Endocrine—Re/ated Cancer (1 999) 6 75-92
`
`Use of aromatase inhibitors in
`
`breast carcinoma
`
`R J Santen and H A Harvey’
`Department of Medicine, University of Virginia Health Sciences Center. Charlottesville, Virginia 22908, USA
`‘Department of Medicine, Penn State College of Medicine, Hershey, Pennsylvania 17033, USA
`(Requests for offprints should be addressed to R J Santen)
`
`Abstract
`Aromatase, a cytochrome P—45D enzyme that catalyzes the conversion of androgens to estrogens, is
`the major mechanism of estrogen synthesis in the post-menopausal woman. We review some of the
`recent scientific advances which shed light on the biologic significance, physiology, expression and
`regulation of aromatase in breast tissue.
`Inhibition of aromatase, the terminal step in estrogen
`biosynthesis, provides a way of treating hormone—dependent breast cancer in older patients.
`Aminoglutethimide was the first widely used aromatase inhibitor but had several clinical drawbacks.
`Newer agents are considerably more selective, more potent, less toxic and easier to use in the clinical
`setting. This article reviews the clinical data supporting the use of the potent, oral competitive
`aromatase inhibitors anastrozole, letrozole and vorozole and the irreversible inhibitors 4—OH andro-
`stenedione and exemestane. The more potent compounds inhibit both peripheral and intra-tumoral
`aromatase. We discuss the evidence supporting the notion that aromatase inhibitors lack cross-
`resistance with antiestrogens and suggest that the newer, more potent compounds may have a
`particular application in breast cancertreatment in a setting of adaptive hypersensitivity to estrogens.
`Currently available aromatase inhibitors are safe and effective in the management of hormone-
`dependent breast cancer in post—menopausal women failing antiestrogen therapy and should now be
`used before progestational agents. There is abundant evidence to support testing these compounds
`asfirst—line hormonal therapy for metastatic breast cancer as well as part of adjuvant regimens in older
`patients and quite possibly in chemoprevention trials of breast cancer.
`Endocrine-Related Cancer (1999) 6 75-92
`
`Introduction
`
`Epithelial cells of the normal breast undergo dramatic
`changes during various events in a woman’s life such as
`puberty, the follicular and luteal phases of the menstrual
`cycle, pregnancy and menopause. The co—ordinated
`interaction of growth factors and steroid hormones
`regulate the proliferation and differentiated function of
`epithelial and stromal cells in the normal mammary gland.
`The key growth factors are insulin-like growth factor-I,
`prolactin, insulin, the fibroblast growth factor family of
`growth factors and growth hormone, and major steroid
`hormones are estradiol, progesterone and testosterone
`(Frantz & Wilson 1998).
`For the process of inducing breast cancer, estrogens
`appear to play a predominant role. These sex steroids are
`believed to initiate and promote the process of breast
`carcinogenesis by enhancing the rate of cell division and
`reducing time available for DNA repair. An emerging new
`
`can be metabolized to
`that estrogens
`is
`concept
`catecholestrogens and then to quinones which directly
`damage DNA. These two processes —
`the estrogen
`rcecptopmcdiatcd, genomic effects on proliferation and
`the receptor-independent, genotoxic effects of estrogen
`metabolites - can act either in an additive or synergistic
`fashion to cause breast cancer (Santen er a1. 1999).
`Breast cancers which arise in patients can be divided
`into two subtypes:
`those which are dependent upon
`hormones for growth and those which grow independently
`of hormonal stimulation (Santen er‘ a1. 1990). In the
`honnone—dependent subtype,
`the role of estrogens as
`modulators ofmitogenesis overrides the influence of other
`factors. These sex steroids stimulate cell proliferation
`directly by increasing the rate of transcription of early
`response genes such as c—1nyc and indirectly through
`stimulation of growth factors which are produced largely
`in response to estrogenic regulation (Dickson & Lippman
`1995).
