`
`Oncogene (2003) 22, 7316-7339
`© 2003 Nature Publishing Group All rights reserved 0950-9232/03 $25.00
`www.nature.c0m/one
`
`Antiestrogen resistance in breast cancer and the role of estrogen receptor
`signaling
`
`Robert Clarke*", Minetta C Liu‘, Kerrie B Bouker‘, Zhiping Guz, Richard Y Lee‘, Yuelin Zhu‘,
`Todd C Skaar3, Bianca Gomez‘, Kerry O’Brien‘, Yue Wang4 and Leena A Hilakivi-Clarkel
`
`’Department of Oncology and Vincent T. Lombardi Cancer Center, Georgetown University School of Medicine, 3970 Reservoir Rd
`NW, Washington, DC 20057, USA,’ 2Celera Genomics, 45 West Gude Drive, Rockville, MD 20850, USA,’ 3Indiana University
`Department of Medicine, Division of Clinical Pharmacology, Indianapolis, IN 46202, USA,’ ‘Department of Electrical Engineering
`and Computer Science, The Catholic University of America, Washington, DC 20064, USA
`
`Antiestrogens include agents such as tamoxifen, toremi-
`fene, raloxifene, and fulvestrant. Currently, tamoxifen is
`the only drug approved for use
`in breast cancer
`chemoprevention, and it remains the treatment of choice
`for most women with hormone receptor positive, invasive
`breast carcinoma. While antiestrogens have been available
`since the early 1970s, we still do not fully understand their
`mechanisms of action and resistance. Essentially,
`two
`forms of antiestrogen resistance occur: de novo resistance
`and acquired resistance. Absence of estrogen receptor
`(ER) expression is the most common de novo resistance
`mechanism, whereas a complete loss of ER expression is
`not common in acquired resistance. Antiestrogen unre-
`sponsiveness appears to be the major acquired resistance
`phenotype, with a switch to an antiestrogen-stimulated
`growth being a minor phenotype. Since antiestrogens
`compete with estrogens
`for binding to ER, clinical
`response to antiestrogens may be affected by exogenous
`estrogenic exposures. Such exposures include estrogenic
`hormone replacement therapies and dietary and environ-
`mental exposures that directly or indirectly increase a
`tumor’s estrogenic environment. Whether antiestrogen
`resistance can be conferred by a switch from predomi-
`nantly ERar to ER]? expression remains unanswered, but
`predicting response to antiestrogen therapy requires only
`measurement of ERar expression. The role of altered
`receptor coactivator or corepressor expression in anti-
`estrogen resistance also is unclear, and understanding
`their
`roles may be confounded by their ubiquitous
`expression and functional redundancy. We have proposed
`a gene network approach to exploring the mechanistic
`aspects of antiestrogen resistance. Using transcriptome
`and proteome analyses, we have begun to identify
`candidate genes that comprise one component of a larger,
`putative gene network. These candidate genes include
`NFKB, interferon regulatory factor-1, nucleophosmin, and
`the X-box binding protein-1. The network also may
`involve signaling through ras and MAPK,
`implicating
`crosstalk with growth factors and cytokines. Ultimately,
`
`*Correspondence: R Clarke; E—mail: clarker@georgetown.edu
`
`signaling affects the expression/function of the prolifera-
`tion and/or apoptotic machineries.
`Oncogene (2003) 22, 7316-7339. doi:10.1038/sj.onc.1206937
`
`tamoxifen; Faslodex; ICI 182,780; estrogen
`Keywords:
`receptor; coregulator
`
`Introduction
`
`Antiestrogens primarily act by competing with estrogens
`for binding to the estrogen receptor (ER) and are the
`most widely administered endocrine agents for
`the
`management of ER—expressing breast cancers. The first
`antiestrogens were generated in the mid— 1950s as fertility
`agents and included ethamoxytriphetol (MER—25) and
`clomiphene. The ability of these compounds to induce
`responses in some breast cancer patients soon became
`apparent (Kistner and Smith, 1960), but the compounds
`induced significant toxicity (Herbst et al., 1964). In the
`early 1970s, the first study in breast cancer patients was
`published with a new antiestrogen tamoxifen (TAM, ICI
`46474) (Cole et al., 1971). Over the next 17 years, the
`total exposure to TAM reached 1.5 million patient years
`(Litherland and Jackson, 1988) and other
`selective
`estrogen receptor modulators
`(SERMs)
`are being
`developed and studied. TAM is now the most frequently
`prescribed antiestrogen, and compelling data have
`demonstrated a significant overall survival benefit with
`the administration of this agent in breast cancer patients
`with endocrine responsive disease (EBCTCG, 1992,
`1998).
