Melding breast and prostate cancers alter egos

Keywords::

• androgens
• coregulators
• estrogens
• hormone receptor
• nuclear transcriptional factor

“…his armour being furbished, his morion turned into a helmet, his hack christened, and he himself confirmed, he came to the conclusion that nothing more was needed now but to look for a lady to be in love with; for a knight-errant without love was like a tree without leaves … or a body without a soul.”

– An excerpt from Cervantes’ El ingenioso don Quijote de la Mancha.

Completed more than 400 years ago, The Ingenious Gentleman Don Quixote of La Mancha ranks among the greatest novels ever published. Even today, people are entranced by the timelessly delightful, although delusional, caricature of one Alonso Quijano. While the quest for honor and glory was at the crux of this chivalrous journey, Quijano’s alter ego is also symbolic of the power and frailty of the human mind. As the don’s sanity teeters between reality and ideality, it is in madness that the ego finds the strength and will to restore order to a world gone askew. Just as the quixotic transformation embodies the purity of the purpose, so does it also mirror the unique, albeit unsettling, way derangements in the genome bring a sense of order to the neoplastic process; and where genomic instability shifts the balance from vulnerability to invincibility.

It is almost counterintuitive to believe that results of scientific research can increase the complexity of already complicated diseases; and nearly as difficult to comprehend is the illusion that looking backward enables us to see forward. Yet both beliefs are apparently true for cancer. Now, almost five decades since the first reports of the estrogen receptor (ER) and androgen receptor (AR) materialized [1–3], research has revealed diverse mechanisms by which hormones regulate a number of dynamic, yet delicate, processes in mammary and prostate glands. In this special focus issue of Expert Review of Endocrinology and Metabolism, numerous authors have contributed articles related to the involvement of sex steroid hormones in breast and prostate cancers. Here, the reader can learn a little about the treatment of these tumors, more about how the science of endocrinology and study of metabolism might improve our understanding of cancer, and perhaps gain a better appreciation of the complexity of cancer genomics.

Over a century ago, an association between estrogens and breast cancer was first established [4,5]; 40 years later, a paper reporting a link between testosterone and prostate cancer was published [6]. While the involvement of estrogens and androgens is consistent with, and apparently largely responsible for, the hormone-dependent nature of these cancers, the precise mechanism by which these hormones induce transformation of normal cells to neoplastic cells is not clear. Histologically, both breast and prostate glands are made up of epithelial cells, myoepithelial cells and the surrounding stroma; the latter contains fibroblasts, smooth muscle cells and endothelial cells. The glandular epithelium is composed of basal and luminal secretory cells. Interestingly, both the ER and AR are expressed in luminal rather than basal epithelial cells [7,8]. The fact that most breast and prostate cancers arise out of the luminal epithelium appears to be consistent with the endocrine-dependent phenotype of these tumors [9]. Perhaps not surprising is the finding that the development of endocrine resistance in both cancers may also be mediated by similar mechanisms, a topic that will be discussed later.

Although sex steroid hormones have been shown to promote tumor growth and survival by affecting nuclear transcriptional activity, much less is known about the entire intracellular signaling network, critical gene mutations and relevant metabolic pathways. It is also not known whether the growth-promoting effects are accomplished only by tumor cells that harbor existing mutations. Moreover, it is conceivable that hormones could enhance the development of mutations in preneoplastic or quiescent tumor cells [10]. The apparent uncertainty suggests the involvement of multiple molecular elements, of which the hormones are merely one component. Several pieces of evidence support this belief. First and foremost, the regulatory effects of estrogens and androgens are mediated by specific receptors that belong to a superfamily of nuclear transcriptional factors [11]. Approximately 25 years after the discovery of ERα, the gene for the AR was cloned [12]. A second ER, ERβ, was identified in 1996, interestingly from prostate tissue [13,14]. Despite differing in homology, the hormone receptors share four functional domains: a transcriptional-activating amino-terminus, which contains constitutively active activation function (AF)1; an amino acid sequence with high affinity to estrogen response element (ERE) or androgen response element (ARE) in target genes; a hormone-binding carboxy-terminus containing ligand-dependent AF2; and a bridge region that connects the DNA- and ligand-binding domains. Although exquisite crystallographic models of the receptors have been constructed, the molecular basis of how the ligand-bound receptor interacts with the nuclear transcriptional elements is slowly becoming a part of biomedical lore.

