Exploring Breast Cancer Estrogen Disposition: the Basis for Endocrine Manipulation

Published OnlineFirst July 26, 2011; DOI: 10.1158/1078-0432.CCR-11-0043

Clinical Cancer Review Research

Exploring Breast Cancer Estrogen Disposition: The Basis for Endocrine Manipulation

Per E. Lønning1,2, Ben P. Haynes3, Anne H. Straume1,2, Anita Dunbier3, Hildegunn Helle1,2, Stian Knappskog1,2, and Mitch Dowsett3

Abstract Although normal breast tissue and breast cancer estrogens are known to be elevated compared with plasma estrogen levels, the mechanism behind this phenomenon has been an issue of debate for 2 decades. If local estrogen aromatization were to be confirmed as the main estrogen source in breast cancer tissue, tissue-specific inhibition of estrogen production, avoiding systemic side effects, would become a potentially attractive option for breast cancer treatment and prevention. Based on recent results from our groups exploring tissue estrogens, together with estrogen-synthesizing and estrogen- regulated gene expression levels, we propose a new model to explain elevated breast tissue estrogen levels. Although local estrogen production may be important, the local contribution is overruled by rapid plasma-to-tissue equilibration, including active uptake of circulating estrogens or enhanced tissue binding. As for breast cancer tissue levels, elevated levels of estradiol may be explained to a large extent by estrogen receptor binding and local conversion of estrone into estradiol. This model þ indicates that effective suppression of benign and malignant tissue estrogens as a treatment for ER breast cancer requires systemic suppression and will not be markedly affected by local enzyme targeting. Clin Cancer Res; 17(15); 4948–58. 2011 AACR.

Introduction ment. The notion that estrogens sustain the growth of ER- þ positive (ER ) cells is confirmed by the dramatic effects þ Estrogens play a pivotal role in breast cancer development achieved through antihormonal therapy for ER tumors in and sustained tumor growth. Observational studies con- both adjuvant (9) and metastatic (10) settings. The stimula- þ ducted before the era of estrogen replacement therapy tion of ER breast cancer cell proliferation in vitro by revealed that early loss of ovarian function reduced breast estrogens, particularly E2, strongly suggests that these clin- cancer risk by 60% to 70% (1, 2). More recently, a number of ical growth effects result from direct effects of estrogens on þ groups reported that high postmenopausal plasma estradiol ER breast cancer and not by intermediary pathways. (E2) levels are associated with a subsequent risk of breast Although multiple estrogen metabolites expressing vary- cancer development (see ref. 3 for references to original ing estrogen-agonistic activity have been identified, in work). Whether this relates only to growth stimulation of general 4 estrogen hormones are considered of interest: already transformed cells or includes promotion of carci- E2, estrone (E1), estriol (E3), and the estrone conjugate nogenesis (4, 5) is not fully understood. Of note, oophor- estrone sulfate (E1S). Although E3 plays a major role in ectomy reduces subsequent breast cancer incidence by 50% pregnancy, it is of little importance in nonpregnant (6) among BRCA1 carriers; however, about 80% of the women and will not be further considered here. Estrone breast cancers that arise in this group are estrogen recep- sulfate is derived by sulfate conjugation of E1 and may be tor–negative (ER ) basal-like (7, 8). This finding supports a deconjugated by sulfatases, leaving conjugated and uncon- role for estrogen during the early stages of cancer develop- jugated estrogens at equilibrium (11, 12). Estradiol is the active hormone that binds the ER (13) with a high affinity (kDa < 1 nmol/L). Although E and E S are biologically Authors' Affiliations: 1Section of Oncology, Institute of Medicine, Uni- 1 1 versity of Bergen, and 2Department of Oncology, Haukeland University inactive, many tissues express enzymes (sulfatase and 3 Hospital, Bergen, Norway; Department of Academic Biochemistry, Royal reductases) that convert E1S and E1 into active E2 (14, 15). Marsden Hospital, London, United Kingdom In premenopausal women, the main estrogen source for Note: Supplementary data for this article are available at Clinical Cancer breast tissue is circulating E , which is secreted by the Research Online (http://clincancerres.aacrjournals.org/). 2 ovaries. At menopause, ovarian production of estrogen Corresponding Author: Per E. Lønning, Department of Oncology, Jonas Lies vei 26, Haukeland University Hospital, N-5021, Bergen, ceases. In postmenopausal women, circulating E1, which Norway. Phone: 47-5597-5000; Fax: 47-5597-2046; E-mail: is produced in different tissues (e.g., skin, soft tissue, [email protected] muscle, and liver) by aromatization of androstenedione doi: 10.1158/1078-0432.CCR-11-0043 taken up from the circulation, is the main unconjugated 2011 American Association for Cancer Research. plasma estrogen. However, due to conflicting evidence, the

4948 Clin Cancer Res; 17(15) August 1, 2011

Breast Cancer Estrogen Levels

Translational Relevance sensitive and specific analyses of plasma, benign breast, and breast tumor estrogen levels. Second, we compare The contribution of local estrogen production to these findings with previous data from in vivo tracer infu- normal breast and breast cancer tissue estrogen levels sion studies. Third, we correlate the expression of hormone remains controversial. If local production is the main synthesizing enzymes in normal breast and breast cancer factor regulating normal breast and breast cancer tissue tissue with hormone levels, with the aim of identifying the estrogen levels, tissue-specific inhibition of estrogen key synthetic pathways regulating tissue estrogen levels. production may become an attractive option for breast Finally, we consider potential correlations between plasma cancer treatment as well as prevention. Tissue estrogen E2 values and intratumor expression of estrogen-regulated measurements obtained by highly sensitive assays, genes. An integration of these lines of evidence provides a new perspective on the importance of different compart- together with studies correlating estrogen-regulating þ enzymes and estrogen-regulated gene expression pro- ments for the estrogenic stimulation of ER breast cancer. files with tissue estrogen levels, are providing new evidence on the subject. Although local estrogen pro- Plasma Estrogen Levels, Production, and duction may be important, the data suggest a state of Metabolic Pathways rapid plasma-to-tissue equilibrium, including active uptake of circulating estrogens or enhanced tissue bind- Although ovarian estrogen synthesis ceases at meno- ing. As for breast cancer tissue levels, intratumor estra- pause, E2,E1, and E1S are all detected in the plasma of diol correlates with the estrogen receptor level and local postmenopausal women. conversion of estrone into estradiol. This model indi- Estrogens arise through aromatization (Fig. 1) of andro- cates that effective suppression of benign and malignant gens only (21). The gene coding for the aromatase enzyme þ tissue estrogens as a treatment for ER breast cancer (CYP19) contains at least 10 promoters, each of which gives requires systemic suppression. rise to mRNAs with unique untranslated first exons but to identical proteins because the translational start site is located in exon II (20). Of note, these promoters are differ- relative contribution from circulating versus locally synthe- entially activated in different tissue compartments, and sized estrogens to intratumor E2 levels in postmenopausal there are differences between benign and malignant tissues women has been controversial for 2 decades. Although derived from the breast; thus, promoter 1.4 seems to act as intratumor aromatase expression has been detected at the major regulator of aromatase expression in benign mRNA (16) and protein (17) levels, and aromatase activity breast tissue, in contrast to breast cancers, in which promo- has been detected in tumor specimens in vitro (18), the ters 1.3, 1.7, and PII seem to be the most active. Each tissue quantitative importance of local estrogen production has its own active promoters [e.g., 1.3, 1.4, and PI in adipose within tumors and adjacent benign tissues versus circula- tissue; 1.1, 1.2, and 2a in the placenta; and 1.4 and 1.6 in tory uptake has not been defined. bone (see ref. 20 for details and references to original work)]. The contribution of locally synthesized versus systemi- Although aromatase has been detected in most normal cally delivered E2 to intratumor hormone levels may tissue compartments (see references in ref. 21), the quanti- have important therapeutic implications. First, if local tative contribution from individual organs and tissue types aromatization is the main source of intratumor E2, to circulating estrogens has not been formally addressed. intratumor aromatase expression should predict sensi- Once thought in error to be a major source of circulating tivity to aromatase inhibitors (19). Second, although estrogens in postmenopausal women, the adrenal cortex is the aromatase protein is similar in all body compart- the main source of circulating androgens, although con- ments, gene transcription is regulated by tissue-specific flicting evidence suggests some minor ovarian contribution promoters (20) that are sensitive to stimulation by to circulating testosterone (22–24). Plasma androstene- different ligands (see below). This raises the possibility dione (A) and testosterone (T) levels average 5 to 7 of using pharmacologic organ- or tissue-specific estrogen nmol/L and 0.5 to 1 nmol/L, respectively, in postmeno- deprivation as an improved means of targeted therapy. pausal women (25). Overall, about 1% to 4% of circulating For example, breast-specific estrogen deprivation could A (and <1% of circulating T) undergoes aromatization (see avoid unwanted side effects in bone and genito-urethral references to original works in refs. 21 and 26); thus, E1 and tissue while suppressing breast cancer estrogen levels. E2 synthesis represents the major and minor pathways, Third, a strong quantitative contribution from alterna- respectively, of estrogen production in postmenopausal tive pathways, such as E1S deconjugation, would in women. Although to a substantial degree E2 may be contrast indicate a role for other classes of drugs, synthesized by reduction of E1 (27), the relative contribu- such as sulfatase or dehydrogenase inhibitors, in breast tion from each of these 2 pathways (aromatization of T to cancer therapy. E2 versus reduction of E1 to E2) to plasma and tissue E2 has The aim of this review is to examine the evidence not been formally assessed (Fig. 1). evaluating the contributions from different sources to Except during the first few days of the menstrual cycle, E2 normal breast tissue and intratumor E2. During this pro- constitutes the major unconjugated plasma estrogen in cess, we focus on 4 topics. First, we discuss data from recent premenopausal women. In postmenopausal women,