`
`Endocrine-Related Cancer (1999) 6 75-92
`1351-0088/99/006-075 © 1999 Society for Endocrinology Printed in Great Britain
`
`Online version via http://vwvw.endocrino|ogy.org
`
`Astrazeneca Ex. 2047 p. 1
`Mylan Pharrns. Inc. v. Astrazeneca AB IPR20l6-01324
`
`

`
`Santen and Harvey.‘ Use of aromatase inhibitors in breast carcinoma
`
`Based upon the concept that estrogen is the proximate
`regulator of cell proliferation, two general strategies were
`developed for treatment of hormone-dependent breast
`cancer: blockade of estrogen receptor action and inhibition
`of estradiol biosynthesis. Antiestrogens such as tamoxifen
`bind to the estrogen receptor and interfere with trans-
`cription of estrogen—induced genes involved in regulating
`cell proliferation Clinical trials showed tamoxifen to be
`effective in inducing objective tumor regressions and to be
`associated with minimal side-effects and toxicity. The
`second strategy, blockade of estradiol biosynthesis, was
`demonstrated to be feasible using the steroidogenesis
`inhibitor, aminoglutethimide, which produced tumor
`regressions equivalent to those observed with tamoxifen
`(Santen ef al. 1990). However, side-effects from amino-
`glutethimide were considerable and its effects on several
`steroidogenie enzymes required concomitant use of a
`glucocorticoid (Santen er
`a1.
`1982). Consequently,
`tamoxifen became the preferred, first—line endocrine agent
`with which to treat advanced breast cancer. However, the
`clinical efficacy of arninoglutethiiude focused attention
`upon the need to develop more potent, better tolerated, and
`more specific inhibitors of estrogen biosynthesis.
`
`Inhibition of estradiol biosynthesis
`
`Multiple strategies could be used to inhibit estradiol
`biosynthesis as a treatment for estrogen-dependent breast
`cancer. Inhibition of several enzymes in the steroidogenie
`pathway,
`including cholesterol side-chain cleavage, 3
`beta—ol—dehydrogenase—delta 4-5 isomerase,
`17—alpha
`hydroxylase,
`l7-beta hydroxysteroid dehydrogenase,
`estrone sulfatase, and aromatase, could be used to reduce
`the biosynthesis of estradiol
`and potentially cause
`horn1one—dependent breast tumor regression. An addition—
`al strategy is the use of exogenous glucocorticoid to inhibit
`release of adrenocoiticotropin (ACTH) and suppress
`estrogen production. Finally, synthetic progestins such as
`megestrol acetate and medroxy—progesterone acetate exert
`glucocorticoid effects and inhibit estradiol synthesis by
`suppressing ACTH.
`
`The ideal strategy would be to block the synthesis of
`estrogen without inhibiting production of other important
`steroids or giving pharmacological amounts of progestins
`or glucocorticoids. For this reason, blockade of the
`tenninal step in estradiol biosynthesis catalyzed by the
`enzyme aromatase is considered a more specific and
`therefore preferable strategy. Several pharmaceutical
`companies sought to develop potent aromatase inhibitors
`designed to specifically block estrogen biosynthesis with-
`out altering glucocorticoid and mineralocorticoid syn-
`thesis, and without requiring addition of large amounts of
`progestins or exogenous glucoeorticoid.
`
`76
`
`Physiology and regulation of aromatase
`Aromatase is a cytochrome l-’-450 enzyme which catalyzes
`the rate—limiting step in estrogen biosynthe sis,
`the
`conversion of androgens to estrogens (Simpson 61‘ al.
`1997, Sasano & Harada 1998). Two major androgens,
`androstenedionc and testosterone, serve as substrates for
`aromatase. The aromatase enzyme consists of a complex
`containing a cytochrome P—450 protein as well as the
`flavoprotein NADPH cytochrome P-450
`reductase
`(Simpson at
`a].
`1997). The gene
`coding for
`the
`cytochrome P-450 protein (P-450 AROM) exceeds 70 kb
`and is
`the largest of the cytochrome P-450 family
`(Simpson er a1. 1993). The cDNA of the aromatase gene
`contains 3.4 kb and encodes a polypeptide of 503 amino
`acids with a molecular weight of 55 kDa. Approximately
`30% homology exists with other cytochrome P-450
`proteins. Because its overall homology to other members
`of the F-450 superfamily is low, aromatase belongs to a
`separate gene family designated CYPI 9.
`Recent studies indicate that the transcription of the
`aromatase gene is highly regulated (Simpson er a]. 1989,
`1993, 1997). The first exon of the aromatase gene is
`transcribed into aromatase message but not translated into
`protein. There exist nine alternative first exons which can
`initiate the transcription of aromatase. Each of these
`alternate exons contains up stream DNA sequences which
`can either enhance or silence the transcription of arom-
`atase. Different tissues utilize specific alternate exons to
`initiate transcription. For example, the placenta utilizes
`alternate exon 1.1, the testis alternate exon H, adipose
`tissue 1.3 and l.4 and brain lf. Enhancers which react with
`upstream elements of these alternate exons markedly
`stimulate the rate of transcription of the aromatase gene.
