The results obtained for the use of COX2-specific inhibitors such as rofecoxib led to a multivariate OR of 0.3. On the other hand, use of acetaminophen, which has little effect on COX2 activity, showed no benefit.
Several in vitro studies have also been published examining the effect of COX2 inhibitors (reviewed in 51) as well as inhibitors of the PGE2 receptors on both the proliferation of breast cancer cells and the adipose stromal cells. In addition, one study used a syngeneic mouse breast cancer model of spontaneous lymphatic metastases (52). In general, these studies, whether they used as endpoints cancer cell migration and invasiveness or else aromatase expression (53), have found that mixed COX1- and COX2- or COX2-specific inhibitors inhibited proliferation, migration, and invasiveness as well as aromatase expression. Inhibitors of the 4 PGE2 receptors (54) have also been examined and in general inhibitors of the EP2 and EP4 receptors that are linked to adenylate cyclase were effective, whereas those targeting the EP3 receptor were less effective or ineffective (52, 55). Therefore, although the in vitro data are strongly suggestive of the efficacy of NSAIDs in terms of therapeutic benefit, the clinical data remain somewhat inconclusive. Perhaps this will remain the case until such time as specific COX2 inhibitors are developed that have no potentially life-threatening contraindications.
Factors that target dysregulated metabolism
As mentioned above, AMPK is now generally recognized to be a master regulator of energy homeostasis. For example, it phosphorylates and thus inhibits acetyl CoA carboxylase 1 as well as HMG-CoA reductase, thus inhibiting both lipid and cholesterol biosynthesis. It also inhibits mTORC1 and thus inhibits angiogenesis, cell growth, and metabolism. However it also plays an important role in the regulation of the tumor suppressor p53 by leading to its upregulation as well as phosphorylation on serine 15. This in turn leads to inhibition of the cell cycle via the release of the sequestered CDK inhibitors p27Kip1 and p21Sip1. Thus AMPK acts to inhibit cellular proliferation by inhibition of protein, lipid, and cholesterol biosynthesis as well as by inhibition of the cell cycle (reviewed in 56). Consequently then, there is much interest in the development of factors that stimulate AMPK, not only in the context of its antidiabetic effects but also in the context of its potential ability to suppress cancer proliferation.
The only drug to stimulate AMPK that is currently in clinical use is the antidiabetic drug metformin and most if not all of the actions of metformin are believed to be mediated by stimulation of AMPK (reviewed in 57). However metformin does not act directly on AMPK and the mechanism of this stimulation is not entirely clear. One major action seems to be the inhibition of complex 1 of the mitochondrial respiratory chain (58). This results in an increase in the ratio of AMP-ADP/ATP, which would bring about an activation of AMPK. This is due to binding of AMP or ADP to the γ-subunit of AMPK, which causes a conformational change, allowing phosphorylation of the α-subunit at T172 by LKB1 or in some cases by CaMKKβ (59, 60).
Because of reported decreases in cancer incidence in diabetic patients treated with metformin, a number of studies have been undertaken or are being undertaken to determine whether this commonly prescribed drug may be useful for the treatment of a number of cancers, including those of the breast, colon, and endometrium. For example, in a preoperative window of opportunity randomized trial, nondiabetic women with operable invasive breast cancer were given metformin for 2 weeks (61). This resulted in a significant decrease in the proliferation marker Ki67, although insulin levels remained stable. Moreover, the treatment was associated with, for example, changes in TNFα signaling as well as the cell-cycle inhibitors. Genome-wide studies in a number of breast cancer cells have also identified that metformin regulates pathways involved in cell proliferation, for example, inhibitory effects on ribosomal proteins and mitosis-related gene family members (62). Metformin has also been shown to inhibit the proliferation of several cancer cells in culture including those of breast, prostate, ovary, colon, and pancreas (63–66). The effect of metformin on specific breast cancer subtypes has also been explored. The expression of LKB1 is required for these actions of metformin because it has been shown that LKB1-deficient MDA-MB-231 cells were unaffected by metformin treatment. On the other hand, in LKB1 expressing breast cancer cells, including MCF7 cells, metformin treatment led to the activation of AMPK and inhibition of the cell cycle via the release of the inhibitors p27 and p21 (64). Moreover, metformin was shown to cause a 30% decrease in global protein synthesis attributed to a decrease in translation initiation as a consequence of mTOR inhibition (67). Metformin also reduces insulin and IGF-1 signaling, suggesting that it may also inhibit tumor growth via these mechanisms (68). Thus, in an orthotopic model of ER-negative breast cancer, treatment with metformin caused a decrease in systemic IGF-1 and inhibition of tumor cell proliferation (66).
Consistent with the action of AMPK to inhibit aromatase expression in human breast adipose stromal cells mentioned above, metformin also significantly inhibited the expression of aromatase in these cells at micromolar concentrations similar to the concentrations present in the blood of women treated with metformin for type II diabetes (69); this inhibition was associated with increased expression of LKB1 and phosphorylation of AMPK at T172 and with a decreased nuclear translocation of CRTC2 (Figure 6) (45, 46). The effects of metformin on aromatase expression were also shown to be promoter-specific, namely use of promoter II (70), implying that treatment may inhibit estrogen production specifically within the breast and thereby prevent side effects associated with current endocrine therapy.