www.aacrjournals.org Clin Cancer Res; 17(15) August 1, 2011 4949

Lønning et al.

Adrenal cortex Cholesterol

2 5678 Pregnenolone Progesterone Aldosterone

3 3

2 56 17-OH-pregnenolone 17-OH-progesterone Hydrocortisone Figure 1. Pathways of estrogen 4 synthesis in postmenopausal O OH women. The numbers in the figure 4 indicate the following enzymes: 1 ¼ desmolase, 2 ¼ 3b-ol- 2 9 DHEA dehydrogenase-d-5 to 4 isomerase, 3 ¼ 17a hydroxylase,

4 ¼ C17,20-lyase, 5 ¼ 21- O O hydroxylase, 6 ¼ 11b- DHEA sulfate Androstenedione Testosterone hydroxylase, 7 ¼ 18-hydroxylase, 8 ¼ 18-hydroxysteroid oxidase, Peripheral conversion 9 ¼ different 17b-hydroxysteroid 10 10 dehydrogenases, 10 ¼ O O OH aromatase, and 11 ¼ estrone sulfatase and sulfotransferase. 11 9

S O HO HO Estrone sulphate Estrone Estradiol

plasma E1 levels average about 70 to 90 pmol/L, with E2 CYP1B1 are the main executors of 2- and 4-hydroxylations, levels of about 15 to 25 pmol/L, although there is substan- respectively (38, 39). In a second step, these metabolites are tial variation among individuals (25, 28). In contrast, after subjected to conjugation (Fig. 2). Although a minor menopause, levels within individuals vary only modestly amount is sulfate-conjugated and excreted in the bile, over a number of years (29). In addition, circulating E1S followed by substantial intestinal deconjugation and recy- concentrations may average from 5 to 600 pmol/L (30). cling (40), the bulk is subsequently glucuronidated and Estrone sulfate is synthesized by sulfate conjugation of E1 excreted through the urine (41). Similarly, circulating E1S (31); no direct secretion of estrone sulfate is known to occur. may be deconjugated by sulfatase and subsequently meta- Estrogens are metabolized and eliminated by a 2-step bolized as unconjugated E1 (34). process (Fig. 2). The first part involves hydroxylations So far, we have gained only a limited understanding of executed by multiple cytochrome P450 enzymes (CYP) which factors determine plasma and tissue estrogen con- (32, 33). The estrogen steroid nucleus may be exposed centrations in postmenopausal women. Several growth to hydroxylations at multiple positions (32). The main factors and cytokines have been shown to regulate aroma- metabolic pathways involve 2-hydroxylations and (to a tase expression in vitro (42), but little is known with respect lesser degree) 4-hydroxylations, generating catechol estro- to aromatase activity regulation in vivo. No feedback gens, and hydroxylations at the 16a position, generating mechanisms regulating plasma estrogen levels have been 16a-hydroxyestrone followed by reduction into E3 (34, identified. Although in vivo aromatization of androstene- 35). Although these enzymes have both E1 and E2 as dione correlates with body weight (see references in ref. substrates, some of them express a selective substrate pre- 21), the correlation between plasma estrogen levels and ference for either E1 or E2 (32, 36, 37). Of note, such in vivo aromatase activity is moderate (43), probably due to hydroxylations take place in different body compartments, interindividual variation in the estrogen elimination rate. including breast cancer and liver tissue, partly executed by Germline mutations in the aromatase gene are rare but different members of the CYP family. Thus, whereas estro- have severe consequences. Although a single germline gain gen-metabolizing CYPs such as CYP1A2 and CYP3A5 play a of function aromatase mutation has been described (44), major role in metabolizing estrogens in the liver (2- and aromatase deficiency due to inactive mutations that reveal a 16a-hydroxylation in particular), in the breast CYP1A1 and clinical phenotype resembling inactivating mutations that

4950 Clin Cancer Res; 17(15) August 1, 2011 Clinical Cancer Research

Breast Cancer Estrogen Levels

O O OH

17β-Hydroxysteroid Sulphatase dehydrogenases

Sulfotransferase S O HO HO Estrone sulphate Estrone Estradiol CYP CYP Minor metabolic CYP pathways (including CYP CYP CYP 4-hydroxylation)

O O OH OH

OH OH HO HO

HO HO HO HO 2-Hydroxyestrone 16-Hydroxyestrone Estriol 2-Hydroxyestradiol COMT COMT O Sulphate Glucuronide OH Glucuronide Sulphate Glucuronide CH3O CH3O