`Thus, each tissue can regulate the amount of aromatase
`transcribed in a highly specific manner (Simpson er a1.
`1993 ).
`Aroinatase expression occurs in many organs, includ-
`ing ovary, placenta, hypothalamus, liver, muscle, adipose
`tissue, and breast cancer itself. Aromatase eatalyyes three
`separate steroid hydroxylations which are involved in the
`conversion of androstenedione to estrone or testosterone
`to estradiol. The first two give rise to l9—hydroxy and 19-
`aldehyde structures and the third, although still contro-
`versial, probably also involves the C-19 methyl group with
`release of formic acid (Fishrnan & Hahn 1987). This
`enzymatic action results in the saturation of the A-ring of
`the steroid molecule to produce an aromatic structure,
`hence the tenn aromatization.
`
`the major source of
`In the premenopausal state,
`aromatase and of its substrates is the ovary. However,
`extra—glandular aromatization of adrenal substrates in
`peripheral
`sites
`such as
`fat,
`liver and muscle also
`contributes substantially to the estrogen pool in the early
`
`Astrazeneca Ex. 2047 p. 2
`
`

`
`Endocrine-Re/ated Cancer( 1999) 6 75-92
`
`AROMATASE INHIBITORS
`
`
`
`Potency
`
`
`-Fiidiloigl
`" 93%
`
`
`
`
`
`
`First
`
`Generation
`
`Second
`
`Generation
`
`Third
`
`Generation
`
`Figure 1 Diagrammatic representation ofthe potency of aromatase inhibitors as reflected by the isotopic kinetic method for
`determining degree ofaromatase inhibition The percent conversion of androstenedione to estrone is measured isotopically,
`correcting for losses ofestrone by giving 14[C] estrone tracer. Values indicated represent percent inhibition oftotal body aromatase.
`
`follicular and late luteal phases of the riieristiual cycle. 111
`the postmenopausal state, the ovary loses its complement
`of aromatase enzyme although it does continue to secrete
`androstenedione. The adrenal subsumes the primary role
`of providing substrate for aromatase by directly secreting
`testosterone and androstenedione. In addition, dehydro-
`epiaiidrosteroiie and its sulfate are secreted by the adrenal
`and convened into the aromatase substrates, andros—
`tenedione and testosterone,
`in peripheral
`tissues. The
`major source of the aromatase enzyme in postmenopausal
`women is peripheral
`tissues and particularly fat and
`muscle.
`Recent studies identified an additional, important site
`of estrogen production, breast tissue itself. Two-thirds of
`breast carcinomas contain aromatase and synthesiae
`biologically significant amounts of estrogen locally in the
`tumor (Abul-Hajj et a1. 1979, Miller’ & O’Neil 1987,
`Santen er a1. 1994). Proof of local estradiol synthesis
`includes measurement of tumor aromatase activity by
`radiometric or product
`isolation assays, by immuno-
`histochemistry, by demonstration of aromatase mRNA in
`tissue, and by arorriatase enzyme assays performed on
`cells isolated from human tumors and grown in cell
`culture. The expression of aroiriatase is highest in the
`strom al eompartm ent of breast tumors (Santen eta]. l 994)
`but is present in epithelial cells as well. In breast tissue
`
`surrounding the tumors, preadipoeyte fibroblasts contain
`aromatase activity that can be detected by biochemiea
`assay or immunohistocheiiiical staining (Miller & O’Nei
`1987, Santen er a1. 1994). Normal breast tissue also
`contains aromatase as documented by iiiimuiioliisto-
`chemistry, by demonstration of aromatase message, an:
`by enzyme assays of cultured cells (Mor et al. 1998,
`Brodie eta], 1999).
`The biologic relevance of in situ estrogen production
`by aromatase has been demonstrated by xenograf
`experiments which compare tumors containing and no
`containing aromatase (Yue et a1. 1998). Human breas
`cancer cells transfected peniianently with the aromatase
`enzyme
`are
`compared with cells
`transfccted with
`irrelevant DNA. In these experiments, tumors containing
`the transfeetcd aromatase enzyme have higher amounts o ‘
`estrogen and grow faster’ than those with transfeetion 0 ‘
`irrelevant DNA. Further,
`these experiments show that
`local production of estradiol in the tumor is a greater
`source of estrogen than uptake from plasma (Yue et al.