`When compared with cytotoxic chemotherapy, anti-
`estrogens are well
`tolerated and are associated with
`mostly minor toxicities (Love, 1989). Common side
`effects associated with TAM therapy include vasomotor
`symptoms, gastrointestinal disturbance, atrophic vagi-
`nitis, and changes in sexual functioning (Day et al.,
`1999). While the frequency and severity of hot flashes
`and other toxicities can be particularly unpleasant for
`some women, remarkably few discontinue TAM be-
`cause of these side effects. Medical indications for the
`
`|nnoPharma Exhibit 10640001
`
`
`
`Antiestrogen resistance
`R Clarke et al
`
`7317
`
`prompt discontinuation of therapy include associated
`venous thromboembolic disease and endometrial cancer
`
`(typically invasive adenocarcinoma, although uterine
`sarcomas have been reported). The incidence of these
`events is very low, and screening methods for both deep
`vein thrombosis and endometrial abnormalities exist.
`
`However, these increased risks must be considered in the
`light of the potential benefits—particularly in the case of
`healthy women considering TAM in the setting of
`chemoprevention as opposed to active treatment. The
`development of both venous thromboembolic disease
`and endometrial cancer is attributed to the estrogenic
`effects of TAM and may be abrogated by the develop-
`ment of more SERMs (e.g., raloxifene) or of pure ER
`antagonists (e.g., ICI 182,780; fulvestrant) (Robertson,
`2001).
`Some antiestrogens produce beneficial effects beyond
`their ability to inhibit existing breast cancers. The most
`convincing evidence supports an association between
`TAM treatment and a marked reduction in the risk of
`
`developing a contralateral breast cancer (EBCTCG,
`1992) and a significant reduction in the incidence and
`severity of osteoporosis in postmenopausal women
`(Freedman er al., 2001; Kinsinger er al., 2002). Several
`early studies
`suggested a reduction in the risk of
`cardiovascular disease with TAM therapy, but this is
`not consistently reported (EBCTCG, 1998; Fisher er al.,
`1998). When observed, the cardiovascular benefit was
`usually attributed to the estrogenic effects of TAM; both
`estrogens and TAM produce apparently beneficial
`changes in serum triglyceride and cholesterol concentra-
`tions (Joensuu er al., 2000), perhaps through effects
`mediated by apolipoprotein E (Liberopoulos er al.,
`2002). However, these findings must be considered in the
`light of recent
`large studies of estrogenic hormone
`replacement therapy (HRT) that either failed to identify
`an HRT—induced reduction in coronary heart disease
`(Hulley er al., 1998; Grady er al., 2002; WHI, 2002) and
`stroke (Viscoli er a[., 2001; WHI, 2002), or demon-
`strated an increase in the risk of these diseases.
`
`An overview of antiestrogen resistance
`
`Despite the relative safety and significant antineoplastic
`and chemopreventive activities of antiestrogens, most
`initially responsive breast
`tumors acquire resistance
`(Clarke er al., 2001b). It
`is unlikely that any single
`mechanism or single gene confers antiestrogen resis-
`tance. Rather,
`several mechanisms likely exist
`that
`encompass pharmacologic,
`immunological, and mole-
`cular events. These mechanisms, none of which are fully
`understood,
`likely vary within tumors.
`Intratumor
`variability in antiestrogen responsiveness will
`reflect
`the presence of multiple cell subpopulations (Clarke
`el al., 1990a). Since breast cancers appear highly plastic
`and adaptable to selective pressures,
`the intratumor
`diversity in antiestrogen responsive subpopulations also
`likely changes over time. Tumors appear capable of
`dynamically remodeling their cell populations in re-
`sponse to changes in host immunity or endocrinology,
`or the administration of local or systemic therapies. This
`
`(some existing
`plasticity is probably both cellular
`populations die out/back while other populations
`become dominant) and molecular (new cell populations
`emerge as
`individual cells/populations adapt
`their
`phenotypes by modifying their
`transcriptomes/pro-
`teomes).