Sulphate HO HO 2-Methoxyestrone 2-Methoxyestradiol

Sulphate Glucuronide Sulphate Glucuronide

Figure 2. The main pathways of estrogen metabolism. Although multiple CYPs and hydroxylations in different sites of the estrone and estradiol nucleus are involved, the key pathways involve 2-hydroxylations and, to a lesser degree 4-hydroxylations, generating catechol estrogens, and the 16a-hydroxylation pathway, generating 16a-hydroxyestrone, and estriol. Following these hydroxylations (which occur mainly in the liver tissue), estrogen metabolites undergo either glucuronidation followed by urinary excretion (the major pathway), or sulfoconjugation followed by biliary excretion,after which they may be subject to enterohepatic circulation. affect the ER receptor gene (45) has been described in a few postmenopausal women. Measuring estrogen in plasma families. Multiple aromatase polymorphisms have been and particularly in tissue is not trivial, and nonoptimized detected, but few of these affect the coding region (46). methodologies have contributed significantly to the vari- Of interest, one of these intronic variants was reported to able data in the literature. Circulating estrogen levels in be associated with improved prognosis in patients treated postmenopausal women are among the lowest of all ster- with letrozole (47), and recently 2 variants, located oids assessed in clinical research or practice. Thus, plasma upstream of the promoter area, were found to be associated estrogen assessment requires radioimmunoassays (RIA) with enhanced aromatase activity (48). specifically derived for that purpose. The different CYPs are subject to multiple polymorphic Over the years, several studies have evaluated plasma and variants (33). However, although plasma E1 and E2 clear- tissue estrogen levels (see below). Due to the relatively low ance rates vary substantially (31, 49), so far no distinct CYP specific activity of 3H-labeled estrogens (50–150 mCi/ polymorphism has been correlated with plasma estrogen mmol), assays based on 3H-isotopes in general do not levels (33). provide the sensitivity required for such purposes. Our groups have made considerable efforts to develop 125I- Analyzing Plasma and Tissue Estrogen Levels based RIAs (specific activity about 2,000 Ci/mmol) for estrogen measurement in the low postmenopausal con- Before discussing data on estrogen concentrations in centration range (28, 50–53). With such methods, we have detail, we must consider certain methodologic limitations achieved sensitivity limits in the range of 0.67 pmol/L, 1.14 related to plasma and tissue estrogen measurements in pmol/L, and 0.55 pmol/L for plasma E2,E1, and E1S,

www.aacrjournals.org Clin Cancer Res; 17(15) August 1, 2011 4951

Lønning et al.

respectively (52). Although these methods for plasma (52) an important comparator for the arguments made with and tissue (51) estrogen measurement are highly sensitive, malignant tissue, and (iii) it may be important in consider- precise, and devoid of cross-reactivity, they are labor-inten- ing prevention strategies for breast cancer. sive because they require multiple prepurification steps. Ithasbeenknownfor3decadesthatE2 levels in breast The need for such sensitive assays, particularly for assessing cancer tissue exceed those in plasma by a factor of about estrogen levels in patients undergoing treatment with aro- 10; however, the mechanism behind this phenomenon matase inhibitors, is illustrated by recent findings (53) remains obscure. In a previous study (30) we measured revealing that the majority of patients who receive treat- tissue estrogen levels using a 125I-based RIA (51) after ment with the third-generation aromatase inhibitors ana- prepurification through several steps, including high- strozole or letrozole achieve plasma estrogen levels below performance liquid chromatography. Although we these sensitivity limits. Recently, commercial assays such as detected lower tissue levels of the different estrogen frac- 125 the DSL-8700 E1 RIA applying I-labeled E1, and the tions compared with previous studies that applied RIAs Spectria estradiol-sensitive RIA and DSL-4800 ultrasensi- without prepurification (62–64), we detected a benign 125 tive RIA using I-labeled E2 have become available. Such tissue/plasma concentration ratio of E1,themainestro- kits in general do not involve any chromatographic gen product in postmenopausal women, of 6 to 7 (30). 14 prepurification procedures, highlighting the need for This is consistent with the benign tissue/plasma CE1 high specificity of the antibody applied. The theoretical ratio of 5 to 6 found by Larionov and colleagues (65) sensitivity limit for DSL-8700 is 4.44 pmol/L, and theore- when they applied presurgical steady-state infusions of 14 tical sensitivity limits for E2 kits are in the range of 5 to C-labeled E1 to breast cancer patients. 8.14 pmol/L. Although such assays may determine plasma The benign tissue/plasma gradient »1 (as will be seen for estrogen levels reliably in the majority of postmenopausal tumor tissue as well) has provoked much discussion about women, some individuals reveal plasma E2 concentrations what the potential mechanism might be. Several explana- even below 5 pmol/L. tions have been proposed, including active uptake from the Many investigators consider methods such as gas chro- circulation, local estrogen synthesis, and enhanced tissue matography/tandem mass spectroscopy (GC/MS-MS) and binding. From their studies infusing 3H-androstenedione 14 3 14 liquid chromatography/tandem mass spectrometry (LC/ and CE1 in concert, by determining the H/ C isotope MS-MS) as gold standards with respect to estrogen level ratio in the tissue compared with plasma E1 fractions, measurements. However, some MS methods lack sensitiv- Larionov and colleagues (65) estimated that local tissue 3 14 ity for determining plasma E2 levels in many postmeno- production (based on an elevated H/ C ratio) may pausal women (54), particularly during treatment with account for a median of 15% of local E1 concentrations. estrogen-suppressing compounds. Nonetheless, using In light of the potential clinical implications of these GC/MS-MS, Santen and colleagues (55) reported average findings, we have explored the robustness of a pharmaco- plasma levels of E2 in postmenopausal women similar to kinetic model (66, 67). Given the specialized nature of the those recorded by our RIA, and others have lower values calculations, we provide a detailed analysis in the Supple- than most of the sensitive RIAs (56). Although such meth- mentary Material and summarize the key findings here. ods, including assays for plasma E1S and tissue estrogen Overall, pharmacokinetic modeling is most consistent levels, are in rapid development (57–59), it is clear that the with a model suggesting rapid plasma/benign tissue choice and validation of assays based on MS are as impor- exchange of the estrogen compounds, in which case the tant as for those based on RIA. actual local contribution may not be determined (66). Method sensitivity problems are even greater when mea- Although an elevated tissue/plasma gradient may be due suring tissue estrogen concentrations. As mentioned above, either to enhanced tissue binding or, alternatively, a mechan- breast tissue contains enzymes involved in estrogen meta- ism of active uptake of circulating estrogens in the tissue, bolism. Although major catechol estrogens, such as 2- regrettably there is no pharmacokinetic model by which hydroxyestrone, are rapidly cleared from the circulation these 2 processes may be evaluated separately in vivo (67). (60), the fact that these metabolites may account for about This model is supported by recent results from our own 50% of total estrogen metabolites in the urine (61) indi- studies measuring plasma and benign breast tissue estro- cates a high production rate. In general, we lack knowledge gens in pre- and postmenopausal women (Tables 1 and 2). regarding tissue concentrations of estrogen metabolites, Average benign tissue levels of E1 and E2 in premenopausal but they are a potential source of interference in RIA that women exceed those in postmenopausal women by a factor must be eliminated during sample preparation for assay. of about 3 and 15, respectively. On the other hand, the tissue/plasma concentration gradient is remarkably similar Benign Breast Tissue Estrogen Concentrations in both groups (E1 averaging 5–6, E2 averaging 1.5–2). and Synthesis These findings are consistent with both active uptake and enhanced tissue binding. In premenopausal women, the Although our focus is on the source of estrogens for great majority of plasma E2 is ovarian in origin. Consider- breast cancers, a brief account of estrogen disposition in ing the difference in tissue E2 concentrations between pre- benign breast tissue is merited because (i) this is potentially and postmenopausal women and the similarity with a source of estrogens for the malignant tissue per se, (ii) it is respect to the tissue/plasma concentration gradient, these

4952 Clin Cancer Res; 17(15) August 1, 2011 Clinical Cancer Research

Breast Cancer Estrogen Levels

findings are consistent with a rapid equilibrium between plasma and benign tissue compartments. Another intriguing finding relates to the tissue/plasma 1216.2) 2566.8) – – 1371.8) 327.2)

– concentration ratios for the different estrogen fractions. A – 184.0) 217.1) – – comparison of E2,E1, and E1S reveals that E1 is the least (308.6 (303.8 (31.6 polar (i.e., the most fat-soluble) compound, and E S is the