`1998). Taken together,
`these
`studies
`support
`the
`importance of in situ estrogen production by breast tumors
`and suggest that aromatase inhibitors in patients must be
`sufficiently potent to block intra-tumoral aromatase.
`Breast tumor tissue aromatase can be regulated by
`several enhancers of aromatase transcription (Simpson et
`
`77
`
`Astrazeneca Ex. 2047 p. 3
`
`

`
`Santen and Harvey.‘ Use of aromatase inhibitors in breast carcinoma
`
`AROMATASE INHIBITORS
`
`Spectrum of Action
`
`
`
`Third
`
`Generation
`
`
`
`First
`
`Second
`
`_.a!s19S*9r9.ne '
`
`Generation
`
`Generation
`
`Figure 2 Diagrammatic representation ofthe spectrum of action of first through third generation aromatase inhibitors. Vtfith
`development of newer inhibitors, the spectrum of action narrows. The third generation aromatase inhibitors act exclusively on the
`aromatase enzyme and do not appearto exert additional effects.
`
`a]. 1997). Dexaniethasone, phorbol esters, cyclic AM?,
`interleukin 6, and prostaglandins can all stimulate aroma-
`tase transcription in cultured breast cancer cells and
`specifically in the stromal components.
`Interestingly,
`products secreted by epithelial cells in the breast ttunors
`appear to stimulate aromatase in the stroma and provide a
`means
`for autoregulation of tumor growth through
`estrogen production. A rather novel means ofregulation of
`aromatase levels was also recently described —
`the
`stabilization of degradation of enzyme (Harada et a1.
`1999). Aromatase inhibitors bind to the active site of the
`enzyme
`and,
`through mechanisms not
`completely
`understood, prevent proteoly sis of the aromatase protein.
`Each of these mechanisms may enhance the amount of
`aromatase in tumor tissue and increase the need for very
`potent aromatase inhibitors.
`
`Development of aromatase inhibitors
`The first aromatase inhibitors were discovered nearly 30
`years ago and included aminoglutethimide and testolo-
`lactone (Santen er al. 1990). Testololaetone was not very
`potent as an inhibitor, and aminoglutethiniide blocked
`several P-450-mediated enzymatic reactions and was
`
`associated with troublesome side—effects. On the other
`hand, aminoglutethimide appeared to be quite effective in
`causing tumor regressions in patients with breast cancer.
`For this reason, pharmaceutical companies and individual
`investigators focused upon developing more potent and
`specific inhibitors. Second and third generation inhibitors
`were developed with 10- to l00OO-fold greater potency
`than aminoglutethimide and greater specificity (Figs l and
`2). The half—lives of the inhibitors increased with synthesis
`of more potent inhibitors. The third generation aromatase
`inhibitors are capable of decreasing the levels of circu-
`lating estrogens to a greater extent than the first and
`second generation inhibitors in postmenopausal women
`with hormone-dependent breast cancer. Hypothetically,
`these highly potent agents could also reduce levels of
`intra-tumoral aromatase activity to a greater extent than
`the earlier inhibitors but this has not yet been examined.
`
`Pharmacologic classification of
`aromatase inhibitors
`into
`inhibitors
`A convenient
`classification divides
`mechanism based or ‘suicide inhibitors’ (Type 1) and
`competitive inhibitors (Type ll) (Brodie 1993). Suicide
`
`78
`
`Astrazeneca Ex. 2047 p. 4
`
`
`
`- 4—0iii-\¥
`
`-Inhibition
`
`aromatase
`
`

`
`Endocrine-Re/ated Cancer( 1999) 6 75-92
`
`DJ0
`
`20
`
`10
`
`
`
`(pmollLestradiolequivalents)
`
`Estrogen
`
`RIA
`
`I Bioassay
`
`
`
`Baseline
`
`Week 6
`
`Week 1 2
`
`Time of Treatment
`
`Figure 3 Inhibition of plasma estrogen levels as assessed by RIA and by an ultrasensitive, recombinant DNA-based bioassay
`(Jones et al. 1992). Basal estradiol levels are approximately threefold lower when measured by the ultrasensitive assay. During
`administration of the aromatase inhibitor, levels fall to 0.05-0.07 pmolll as assessed by the ultrasensitive assay and to 2-5 pg/ml
`with the standard RIA.