Second, while the concept that hormonal action is mediated by receptors is simplistically correct, the process by which the hormone-receptor complex regulates gene expression is extremely complicated. Ultimately, transcriptional activation depends on recruitment of distinct coregulators as well as certain epigenetic proteins. Whereas many of the genes that affect tumor cell cycle, proliferation and survival have been identified, the manner in which this takes place depends on surreptitiously precise interactions with a number of transcriptional coactivators [15,16]. For example, estrogen (or androgen) binding at the C-terminal end induces conformational changes in the receptor, which sequentially leads to activation of intrinsic AF2, dissociation from heat-shock proteins, receptor phosphorylation and homodimerization [17]. The formed homodimer complexes bind to EREs (or AREs) in the promoter region of DNA. Activation of targeted genes necessitates recruitment of two chromatin remodeling complexes known as switch/sucrose nonfermentable (SWI/SNF)-A and SWI/SNF-B, as well as coactivators of the p160 family, including thyroid receptor-associated protein and vitamin D3 receptor-interacting protein [18–20]. These events require participation of two additional cofactors, steroid receptor coactivator-1 (SRC-1) and glucocorticoid receptor-interacting protein 1 (GRIP1). Both SRC-1 and GRIP1 possess two activation domains, AD1 and AD2, which serve as binding sites for p300 and coactivator-associated arginine methyltransferase 1 [21]. In essence, an intrinsic cellular program initiated by hormones follows a blueprint for the repetitive remodeling of chromatin, regulates the process of gene transcription and determines specific endocrine effects on target tissue.

Third, the importance of receptor coactivators notwithstanding, the full effect of hormones on target tissue also depends on the presence of corepressors, and perhaps the balance between corepressors and coactivators. This concept is relevant in cancer because receptor-mediated effects appear to result from both upregulation of specific genes that contribute to tumor growth and survival, as well as downregulation of genes that possess repressive effects on proliferation and apoptosis. Indeed, ERα- and AR-targeted gene expression can be downregulated in the presence of receptor corepressors that consequently blunt the effect of hormones on tumor cell proliferation [22,23]. While the impact of corepressors may be restricted to receptor-expressing (i.e., ER-positive) and hormone-dependent tumor cells, the development of hormone-independent breast and castrate-resistant prostate cancers could also be partly linked to diminished corepressor activity. This conclusion is supported by the following:

• • It has been shown that corepressors, such as nuclear receptor corepressor and silencing mediator for retinoic acid and thyroid hormone receptor, bind to unliganded ERα and are dislodged by estradiol [24];

• • These same corepressor complexes, which are recruited by anti-estrogens, could conceivably contribute to some of tamoxifen’s anti-tumor effect [25].

An extension of this concept was the finding that estrogens modulate an extensive repertoire of genes, not only involved in stimulating and/or promoting tumor cell growth (i.e., Cyclin D1, proliferating cell nuclear antigen [PCNA], VEGF and IGF-binding protein 4) and inhibiting cell death (i.e., Survivin, Caspase 9 and Inhibin, βB), but also genes for critical signaling pathways components, such as the chemokine ligand SDF-1 and its receptor CXCR4. Estrogens were also able to upregulate the oncogene c-myb and two familial transcription factors, A-Myb and B-Myb [26]. What these data revealed (in an ER-positive breast cancer cell line) was somewhat surprising. In contrast to the 130 genes that were upregulated by estradiol, nearly 70% (∼300) of the genes were downregulated by estrogen. In addition, these findings highlight one hallmark concept related to hormone-dependent normoplasia and hormone-dependent neoplasia. Cellular integrity in the former is based, in part, on a critical balance of coregulators that provides safeguards against the development of cancer; alternatively, these same coregulators can be manipulated in ways that favor accumulation of mutations, which may promote the development of the latter.