(53.8 (44.8 1 c (c) (c) e c most polar (i.e., the most water-soluble). The tissue/ S (tumor) 1 plasma concentration ratios for E1,E2, and E1S are about 5–6, 1.5–2, and 0.1, respectively. Lipophilicity and small- molecule size both predict for rapid membrane transfer (67). Because of its sulfate conjugation, E1S behaves as a 1230.3) 883.1 730.1)967.4) 612.6 447.8 (146.2 – – – weak acid, a group of compounds known in general to 90.4)107.4)140.5) 99.5 98.6 101.6 – – – reveal a small volume of distribution as compared with basic drugs (67). In contrast, E1 and E2 both behave as

S (benign) E neutral compounds at physiologic pH. Taken together, 1

E these findings are consistent with (but do not definitely prove) the notion that tissue/plasma estrogen concentration ratios may be explained by the physiochemical prop- 280.5) 56.1 (29.3 344.5)514.5) 394.4 (213.0 331.6 (113.7

– – – erties of the respective molecule. 221.3) 53.8 (32.0 252.3) 48.8 (17.0 471.6) 482.8 (189.5 – – – These results have potential implications for the future

(93.3 (100.1 (39.9 (109.9 (117.0 (46.7 clinical development of estrogen-modulating com-

e e c e c c pounds. Because of its high plasma levels, the estrogen (tumor)

1 conjugate E1S has received much attention as a potential E source of intratumoral estrogens (68). Thus, the findings from previous studies of high levels of plasma (25) and tissue E1S (63) as compared with unconjugated E1 and E2 1986.6) 245.4 1859.5) 194.6 3167.4) 148.5 – – – suggest that circulating E S, taken up by the tissue with 620.9) 143.7 646.4) 167.6 876.8) 100.3 1 – – – subsequent deconjugation and reduction, is an important 0.05.

< source of intratumoral E2 (68). Although our own studies P

c (69, 70) have confirmed high plasma levels of E1S, our (benign)

1 findings of low tissue E1S and a low tissue/plasma con- E 500.3 (285.4

0.10, centration ratio argue against the hypothesis that this < conjugate may be a major contributor to tissue E2. Second, P

< although we cannot directly assess benign breast tissue

3047.9) 1144.2 (659.0 estrogen synthesis in postmenopausal women, rapid equi- – 446.2) 467.6 (338.2 1646.3) 1233.1 (817.5 0.05 – – libration between plasma and tissue compartments, which 232.0) 477.2 (366.7 (c) 1207.5) 1345.2 (571.5 – – is supported by the highly significant correlation between (863.2 61.5) – (159.9 (230.0

(63.6 plasma and tissue estrogen concentrations for E1 and E2 a,d e b,e (c) (Table 2), indicates that plasma estrogen levels predict (tumor) subgroup of patients:

2 benign breast tissue levels well in both pre- and postme- – E nopausal women. Of even more importance, whatever the magnitude of local estrogen production within benign breast tissue, rapid equilibration of breast tissue with the 653.4)1092.4) 1621.8 198.7 (32.7 652.8) 615.3 – – – plasma compartment indicates the role of local synthesis 49.7) 267.1 44.1) 121.5 – – 69.6) 19.3 (6.1 to be of minor importance. Whether there is a substantial – local synthesis (or not), this will have little impact on local

(benign) estrogen levels in benign tissues as long as local concen- 2

E trations are at equilibrium with plasma estrogen levels.

Breast Cancer Estrogen Concentrations and 30 29.2 (19.3 13 453.0 (314.3 219 31.1 (19.5 25.1 (9.1 76 398.7 (243.2 525.9 (253.2 Whitney test). Synthesis – ¼ ¼ ¼ ¼ ¼ ¼ n n n n n n In contrast to the finding that intratumoral concentra- Tissue estrogen levels (fmol/g wet weight) given as geometric mean levels with 95% confidence interval in benign and malignant breast tions of E2 in ER tumors resembled concentrations in benign tissue (30), we found substantially elevated levels 0.001 (Mann 0.01. þ þ 0.01. 0.001 (Wilcoxon signed ranks test). þ < < < < of E2 in ER tumors as compared with normal tissue from Total group ER Total group ER ER ER P P P P samples from the same breast in correlation with menopausal status and ER status Table 1. ER/PGR a b d e Data from Lønning et al. (30). Statistically significant difference compared with benign tissue: Postmenopausal women Statistically significant difference compared with ER Premenopausal women the same breast in postmenopausal women (Tables 1 and 2; Fig. 3). Surprisingly, however, we detected tumor levels

www.aacrjournals.org Clin Cancer Res; 17(15) August 1, 2011 4953

Lønning et al.

Table 2. Ratios of diverse plasma and tissue estrogen fractions in subgroups of patients (geometric mean with 95% confidence interval)

Postmenopausal women Premenopausal women

ER status ERþ ER ERþ ER n ¼ 21 n ¼ 9 n ¼ 7 n ¼ 6

NT-E2/PL-E2 2.0 (1.3–3.2) 1.3 (0.5–3.2) 1.6 (0.7–3.5) 2.2 (1.7–3.0)

NTE1/PL-E1 5.9 (5.0–7.0) 6.3 (4.4–9.0) 4.4 (2.8–6.9) 7.4 (4.7–11.5)

NTE1S/PL-E1S 0.1 (0.0–0.2) 0.1 (0.0–0.4) 0.1 (0.0–0.3) 0.1 (0.0–0.3) c a TU-E2/PL-E2 16.4 (10.8–24.9) 1.2 (0.3–5.3) 6.6 (2.2–20.5) 0.5 (0.0–9.7)

TU-E1/PL-E1 2.1 (1.5–3.0) 1.4 (0.5–4.2) 1.0 (0.3–2.7) 0.7 (0.2–2.6)

TU-E1S/PL-E1S 0.2 (0.1–0.4) 0.2 (0.0–0.8) 0.1 (0.0–0.3) 0.2 (0.1–0.6) b a TU-E2/TU-E1 1.5 (0.9–2.6) 0.2 (0.0–1.1) 7.1 (4.5–11.3) 0.8 (0.1–8.5)

Abbreviations: E1, estrone; E2, estradiol; E1S, estrone sulfate; ER, estrogen receptor; NT, normal (benign) tissue of the same breast; PL, plasma; TU, tumor tissue. Statistically significant difference compared with ER subgroup of patients: aP < 0.05. bP < 0.01. cP < 0.001 (Mann–Whitney test). Data from Lønning et al. (30).

