IPCS INCHEM Home


    INTERNATIONAL PROGRAMME ON CHEMICAL SAFETY

    WORLD HEALTH ORGANIZATION





    TOXICOLOGICAL EVALUATION OF CERTAIN
    VETERINARY DRUG RESIDUES IN FOOD



    WHO FOOD ADDITIVES SERIES: 43





    Prepared by the Fifty-second meeting of the Joint FAO/WHO
    Expert Committee on Food Additives (JECFA)





    World Health Organization, Geneva, 2000
    IPCS - International Programme on Chemical Safety


    PRODUCTION AIDS



    ESTRADIOL-17, PROGESTERONE, AND TESTOSTERONE

    First draft prepared by
    J. Leighton, S. Franceschi, G. Boorman, D.W. Gaylor, and J.G. McLean

    Estradiol-17
         Explanation
         Biological data
              Absorption, distribution, and elimination
              Biotransformation
                   Hydroxylation
                   Conjugation
              Biochemical parameters
                   Synthesis
                   Mechanism of action
         Toxicological studies
              Acute toxicity
              Short-term studies of toxicity
              Long-term studies of toxicity and carcinogenicity
              Genotoxicity
              Reproductive toxicity
              Special studies on mechanism of action
         Observations in humans 
              Therapeutic use 
              Estradiol-related genetic markers of carcinogenicity
    Progesterone 
         Explanation 
         Biological data 
              Absorption, distribution, and excretion 
              Biotransformation 
              Biochemical parameters 
                   Synthesis
                   Mechamism of action
         Toxicological studies
              Acute toxicity 
              Short-term studies of toxicity 
              Long-term studies of toxicity and carcinogencity
              Genotoxicity 
              Reproductive toxicity
         Observations in humans 
    Testosterone 
         Explanation 
         Biological data 
              Absorption, distribution, and elimination 
              Biotransformation
              Biochemical parameters 
                   Synthesis
                   Mechamism of action
         Toxicological studies
              Acute toxicity
              Short-term studies of toxicity 

              Long-term studies of toxicity and carcinogenicity 
              Genotoxicity 
              Reproductive toxicity 
         Observations in humans 
    Epidemiological studies of women exposed to postmenopausal estrogen
    therapy and hormonal contraceptives
         Methods
         Postmenopausal oestrogen therapy 
              Exposure 
              Human carcinogenicity 
                   Breast cancer 
                   Endometrial cancer
                   Cervical cancer
                   Ovarian cancer
                   Cancers of the liver and biliary tract
                   Colorectal cancer
                   Cutaneous malignant melanoma
                   Thyroid cancer
                   Summary and conclusions
                   Cardiovascular disease
                   Osteoporosis
                   Overall mortality
         Hormonal contraceptives
              Exposure
              Human carcinogenicity
                   Breast cancer
                   Endometrial cancer
                   Cervical cancer
                   Ovarian cancer
                   Cancers of the liver and biliary tract
                   Colorectal cancer
                   Cutaneous malignant melanoma
                   Thyroid cancer
                   Summary and conclusions
              Cardiovascular disease
                   Acute myocardial infarct
                   Stroke
                   Venous thromboembolism
                   Overall mortality
                   Meat intake and cancer risk
         Comments and evaluation
              Estradiol-17
              Progesterone
              Testosterone
         References

         The purpose of this monograph is to provide a review and summary
    of the scientific information relative to a toxicological assessment
    of the safety of three endogenous hormones, estradiol-17,
    progesterone, and testosterone, with emphasis on information published
    since the review of the Committee at its thirty-second meeting (Annex
    1, reference 80). The biology and toxicology of the compounds and

    metabolites formed endogenously and ingested orally are summarized. As
    the pharmacokinetics and pharmaco-dynamics of synthetic steroidal and
    nonsteroidal substances (e.g. diethyl-stilbestrol) differ
    substantially, only a limited discussion of the pharmacology of these
    compounds is presented. This review is not intended to be exhaustive
    but to highlight the scientific literature that may be relevant to use
    of the hormones from the point of view of food safety. The Committee
    at its thirty-second meeting did not prepare toxicological monographs
    on the natural hormones estradiol-17, progesterone, and testosterone.

    1.  ESTRADIOL-17b

    1.1  Explanation

         Estradiol benzoate (10-28 mg) or estradiol-17 (estradiol; 8-24
    mg) are administered to cattle as an ear-implant formulation to
    increase the rate of weight gain (i.e. growth promotion) and to
    improve feed efficiency. Estradiol valerate is administered by
    subcutaneous or intramuscular injection to synchronize estrus in
    cattle. Esters of estradiol are rapidly cleaved to estradiol  in vivo
    and are thus also considered to be endogenous substances, as the
    residues produced are structurally identical to the estradiol produced
    in animals and humans after hydrolysis. 

         Estradiol was reviewed previously by the Committee, at its
    thirty-second meeting (Annex 1, reference 80), when it concluded that
    the establishment of an acceptable residue level and an ADI was
    'unnecessary'. This conclusion was based on studies of the patterns of
    use of estradiol for growth promotion in cattle, the residues in
    animals, analytical methods, toxicological data from studies in
    laboratory animals, and clinical findings in human subjects. The
    Committee further concluded that estradiol residues resulting from its
    use for growth promotion in accordance with good husbandry practices
    were unlikely to be a hazard to humans. 

    1.2  Biological data

    1.2.1  Absorption, distribution, and excretion

         Estradiol is generally considered to be inactive when
    administered orally due to gastrointestinal and/or hepatic
    inactivation. In a study to monitor its oral availability and to
    identify the sites of metabolism, 14C-estradiol was infused into
    selected portions of the gastrointestinal tract of gilts, and blood
    samples were collected from the jugular and portal veins. The
    concentration of free estrogens in the jugular vein was low (< 1%) at
    all times after instillation of labelled estradiol, and it was
    detectable only briefly. The concentration of conjugated estrogens in
    the jugular vein peaked rapidly after instillation, particularly when
    instilled into the lower gut. Approximately 60-90% of the radiolabel
    in blood was present as glucuronide conjugates; smaller amounts of
    sulfated compounds were detected, and approximately 1% as

    diconjugates. The principal steroid identified after cleavage by
    b-glucuronidase and sulfatase was estrone. The authors concluded that
    conjugation occurs as estradiol crosses the mucosa of the
    gastrointestinal tract, and free estradiol in the portal plasma is
    conjugated during the first pass through the liver (Moore et al.,
    1982). In companion studies, the authors concluded that the limiting
    factor in absorption of conjugates was hydrolysis to free estrogen
    (Pohland et al., 1982) and that a possible dose-limiting rate of
    absorption was observed at the highest dose (4 mmol 3H-estradiol
    glucuronide) (Coppoc et al., 1982).