`Since the major pharmacologic and immunologic
`mechanisms of antiestrogen resistance have been pre-
`viously reviewed (Clarke el al., 2001b), we will focus on
`the role of molecular signaling through ER—mediated
`activities in antiestrogen responsiveness. Antiestrogen
`resistance can be either de novo or acquired. The most
`common and best defined mechanism of de novo
`
`resistance is the absence of both ER and progesterone
`receptor (PR) expressions. However, we fail to predict
`response to antiestrogens in approximately 25% of
`ER——/PR+, 66% of ER+/PR—, and 55% of ER—/
`PR —— breast tumors (Honig, 1996). Many ER+ and/or
`PR —— breast tumors are already resistant by the time of
`diagnosis and the resistance mechanism in these tumors
`is unknown.
`
`Overall, a loss of antiestrogen responsiveness by
`initially responsive tumors is
`likely to be the most
`common acquired resistance phenotype. Most initially
`antiestrogen responsive tumors retain levels of ER
`expression at recurrence on antiestrogen therapy that
`would still define them as being ER+ (Encarnacion
`er a[., 1993; Kuukasjarvi er al., 1996; Bachleitner-
`Hofmann er al., 2002). Most data are for TAM
`treatment; ICI 182780, which causes degradation of
`ER (Dauvois er al., 1992), may have a greater potential
`for producing ER— tumors (Kuukasjarvi er al., 1996).
`From our in vizro studies, loss of ER is not required to
`achieve resistance to either
`ICI 182,780 or TAM
`(Brunner er al., 1993b, 1997). The loss of ER expression
`upon recurrence despite adjuvant TAM therapy has
`been reported in less than 25% of tumors (Kuukasjarvi
`er al., 1996; Bachleitner—Hofmann er al., 2002). Overall,
`a loss of ER expression does not seem to be the major
`mechanism driving acquired antiestrogen resistance.
`A different resistance phenotype has been described in
`human breast cancer xenografts that exhibit a switch to
`a TAM—stimulated phenotype. This mechanism of
`clinical but not pharmacologic resistance may not be
`the dominant antiestrogen resistance phenotype. If the
`prevalence of acquired resistance phenotypes in ER+
`tumors broadly reflects what
`is
`seen in de novo
`resistance, then the dominant resistance phenotype is a
`loss of antiestrogen responsiveness.
`Whether the continued expression of ER is required
`for antiestrogen—resistant tumor growth or survival is
`not known. However, responses to aromatase inhibitors
`after an initial response and then failure on TAM are
`common (Buzdar and Howell, 2001) and strongly
`suggest that some TAM—resistant tumors retain a degree
`of estrogen responsiveness. Where durations of re-
`sponses to second—line endocrine manipulations are
`short, truly estrogen—independent cell populations are
`either already present at the time of recurrence and/or
`many cells in the tumor are able to adapt rapidly to
`further changes in their endocrine environment. Very
`
`Oncogene
`
`|nnoPharma Exhibit 1064.0002
`
`
`
`7318
`
`Antiestrogen resistance
`R Clarke et al
`
`short response durations or disease stabilization may
`reflect the withdrawal of a mitogenic stimulus that is not
`required for the survival or basal proliferation of most
`cells in the tumor.
`
`issue may be addressed with the concurrent use of
`bisphosphonates or other therapies for osteoporosis.
`Clinical experience with ICI 182,780 has been reviewed
`by Howell (2001).
`
`Antiestrogens
`
`TAM is a triphenylethylene and its triaryl structure has
`been widely copied in the design of new compounds.
`Several TAM derivatives are already available, includ-
`ing toremifene (chloro—tamoxifen) and droloxifene (3-
`hydroxytamoxifen). Not surprisingly, both drugs are
`essentially equivalent
`to TAM in terms of
`their
`antitumor activities and toxicities (Roos et al., 1983;
`Pyrhonen et al., 1999), so neither is widely used in
`clinical practice.