of E1 to be consistently lower as compared with tissue in tumor cells (17). Based on these conflicting results, an concentrations of E1 in benign tissue from the same breasts international consortium was formed with the goal of independently of tumor ER status. Consistent with this developing and validating an antibody that could provide observation, in their tracer infusion studies, Larionov and uniform results. This has now been achieved with the 14 colleagues (65) found a lower concentration of C-E1 in ARO677 antibody, which has been validated by different tumor than in benign tissue. Of interest, they recorded an laboratories (73). 3 14 elevated E1 H/ C ratio in breast cancer compared with No significant correlation of ARO677 immunostaining benign breast tissue and plasma, indicating intratumor to either intratumoral E1 or E2 levels has been recorded. þ aromatization of 3H-labeled androstenedione. Although When ER tumors were analyzed separately, statistically individual variation was substantial, the results revealed higher intratumoral E2 but not E1 levels were recorded þ a median contribution from local aromatization account- in the group of ARO677 tumors as compared with ing for about 30% of intratumor E1 levels, compared with a ARO677 tumors (74). In the same material (75), we median value of 15% in benign tissue. On the other hand, detected low levels of aromatase mRNA in breast cancer local synthesis was found to account for more than 50% of and normal breast tissue. No correlations between benign intratumor E1 in only 3 of 24 tumors. The variation and malignant tissue aromatase mRNA expression and between tumors, however, was substantial. This may be either E2 (Spearman rs ¼0.02 and 0.05, respectively) or due not only to variable local synthesis but also to variation E1 (Spearman rs ¼0.11 for both) levels were recorded. in diffusion rates between compartments (see Supplemen- Although enzymatic aromatase activity may be detected tary Material). Similar results were obtained in a smaller in most fresh breast cancer samples (19), these findings, study by Reed and colleagues (71). Although this difference in similarity to the data presented above, do not address between breast cancer and normal breast tissue could the quantitative role of tumor aromatization in vivo. indicate increased intratumor estrogen synthesis, alterna- A striking observation is the elevated tumor/benign- þ tively it may be due to a slower uptake of E1 from the tissue E2 ratio in ER breast cancer tissue. In contrast, circulation into the tumor compared with a normal breast tumor tissue levels of E1 are reduced as compared with tissue compartment (see details outlined in Supplementary benign tissue concentrations (Fig. 3). The balance between Material). E2 and E1 is regulated through multiple dehydrogenase The potential impact of intratumor steroid synthesis or isoforms. Some of these isoforms (e.g., HSD17B1, B5, B7, modulation may have significant implications. Immuno- and B12) act reductively in catalyzing activation of E1 into staining for aromatase protein expression remained a con- E2 (76–79), whereas others (e.g., HSD17B2, B10, and B14) troversial subject for 2 decades. For several years, favor oxidation, converting E2 into E1 in different tissues, conflicting results were obtained with 2 antibodies: One including benign breast tissue (80–82). Of interest, both B2 (72) revealed major protein staining related to stromal and B7 were found to be elevated in breast cancer as cells, and the other antibody detected aromatase staining compared with normal breast tissue (75), and intratumor

4954 Clin Cancer Res; 17(15) August 1, 2011 Clinical Cancer Research

Breast Cancer Estrogen Levels

wet weight of tissue or 100 pmol/L, the ER concentration of tumors considered positive by this cutoff would far A *** þ exceed the concentration of E2 in ER tumors. The estradiol 14 binding affinity for the ER is less than 1 nmol/L (13); thus, at the concentrations measured, on average we may envi- 12 sion the bulk of intratumor E2 to be receptor bound. Taken together, the findings with respect to intratumor 10 E1 and E2 versus plasma levels differ significantly from what we observed with respect to benign tissue hormonal con- 8 centrations. Although plasma E1 levels in postmenopausal 6 women correlated with intratumor E1 and E2 levels, there were no significant correlations in pre- or postmenopausal ** tissue concentration

Ratio malignant/benign 4 women for plasma E2 and intratumor estrogen levels, or for plasma E1 in premenopausal women. This suggests that the 2 **** concentration equilibrium seen for benign tissue may not apply to intratumor levels. Although we cannot estimate 0 the quantitative contribution from intratumoral aromati- E2 E1 E1S zation, it is notable that the lack of correlation between + − ER n = 21 ER n = 9 aromatase expression at the mRNA and protein levels (immunostaining) contrasts with the correlations between B 7 * intratumor E2 and dehydrogenase enzymes as well as ER level expression. On the contrary, the observation that 6 tumor E2 concentration correlated positively to HSD17B7 but negatively to HSD17B2 expression strongly supports a 5 role for intratumor conversion of E1 into E2 as a source of intratumor E . ** 2 4 No correlations between intratumor levels of either E1, E2,orE1S and plasma E1S levels or intratumor estrone 3 sulfatase levels were observed (30, 75). Furthermore, the average expression level of estrone sulfatase in tumor tissue

tissue concentration 2

Ratio malignant/benign was about one third the level recorded in benign breast *** * tissue. Taken together, these findings do not suggest a 1 major contribution from circulating E1S to intratumor E2 levels. 0 E E E S 2 1 1 Plasma Estrogen Concentrations and + − ER n = 7 ER n = 6 Intratumor Expression of Estrogen-Related Genes

The development of multigene expression assays (84) Figure 3. Ratio between intratumor and benign breast tissue has allowed investigators to study tumor gene expression estrogen levels related to ER expression and menopausal status. A, postmenopausal women. B, premenopausal women. *, P < 0.05, profiling in relation to estrogen suppression with aroma- **, P < 0.01, and ***, P < 0.001. Reproduced from Lønning et al. (30) with tase inhibitors (85), as well as to explore potential correla- permission. tions between circulating estrogen levels and expression of estrogen-regulated genes within tumor tissue. Although levels of E2 were found to be negatively correlated with germline polymorphisms in estrogen-synthesizing and HSD17B2 expression but positively correlated with B7 metabolizing genes (33) should affect all body compart- expression. These findings support an additional mechan- ments, as stated above, this may also have tissue- or ism of intratumoral E2 synthesis, i.e., reduction of E1 into plasma-specific effects due to organ-specific expression. E2, which would explain elevated levels of tumor E2 as Although the CYP19 SNPs rs10046 and TCT were both compared with benign tissue E2 levels. found to be significantly associated with plasma E2 levels, Finally, we observed a strong correlation between they explained only a small percentage of interindividual E2 intratumoral E2 concentration and ER expression levels, variation (86). suggesting an affinity effect of the ER (75). Before immu- In a recent study, some members of our group (87) nostaining became standard practice for ER assessment, the assessed intratumor expression profiles of genes under general cut-off limit for ER positivity in ligand-binding estrogen regulation in postmenopausal women. Genes assays was 10 fmol/mg protein (83). Assuming that 1 upregulated or downregulated by estrogen stimulation of þ fmol/mg protein may correspond to about 100 fmol/g ER breast cancer cells were previously indentified from

www.aacrjournals.org Clin Cancer Res; 17(15) August 1, 2011 4955

Lønning et al.

in vitro experiments (88). When 104 breast cancers were gradient is similar in pre- and postmenopausal women studied (87), a strong positive correlation between plasma supports the hypothesis that this gradient may be due E2 levels and intratumor ER expression on the one hand either to active tissue uptake or simply to enhanced tissue and intratumor expression levels of several classical E2- steroid binding or depot (fat) storage. High intratumor E2 regulated genes on the other hand were found. When genes levels may be due mainly to a concerted mechanism such as trefoil factor 1 (TFF1), GREB1, PDZK1, and PgR related to uptake of circulatory (and potentially benign were combined into an index, a correlation coefficient of tissue) E1 with intratumor reduction of E1 into E2 and, to (rs) of 0.51 indicated that plasma E2 levels on average a greater degree, an affinity effect of the ER on E2.Onthe predicted intratumor E2 levels by 27%. In this study, contrary, indirect evidence suggests that intratumor aro- neither breast cancer nor benign tissue estrogen concentra- matization plays a minor role, and there is no evidence tions were available. The data, however, provide strong advocating a role of E1S as a potential source of intratu- supporting evidence for the hypothesis that plasma/ moral E2. benign-tissue estrogens exert a strong biological influence Our data suggest that local enzyme inhibition of on breast cancer tissue. the aromatase enzyme or sulfatase is less important than inhibition of total-body estrogen synthesis. Although Integrative Model inhibition of local tumor dehydrogenases could potentially have a role in breast cancer therapy, the evidence The topic of the contribution of intratumor synthesis argues against targeting local aromatization or sulfate versus circulating estrogens to intratumor E2 levels and deconjugation as a therapeutic strategy in breast cancer. ER stimulation in breast cancer has been controversial for decades. We believe that the recent evidence as reviewed Disclosure of Potential Conflicts of Interest here adds a significant contribution to our understanding of the topic. Although we are not formally able to assess P. E. Lønning received honoraria from Novartis, AstraZeneca, and Pfizer (and probably will not be able to in the near future) the Inc. for advice and lecture fees. M. Dowsett received honoraria from quantitative contribution from each individual organ to AstraZeneca for advice and lecture fees. He also received grants from AstraZeneca and Novartis. plasma estrogen levels in postmenopausal women, we believe that a relatively straightforward model can explain our recent observations and clarify previous con- Grant Support tradictory evidence with respect to plasma as well as Norwegian Cancer Society; Western Health Region of Norway; Innovest benign and malignant breast tissue estrogen concentra- Program of Excellence; Breast Cancer Research Foundation; Mary-Jean tions. As for benign tissue estrogen concentrations, due to Mitchell Green Foundation; Breakthrough Breast Cancer; National Institute rapid equilibration between the tissue and plasma com- for Health Research (to the Royal Marsden National Institute for Health Research Biomedical Research). partments, plasma estrogens become good surrogate markers for benign tissue levels. Although a tissue/ Received January 6, 2011; revised April 5, 2011; accepted April 15, 2011; plasma gradient for E2 and E1 exists, the fact that this published OnlineFirst July 26, 2011.