         Crystalline estradiol (10 mg in cocoa butter) was placed in the
    stomachs of prepubertal gilts that had been held without food for 26
    h, and blood samples were taken from the jugular and hepatic portal
    veins for hormone measurements. The concentrations of estradiol,
    estrone, estradiol glucuronide, and estrone sulfate in the hepatic
    portal vein rose within 5 min and remained elevated for several hours.
    Estradiol represented only 6% of the total estrogen measured during
    the sampling period, indicating extensive pre-hepatic metabolism of
    estradiol. In the periphery, the concentrations of estradiol
    glucuronide, estrone glucuronide, and estrone sulfate, but not those
    of estradiol or estrone, rose in the jugular vein, indicating that
    most of the estradiol and estrone had been removed by the liver.
    Infusion of bile containing estrogens into the duodenum resulted in
    peaks of estrogen glucuronide and estrone glucuronide in the hepatic
    portal and jugular veins within a few minutes, followed by a second
    rise 180 min later. The first peak did not occur in bile extracted
    with ether to remove free estradiol and estrone, and the second peak
    did not occur in gilts given oral antibiotics before bile infusion.
    The authors concluded that estrogens administered orally are
    conjugated by the gut wall and pass to the liver, where they enter
    either the bile pool for enterohepatic circulation or the bloodstream
    (Ruoff & Dziuk, 1994).

         Oral administration of 0.5 mg fine-particle estradiol in the
    early follicular phase of the menstrual cycle to six fasting, female
    volunteers resulted in a peak mean estradiol concentration of 211
    pg/ml 4 h after administration (mean basal estradiol concentration,
    138 pg/ml). The serum estrone concentrations also peaked at this time,
    when the peak:baseline ratio of estrone was greater than that of
    estradiol. Peaks were observed 4 h after dosing for estrone sulfate
    and 6 h after dosing for estradiol sulfate; the peak for estrone
    sulfate was always higher than that for estradiol sulfate. The
    predominance of estrone over estradiol in serum after oral
    administration of estradiol and comparison with serum concentrations
    reached after vaginal administration indicate extensive first-pass,
    probably intestinal, metabolism (Nahoul et al., 1993).

         The distribution of estradiol in female Wistar rats was measured
    in heart, liver, kidney, brain, and plasma by radioimmunoassay for 24
    h after intravenous administration of 0.1 mg/kg bw or after
    intragastric administration of 10 mg/kg bw. The concentration of
    estradiol in liver was 20 times higher after intragastric than after
    intravenous administration when equivalent plasma concentrations of
    hormone were evaluated. Negligible differences were seen in the
    estradiol concentrations of other tissues. The tissue concentrations
    of estradiol were higher than those in plasma at all times. The
    absolute bioavailability, as measured by comparison of the
    dose-corrected values for the area under the integrated
    concentration-time curve (AUC), was 8.3% after an intragastric dose of
    10 mg/kg bw. The total clearance was 154 ml/min per kg bw. The
    half-life of estradiol in liver was 2.6 h (Schleicher et al., 1998).
    The uptake of estradiol by adipose tissue, a reservoir for estrogens,
    was not investigated in this study.

         Fourteen young women received a single dose of 2, 4, or 8 mg
    estradiol orally or 0.3 mg intravenously. The 8-mg dose of estradiol
    resulted in a 70-78% reduction in the AUC relative to expected
    values for estradiol and for free and total estrone, suggesting
    incomplete absorption at this dose. The absolute bioavailability of
    the 4-mg dose was calculated to be 5%. The mean ratio of free
    estrone:estradiol was 1 after intravenous injection and and 20 after
    oral administration. In a two-comparment model, the AUC for young
    women given a 0.3-mg dose intravenously was 4000 pg-h/ml; total
    clearance was 22 ml/min per kg bw. Pharmacokinetic parameters showed
    high intraindividual and interindividual variation,which limits the
    therapeutic usefulness of oral preparations (Kuhnz et al., 1993).

         Circulating estradiol is bound to sex hormone-binding globulin
    (SHBG) and, to a lesser extent, serum albumin. Only 1-2% of
    circulating estradiol is unbound; 40% is bound to SHBG and the
    remainder to albumin (Carr, 1998). Plasma SHBG is secreted from the
    liver; a similar, non-secretory form is present in many tissues,
    including reproductive tissues and the brain. Adult rodent livers do
    not produce the secretory form of SHBG (Reventos et al., 1993). Some
    estrogen metabolites (2-methoxyestrone and 2-methoxy-estradiol) have
    higher binding affinities for SHBG than estradiol itself (Philip &
    Murphy, 1986), and other estrogens (estrone and estriol) do not bind
    to this serum protein in humans (Renoir et al., 1980). Estradiol binds
    to human SHBG with lower affinity than testosterone.

         The plasma concentrations of SHBG are regulated; they may be
    increased 5-10-fold by estrogens and decreased twofold by testosterone
    (Griffin & Wilson, 1998). Thus, a 20-fold higher concentration of
    total testosterone in men than in women results in a 40-fold
    difference in free testosterone (Grumbach & Styne, 1998). Unliganded
    plasma SHBG binds to either steroid or to SHBG-receptor; SHBG must
    first bind to the receptor and then the steroid in order to act: SHBG
    that is liganded to steroid cannot bind to the receptor (Hyrb et al.,
    1990). The SHBG-receptor complex present on the membranes of target

    tissues may be responsible for the interaction between the steroid
    hormone and cAMP pathways (Rosner, 1991). These observations provide a
    mechanistic explanation for the finding that some estrogenic effects
    are rapid (milliseconds) and are possibly mediated in a non-genomic
    manner. The intracellular form of the SHBG protein may sequester or
    direct hormone to the target tissue.