`The characteristic of raloxifene that has attracted the
`
`most interest is its apparent lack of estrogenic effects in
`the uterus,
`resulting in great
`interest
`in this drug’s
`potential role in breast cancer chemoprevention. Sub-
`group analysis of the data from the Multiple Outcomes
`of Raloxifene (MORE) trial revealed that administra-
`tion of raloxifene was associated with a 75% reduction
`in the incidence of invasive breast cancer without a
`concurrent
`increase in the incidence of endometrial
`
`cancers (Cummings et al., 1999). This finding has led to
`the ongoing randomized study of TAM and raloxifene
`(STAR) in breast cancer prevention. Raloxifene still acts
`as an antiestrogen in the brain, increasing the incidence
`of hot flashes (Davies et al., 1999). A high incidence of
`severe hot
`flashes is problematic for a drug to be
`administered for approximately 5 years to otherwise
`apparently healthy women. Raloxifene was recently
`approved by the Food and Drug Administration for the
`treatment and prevention of osteoporosis in postmeno-
`pausal women. While a benzothiophene,
`raloxifene
`(keoxifene; LY 156,758) has a three—dimensional struc-
`ture broadly similar to the triphenylethylenes.
`ICI 182,780 (Faslodex; Fulvestrant) is among the
`more promising new antiestrogens. Unlike TAM, ICI
`182,780 is a steroidal ER inhibitor
`that
`is often
`described as a ‘pure’ antagonist with no estrogenic
`activity. This is in comparison to the triphenylethylene
`and benzothiophene antiestrogens, which are nonster—
`oidal, competitive ER inhibitors with partial agonist
`activity. The pure antagonist
`is characterized by
`antineoplastic activity in breast cancer and is devoid of
`uterotropic effects. However, the lack of agonist activity
`limits beneficial effects in bone. Whether ICI 182,780
`also will
`increase hot flashes depends on whether it
`reaches adequate concentrations in the brain. Unlike
`TAM (Clarke et a[., 1992), ICI 182,780 appears to be a
`substrate
`for
`the
`P—glycoprotein
`efflux
`pump
`G)e Vincenzo et al., 1996), a major contributor to the
`blood—brain barrier (Cordon—Cardo et al., 1989). Con-
`sistent with this observation, initial studies suggest that
`this antiestrogen does not enter
`the brain in high
`concentrations (Howell et al., 1996). Pure antagonists
`may further exacerbate bone loss, a concern that also
`applies to aromatase inhibitors (Dowsett, 1997), but this
`
`Antiestrogens and breast cancer treatment
`
`Antiestrogens are effective in the adjuvant, metastatic,
`and chemopreventive settings and clearly induce sig-
`nificant
`increases in overall survival
`in some breast
`
`cancer patients (EBCTCG, 1992, 1998). Unlike aroma-
`tase inhibitors (inhibit estradiol biosynthesis), which are
`administered as
`single agents only to women with
`nonfunctioning ovaries, TAM can be given irrespective
`of menopausal status. In the adjuvant setting, TAM is
`administered at a daily oral dose of 20 mg, and several
`studies have now shown that the optimal duration of
`treatment is 5 years. While shorter (2 years) and longer
`(10 years)
`treatment durations produce notable re-
`sponses, the risk : benefit ratios are strongly in favor of
`5 years of treatment (Stewart et a[., 1996; EBCTCG,
`1998).
`While molecular predictors of tumor responsiveness
`are rare for most breast cancer treatments, expressions
`of ER and PR strongly predict for a response to
`antiestrogens. Up to 75% of breast tumors expressing
`both receptors (ER+/PR+) respond to TAM. Re-
`sponse rates are somewhat lower in ER+ /PR— tumors
`(~34%) and ER—/PR+ tumors (45%). The response
`rate in ER—/PR+ may be an overestimate; relatively
`few tumors with this phenotype have been evaluated and
`the ER— assessment may include false—negative ER
`measurements. Only a small proportion of ER—/PR—
`tumors respond to antiestrogens (< 10%), perhaps also
`reflecting false—negative ER measurements. Indeed, the
`most recent meta—analysis from the Early Breast Cancer
`Trialists Collaborative Group (EBCTCG)
`found no
`significant reduction in recurrence rates in patients with
`ER—poor
`tumors who
`received
`adjuvant TAM
`(EBCTCG, 1998).
`Results of the 1998 EBCTCG meta—analysis found
`limited evidence for a TAM—induced increase in the risk
`
`of death from any cause in women with ER—poor
`tumors. Why TAM might be detrimental
`to some
`women is unclear. However, ER— tumors are known
`to exhibit a more aggressive phenotype associated with
`lower rates of overall survival (Aamdal et al., 1984) and
`would be expected to recur earlier and more frequently.
`Estrogenic effects of TAM in these women also could
`have increased the number of deaths from cardiovas-
`
`cular disease and stroke, reflecting the data noted above
`from recent studies of estrogenic HRT use (Viscoli et al.,
`2001; WHI, 2002).