References 1. Trichopoulos D, MacMahon B, Cole P. Menopause and breast cancer survival: an overview of the randomised trials. Lancet 2005;365: risk. J Natl Cancer Inst 1972;48:605–13. 1687–717. 2. Feinleib M. Breast cancer and artificial menopause: a cohort study. J 10. Lønning PE. Aromatase inhibitors in breast cancer. Endocr Relat Natl Cancer Inst 1968;41:315–29. Cancer 2004;11:179–89. 3. Key T, Appleby P, Barnes I, Reeves G Endogenous Hormones and 11. Ruder HJ, Loriaux L, Lipsett MB. Estrone sulfate: production rate and Breast Cancer Collaborative Group. Endogenous sex hormones and metabolism in man. J Clin Invest 1972;51:1020–33. breast cancer in postmenopausal women: reanalysis of nine prospec- 12. Longcope C. The metabolism of estrone sulfate in normal males. J tive studies. J Natl Cancer Inst 2002;94:606–16. Clin Endocrinol Metab 1972;34:113–22. 4. Russo J, Russo IH. Genotoxicity of steroidal estrogens. Trends 13. Edwards DP, McGuire WL. 17 alpha-Estradiol is a biologically active Endocrinol Metab 2004;15:211–4. estrogen in human breast cancer cells in tissue culture. Endocrinology 5. Clemons M, Goss P. Estrogen and the risk of breast cancer. N Engl J 1980;107:884–91. Med 2001;344:276–85. 14. Suzuki M, Ishida H, Shiotsu Y, Nakata T, Akinaga S, Takashima S, 6. Rebbeck TR, Kauff ND, Domchek SM. Meta-analysis of risk reduction et al. Expression level of enzymes related to in situ estrogen synthesis estimates associated with risk-reducing salpingo-oophorectomy and clinicopathological parameters in breast cancer patients. J Ster- in BRCA1 or BRCA2 mutation carriers. J Natl Cancer Inst 2009; oid Biochem Mol Biol 2009;113:195–201. 101:80–7. 15. Santner SJ, Feil PD, Santen RJ. In situ estrogen production via the 7. Carey LA. Targeted chemotherapy? Platinum in BRCA1-dysfunctional estrone sulfatase pathway in breast tumors: relative importance versus breast cancer. J Clin Oncol 2010;28:361–3. the aromatase pathway. J Clin Endocrinol Metab 1984;59:29–33. 8. Sørlie T, Perou CM, Tibshirani R, Aas T, Geisler S, Johnsen H, et al. 16. Price T, Aitken J, Head J, Mahendroo M, Means G, Simpson E. Gene expression patterns of breast carcinomas distinguish tumor Determination of aromatase cytochrome P450 messenger ribonucleic subclasses with clinical implications. Proc Natl Acad Sci U S A acid in human breast tissue by competitive polymerase chain reaction 2001;98:10869–74. amplification. J Clin Endocrinol Metab 1992;74:1247–52. 9. Abe O, Abe R, Enomoto K, et al. Effects of chemotherapy and 17. Esteban JM, Warsi Z, Haniu M, Hall P, Shively JE, Chen S. Detection of hormonal therapy for early breast cancer on recurrence and 15-year intratumoral aromatase in breast carcinomas. An immunohistochemical