`‘P< 0.01 vs baseline.
`
`inhibitors initially compete with natural substrates (i.e.
`androstenedione and testosterone) for binding to the active
`site of the enzyme. The enzyme, then, specifically acts
`upon the inhibitor to yield reactive alkylating species
`which form covalent bonds at or near the active site of the
`enzyme. Through this mechanism, the enzyme is irrever-
`sibly inactivated. Competitive inhibitors, on the other
`hand, bind reversibly to the active site of the enzyme and
`prevent product formation only as long as the inhibitor
`occupies the catalytic site. Whereas mechanism—based
`inhibitors are exclusively steroidal in type, competitive
`inhibitors consist both of steroidal and non-steroidal
`compounds (Brodie 1993).
`
`Methods used to demonstrate
`aromatase inhibition
`
`The standard method to study aromatase inhibitors in
`patien s is to measure either plasma or urinary estrogen by
`RIA. Early studies demonstrated 50-80% inhibition of
`plasma or urinary estrone or estradiol (Santen e1 21]. 1978,
`
`1981, 1982). Another method involved measurement of
`each estrogen metabolite in urine with calculation of total
`aromatized product. This technique provided results
`similar to those from measurements of urinary estrone or
`estradiol (Lipton 61 a1. 1995). Using these plasma or
`urinary methods, each agent appeared to suppress estrogen
`levels to concentrations approaching the sensitivity of the
`RlAs used. To gain greater specificity and sensitivity,
`investigators utilized the isotopic kinetic technique of
`Siiteri el al. to measure total body aromatase (Grodin ei a1.
`1973, Santcn er a1. 1978, Jones et a1. 1992, Dowsctt er a1.
`1995). This required adn1inist1’ation of tritiated andros-
`tenedione and 14[C]-estrone to patients under steady-state
`conditions and measurement of radiochemically pure
`tritiated estrone and estradiol (Santen er a]. 1978). The
`14[C]-estrone
`allowed correction for
`losses during
`multiple purification steps. Using this technique,
`the
`degree of inhibition with various inhibitors ranged from
`90 to 99%.
`From these observations, it was recognized that more
`sensitive plasma assays of estradiol were needed. One
`
`79
`
`Astrazeneca Ex. 2047 p. 5
`
`

`
`Santen and Harvey.‘ Use of aromatase inhibitors in breast carcinoma
`
`approach was tl1e use of the plasma estrone sulfate assay
`since basal levels of this conjugate in postmenopausal
`women are tenfold higher than the levels of uneonjugated
`estrone and estradiol (Sarnojlik eta]. 1982, Lonning et a1.
`1997). With this measurement, suppression to 85% of
`basal values was observed with most inhibitors. Finally, an
`ultrasensitive bioassay of plasma estradiol which was 50-
`to 100-fold more sensitive than RIA was developed
`(Oerter—Kleir1 el al. 1995). Surprisingly, with this assay,
`one could demonstrate suppression to levels of estradiol of
`0.05-0.07 pg/ml, concentrations substantially lower than
`the 2-5 pg/ml suppressed levels detected by RIA (Fig. 3).
`As observed with use of other highly sensitive plasma
`hormone assays, for example for luteinizing hormone
`(TH), follicle-stimulating hormone (FSH),
`thyrotropin
`(TSH), and growth hormone, the levels measured under
`basal conditions and during suppression with these assays
`reveals much lower values than with insensitive RIAS.
`This probably reflects the fact that insensitive assays are
`measuring a substantial fraction of ‘blank’ or non-specific
`assay artifact. With the use ofhighly sensitive assays, this
`artifactual measurement
`is eliminated and the actual
`values measured are much lower. Thus with the ultra-
`sensitive estradiol bioassay,
`the basal
`levels in post-
`menopausal women average 1-3 pg/1111 (vs 5-20 pg/ml
`with RIA) (Oerter-Klein et a1. 1995). During development
`of the second and third generation aromatase inhibitors,
`each of these methods has been used to demonstrate the
`magnitude of suppression of enzymatic activity. For these
`measurements, the isotopic kinetic technique is consid-
`ered the “gold standard’ since it is highly sensitive and
`allows comparison among various inhibitors (Fig. 1).