`
`Antiestrogens and breast cancer chemoprevention
`
`TAM’s ability to inhibit contralateral breast cancers and
`relatively low incidence of serious side effects led to
`studies into its potential use as a chemopreventive agent
`for patients with a high breast cancer risk. Three large,
`randomized, chemoprevention studies with TAM have
`
`Oncogene
`
`|nnoPharma Exhibit 10640003
`
`
`
`R Clarke et al
`Antiestrogen resistance
`
`@
`7319
`
`been performed to date: the NSABP P-1 trial (n : 13 388
`participants) (Fisher er al., 1998), the Royal Marsden
`Trial (n:247l participants) (Powles er al., 1998), and
`the Italian Chemoprevention Trial (n:5408 partici-
`pants) (Veronesi er al., 1998). Outcomes have been
`mixed: no significant reduction in risk was seen in the
`initial reports of either the UK or Italian trials, whereas
`the P—1
`trial
`reported significant
`reductions in the
`incidence of both noninvasive (50%) and invasive
`(49%) breast cancers. A recent update on the Italian
`Trial reports an 82% TAM—induced reduction in the
`breast cancer risk among women at high risk for ER +
`breast cancer (Veronesi er a[., 2003). In the NSABP trial,
`reductions in breast tumor incidence were seen only in
`the incidences of ER+ tumors (Fisher er al., 1998).
`Reasons for the disparities among the trials have been
`widely discussed; these tend to focus on differences in
`patient populations, subject eligibility criteria, and study
`size. Results from the NSABP P-1 trial, which are
`broadly consistent with the 39% reduction in contral-
`ateral breast cancer incidence reported for TAM use
`(EBCTCG, 1992), are usually considered the more
`definitive. These data contributed to the decision by
`the Federal Drug Administration (USA) in October
`1998 to allow the use of TAM as a chemopreventive
`agent for breast cancer. More recently, NSABP has
`reported TAM—induced reductions
`in the risks of
`adenosis, fibrocystic disease, hyperplasia, metaplasia,
`fibroadenoma, and fibrosis in the P—1 trial (Tan—Chiu
`er al., 2003).
`
`Estrogens and breast cancer
`
`Since antiestrogen action and resistance are intimately
`affected by estrogen exposure, we briefly address the
`role of estrogens in breast cancer. An association
`between parity and breast cancer risk was observed by
`the 16th century Italian physician Bernadino Ramazzini
`(l633—1714) in his ‘De Morbis Artzficium’ published in
`1700. The ability of ovariectomy to induce remissions in
`premenopausal breast cancer patients was shown by the
`Scottish physician George Beatson,
`the first clear
`evidence of an effective endocrine therapy for this
`disease (Beatson, 1896). More recent epidemiologic
`data show clear associations of early age at menarche,
`late age at menopause (Nishizuka, 1992), pregnancy
`(Hsieh er al., 1994), obesity (Hulka and Stark, 1995),
`serum estrogen concentrations (EHBCCG, 2002), and
`use of estrogenic HRTs
`(Magnusson er al., 1999;
`Schairer er al., 1999, 2000) or oral contraceptives
`(Berger er al., 2000) with an increase in the risk of
`developing breast cancer. Risk appears related to the
`timing of exposure and whether the cancer develops
`during the premenopause or postmenopause (Hilakivi—
`Clarke er al., 2002).
`risk
`Precisely how estrogens affect breast cancer
`remains controversial and outcome may be dependent
`upon the timing and duration of exposure. During the
`postmenopausal years, estrogenic stimuli are more
`closely associated with an increased breast cancer risk.
`
`However, we have recently reviewed evidence consistent
`with the hypothesis that, depending on the timing of
`exposure,
`increased estrogenic exposure can be asso-
`ciated with a reduced risk of breast cancer (Hilakivi—
`Clarke er a[., 2002). For example, estrogenic stimuli
`during childhood or the premenopausal years may affect
`breast development such that the breast is less suscep-
`tible to transformation. Estrogens may reduce breast
`cancer incidence in some women by altering mammary
`gland development and inducing the expression of genes
`involved in DNA repair (Hilakivi—Clarke er al., 1999a;
`Hilakivi—Clarke, 2000).