         Estrogens are eliminated in faeces and urine. The principal
    metabolites found in urine are polyhydroxylated forms conjugated at C3
    to glucuronic acid or sulfate. Elimination in bile is subject to
    enterohepatic circulation, and 20% of estrogens may be lost through
    faecal elimination. A high-fibre diet has been implicated in increased
    elimination of estrogens by this route, probably by decreasing gut
    transit time (Lewis et al., 1997). A high-fibre diet nonsignificantly
    lowered the serum estradiol AUC in human volunteers given an oral dose
    of estradiol glucuronide (Lewis et al., 1998).

         Urinary and faecal metabolites of estrogens in animals and humans
    have been studied for use as possible indicators of risk for
    hormone-dependent cancers or for infertility. Quantitative and
    qualitative differences between low-and high-risk populations and
    alterations in metabolite profiles due to diet have been reported
    (Michnovich & Bradlow, 1990; Aldercreutz et al., 1994; Ursin et al.,
    1997). There is at present no consensus about the importance of
    specific metabolites or metabolite ratios as prognostic factors, with
    the possible exception of estriol as a marker of the well-being of the
    feto-placental unit.

         The terminal plasma half-life of estradiol after intravenous
    adminis-tration to humans was 27 min; the volume of distribution was
    calculated to be 0.082 l/kg bw (White et al., 1998). Elsewhere, the
    plasma half-life of estradiol has been reported to be approximately 30
    min (Wingard et al., 1991).

    1.2.2  Biotransformation

         The major metabolites of estradiol, progesterone, and
    testosterone are shown in Figure 1.

    1.2.2.1  Hydroxylation

         Concern about the carcinogenicity of estrogens and, more
    recently, the possible genotoxicity of estrogen metabolites has
    sparked interest in establishing the pathways of estradiol metabolism,
    and extensive reviews have been published (IARC, 1979; Zhu & Conney,
    1998a). The two main competing, irreversible pathways for estradiol
    hydroxylation are 2-or 4-hydroxylation and 16 alpha-hydroxylation
    (Michnovicz et al., 1989), which have been implicated in both the
    pathophysiology and the protective characteristics of estrogens. Minor
    pathways of hydroxylation at other sites in the steroid metabolic
    pathway have also been identified (Zhu & Conney, 1998a).

         Hepatic hydroxylation of estradiol in humans and most other
    species leads primarily to the formation of 2-hydroxyestradiol or
    2-hydroxyestrone, with subsequent methylation; 4-hydroxy estrogens are
    also formed, although to a lesser extent. In the alternative pathway,
    the principal products are 16 alpha-hydroxyestrone and estriol, both of
    which are estrogen agonists. The pathway of estradiol metabolism in
    vitro was shown to be concentration-dependent; in hamster liver
    microsomes, 16alpha-hydroxylation predominates at low (< 25 mol/L)
    concentrations, whereas 16 alpha-and C2-hydroxylation contributed 
    equally to estradiol metabolism at higher concentrations (Butterworth 
    et al., 1996). Human forms of cytochrome P450s (CYP) which catalyse 
    the 2-or 4-hydroxylation of estradiol and estrone include CYP1A2 and, 
    to a lesser extent, CYP3A4 and CYP2C9 (Shou et al., 1997; Yamazaki et 
    al., 1998). Estradiol and estrone 16a-hydroxylation is catalysed by 
    CYP1A2 (estradiol) and CYP3A4 (estradiol and estrone). CYP1B1 
    catalyses the 4-hydroxylation of estrone and estradiol and may be the 
    dominant enzymatic pathway for estrogen metabolism in some 
    extrahepatic tissues, particularly steroidogenic tissues and their 
    respective targets (Larsen et al., 1998; Zhu & Conney, 1998a).

         While most estrogen metabolism occurs in the liver as
    2-hydroxylation, extrahepatic metabolism occurs as well. Conflicting
    reports have been published on the predominance of 2-and
    4-hydroxylation of estradiol in Syrian hamster kidney. The major route
    appears to be 2-hydroxy formation after catalysis by CYP1A1/2 and
    CYP3A expressed in this tissue (Hammond et al., 1997; Sarabia et al.,
    1997). Alternatively, 4-hydroxylation has been shown to predominate
    over 2-hydroxylation in the hamster kidney (Weisz et al., 1992). The
    4-hydroxy estradiol formed in this tissue is thought to be due to the
    lack of specificity of the responsible CYPs, as a specific estrogen
    4-hydroxylase (presumably CYP1B1) was not found in this tissue. CYP1B1
    protein was also not found in human renal adenocarcinoma cells (Spink
    et al., 1997). Rat pituitary, mouse, and human uterus and human
    mammary gland are other tissues that express high levels of estrogen
    4-hydroxylase (Liehr et al., 1995; Liehr & Ricci, 1996; Yager & Liehr,
    1996; Larsen et al., 1998).

         Significant differences in steroid metabolism are seen between
    rodents and humans (IARC, 1979). Sex-specific regulation of CYPs has
    been observed in rodent but not human liver, although sex differences
    in the metabolism of xenobiotics are found in humans (Kedderis &
    Mugford, 1998). Human but not mouse CYP1B1, recently identified as an
    estrogen 4-hydroxylase, metabolizes estradiol (Savas et al., 1997).

         Several mechanisms of CYP-mediated aromatic hydroxylation of
    estrogens (estradiol and estrone) have been proposed, including
    epoxide formation, direct oxygen insertion, and hydrogen abstraction.
    Hydroxylation by hydrogen abstraction, electron delocalization, and
    subsequent hydroxy radical addition has been proposed on the basis of
    electronic considerations of oxidation of estrone and substrates with
    additional aromaticity (2-napthol and equilenin) (Sarabia et al.,
    1997).

    FIGURE 3

         Hydroxyestrogens may be further modified by the action of
    catechol-O-methyltransferase (COMT). High activity of COMT is found in
    many tissues, including liver and kidneys, blood cells, endometrium,
    and breast. A genetic polymorphism for this enzyme results in a
    trimodal distribution of activity, but epidemological studies of the
    polymorphism in relation to breast cancer risk have yielded
    conflicting results (Lavigne et al., 1997; Millikan et al., 1998;
    Thompson et al., 1998). The methylation of catecholestrogens
    effectively prevents these compunds from entering the redox cycling
    pathway, and 2-methoxyestradiol may be an antitumour agent (Zhu &
    Conney, 1998a,b).