`
`First generation aromatase inhibitors
`The first aromatase inhibitor to be widely used in the
`treatment of metastatic breast cancer i11 postmenopausal
`women was the drug aminoglutethimide (Santen ct ai.
`1978,
`1981,
`1982,
`1990).
`lsotopic kinetic
`studies
`demonstrated a 90-95% inhibition of aromatase activity
`(Santen er a1. 1978). Plasma estrone and estradiol levels
`and urinary estrogens fell by 50-80% in response to this
`aromatase inhibitor. An additional effect, described by
`Lom1ing
`and colleagues, was
`the
`acceleration of
`metabolism of estrogen sulfate (Geisler et a1. 1997). This
`effect resulted in further lowering of free estrogen levels
`in plasma and in urine. With further study of a1nino—
`glutethimide, multiple metabolic effects were demon-
`strated,
`including inhibition of 11-beta hydroxylase,
`aldosterone synthase, and thyrcxine synthesis as well as
`induction of enzymes metabolizing synthetic g1ucocorti-
`coids and aminoglutethimide itself (Santen er a1. 1990).
`When aminoglutethimide was combined with a
`corticosteroid such as hydrocortisone,
`the regimen
`
`80
`
`produced durable clinical responses in 30-50% of patients
`(Santen er a1. 1990). This approach, however, had several
`important drawbacks. First, aminoglutethimide was asso-
`ciated with troublesome side-effects, including drowsi-
`ness, skin rash, and ataxia. Secondly, standard doses of
`1000 mg aminoglutethimide daily could also inhibit other
`cytochrome P-450-mediated steroid hydroxylations, par-
`tieularly those involving the cholesterol
`side-chain
`cleavage enzymes (Santen et a1. 1990, Cocconi 1994).
`This non-selectivity for aromatase led to inhibition of the
`biosynthesis of cortisol, aldosterone and also of thyroid
`hormone. This necessitated co-administration of the
`glueocorticoid, hydroeortisone, and in about 5% of
`patients, thyroxine.
`Four randomized. controlled clinical trials compared
`aminoglutethimide in combination with hydrocortisone
`with tamoxifen in advanced breast cancer. (Smith eta].
`1981, Lipton et a1. 1982, Alonso—Mun0z et a]. 1988, Gale
`et a1. 1994). The antiestrogen tamoxifen and the inhibitor
`of estrogen biosynthesis, aminoglutethimide/hydroeorti-
`sone produced similar rates of objective disease regression
`and duration of response (Santen et a1. 1990, Gale et a1.
`1994). Tamoxifen produced many fewer side-effects than
`did aminoglutethimide/hydrocortisone. Cross-over
`re-
`sponses to aminoglutethimide/hydrocortisone in patients
`relapsing on tamoxifen were substantial, ranging from 25
`to 50% and 36% in the largest randomized study (Gale at
`al. 1994). In marked contrast, patients initially treated with
`aminoglutethimide/hydrocortisone responded less
`fre-
`quently when crossed over to tamoxifen (19%) (Gale eta].
`1994). This observation reinforced the concept that the
`antiestrogens be used as
`first-line
`agents and the
`aromatase inhibitors as second- or third-line therapies.
`With the development of better aromatase inhibitors,
`aminoglutethimide is now of historical interest only.
`
`Second generation aromatase inhibitors
`
`Fadrozole
`
`4—(5,6,7,8—tetrahy dro-
`l6949A',
`(CGS
`Fadrozole
`imidazo[1,5a]-pyridin-5yl)
`bcnzonitrile monohydro-
`chloride) is a fairly potent inhibitor of aromatase witl1 an
`inhibitory constant
`(IQ) of 0.19 nM (vs 600 nM for
`aminoglutethimide) (Harvey et a1. 1994, Harvey 1996).
`Cholesterol side-chain cleavage activity is minimal but C-
`11 hydroxylase inhibitory effects are observed in I/Itno at
`high drug concentrations.