`For the purposes of this review, we will focus on the
`aspects of estrogen exposure that are associated with
`increased breast cancer risk and the survival/prolifera-
`tion of established neoplastic breast
`cells. Hence,
`estrogens can be considered to act either as promoters
`(factors that stimulate the growth and/or survival of
`existing transformed cells) or as initiators (factors that
`induce the genetic damage that
`leads
`to cellular
`transformation). Evidence that estrogens are tumor
`promoters is well established from both experimental
`and clinical observations. For example, the growth of
`several human breast cancer cell lines in vitro and in viva
`is stimulated by estrogenic supplementation.
`Indeed,
`such estrogenic supplementation is effective whether
`administered as
`classical estrogens
`(e.g., estradiol,
`estrone, or estriol) or plant—derived phytoestrogens such
`as the isoflavone genistein (Hsieh er al., 1998).
`In
`addition, antiestrogens, aromatase inhibitors, leutinizing
`hormone releasing hormone agonists/antagonists, and
`ovariectomy are effective in the treatment of some
`breast cancer patients, all of which limit the interaction
`between a promotional (estrogenic) stimulus and cancer
`cells.
`the effects of estrogens are
`As tumor promoters,
`related to the duration and timing of exposure. With-
`drawal of an estrogenic stimulus that acts as a promoter
`could produce an eventual reduction in risk because it
`no longer promotes the growth or survival of existing
`cancer cells. Pregnancy produces a natural and sig-
`nificant
`increase in circulating estrogens, but only a
`transitory increase in breast cancer
`risk in young
`women. Indeed, if the first pregnancy was at a young
`age,
`the short—term increase may eventually translate
`into a lifetime reduction in breast cancer risk (Hsieh
`er a[., 1994). The increased breast cancer risk associated
`with either oral contraceptive or estrogenic HRT use is
`also related to the recency of use. Risk begins to reduce
`with the cessation of use and is highest in current users
`(CGHFBC, 1996; Schairer etaZ.,2000).
`Evidence that estrogens act as chemical initiators is
`more controversial. Estrogens can exhibit carcinogenic
`activity in some animal models; perhaps the best—known
`example is
`the ability of estrogens to induce renal
`cancers in Syrian hamsters (Kirkman, 1972). However,
`compelling evidence that estrogens initiate mammary
`cancer
`in animals
`is hard to find.
`In the 1930s,
`Lacassagne (1932) performed several studies in male
`mice and showed that administration of large doses of
`estrone can induce mammary tumors. While consistent
`
`Oncogene
`
`|nnoPharma Exhibit 1064.0004
`
`
`
`7320
`
`Antiestrogen resistance
`R Clarke et al
`
`with an estrogen—mediated initiation of mammary
`cancer, it is possible that the mice were infected with
`the mouse mammary tumor virus (MMTV). Other than
`some transgenic/null mouse models, only in the ACI rat
`does estrogen administration reproducibly produce a
`high incidence of mammary tumors (Cavalieri and
`Rogan, 2002).
`Reactive estrogen semiquinone/quinone intermedi-
`ates, produced by the redox cycling of estrogen
`metabolites hydroxylated at the C3 and C4 positions
`of the aromatic A—ring, are the most likely estrogen
`initiators (Cavalieri et al., 1997; Bishop and Tipping,
`1998; Cavalieri and Rogan, 2002). These reactive species
`can generate a substantial intracellular oxidative stress
`and directly damage DNA through the production of
`DNA adducts. Such events could define
`reactive
`
`estrogen metabolites as initiators, rather than as merely
`promoters of carcinogenesis. Recently,
`the National
`Toxicology Program (2003) listed, for the first
`time,
`steroidal estrogens as carcinogens.
`
`Estrogen independence and antiestrogen resistance
`
`Estrogen independence and antiestrogen resistance are
`often considered to be synonymous, which is not
`surprising since ER— tumors are definitively estrogen-
`independent and very rarely respond to antiestrogens,
`ovariectomy, or aromatase inhibitors. Nonetheless,
`several observations suggest that various forms of both
`estrogen independence and antiestrogen resistance exist
`and that these may be biologically and clinically very
`different. For example, second—line responses to aroma-
`tase inhibitors after response and recurrence on TAM
`are common (Goss et al., 1995; Buzdar et al., 1996).