         Methylation of 4-hydroxyestradiol by COMT is inhibited by
    2-hydroxy-estradiol (Roy et al., 1990). Interestingly, tissues which
    develop estradiol-induced tumours (rat pituitary, male Syrian hamster
    kidney and mouse uterus) have very high concentrations of endogenous
    catecholamines (up to 50-fold relative to other strains or species and
    non-target tissues). Catecholamines in target tissues may inhibit or
    compete for COMT-catalysed methylation, thus leading to increased
    concentrations of hydroxylated metabolites of estradiol (Zhu and
    Conney, 1998a). 

    1.2.2.2  Conjugation

         Studies of the conjugation of estrogens with glucuronic acid or
    sulfate have been reviewed in detail (IARC, 1979). Lysosomes from male
    Syrian hamster livers and kidneys can catalyse the deconjugation of
    estradiol and estrone glucuronides. The rates of deconjugation of
    estrogen glucuronides were higher in kidney than in liver, by 56% for
    estrone and 34% for estradiol. Treatment of hamsters for nine days
    with subcutaneous implants containing 25 mg estradiol (releasing 61
    g/day) increased lysosomal estrone and estradiol 3-glucuronidase
    activity in kidney by 15 and 25%, respectively, and by about 100% in
    liver. Estradiol was deconjugated at negligible rates in both liver
    and kidney (Zhu et al., 1996). Human liver microsomal sulfatases
    convert estrone sulfate to estrone before 16 alpha-hydroxylation
    (Huang et al., 1998). Estrone sulfate, the most abundant estrogen in
    blood, and other estrogen conjugates may serve as a circulating
    reservoir of estradiol, and regulation of deconjugation reactions may
    affect intracellular estradiol concentrations.

         Demethylation of catechol estrogens has also been reported. The
    rates of demethylation of 2-and 4-methoxyestradiol were about equal in
    kidney microsomes, but the rate of 2-methoxyestradiol demethylation in
    liver was fivefold higher than that of 4-methoxyestradiol. Estradiol
    treatment decreased hepatic 2-methoxyestradiol demethylation by about
    20% relative to controls with little effect on 4-methoxyestradiol
    demethylation, whereas the opposite was observed in kidney (Zhu
    et al., 1996).

         In the absence of conjugation, a pathway for further catalysis of
    catechol estrogens has been suggested. Redox cycling of catechol
    (hydroxyquinone) to quinone through semiquinone intermediates is
    catalysed by oxidation of catechol estrogens by peroxidases or CYP1A1
    lipid hydroperoxide cofactors. Reduction of the quinone to the
    hydroquinone is catalysed by NADPH-dependent P450 reductase and other
    enzymes. Oxygen radicals formed in this redox process may increase the
    carbonyl content of proteins, formation of DNA 8-hydroxydeoxyguanosine
    adducts, and lipid peroxidation. The authors concluded that redox
    cycling is a critical step in estrogen-mediated carcinogenesis (Wang &
    Liehr, 1994; Yager & Liehr, 1996).

    1.2.3  Biochemical parameters

    1.2.3.1  Synthesis

         In mammals, estradiol, estrone, and estriol are synthesized from
    steroid precursors in the gonads, adrenal cortex, and placenta or
    through peripheral conversion of androgens in other tissues of
    mammals. Cholesterol, obtained primarily from circulating low-density
    lipoprotein, serves as the precursor for steroid biosynthesis,
    although steroidogenic cells are capable of local cholesterol
    synthesis  de novo. In non-pregnant premenopausal women, the
    principal estrogens found in the blood are estradiol and estrone.
    Estradiol is synthesized and secreted primarily from ovarian granulosa
    cells, whereas most estrone (approximately 40% of total estrogens) is
    formed peripherally from estradiol and androstenedione. Estradiol and
    estrone may be intra converted through the action of the enzyme
    17-hydroxysteroid dehydrogenase. Estrone may be further metabolized
    to estriol, primarily in the liver. Estriol is also formed in the
    fetal liver and in the placenta from
    16alpha-hydroxydehydroepiandrosterone sulfate, which is secreted from
    the fetal adrenal and circulating dehydroepiandrosterone sulfate. At
    least 90% of urinary estriol is derived from fetal sources (Carr,
    1992). In men and postmenopausal women, the source of serum estradiol
    is peripheral conversion of androgens by the enzyme aromatase. In men,
    approximately 0.3% of plasma testosterone is aromatized to estradiol;
    an additional contribution of approximately 25% of the total estradiol
    may be due to testicular secretion (Griffin & Wilson, 1998). Estrone
    is the predominant circulating estrogen in postmenopausal women,
    formed by peripheral conversion of adrenal androgens in adipose
    tissue.

         Gonadal synthesis of estradiol is regulated by luteinizing and
    follicle stimulating hormones secreted by the anterior pituitary
    gland. The secretion of these two hormones is regulated by
    gonadotropin-releasing hormone secreted by the hypothalamus, steroid
    hormones, and other factors in a complex feedback loop which
    effectively regulates the serum concentrations of hormone within a
    physiological concentration range, which is particularly variable in
    premenopausal women. The feedback loop is controlled by the dominant
    circulating hormone (estradiol and progesterone in women, testosterone

    in men). Feedback control for estradiol in men and for testosterone in
    women is therefore not operative (Wilson et al., 1998). There is
    evidence that this feedback loop exists in prepubertal children but is
    quiescent.

         In humans, plasma estradiol concentrations generally remain low
    during the first 12 years of life. Around the time of menarche, rising
    plasma concentrations of gonadotropins from the anterior pituitary
    stimulate the ovary to produce estradiol. During a normal menstrual
    cycle, plasma estradiol concentrations change very little throughout
    the first half of the follicular phase but increase as the follicles
    develop, reaching serum concentrations that are up to nine times
    greater than the basal concentrations near mid-cycle. After the
    mid-cycle surge, the estradiol concentrations fall precipitously.
    During the luteal phase, the serum estradiol concentrations rise to a
    plateau for 8-10 days, before declining. Should fertilization occur,
    the corpus luteum formed from the dominant follicle after ovulation
    remains active as the principal source of estradiol for the first 6-8
    weeks of pregnancy. The corpus luteum is later supplanted by the
    placenta as the site of estrogen synthesis. As the placenta lacks
    17 alpha-hydroxylase, fetal and maternal circulating androgens are
    necessary for placental estrogen synthesis. During pregnancy, the
    feto-placental unit secretes a large quantity of estriol into the
    maternal circulation, which is ultimately excreted in the urine.