`lnitial dose-seeking studies conducted in patients
`demonstrated effective aromatase inhibition at doses of
`1.8-4.0 mg daily (Harvey et a1. 1994). A phase 11 study
`then compared doses of 0 6 mg three times daily,
`1 mg
`twice daily, and 2 mg twice daily. Maximal suppression of
`plasma and urinary estrogens occurred at a dose of 1.0 mg
`
`Astrazeneca Ex. 2047 p. 6
`
`

`
`Table 1 Comparison ofthird generation aromatase inhibitors with progestin therapy
`
`Endocrine-Re/ated Cancer( 1999) 6 75-92
`
`Meg ace vs vo rozo|e*
`
`Megace vs anastrozole (1 mg)
`
`Megace vs letrozole (2.5 mg)
`
`Response
`parameters Megace
`Overall
`28.7
`survival
`months
`
`Vorozole
`26
`months
`
`P
`NS
`
`Megace
`22.5
`months
`
`Anastrozole
`26.7
`months
`
`P
`0.02
`
`Megace
`215
`months
`
`Letrozole
`25.3
`months
`
`P
`0.15
`
`Objective
`response
`rates
`(CR+PR)
`
`Clinical
`benefit
`(CR+PR+
`stab|e>6
`months)
`
`Timeto
`progression
`
`Number in
`study
`
`7.6%
`
`10.5%
`
`NS
`
`7.9%
`
`10.3%
`
`NS
`
`16.4%
`
`23.6%
`
`0.04
`
`Not
`reported
`
`Not
`reported
`
`26.1%
`
`35%
`
`NS
`
`32%
`
`35%
`
`NS
`
`3.6
`months
`
`2.7
`months
`
`452
`
`NS
`
`5
`months
`
`5
`months
`
`764
`
`NS
`
`5.5
`months
`
`0.07
`
`5.6
`months
`
`551
`
`NS, not significant.
`"Goss (1998).
`Megace, megestrol acetate
`
`twice daily and minimal effects on cortisol secretion were
`observed. Basal cortisol and AC TH levels were unaffected
`and cortisol
`levels increased appropriately after exog-
`enous syntlietic ACTH (cortrosyn) administration in all
`patients. Basal levels of aldosterone also remained stable
`following administration of all three drug doses. There
`were no changes
`in urinary or plasma sodium or
`potassium, nor in standing blood pressure to suggest a
`clinical
`state of aldosterone
`deficiency. However,
`cortrosyn- stimulated aldo sterone levels were significantly
`blunted at all three doses. (Santen 61 a1. 1991). Based on
`several phase 11 trials, toxicity attributed to this agent was
`mild and consisted mainly of nausea, anorexia, fatigue,
`and hot
`flashes. The potency of the compound,
`its
`relatively specific effects on aromatase and its lack of
`toxicity
`suggested that
`it might provide
`a major
`improvement over aniinoglutcthimidc for treatment of
`patients with breast cancer.
`Two large multicenter phase 111 trials in the USA
`comparing fadrozole hydro chloride to niegestrol acetate in
`patients who had received only tamoxifen as prior
`hormonal therapy have now been completed (Buzdar er a].
`19961), Trunet er a1. 1997). These two studies accrued a
`total of 672 patients. Final clinical results showed that
`there were no significant differences between the two
`treatment arms of the trials with respect
`to time to
`progression, objective response rates, response duration or
`
`overall survival. In these two trials, responses to mcgestrol
`acetate were somewhat lower than expected from previous
`studies with objective response rates of 11 and 13%
`respectively. Randomized patients receiving fadrozole
`experienced objective responses of 11 and 16% which did
`not differ significantly from those with megestrol. Stable
`disease for more than 6 months occurred in 25% of
`patients receiving fadrozole and 20% taking inegestrol
`acetate. Nausea was more frequent for fadrozole than
`megestrol acetate in both trials (22 vs 13% and 36% vs
`11% respectively). lncontrast, edema was commoner with
`megestrol acetate (21 vs 12% and 19 vs 12%) as was
`weight gain.
`compared fadrozole with tamoxifen
`Two trials
`(Falkson & Falkson 1996, Thurlimann et al. 1996). In the
`first, 1 mg fadrozole twice daily was compared with 20 mg
`tamoxifen daily in 212 postmenopausal patients with
`riietastatic breast cancer. Response rates to tamoxifen
`(27%) and to fadrozole (20%) did not differ significantly
`nor did response durations (20 months vs 15 months).
`However, tamoxifen achieved a significantly longer time
`to treatment failure (8.5 months vs 6 months, P<0.05). In
`the second study. fadrozole was compared with tamoxifen
`as first—line therapy in a randomized, controlled trial
`conducted in South Africa. Response rates to tamoxifen
`were 48% vs 43% with fadrozole (P=riot significant).