4956 Clin Cancer Res; 17(15) August 1, 2011 Clinical Cancer Research

Breast Cancer Estrogen Levels

study with clinicopathologic correlation. Am J Pathol 1992; 140: 38. Huang ZQ, Fasco MJ, Figge HL, Keyomarsi K, Kaminsky LS. Expres- 337–43. sion of cytochromes P450 in human breast tissue and tumors. Drug 18. Miller WR. Aromatase activity in breast tissue. J Steroid Biochem Mol Metab Dispos 1996;24:899–905. Biol 1991;39[5B]:783–90. 39. Cribb AE, Knight MJ, Dryer D, Guernsey J, Hender K, Tesch M, et al. 19. Miller WR, Anderson TJ, Jack WJL. Relationship between tumour Role of polymorphic human cytochrome P450 enzymes in estrone aromatase activity, tumour characteristics and response to therapy. J oxidation. Cancer Epidemiol Biomarkers Prev 2006;15:551–8. Steroid Biochem Mol Biol 1990;37:1055–9. 40. Adlercreutz H, Hockerstedt€ K, Bannwart C, Bloigu S, Ham€ al€ ainen€ E, 20. Bulun SE, Sebastian S, Takayama K, Suzuki T, Sasano H, Shozu M. Fotsis T, et al. Effect of dietary components, including lignans and The human CYP19 (aromatase P450) gene: update on physiologic phytoestrogens, on enterohepatic circulation and liver metabolism of roles and genomic organization of promoters. J Steroid Biochem Mol estrogens and on sex hormone binding globulin (SHBG). J Steroid Biol 2003;86:219–24. Biochem 1987;27:1135–44. 21. Lønning PE, Dowsett M, Powles TJ. Postmenopausal estrogen synth- 41. Gurpide E. Metabolic influences on the action of estrogens: thera- esis and metabolism: alterations caused by aromatase inhibitors used peutic implications. Pediatrics 1978;62:1114–20. for the treatment of breast cancer. J Steroid Biochem 1990;35: 42. Bulun SE, Simpson ER. Regulation of aromatase expression in human 355–66. tissues. Breast Cancer Res Treat 1994;30:19–29. 22. Dowsett M, Cantwell B, Lal A, Jeffcoate SL, Harris AL. Suppression of 43. Jacobs S, MacNeill F, Lønning P, Dowsett M, Jones A, Powles T. postmenopausal ovarian steroidogenesis with the luteinizing hor- Aromatase activity, serum oestradiol and their correlation with demo- mone-releasing hormone agonist goserelin. J Clin Endocrinol Metab graphic indices. J Steroid Biochem Mol Biol 1992;41:769–71. 1988;66:672–7. 44. Shozu M, Sebastian S, Takayama K, Hsu WT, Schultz RA, Neely K, 23. Couzinet B, Meduri G, Lecce MG, Young J, Brailly S, Loosfelt H, et al. et al. Estrogen excess associated with novel gain-of-function muta- The postmenopausal ovary is not a major androgen-producing gland. tions affecting the aromatase gene. N Engl J Med 2003;348:1855–65. J Clin Endocrinol Metab 2001;86:5060–6. 45. Smith EP, Boyd J, Frank GR, Takahashi H, Cohen RM, Specker B, 24. SluijmerAV,HeinemanMJ,DeJongFH,EversJL.Endocrine et al. Estrogen resistance caused by a mutation in the estrogen- activity of the postmenopausal ovary: the effects of pituitary receptor gene in a man. N Engl J Med 1994;331:1056–61. down-regulation and oophorectomy. J Clin Endocrinol Metab 46. Ma CX, Adjei AA, Salavaggione OE, Coronel J, Pelleymounter L, Wang 1995;80:2163–7. L, et al. Human aromatase: gene resequencing and functional geno- 25. Lønning PE, Helle SI, Johannessen DC, Adlercreutz H, Lien EA, Tally mics. Cancer Res 2005;65:11071–82. M, et al. Relations between sex hormones, sex hormone binding 47. Colomer R, Monzo M, Tusquets I, Rifa J, Baena JM, Barnadas A, et al. globulin, insulin-like growth factor-I and insulin-like growth factor A single-nucleotide polymorphism in the aromatase gene is asso- binding protein-1 in post-menopausal breast cancer patients. Clin ciated with the efficacy of the aromatase inhibitor letrozole in Endocrinol (Oxf) 1995;42:23–30. advanced breast carcinoma. Clin Cancer Res 2008;14:811–6. 26. Geisler J, Lønning PE. Aromatase inhibition: translation into a suc- 48. Wang LW, Ellsworth KA, Moon I, Pelleymounter LL, Eckloff BW, Martin cessful therapeutic approach. Clin Cancer Res 2005;11:2809–21. YN, et al. Functional genetic polymorphisms in the aromatase gene 27. Lønning PE, Geisler J, Johannessen DC, Ekse D. Plasma estrogen CYP19 vary the response of breast cancer patients to neoadjuvant suppression with aromatase inhibitors evaluated by a novel, sensitive therapy with aromatase inhibitors. Cancer Res 2010;70:319–28. assay for estrone sulphate. J Steroid Biochem Mol Biol 1997;61: 49. Lønning PE, Johannessen DC, Thorsen T. Alterations in the produc- 255–60. tion rate and the metabolism of oestrone and oestrone sulphate in 28. Dowsett M, Goss PE, Powles TJ, Hutchinson G, Brodie AM, Jeffcoate breast cancer patients treated with aminoglutethimide. Br J Cancer SL, et al. Use of the aromatase inhibitor 4-hydroxyandrostenedione in 1989;60:107–11. postmenopausal breast cancer: optimization of therapeutic dose and 50. Lønning PE, Ekse D. A sensitive assay for measurement of plasma route. Cancer Res 1987;47:1957–61. estrone sulphate in patients on treatment with aromatase inhibitors. J 29. Hankinson SE, Manson JE, Spiegelman D, Willett WC, Longcope C, Steroid Biochem Mol Biol 1995;55:409–12. Speizer FE. Reproducibility of plasma hormone levels in postmeno- 51. Geisler J, Berntsen H, Lønning PE. A novel HPLC-RIA method for the pausal women over a 2-3-year period. Cancer Epidemiol Biomarkers simultaneous detection of estrone, estradiol and estrone sulphate Prev 1995;4:649–54. levels in breast cancer tissue. J Steroid Biochem Mol Biol 2000; 30. Lønning PE, Helle H, Duong NK, Ekse D, Aas T, Geisler J. Tissue 72:259–64. estradiol is selectively elevated in receptor positive breast cancers 52. Geisler J, Ekse D, Helle H, Duong NK, Lønning PE. An optimised, while tumour estrone is reduced independent of receptor status. J highly sensitive radioimmunoassay for the simultaneous measure- Steroid Biochem Mol Biol 2009;117:31–41. ment of estrone, estradiol and estrone sulfate in the ultra-low range 31. Lønning PE, Kvinnsland S, Thorsen T, Ueland PM. Alterations in the in human plasma samples. J Steroid Biochem Mol Biol 2008;109: metabolism of oestrogens during treatment with aminoglutethimide in 90–5. breast cancer patients. Preliminary findings. Clin Pharmacokinet 53. Geisler J, Helle H, Ekse D, Duong NK, Evans DB, Nordbø Y, et al. 1987;13:393–406. Letrozole is superior to anastrozole in suppressing breast cancer 32. Lee AJ, Cai MXX, Thomas PE, Conney AH, Zhu BT. Characterization tissue and plasma estrogen levels. Clin Cancer Res 2008;14:6330–5. of the oxidative metabolites of 17beta-estradiol and estrone formed 54. Xu X, Roman JM, Issaq HJ, Keefer LK, Veenstra TD, Ziegler RG. by 15 selectively expressed human cytochrome p450 isoforms. Quantitative measurement of endogenous estrogens and estrogen Endocrinology 2003;144:3382–98. metabolites in human serum by liquid chromatography-tandem mass 33. Zhou SF, Liu JP, Chowbay B. Polymorphism of human cytochrome spectrometry. Anal Chem 2007;79:7813–21. P450 enzymes and its clinical impact. Drug Metab Rev 2009;41: 55. Santen RJ, Demers L, Ohorodnik S, Settlage J, Langecker P, Blan- 89–295. chett D, et al. Superiority of gas chromatography/tandem mass 34. Bolt HM. Metabolism of estrogens—natural and synthetic. Pharmacol spectrometry assay (GC/MS/MS) for estradiol for monitoring of aro- Ther 1979;4:155–81. matase inhibitor therapy. Steroids 2007;72:666–71. 35. Martucci CP, Fishman J. P450 enzymes of estrogen metabolism. 56. Lee JS, Ettinger B, Stanczyk FZ, Vittinghoff E, Hanes V, Cauley JA, Pharmacol Ther 1993;57:237–57. et al. Comparison of methods to measure low serum estradiol 36. Lee AJ, Conney AH, Zhu BT. Human cytochrome P450 3A7 has a levels in postmenopausal women. J Clin Endocrinol Metab 2006;91: distinct high catalytic activity for the 16alpha-hydroxylation of estrone 3791–7. but not 17beta-estradiol. Cancer Res 2003;63:6532–6. 57. Corona G, Elia C, Casetta B, Da Ponte A, Del Pup L, Ottavian E, et al. 37. Hershcopf RJ, Bradlow HL, Fishman J. Differential hydroxylations of Liquid chromatography tandem mass spectrometry assay for fast and estrone and estradiol in man. J Clin Endocrinol Metab 1986;62: sensitive quantification of estrone-sulfate. Clin Chim Acta 2010; 170–3. 411:574–80.

www.aacrjournals.org Clin Cancer Res; 17(15) August 1, 2011 4957

Lønning et al.