         The concentrations of circulating estrogens, their daily
    production, and their metabolic clearance rates can be found in
    previous reviews (IARC, 1979, specifically 'General remarks on sex
    hormones') and in most textbooks of endocrinology (e.g. Braunstein,
    1994; Goldfien & Monroe, 1994; Carr, 1998; Griffin & Wilson, 1998; see
    also textbook appendices for summary tables). They are summarized in
    Table 1. Somewhat different values can be calculated for the basal
    production rate of estradiol in prepubertal boys and girls
    (Angsusingha et al., 1974; IARC, 1979) when measurements of hormone in
    urine or plasma are used as the basis for the calculation (4 or 12
    mg/day).

    1.2.3.2  Mechanism of action

         The conventional view of the action of steroid sex hormones
    involves interaction of the sex hormone with specific intracellular
    receptors, which subsequently bind tightly to specific DNA sequences
    in the genome (Malayer & Gorski, 1993). This tight nuclear binding
    initiates transcription of specific genes, which ultimately leads to
    physiological events. These include include development of
    reproductive tissues, maturation of the ovarian follicle, development
    of the uterus and vagina, and ductal development in the breast.
    Estrogen withdrawal results in menstruation. In non-reproductive
    tissues, estrogens may affect bone growth and prevent bone resorption,
    and effects on the plasma lipid profile through action in the liver.
    Estrogens typically promote cell growth or cell proliferation in
    responsive cells in culture.

        Table 1. Levels of circulating estrogens, metabolic clearance rates, and
    daily production
                                                                                     

    Sex          Age or phase         Serum              Metabolic       Total daily
                                      concentration      clearance       production
                                      (pg/ml)            (L/day)         (mg/day)
                                                                                     

    Male                                                 1400
                 Prepubertal          < 10                               < 0.014
                 12-16 years          < 23                               < 0.031
                 > 16 years           20-50                              0.027-0.068

    Female                                               1400
                 < 8 years            < 7                                < 0.01
                 2-12                 8-18                               0.01-0.024
                 12-14                16-34                              0.02-0.09
                 14-16                20-68                              0.03-0.09
                 Early follicular     20-100                             0.03-0.14
                 Preovulatory         100-350                            0.14-0.47
                 Luteal               100-350                            0.14-0.47
                 Late pregnancy       18 000                             24
                 Postmenopausal       10-30                              0.01-0.04
                                                                                     
    
         It has been proposed that the carcinogenicity of estrogens is
    distinct from its hormonal properties; however, since that hypothesis
    originated, information on estrogen-receptor and ligand interactions
    has been re-evualuated in the light of recent identification of
    additonal estrogen receptor proteins. Three specific receptors have
    now been identified for the endogenous ligand estradiol: the classical
    receptor ER alpha, ER, and ER2. Analyses of ER alpha and ER RNA
    indicate that ER alpha is widely distributed and that ER  is
    prominent in prostate, ovary (localized to the granulosa cells of the
    maturing ovary), epididymis, urinary bladder, uterus, lung, thymus,
    colon, small intestine, vessel walls, pituitary glan, hypothalamus,
    cerebellum, and brain cortex (Couse et al., 1997a; Kuiper et al.,
    1998). These receptor isoforms differ in their agonist and antagonist
    reactions to agents such as tamoxifen and to classes of agents
    variously referred to as 'endocrine disruptors' or 'xenoestrogens' and
    to endogenous estrogen metabolites such as 2-hydroxyestradiol.
    Moreover, alpha-estrogen receptor knockout (alpha ERKO; Couse et al.,
    1997a) and ER-/- female mice (Krege et al., 1998) have very different
    phenotypes. Alpha ERKO females lack the ability to complete
    folliculogenesis, are infertile, and have multiple cystic,
    haemorrhagic, and atretic follicles, whereas ER-/- female mice
    develop normally and are indistinguishable from their wild-type
    littermates. These mice are fertile, but their fertility is

    compromised, as demonstrated by reductions in litter size. Mammary
    gland development is normal in ER-/- mice, in contrast to the absence
    of breast development beyond that of prepubertal females in alpha ERKO
    mice. Male mice lacking ER are also fertile (alpha ERKO males are
    infertile) but develop prostate and bladder hyperplasia as they age.
    These ERKO mice and the receptor-binding properties of various ligands
    demonstrate that ER alpha and ER have different functional
    responsibilities.

         Abundant evidence exists that hormonal carcinogenicity is linked
    to the relative balance of various estrogens. Proliferation of mammary
    glands and other reproductive (i.e. target) tissues is inextricable
    linked to hormonal changes during the menstrual cycle, pregnancy, and
    the initiation or cessation of cyclic menstruation. The cyclic
    proliferation of tissues is correlated to the cellular content of the
    estrogen receptor. In cultured cells, the growth of cells with
    estrogen receptors (e.g. MCF-7 cells) but not those without estrogen
    receptors (e.g. MDA-MB-231 cells) is dependent on the estradiol
    concentration in serum. Molecular genetic studies of human cancer
    indicate that the progression from a normal to malignant phenotype
    requires activation of one or several oncogenes and/or inactivation of
    tumour suppressor genes. These events require cell division (reviewed
    by Bernstein & Ross, 1993; Russo & Russo, 1996; Tsai et al., 1998).

         A functional role for catechol estrogens distinct from that for
    estradiol has been suggested (reviewed by Zhu & Conney, 1998a,b).
    Because of the extremely short half-lives of these compounds
    (Schneider et al., 1984), they are unlikely candidates for circulating
    hormones. Locally, 2-hydroxyestradiol reportedly stimulates
    progesterone production in ovarian granulosa cells (Spicer et al.,
    1990), and 2-hydroxyestrone has been shown to inhibit MCF-7 cell
    proliferation in the presence of quinalizarin, a potent inhibitor of
    COMT. This challenge can be overcome with physiological concentrations
    of estradiol. No effects on cell growth were observed with
    concentrations of 10-9 to 10-6 mol/L of 2-hydroxyestradiol in the
    absence of methyltransferase inhibition (Schneider et al., 1984).