`However, response duration was significantly longer with
`
`81
`
`Astrazeneca Ex. 2047 p. 7
`
`

`
`Santen and Harvey.‘ Use of aromatase inhibitors in breast carcinoma
`
`tamoxifen (median duration not reached vs 343 days,
`P<0.009) as was overall
`survival
`(34 months
`for
`tamoxifen vs 26 months for fadrozole, I’<0.0-16).
`Taken together,
`these studies demonstrated that
`fadrozole may be inferior to tamoxifen in efficacy and no
`better tolerated than rnegestrol acetate. Based upon these
`findings,
`the second generation aromatase inhibitor,
`fadrozole, would likely find its place as third-line therapy.
`Fadrozole has been approved for
`the treatment of
`advanced breast cancer in postmenopausal women in
`Japan. This agent is not likely to be further developed in
`the United States since both anastrozole and letrozole
`appear to be more potent and more selective aroinatase
`inhibitors.
`
`Careful analysis of the fadrozole/inegestrol acetate
`trials raises the concern that responses to endocrine
`therapies appeared to be less frequent than observed in
`prior studies. For example, the randomised comparison of
`the first generation aromatase inhibitor, aminoglute—
`thimide, with surgical
`adrenalectomy demonstrated
`responses of 40-50% in patents previously treated with
`tamoxifen (Santen et ai. 1981). Other
`studies with
`megestrol acetate as second—line therapy demonstrated
`responses ranging from 30 to 50%. Several possibilities
`could explain the low response rates. In recent studies,
`more stringent criteria have been used than in previous
`trials. For example, reealcification of mixed lytic/blastic
`metastases were previously considered objective evidence
`of partial responses. Such lesions are now considered non-
`assessable, non-measurable disease. External review of
`cases probably also increases the stringency of assess-
`ment.
`lt should be noted that
`in a previous study
`comparing tamoxifen alone vs tamoxifen and fluoxy-
`mestrone, the objective response rate for tam oxifcn alone
`was only 10% (Swain et a1. 1988). These considerations
`lead to the conclusion that one can only compare new
`agents with established ones such as tamoxifen and assess
`the relative differences between them. It is inappropriate
`to compare the percent of objective responses to those
`observed in historical controls.
`
`4-Hydroxyandrostenedione (4-OHA)
`Formestane (Lentaron, 4-OHA‘, 4-l1ydroxyandrost-4-ene-
`3,17—dione) is a structural analog of androstenedione and
`is thus a highly specific aromatase inhibitor (Lonning
`1998).
`It was the first steroidal suicide—type (Type 1)
`aromatase inhibitor to enter clinical
`trials and is now
`commercially available in Europe. Using the in Vjtm
`placental aromatase assay system, 4-OHA was shown to
`be 60-fold more potent than aininoglutethimide (Ki=-1.1
`uM). Extensive studies revealed no estrogenic, anti-
`estrogenic, or antiandrogenic properties (Brodie & V1/ing
`1987). However, transformation to 4—hydroxytestosterone
`
`82
`
`occurs and androgenic effects can be demonstrated under
`certain circumstances (Brodie er a1. 1981).
`4-OHA (Lentaron) has been studied extensively in
`Europe in postmenopausal women with breast cancer.
`Data from four phase ll clinical
`trials of 4-OHA
`demonstrated a 33% objective regression rate of breast
`cancer in postmenopausal patients previously treated with
`multiple endocrine therapies. Toxicity included six
`patients with sterile abscesses due to intramuscular
`injections, two of sufficient severity to warrant discon-
`tinuation of therapy. No androgenic effects were observed
`(Goss eta]. 1986).
`Hoffken eta]. ( 1990) conducted a large trial of 4-OHA
`in postmenopausal women. Patients initially received 500
`mg intramuscularly every two weeks for 6 weeks and then
`250 rrrg every 2 weeks thereafter. Plasrrra estradiol levels
`fell from baseline values of 10-11 pg/ml to levels of
`approximately 4 pg/ml for up to 7 months of therapy. The
`drug appeared specific since no reduction of cortisol or
`symptoms of cortisol deficiency were observed. Of 86
`evaluable patients, there were 2 complete and 19 partial
`remissions (24%) and 26 with disease stabilization (30%).
`Side-effects included minor systemic symptoms in 11%
`(hot flashes, constipation, alopecia, and pruritns) and local
`symptoms in 8% (pruritus,
`local pain, and erytherna).
`These side-effects resulted in discontinuation oftherapy in
`only 2% of patients.