While undeniably already strong, further evidence of the intimate relationship between the sex hormones and breast cancer is the association between circulating estrogens and disease risk among postmenopausal women [27]. Interestingly, the importance of these endogenous hormones was also strongly correlated with tumor receptor status and in situ disease. In addition, the data indicated that higher circulating androgen levels were associated with an increased risk of developing hormone receptor-positive breast cancer, most notably in women with lower risk-predicted Gail scores (i.e., ≤1.66%). This particular finding builds upon two previously reported studies in which similar associations were observed [28,29]. The fact that these outcomes show a direct relationship between circulating androgens and postmenopausal ER-positive breast cancer, irrespective of perceived risks based on multivariate risk models, suggests that testosterone levels may be useful, either independently or additively (with other known risk factors), to identify women at high risk for the disease. While these data are indeed compelling, it should be mentioned that not all studies reported similar associations between circulating sex hormones and breast cancer in a high-risk group of postmenopausal women [30].

Data regarding serum testosterone levels and prostate cancer are in apparent contradistinction, as high levels of endogenous androgens have not been correlated with an increased risk of developing the disease [31,32]. In fact, an apparent paradox has been observed in that the risk of prostate cancer increases with age-related declines in serum testosterone levels [33,34]. Despite some limitations, the results of a number of studies indicate that an inverse relationship exists between testosterone levels and tumor features (i.e., grade, pathological stage, Gleason score and even survival) [35–38]. Nevertheless, the etiologic relationship between estrogens and breast cancer is no more widely accepted than the putative role of testosterone in prostate cancer as hormone-deprivation therapy can induce tumor regression in both cancer types.

While the hormonal data may merely be correlative, a causative role cannot be completely ruled out. Evidence, albeit circumstantial, which supports hormone involvement in the etiologies of breast and prostate cancers comes from chemoprevention and adjuvant trials of tamoxifen, the aromatase inhibitors and finasteride [39]. In both settings, the mode of operation of each class of agents is the induction of a hormone-deprived state. However, the outcomes are worth noting. In the P1 prevention trial, patients in the tamoxifen arm had a 49% reduction in risk of developing invasive breast cancer compared with women receiving placebo [40]; when used as adjuvant therapy and after a median follow-up of 8 years, anastrozole was superior to tamoxifen in all clinical end points (i.e., prolonged disease-free survival [p = 0.01] and time to recurrence [p = 0.0005]; and reduced distant metastases [p = 0.04] and contralateral breast cancers [p = 0.01] [41]), and a lower incidence of prostate cancer (18.4 vs 24.8%; p < 0.001) was observed among males treated with finasteride compared with men receiving placebo in the Prostate Cancer Prevention Trial [42].

However, these findings do not necessarily prove causation. Rather than the hormones themselves, is it possible that tumor initiation and progression are promoted primarily by alterations of hormone receptor expression and/or their functional activity? The following data suggest that this is indeed conceivable. In addition to its association with more aggressive disease, low serum testosterone levels have also been correlated with higher expression of the AR [43]. While low testosterone levels in the previous example were noted in men with newly diagnosed prostate cancer prior to surgery, low testosterone as a result of finasteride therapy was also associated with an increased risk of developing high-grade tumors [42]. The corollary in breast cancer is related to the finding that tamoxifen chemoprevention reduces the risk of developing hormone receptor-positive breast cancers; the effect on ER-negative tumors being much less actually reduces the reduction in risk of the disease from nearly 50 to 38%. An additional concern related to chemoprevention (perhaps more so with aromatase inhibitors than selective ER modulators) is the real, although as-yet unproven, possibility that estrogen-deprivation could increase the risk (or promote the development) of the poorer-prognosis HER2 or triple-negative phenotypes [44].

For reasons that are not completely understood, both breast and prostate cancers become hormone independent and hence refractory to endocrine manipulation. The fact that hormone resistance develops despite the maintenance of a functional receptor is an intriguing paradox [45,46]. Although multiple mechanisms have been proposed, one relates to the efficacy of hormonal therapies themselves as anti-estrogen (i.e., tamoxifen) and anti-androgen (i.e., flutamide) may select for mutant receptors in which they act as agonists, thus stimulating tumor growth [46,47]. Hormone deprivation is also associated with the upregulation of alternative growth factor signaling pathways, which results in phosphorylation of both receptors (independent of hormone binding) and induction of transcriptional activity of ER and AR target genes [48–51]. There is also evidence to suggest that endocrine resistance may be an intrinsic characteristic of some tumor cells. Despite the relative importance of the luminal cells, one other noteworthy aspect regarding breast and prostate cancers may reside in the basal epithelial cells. The observation that basal-like tumors have the worst overall prognosis may be partially explained by the absence of hormone receptor expression and a genomic profile characterized by high expression of proteins that mediate proliferation, invasion and survival [52]. However, an equally important reason may relate to the finding that basal-like tumors tend to be less differentiated, and even appear to possess stem cell-like properties [53]. Hence, the ‘emergence’ of resistance may be due to a subset of hormone-independent tumor cells that were present at the time of diagnosis.