Because of these beneficial actions of metformin mediated by AMPK, albeit indirectly, there is considerable interest in developing specific activators of AMPK. However this turns out to be complicated because AMPK is a heterotrimeric enzyme composed of α, β, and γ subunits and multiple isoforms of each subunit have been identified, each of which is encoded by a distinct gene and these are expressed differentially in a tissue-specific fashion. Nevertheless, Lee et al (71) described that the AMPK activator OSU53 significantly inhibited the viability and clonogenic growth of triple negative breast cancer cells in vitro and in vivo, leaving the nonmalignant MCF10A cells unaffected. It also caused an almost 50% decrease in the growth of MDA-MB-231 cells in tumor-bearing mice. AICAR is a commonly used AMPK activator in in vitro studies, which is dependent on its conversion to ZMP, which is an AMP analog. Nevertheless this compound is unsuitable for clinical use.
Another compound of potential interest is resveratrol, a polyphenol found in red wine, which has been reported to be a calorie restriction mimetic with potential anti-aging and anti-diabetogenic properties in mouse models of diet-induced obesity (72) and in obese humans with impaired glucose tolerance (73), but not in nonobese women with normal glucose tolerance (74). The action of resveratrol has been thought primarily to be mediated by activation of SIRT1; however, recently evidence has been presented that its primary action is to activate AMPK. This results from the inhibition of cyclic AMP degrading phosphodiesterases, leading to elevated cyclic AMP levels. This results in the activation of EPAC1, a cyclic AMP effector protein, which in turn activates the CAMKKβ/AMPK pathway (75). AMPK is known to increase the cellular concentrations of NAD+ due to an increase in nicotinamide phosphoribosyltransferase activity and NAD+ in turn is an activator of SIRT1, thus apparently explaining the ability of resveratrol to activate SIRT1. Interestingly in this context, resveratrol has been shown to inhibit proliferation of breast cancer cells in an AMPK-dependent manner and independent of hormone receptor status (76) and also inhibits aromatase activity in breast cancer cells (77).
Conclusions
Interest in metabolism has revived dramatically in the last decade or so with the development of techniques to study its regulation and with the realization that dysregulated metabolism is a key player on center stage in carcinogenesis. Thus there is hardly a current issue of journals such as Cancer Cell and Nature Reviews Cancer that does not feature some aspect of the relationship between these 2 processes. Likewise, inflammation has also emerged as a leading role in cancer biology. Obesity provides a direct link between inflammation and dysregulated metabolism and, not surprisingly therefore, has an emergent role in the etiology of numerous cancers.
Although many factors play a part in this linkage, the role of obesity in postmenopausal breast cancer must also be seen in the context that estrogen plays a dominant role in driving this disease. It would seem plausible therefore that obesity should also play a role in the regulation of estrogen biosynthesis in adipose tissue, and this indeed is the case, metabolically in terms of the interplay between leptin and adiponectin and also because of the role of inflammatory mediators as stimulators of aromatase expression, especially PGE2, produced both in the adipose itself and in the tumor. Furthermore, evidence is emerging that these factors play a role in endometrial cancer, which is also estrogen-dependent and linked to obesity. As interest in obesity and carcinogenesis gains momentum, it is likely that we are seeing only the tip of the iceberg in terms of new knowledge and new facets of this deadly connection.
Acknowledgments
This work was funded by National Health and Medical Research Council (NHMRC, Australia) Project Grant GNT1005735 (to K.A.B. and E.R.S.), the Victorian Government, through the Victorian Cancer Agency funding of the Victorian Breast Cancer Research Consortium (to E.R.S. and K.A.B.), and by the Victorian Government Operational Infrastructure Support Program. E.R.S. is supported by an NHMRC (Australia) Senior Principal Research Fellowship GNT0550900. K.A.B. is supported by an NHMRC (Australia) Career Development Award GNT1007714.
Disclosure Summary: The authors have nothing to disclose.
Abbreviations
- AMPK
AMP-activated protein kinase
- COX2
cyclooxygenase-2
- CREB
CRE-binding protein
- CRTC2
CREB regulated transcription coactivator 2
- EP
epinephrine
- ER
estrogen receptor
- HIF1α
hypoxia-inducible factor 1-α
- LKB1
liver kinase B1; MCF
- mTOR
mammalian target of rapamycin
- mTORC
mammalian target of rapamycin complex 1
- NAD
nicotinamide adenine dinucleotide
- NFκB
nuclear factor κB
- NSAIDs
nonsteroidal anti-inflammatory drugs
- OR
odds ratio
- PGE2
prostaglandin E2
- PI3K
phosphoinositide 3-kinase
- TCA
trichloroacetic acid
- TNFα
tumour necrosis factor α.