`Crossover between more similar compounds, such as
`other nonsteroidal antiestrogens, rarely produces sec-
`ondary responses (Johnston, 2001), although crossover
`to structurally different antiestrogens can produce
`secondary responses in patients. Tumors that respond
`first
`to TAM (triphenylethylene)
`show a marked
`response to ICI 182,780 (steroidal) administered upon
`failure of the TAM therapy (Howell et al., 1995). Similar
`patterns of responses were seen previously in experi-
`mental models (Brunner et al., 1993b). For example,
`MCF—7 human breast cancer cells were selected for the
`
`ability to grow in the absence of estrogens (Clarke et al.,
`1989a). The selected cells are estrogen—independent
`because they no longer require estrogens for growth
`either in cell culture or as xenografts in athymic nude
`mice. However, when exposed to either 4—hydroxyta—
`
`moxifen or ICI 182,780, the cells are growth inhibited
`both in vitro and in vivo (Clarke et al., 1989a; Brunner
`et al., 1993a, b).
`These observations strongly imply that the ability of
`breast cancer cells to grow in a low or nonestrogenic
`environment is not always synonymous with antiestro-
`gen resistance. Four antiestrogen resistance phenotypes
`have been defined (Clarke and Brunner, 1995) and are
`shown in Table 1. The clinical applicability of these
`phenotypes remains to be determined but they are useful
`for defining resistance phenotypes
`in experimental
`models.
`
`Intratumor estrogens and antiestrogens and exogenous
`estrogenic exposures
`
`Antiestrogens act within cells, primarily to compete with
`available estrogens
`for binding to ER. Thus,
`the
`antiestrogenic potency of any compound is related to
`its affinity for ER relative to that of any estrogens
`present and the concentrations of both the antiestrogens
`and estrogens. The data in Table 2 show the relative
`affinities of the primary estrogens, antiestrogens and
`their major metabolites, and selected environmental
`estrogens and phytoestrogens.
`Intratumor estrogen
`concentrations are affected by several factors including
`serum estrogen concentrations
`and local estrogen
`production within the breast. Serum estrogen concen-
`trations are affected by the presence or absence of
`functional ovaries and exogenous estrogen use such as
`HRT, some oral contraceptives, and various dietary
`components.
`Passive diffusion into cells across the plasma mem-
`brane appears to be TAM’s and estradiols’s primary
`method of entry into cells. However, both TAM and
`estrogens are extensively bound to serum proteins and
`probably also to cellular proteins in tumor/nontumor
`cells within the breast (Clarke et al., 2001b). Release
`from serum proteins likely occurs within the tumor
`vasculature, with both estrogens and antiestrogens being
`subsequently sequestered within tumor/nontumor cells
`by intracellular proteins. The lipophilicity of both
`hormone and drug, and the significant amount of
`adipose tissue in the breast, may produce a local
`reservoir for both estrogens and antiestrogens. How-
`ever,
`the concentration of free drug/hormone within
`cells and serum may be relatively low. Intracellular
`sequestration of drug/hormone in tumor and stromal
`cells could produce a concentration gradient favoring
`
`Antieszrogen resistance
`
`Phenotype
`
`Table 1
`
`Antiestrogen resistance phenotypes
`
`Type 1
`Type 2
`
`Type 3
`Type 4
`
`Fully responsive to antiestrogens and aromatase inhibitors
`Resistant“ to nonsteroidal antiestrogens but responsive to ICI 182,780 and aromatase inhibitors
`(or resistant to ICI 182,780 but responsive to nonsteroidal antiestrogens and aromatase inhibitors)
`Resistant to all antiestrogens but potentially responsive to aromatase inhibitors
`Multihormone—resistant (resistant to all endocrine therapies and includes ER— and PR— tumors)
`
`“Resistance can be considered as unresponsiveness and antiestrogen—stimulated phenotypes
`
`Oncogene
`
`|nnoPharma Exhibit 1064.0005
`
`
`
`Table 2 Relative binding affinities (approximate) of selected estro—
`gens, antiestrogens, and environmental estrogens and phytoestrogens“
`Cgmpmmd
`Regan-W binding affimy
`(17/3—estradiol:100)
`ER
`ER
`
`0‘
`
`3
`
`Estrogens
`Estrone
`E33101
`.