Another truly remarkable finding (at least in breast tumors) is that long-term estrogen deprivation also appears to sensitize hormone-resistant tumor cells to the apoptotic effects of estrogen [54]. A corollary process in prostate cancer may be more difficult to demonstrate, which may be partly due to the degree and uniformity of androgen deprivation in the tumor microenvironment following surgical or medical castration [55]. This finding is also consistent with marginal suppression of AR gene transcription [56].

The next step in this extraordinary journey to better understand these tumors must incorporate the visionary expectations of genomics. It is now well appreciated (although not necessarily completely understood) that the complexity of the gene network in breast cancer (and probably in prostate cancer too) appears to have clinical importance. For example, hormone receptor-positive breast cancers are not homogeneous clinically or molecularly. In groundbreaking research, Sørlie and colleagues were able to further separate what was previously defined as ER-positive (luminal) tumors into two subgroups: luminal A (highest ERα gene expression) and luminal (B + C) [52]. Although both luminal subgroups expressed the ER, clinical outcomes, especially relapse-free survival, were significantly poorer among those labeled luminal (B + C). It was interesting that in the latter subgroup there were similarities in expression of some of the genes (i.e., mutant p53) that are characteristic of the poorer prognosis basal-like and HER2/neu-positive subtypes. It is evident that classification of tumors based on gene-expression patterns has provided deeper insight into the combinatorial effects various genes have on clinical outcomes. Partial validation of this belief is ensconced in a genomic test called Oncotype^® DX, which analyzes 21 genes collectively to predict the likelihood of disease recurrence and guide pharmacologic management of patients with early ER-positive, node-negative breast cancers.

Breast cancer (and, to a much lesser extent, prostate cancer) has truly benefited from the science of genomics. Will the future depend on advances made in our ability to perform (and replicate the findings of) a global analysis of the proteome? The hope for proteomics in cancer is to identify unique protein signatures that provide additive (or even more accurate) information about the risk or presence of early disease. The concept that this will not be an easy task is based on our current knowledge of the human genome, which consists of approximately 25,000 genes. While not trifle, this number corresponds to only 1–2% of the bases that encode proteins; even less is known about the noncoding, yet potentially important, regions of DNA [57]. What is also uncertain is the functional relevance of the 10 million common single-nucleotide polymorphisms (SNPs) and the 20 billion genotypes that would result if all DNA were analyzed for all SNPs. One example that exemplifies such a challenge is the persuasive evidence of a positive association between circulating estrogens and risk of breast cancer in postmenopausal women. The relevance of this finding relates to the hypothesis that polymorphisms of genes (and their encoded enzymes involved in estrogen synthesis and/or metabolism) could affect, or alter, breast cancer risk. However, investigations of what may be aptly termed ‘estrogenetics’ have provided, at best, inconsistent evidence to support this rational idea. The results of a relatively large prospective study that included American women of different ethnic backgrounds were published a few years ago [58]. Perhaps contrary to expectation, there was no evidence of an association between polymorphisms of a number of protein-encoding genes, including catecho-O-methyltransferase (COMT) Val⁵⁸Met, CYP1A1*2A, CYP3A4*1B, CYP1B1 Leu⁴³²Val, SULT1A1 Arg²¹³His and AHR Arg⁵⁵⁴Lys, and the disease. What the investigators did find was an inverse association between CYP1A2*1F and risk, which was stronger for hormone receptor-negative than receptor-positive breast cancer.

Another study investigated the association between genetic polymorphisms of several other genes involved in the same pathway and breast cancer risk in Thailand [59]. Data obtained from over 1000 women who were included in this study indicated that specific SNPs in CYP1A2 (homozygote carriers of rs762551), CYP2C19 (heterozygote carriers of rs4917623), AhR (heterozygote carriers of rs2066853) and ERRG (homozygote carriers of rs945453) may play an important role in estrogen metabolism and modify individual susceptibility to breast cancer in Thai women. When stratified by menopausal status, an association between CYP1A2 (rs762551) and CYP17 (rs743572) polymorphisms and breast cancer risk was evident in premenopausal women.