58. Blonder J, Johann DJ, Veenstra TD, Xiao Z, Emmert-Buck MR, Ziegler 74. Geisler J, Suzuki T, Helle H, Miki Y, Nagasaki S, Duong NK, et al. RG, et al. Quantitation of steroid hormones in thin fresh frozen tissue Breast cancer aromatase expression evaluated by the novel antibody sections. Anal Chem 2008;80:8845–52. 677: correlations to intra-tumor estrogen levels and hormone receptor 59. Arai S, Miyashiro Y, Shibata Y, Kashiwagi B, Tomaru Y, Kobayashi M, status. J Steroid Biochem Mol Biol 2010;118:237–41. et al. New quantification method for estradiol in the prostatic tissues of 75. Haynes BP, Straume AH, Geisler J, A’Hern R, Helle H, Smith IE, et al. benign prostatic hyperplasia using liquid chromatography-tandem Intratumoral estrogen disposition in breast cancer. Clin Cancer Res mass spectrometry. Steroids 2010;75:13–9. 2010;16:1790–801. 60. Longcope C, Femino A, Flood C, Williams KIH. Metabolic clearance 76. Poutanen M, Miettinen M, Vihko R. Differential estrogen substrate rate and conversion ratios of [3H]2-hydroxyestrone in normal men. J specificities for transiently expressed human placental 17 beta-hydro- Clin Endocrinol Metab 1982;54:374–80. xysteroid dehydrogenase and an endogenous enzyme expressed in 61. Lønning PE, Skulstad P. Alterations in the urine excretion of estrogen cultured COS-m6 cells. Endocrinology 1993;133:2639–44. metabolites in breast cancer women treated with aminoglutethimide. 77. Penning TM, Burczynski ME, Jez JM, Hung CF, Lin HK, Ma H, et al. J Steroid Biochem 1989;33[4A]:565–71. Human 3alpha-hydroxysteroid dehydrogenase isoforms (AKR1C1- 62. Vermeulen A, Deslypere JP, Paridaens R, Leclercq G, Roy F, Heuson AKR1C4) of the aldo-keto reductase superfamily: functional plasticity JC. Aromatase, 17 b-hydroxysteroid dehydrogenase and intratissular and tissue distribution reveals roles in the inactivation and formation of sex hormone concentrations in cancerous and normal glandular male and female sex hormones. Biochem J 2000;351:67–77. breast tissue in postmenopausal women. Eur J Cancer Clin Oncol 78. Krazeisen A, Breitling R, Imai K, Fritz S, Moller€ G, Adamski J. Deter- 1986;22:515–25. mination of cDNA, gene structure and chromosomal localization of the 63. Chetrite GS, Cortes-Prieto J, Philippe JC, Wright F, Pasqualini JR. novel human 17beta-hydroxysteroid dehydrogenase type 7(1). FEBS Comparison of estrogen concentrations, estrone sulfatase and aro- Lett 1999;460:373–9. matase activities in normal, and in cancerous, human breast tissues. J 79. Luu-The V, Tremblay P, Labrie F. Characterization of type 12 17beta- Steroid Biochem Mol Biol 2000;72:23–7. hydroxysteroid dehydrogenase, an isoform of type 3 17beta-hydro- 64. van Landeghem AA, Poortman J, Nabuurs M, Thijssen JHH. Endo- xysteroid dehydrogenase responsible for estradiol formation in genous concentration and subcellular distribution of estrogens in women. Mol Endocrinol 2006;20:437–43. normal and malignant human breast tissue. Cancer Res 1985; 80. Wu L, Einstein M, Geissler WM, Chan HK, Elliston KO, Andersson S. 45:2900–6. Expression cloning and characterization of human 17 beta-hydro- 65. Larionov AA, Berstein LM, Miller WR. Local uptake and synthesis of xysteroid dehydrogenase type 2, a microsomal enzyme possessing oestrone in normal and malignant postmenopausal breast tissues. J 20 alpha-hydroxysteroid dehydrogenase activity. J Biol Chem Steroid Biochem Mol Biol 2002;81:57–64. 1993;268:12964–9. 66. Gurpide E, MacDonald PC, Wiele RLV, Lieberman S. Measurement of 81. He XY, Wegiel J, Yang YZ, Pullarkat R, Schulz H, Yang SY. Type 10 the rates of secretion and of peripheral metabolism of two intercon- 17beta-hydroxysteroid dehydrogenase catalyzing the oxidation of vertible compounds: dehydroisoandrosterone-dehydroisoandroster- steroid modulators of gamma-aminobutyric acid type A receptors. one sulfate. J Clin Endocrinol Metab 1963;23:346–54. Mol Cell Endocrinol 2005;229:111–7. 67. Rowland M, Tozer TN. Clinical pharmacokinetics and pharmacody- 82. Lukacik P, Keller B, Bunkoczi G, Kavanagh KL, Lee WH, Adamski J, namics. 4th ed. Baltimore: Lippincott Williams & Wilkins; 2011. et al. Structural and biochemical characterization of human orphan 68. Santner SJ, Leszczynski D, Wright C, Manni A, Feil PD, Santen RJ. DHRS10 reveals a novel cytosolic enzyme with steroid dehydrogen- Estrone sulfate: a potential source of estradiol in human breast cancer ase activity. Biochem J 2007;402:419–27. Erratum in Biochem J tissues. Breast Cancer Res Treat 1986;7:35–44. 2007;403:615. 69. Geisler J, King N, Dowsett M, Ottestad L, Lundgren S, Walton P, et al. 83. Everson RB, Lippman ME, Thompson EB, McGuire WL, Wittliff JL, De Influence of anastrozole (Arimidex), a selective, non-steroidal aroma- Sombre ER, et al. Clinical correlations of steroid receptors and male tase inhibitor, on in vivo aromatisation and plasma oestrogen levels in breast cancer. Cancer Res 1980;40:991–7. postmenopausal women with breast cancer. Br J Cancer 1996; 84. Perou CM, Sørlie T, Eisen MB, van de Rijn M, Jeffrey SS, Rees CA, 74:1286–91. et al. Molecular portraits of human breast tumours. Nature 2000; 70. Geisler J, Haynes B, Anker G, Dowsett M, Lønning PE. Influence of 406:747–52. letrozole (Femara) and anastrozole on total body aromatization and 85. Miller WR, Larionov A, Renshaw L, Anderson TJ, Walker JR, Krause A, plasma estrogen levels in postmenopausal breast cancer patients et al. Gene expression profiles differentiating between breast cancers evaluated in a randomized, cross-over-designed study. J Clin Oncol clinically responsive or resistant to letrozole. J Clin Oncol 2009; 2002;20:751–7. 27:1382–7. 71. Reed MJ, Owen AM, Lai LC, Coldham NG, Ghilchik MW, Shaikh NA, 86. Dunning AM, Dowsett M, Healey CS, Tee L, Luben RN, Folkerd E, et al. et al. In situ oestrone synthesis in normal breast and breast tumour Polymorphisms associated with circulating sex hormone levels in tissues: effect of treatment with 4-hydroxyandrostenedione. Int J postmenopausal women. J Natl Cancer Inst 2004;96:936–45. Cancer 1989;44:233–7. 87. Dunbier AK, Anderson H, Ghazoui Z, Folkerd EJ, A'hern R, Crowder 72. Sasano H, Harada N. Intratumoral aromatase in human breast, RJ, et al. Relationship between plasma estradiol levels and estro- endometrial, and ovarian malignancies. Endocr Rev 1998;19:593– gen-responsive gene expression in estrogen receptor-positive 607. breast cancer in postmenopausal women. J Clin Oncol 2010;28: 73. Sasano H, Anderson TJ, Silverberg SG, Santen RJ, Conway M, 1161–7. Edwards DP, et al. The validation of new aromatase monoclonal 88. Frasor J, Stossi F, Danes JM, Komm B, Lyttle CR, Katzenellenbogen antibodies for immunohistochemistry—a correlation with biochemical BS. Selective estrogen receptor modulators: discrimination of ago- activities in 46 cases of breast cancer. J Steroid Biochem Mol Biol nistic versus antagonistic activities by gene expression profiling in 2005;95:35–9. breast cancer cells. Cancer Res 2004;64:1522–33.

4958 Clin Cancer Res; 17(15) August 1, 2011 Clinical Cancer Research

Exploring Breast Cancer Estrogen Disposition: The Basis for Endocrine Manipulation

Per E. Lønning, Ben P. Haynes, Anne H. Straume, et al.

Clin Cancer Res 2011;17:4948-4958. Published OnlineFirst July 26, 2011.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-11-0043

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2011/08/08/1078-0432.CCR-11-0043.DC1

Cited articles This article cites 87 articles, 22 of which you can access for free at: http://clincancerres.aacrjournals.org/content/17/15/4948.full#ref-list-1

Citing articles This article has been cited by 6 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/17/15/4948.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/17/15/4948. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.