         The endogenous metabolite 2-methoxyestradiol has been suggested
    to be an antiangiogenic factor and tumour suppressor (Fotis et al.,
    1994; Zhu & Conney, 1998a); however, the concentrations required to
    induce apoptotis in cultured cells are about 10 times greater than
    those observed in humans (Yue et al., 1997). Those authors suggested
    that unless the concentration of 2-methoxyestradiol in the lipid phase
    (i.e. membranes) exceeds that in the aqueous phase, as reported for
    some lipophilic calcium blockers, the antiangiogenic properties of
    2-methoxyestradiol are of little physiological relevance.

         A role for catechol estrogens in implantation in the mouse uterus
    has been suggested on the basis of the observation of increased
    activity of 4-hydroxylase activity on day 4 of pregnancy (Paria et
    al., 1990). 2-and 4-Hydroxyestradiol bind to the estrogen receptor

    with reduced relative affinities of 23 and 26%, respectively (Schultze
    et al., 1994). Additionally, a distinct signaling pathway separate
    from the estrogen receptor may exist for 4-hydroxyestradiol. In
    alpha ERKO mice, 4-hydroxyestradiol stimulated up-regulation of
    lactoferrin mRNA and water imbibition (Das et al., 1997). In these
    mice, the effects of 4-hydroxyestradiol could not be mimicked by
    estradiol, nor could the effects be blocked by the anti-estrogen ICI
    182,780; however, it has been postulated that the estrogenic effects
    in the uterus in alpha ERKO mice are mediated through ER (Krege et
    al., 1998).

         The physiological effects of estradiol that are reported not to
    be receptor-mediated (i.e. not mediated via the classical receptor
    mechanism) include those on myometrial, neuronal, pituitary, maturing
    oocyte, and granulosa cells (Wehling, 1997). Estradiol may protect
    against oxidative damage in neuronal cells (Behl et al., 1995), but
    very high concentrations of estradiol were required for this effect:
    10-5 but not 10-7 mol/L was effective in preventing cell death. At
    physiological concentrations in blood, the antioxidant properties of
    estradiol protect against oxidation of low-density lipoprotein (Rifici
    & Kachadurian, 1992; Hoogerbrugge et al., 1998). Novel so-called
    'scavestrogens', structurally related to 17 alpha-estradiol, have
    antioxidant properties that protect against the radical-mediated death
    of cultured cells (Blum-Degen et al., 1998). Other studies indicate
    that the pro-oxidant and antioxidant properties of estrogens may be
    dependent on structure and concentration. Estradiol, estriol, and
    methoxyestrogen metabolites had only antioxidant properties. Catechol
    estrogens showed pro-oxidant properties at low concentrations (5
    pmol/L to 100 nmol/L), but antioxidant properties dominated at high
    concentrations (0.5-50 mol/L) (Markides et al., 1998). In addition to
    oxidant and antioxidant effects, a rapid, non-genomic effect of
    estradiol, possibly mediated by cAMP, Ca++, or ion channel gating,
    has been postulated (reviewed by Moss et al., 1997).

    1.3  Toxicological studies

    1.3.1  Acute toxicity

         The therapeutic dose of fine-particle estrogen given orally is    
    0.5-2 mg/day. No adverse effects were reported in children after
    accidental ingestion of large doses of estrogen-containing oral
    contraceptives (Physician's Desk Reference, 1999). An
    electroencephalogram of a young woman who took 160 mg of estradiol
    valerate (80 tablets of 2 mg), which is converted to estradiol-
     in vivo, showed traces typical of subcortical disturbance on the
    first day; however, the recording was normal one week later (Punnenon
    & Salmi, 1983).

    1.3.2  Short-term studies of toxicity

         Estradiol was administered in the diet to female Crl:CD BR rats
    at doses equal to 0. 0.003, 0.17, 0.69, or 4.1 mg/kg bw per day for 90
    days. The end-points were chosen to evaluate both short-term and
    reproductive toxicity and several mechanistic and biochemical
    parameters. Administration of doses > 0.17 produced dose-dependent
    increases in body weight, food consumption, and feed efficiency. At
    0.69 and 4.1 mg/kg bw per day, minimal to mild non-regenerative
    anaemia, lymphopenia, decreased serum cholesterol (at the high dose
    only), and altered splenic lymphocyte subtypes were observed. Changes
    in the weights of several organs were noted. Evidence of ovarian
    malfunction (reduced corpora lutea and large antral folicles) was
    found at doses > 0.17 mg/kg bw per day. Pathological changes in
    males and females fed 0.69 or 4.1 mg/kg bw per day included
    centrilobular hepatocellular hypertrophy, diffuse hyperplasia of the
    pituitary gland, feminization of the male mammary gland, mammary gland
    hyperplasia in females, cystic ovarian follicles, hypertrophy of the
    endometrium and endometrial glands in the uterus, degeneration of the
    seminiferous epithelium, and atrophy of the testes and acessory sex
    glands (Biegel et al., 1998b).

    1.3.3  Long-term studies of toxicity and carcinogenicity

         The toxicity of estrogens, including estradiol and related
    esters, has been reviewed extensively by working groups convened by
    the IARC (1979, 1987). After reviewing studies of estradiol
    administered orally to mice and by subcutaneous injection or
    implantation in mice, rats, hamsters, guinea-pigs, and monkeys, the
    group concluded that there is sufficient evidence for the
    carcinogenicity of estradiol in experimental animals, noting that:
    "Administration to mice increased the incidences of mammary,
    pituitary, uterine, cervical, vaginal, testicular, lymphoid and bone
    tumours. In rats, there was an increased incidence of mammary and/or
    pituitary tumours. estradiol-17b produced a statistically
    nonsignificant increase in the incidence of foci of altered
    hepatocytes and hepatic nodules induced by partial hepatectomy and
    administration of  N-nitrosodiethylamine in rats. In hamsters, a high
    incidence of malignant kidney tumours occurred in intact and castrated
    males and in ovariectomized females, but not in intact females. In
    guinea-pigs, diffuse fibromyomatous uterine and abdominal lesions were
    observed.' The IARC working group concluded that the carcinogenic
    effects of estrogens and progestogens were inextricably linked to the
    hormonal milieu and to dose-effect relationships (IARC, 1987).
    Hormonal effects on non-cancer end-points were not evaluated by the
    groups. 