The frequency of polymorphisms related to estrogen metabolism (i.e., COMT Val¹⁵⁸Met; CYP17 (5´ untranslated region [5´UTR], T27C); HSD17 β1 Gly³¹³Ser; and MnSOD Val¹⁶Ala) was also investigated in Xavante Indians, a breast cancer-resistant population and compared with a breast cancer case-control population of Portuguese women [60]. The results of this study did not demonstrate an association between any of the unique genotypes and breast cancer risk. However, the MnSOD (Val¹⁶Ala) polymorphism appeared to have a protective role against the disease (adjusted odds ratio [OR]: 0.575; 95% CI: 0.327–1.011) among patients that never breastfed.

Polymorphisms in the CYP19 gene have been of particular interest because the gene product (aromatase) catalyzes the conversion of androgens to estrogens. The potential impact of this enzyme on breast cancer risk is supported by findings of increased activity and/or expression in experimental and clinical settings [61,62]. As such, and with some caveats, numerous investigations have implicated various CYP19 polymorphisms as a breast cancer risk factor. Allele frequencies of polymorphic tetranucleotide (TTTA_n repeats have been studied. Only two, the 10 (TTTA)₁₀ and 12 (TTTA)₁₂ repeat alleles, were found more frequently in some Caucasian populations but not in others [63–65]. Another disparate finding involves the COMT gene, which encodes a protein that metabolizes estrogen. Even though the relatively low activity allele COMT^Met genotype is more prevalent among Caucasians, this particular polymorphism confers a higher risk of breast cancer in Taiwanese women [66]. However, in American Caucasian females COMT^Met was associated with increased premenopausal risk (OR: 2.1; 95% CI: 1.4–4.3) but decreased postmenopausal risk (OR: 0.4; 95% CI: 0.2–0.7). Furthermore, the increased risk correlated with weight and menopausal status, and was strongest among the heaviest premenopausal women (OR: 5.7; 95% CI: 1.1–30.1) and leanest postmenopausal women (OR: 0.3; 95% CI: 0.1–0.7) [67]. Hence, polymorphisms of the same genes involved in the synthesis or metabolism of estrogens may differ (or be influenced) by ethnicity, lifestyle and environmental factors.

Finally, the importance of the ER and AR should not be underestimated as the receptors are the interface between hormone and tumor. Within the amino-terminus of the AR is a polymorphic CAG (polyglutamine) repeat region. Interestingly, in both breast and prostate cancer, CAG repeat polymorphisms have been associated with an increased risk of disease or tumor progression [68,69]. Because the number of CAG repeats vary between eight and 30, numerical length appears to be of importance but the implications differ by disease, gender and ethnicity. For example, longer CAG repeats among African–American women who have a first-degree relative with breast cancer significantly increases their risk for the disease [70]. However, inverse relationships between CAG repeats and risk for disease have also been observed in African–American men, who harbor the shortest contiguous repeats yet have the highest incidence of prostate cancer [71]. Short CAG repeat length in prostate cancer has also been correlated with earlier age and more advanced stage at the time of diagnosis [72,73]. By contrast, Asians with predominantly longer CAG repeats have the lowest incidence of the disease [74].

Although the articles in this journal add to our understanding of hormone-dependent tumors, this special focus issue also exposes how astonishingly little is known about the functional genome and its consummate impact on breast and prostate cancers. While the hope of estrogenetics and androgenetics in cancer is to prevent (or at least decrease the risk of) the diseases, the reality is that both fields of study remain a chimeric dream; part fantasy and part prophecy. Similarly, this paper is a chimerism of opinions as well as facts. As such, any case where interpretation of data made speculation and reality appear inseparable was purely coincidental. However, in a novel, yet still factual, way, Quixote’s quest to mend a fractured society is analogous to the relentless pursuit of researchers today who strive to improve our understanding of the deranged mechanisms of two endocrine-related cancers that could be alter egos of one another.

Financial & competing interests disclosure

The author has no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.