`Antzestrogens
`Tamoxifen
`4—Hydroxytamoxifen
`b
`Nafoxidinfi
`ICI 16.4384
`Raloxifene
`Clomiphene
`
`60
`14
`
`7
`178
`44
`85
`69
`25
`
`37
`31
`
`6
`339
`16
`166
`16
`12
`
`Environmental estrogens andphytoestrogens
`genistein 1
`
`0,1)/_DDE 2(2_Ch1Om_phenyD_2_
`(4—chlorophenyl)—l,ldichloroethylene
`1318131161101 A
`
`1 15 104
`< ' 7X
`<0_01
`
`1 636 104
`< ' 5X
`<0_01
`
`0-01
`
`0-01
`
`.
`.
`'AdaptCd from Kulper 81 “Z” 09.98)’ Kulper 8.1”]: (1997) and Bowers
`et al. (2000); the methods for estimating ER binding are not the same
`across these studies but all three express binding relative to the values
`estimated for l7/i—estradiol. "ICI 182,780 is an analog of ICI 164,384
`
`diffusion into local tissues. If the affinity and capacity of
`tissue for drug/hormone exceed that of blood, significant
`accumulation within tumors would likely occur. Data in
`Table 3 (adapted from Clarke et al., 2001b) illustrate
`
`Antiestrogen resistance
`R Clarke et al
`
`732 1
`
`several points regarding the pharmacokinetics of estro-
`ggns and antigstfoggns. For gxalnplg’ intratumor Con-
`centrations of both estradiol and TAM are much higher
`than their respective concentrations in the serum. For
`estrogens, where the primary estrogen present in tumors
`is l7[3—estradiol, both biosynthesis within the tumor and
`significant uptake from blood occur.
`The ability of estrogens and antiestrogens to compete
`for binding to ER is
`likely to reflect
`intracellular
`availability. While their respective free concentrations
`.
`.
`are largely unknown, the data in Tables 2 and 3 imply
`that many breast tumors should accumulate a sufficient
`.
`.
`.
`excess of TAM and its major antiestrogenic metabolites
`to compete readily with intratumor estrogens. If the
`t.
`t
`f
`t d.
`1
`t
`t.
`1 29
`d th
`CS 11113. C
`01' CS 1'3. 10. CO1’1C€1’1 1'3.101’1S (
`.
`3.11
`. 6
`reported concentrations
`for TAM and its major
`metabolites (~ 3 ,uM TAM Jr ~ 7 ,uM N—desmethyltamox-
`ifen + ~0.2 ,aM 4—hydroxytamoxifen) in tumors are good
`approximations (Table 3), antiestro4genic metabolites
`may accumulate to levels up to 10 —fold higher than
`estradiol. While TAM and N—desmethyltamoxifen have
`relative ER binding affinities about 10% that of
`estradiol (Table 2), overall, antiestrogenicity may exceed
`estrogenicity in most TAM—treated breast tumors by
`100 f Id
`.
`.
`1
`t
`.1 b.1.t
`'
`.(aSSummg.equ1Ya en 3'.‘/al 3'
`1 1
`.
`.
`.
`This
`interpretation is consistent with the initial
`antiestrogenic activity of TAM seen in most ER +
`breast cancers. No compelling evidence shows that
`TAM becomes extensively metabolized to purely estro-
`genic metabolites in patients with antiestrogen—resistant
`cancer. Furthermore, little evidence has been produced
`to suggest that the balance of TAM metabolism is such
`
`Table 3 Serum and intratumor estrogen and tamoxifen concentrations“
`
`Serum concentrations
`Mean estimates of estrogen concentrations
`Follicular phase
`Luteal phase
`< 0.28 nM
`< l.l nM
`Pregnancy
`< 150 nM
`Breast cancer
`0.114 nM
`
`Controls
`0.093 nM
`
`Estimates of tamoxifen concentrations
`Concentration
`Drug/metabolite
`< l.l ,aM
`Tamoxifen+metabolites
`<4.0 ,aM
`Tamoxifen
`< 6.0 ,aM
`N—desmethyltamoxifen
`
`Intratumor concentrations
`Mean estimates of estrogen concentrations
`Breast tumors
`Non—neoplastic
`1.29 nM
`0.76 nM
`
`Mean estimates of tamoxifen concentrations
`Concentration
`Drug/metabolite
`
`< 3.0 ,aM
`<4.0 ,aM
`< 7.0 ,aM
`< 8.0 ,aM
`<0.2,aM
`
`Tamoxifen
`Tamoxifen
`N—desmethyltamoxifen
`N—desmethyltamoxifen
`4—Hydroxytamoxifen
`
`Comments
`
`Normal menstrual cycle
`
`Normal third