         The induction of renal tumours by various steroidal and
    non-steroidal estrogens was examined in castrated male Syrian hamsters
    treated subcutaneously for nine months with a capsule that released an
    average of 110 g of the hormone per day. The tumour incidences

    associated with the hormonal activity of the substances tested are
    presented in Table 2. These data demonstrate a good correlation among
    the hormonal parameters progesterone receptor induction and serum
    prolactin and relative estrogenic potency (estrogen receptor binding)
    in hamster kidney. All animals trested with estrone, equilin plus
    d-equilin, or Premarin(R) developed renal tumours, with combined 
    numbers of tumour foci in both kidneys of 15, 18, and 16, respectively 
    (Li et al., 1995). 

         Castrated adult male Syrian hamsters were treated for eight
    months with subcutaneous pellets containing 20 mg estrogen or
    anti-estrogen; the release rates in g/day were as follows:
    diethylstilbestrol, 156; ethinyl-estradiol, 215; estradiol, 96;
    estradiol-17 alpha, 104; 17 alpha-ethinyl-11-methoxy-estradiol
    (moxestrol), 104; and tamoxifen, 183. Treatment with ethinylestradiol
    resulted in progressive dysplasia but no renal tumours, but dysplasia
    was observed in the proximal tubules of the renal cortex, which was
    uncommon in animals treated with diethylstilbestrol or estradiol.
    Animals treated with estradiol, diethylstilbestrol, or moxestrol had a
    tumour incidence of 100%, which was completely abolished by
    concurrrent treatment with ethinylestradiol (Table 3). Simultaneous
    administration of diethylstilbestrol and estradiol-17 or estradiol-17
    alpha (a noncarcinogenic estrogen) did not mitigate the
    carcinogenicity of diethylstilbestrol (Li et al., 1998).

         The carcinogenicity of estradiol and its metabolites was
    investigated in groups of 20-35 B6C3F1 mice of each sex given a
    single daily intraperitoneal injection of the compound on four
    consecutive days starting at 12 days of age. They were then maintained
    for approximately 18 months and killed. Of the catechol estrogens and
    their quinones tested, only estrone-3,4-quinone was significantly
    carcinogenic in the livers of male mice (Table 4). It was also highly
    toxic, as most of the mice died from unknown causes shortly after
    treatment. Estrone was protective against liver tumour formation in
    this system, and few tumours were induced in female mice (Cavalieri et
    al.,1997). 

         The tumour formation in the kidneys of male Syrian hamsters given
    25-mg pellets containing estrogen or catechol estrogen:cholesterol
    (90:10) by subcutaneous implantation and killed 175 days later was
    studied histologically. Estradiol and 4-hydroxyestradiol each induced
    renal tumours in four of five animals, whereas neither
    2-hydroxyestradiol nor 2-methoxystradiol was carcinogenic. The lack of
    carcinogenicity of 2-hydroxyestradiol was not due to failure of the
    hormone to stimulate cell growth  in vivo, as estradiol,
    4-hydroxyestradiol, and 2-hydroxyestradiol supported the growth of
    estrogen-dependent H-301 cells injected into male hamsters.
    2-Methoxyestradiol was not effective in stimulating these cells. The
    authors note that none of the three compounds was mutagenic in
     Salmonella typhimurium TA100 strain (see below; Liehr et al., 1986).

         The role of estrogen and its metabolites in tumour formation was
    examined in castrated male Syrian hamsters implanted for 9-10 months
    with pellets containing various estrogens (doses not given). The
    results are shown in Table 5. The authors suggested that the
    neoplastic changes seen in hamster kidney after continuous exposure to
    estrogens are due to synergistic action between hormonal and
    carcinogenic factors (Li & Li, 1989).

         Castrated male Syrian hamsters received subcutaneously implanted
    pellets containing diethylstilbestrol, alpha-dienestrol, hexestrol,
    diethylstilbestrol 3,4-oxide, estradiol, estrone, ethinylestradiol,
    equilin or (+)-equilenin, which released 100-210 g/day, for a total
    of nine months. The tumour incidence was 75-100%, except with
    ethinylestradiol (20%) and (+)-equilenin (0%). The ability of
    estrogens to cause renal tumours correlated well with their ability to
    compete for estrogen receptor binding, with the notable exception of
    ethinylestradiol (Li et al., 1983). The tumours induced by
    ethinylestradiol are of a different origin than other hormone-induced
    renal tumours (Oberley et al., 1991). The presence of estrogen
    receptors, probably ER alpha, in interstitial cells of control and
    estrogen-treated hamsters was confirmed by immuno-histochemical
    staining and northern blotting. Receptors were also found in renal
    corpuscles, arterial cells, and interstitial capillaries but not in
    the tubular epithelia of the cortex, further indicating that the
    tumours have an interstitial or mesenchymal origin (Bhat et al.,
    1993). The presence of ER, which was identified after the study, has
    not been investigated.

         Virgin female C3H/HeJ mice with a high titre of antibodies to the
    mouse mammary tumour virus were fed diets that provided doses
    equivalent to 0.015, 0.15, or 0.75 mg/kg bw per day estradiol from
    week 6 to week 110 of age (Highman et al., 1978). The microscopic
    findings in animals sacrificed after 52 weeks of feeding are shown in
    Table 6. In a continuation of this study, the authors reported the
    preneoplastic and neoplastic findings in mice sacrificed after up to
    104 weeks on the estrogenic diets (Highman et al., 1980). The results
    are shown in Tables 7 and 8. Uterine cervical adenosis may be a
    precursor of cervical adenocarcinoma in C3H/HeJ mice and may therefore
    serve as an early indicator of the uterine carcinogenicity of a
    compound. In these experiments, high doses of estradiol increased the
    incidence of adenosis but did not affect the incidence of ovarian
    tubular adenomas. After 66-91 weeks of treatment, high doses of
    estradiol also increased the incidence of mammary gland hyperplastic
    alveolar nodules but not mammary adenocarcinomas. The authors reported
    the occurrence of several other sporadic tumours at different sites in
    both control and treated animals. They concluded that the incidence of
    lesions in mice given estradiol was generally dose-dependent,
    indicating that this compound either induces or facilitates the
    development of these lesions (Highman et al., 1980)


        Table 2. Estrogenicity and carcinogenicity of various steroidal and stilbene estrogens in Syrian hamster kidney
                                                                                                                           

    Estrogen                        % of competitive      Induction of          Serum         % of animals    No. of tumour
                                    binding to estrogen   progesterone          prolactin     with tumours    foci in both
                                    receptors due to      receptors in kidney   (ng/ml)                       kidneys
                                    renal tumours         (fmol/mg protein)
                                                                                                                           

    Steroidal

    Estradiol-17                   55                    48                    390           100             17
    11-Methoxy ethinyloestradiol   52                    60                    330           100             22
    16 alpha-Hydroxyestrone         48                    45                    390           38              3
    11-Methoxyestradiol            30                    35                    390           25              2
    11-Methylestradiol             14                    18                    150           0               0
    Estradiol-17 alpha              34                    6                     130           0               0
    Deoxestrone                     14                    8                     94            0               0

    Stilbene

    Diethylstilbestrol              46                    50                    450           100             19
    Indenestrol B                   46                    49                    280           100             11
    Indanestrol                     10                    29                    100           0               0
                                                                                                                           

    From Li et al. (1995)
    

        Table 3. Prevention of the carcinogenicity of estrogens by ethinylestradiol
                                                                                       

    Estrogen                                   % tumour induction      No. of tumour 
                                                                       nodules in both
                                                                       kidneys
                                                                                       

    Diethylstilbestrol                         100                     15

    Estradiol-17                              100                     13

    Ethinylestradiol                           10                      2

    Moxestrol                                  100                     18

    Diethylstilbestrol + ethinylestradiol      0                       0

    17-Estradiol + ethinylestradiol           0                       0

    Moxestrol + ethinylestradiol               0                       0

    Diethylstilbestrol + diethylstilbestrol    100                     12

    Diethylstilbestrol + 17-estradiol         100                     12

    Diethylstilbestrol + 17 alpha-estradiol    100                     9
                                                                                       

    From Li et al. (1998) 
    
         Intact and ovariectomized or hysterectomized nulliparous female
    C3H/HeJ mice, five months of age, received either estradiol at a dose
    of 0.5 mg/L in drinking-water for one year or estradiol plus daily
    injections of 0.1 mg 2-bromo-alpha-ergocryptine, an effective
    suppressor of prolactin secretion. All mice were examined weekly for
    mammary tumours. One year after the onset of treatment, all surviving
    mice were sacrificed. The formation of mammary hyperplastic nodules
    was highly significantly suppressed in mice that received both
    estradiol and 2-bromo-alpha-ergocryptine, and the mammary tumour
    incidence was slightly but significantly reduced in comparison with
    that in animals receiving estradiol alone. The tumour incidence in
    concurrent controls was not reported. In a separate study, nulliparous
    30-day-old mice received estradiol in the drinking-water with or
    without a daily injection of 0.1 mg 2-bromo-alpha-ergocryptine for
    19-20 months. The effects on the mammary gland are shown in Table 9.
    The authors concluded that at least a portion of the oncogenic
    activity of estrogenic steroids on the mammary gland in rodents is
    manifested through a stimulatory effect on prolactin secretion
    (Welsch, 1976).

         The results of short-and long-term studies of the carcinogenicity
    of estrogens are summarized in Table 10.

        Table 4. Carcinogenicity of estradiol and its metabolites in male
    B6C3F1 mice
                                                                            

    Compound                     Dose             No. of mice with tumours/
                                 (mol/kg bw)     total no. of animals (%)
                                                                            

    Estradiol-17                30               6/20            (30)
    2-Hydroxy-17-estradiol      30               10/28           (36)
    4-Hydroxy-17-estradiol      30               10/24           (42)
    17-Estradiol-2,3-quinone    7.5              8/26            (31)
    17-Estradiol-3,4-quinone    7.5              4/21            (19)
    Estrone                      30               3/32            (9.4)
    2-Hydroxyestrone             30               9/30            (30)
    4-Hydroxyestrone             30               8/33            (24)
    Estrone-2,3-quinone          7.5              12/25           (48)
    Estrone-3,4-quinone          3.7              6/10            (60)
    Benzo[a]pyrene               60               12/12           (100)
    Solvent                                       7/19            (37)
    Untreated                                     11/33           (33)
                                                                            

    From Cavalieri et al. (1997) 
    
    1.3.4  Genotoxicity

         A working group convened by IARC evaluated the results of tests
    for genetic toxicity conducted with estradiol and concluded:
    "Oestradiol-17 did not induce chromosomal aberrations in bone-marrow
    cells of mice treated  in vivo. Unusual nucleotides were found in
    kidney DNA of treated hamsters. It induced micronuclei but not
    aneuploidy, chromosomal aberrations or sister chromatid exchanges in
    human cells  in vitro. In rodent cells  in vitro, it induced
    aneuploidy and unscheduled DNA synthesis but was not mutagenic and did
    not induce DNA strand breaks or sister chromatid exchanges.
    Oestradiol-17 was not mutagenic to bacteria." The group also
    concluded that the limited data available on estrone and estriol were
    indicative of genotoxicity (IARC, 1987).

         A mechanism has been proposed by which catechol estrogens
    interact with DNA. Nonmethylated catechol estrogens can be oxidized to
    a quinone which can bind to DNA; thus, 2-and 4-hydroxy estradiol
    produce 2,3-and 3,4-quinones, respectively. Reaction of the
    3,4-quinone of estrone or estradiol with deoxyguanosine (dG) at N7
    resulted in loss of the deoxyribose moiety and thus induced
    depurinating adducts. No adducts were observed after reaction of the
    3,4-quinone with deoxyadenosine (dA). Reaction of estrone-2,3-quinone
    with dG and dA produced a stable N2-dG or N6-dA adduct, the
    deoxyribose group remaining intact. Formation of depurinating and
    stable adducts in calf thymus DNA by activated catechol estrogens and
    in mammary glands of female Sprague-Dawley rats injected with 200 nmol
    4-hydroxy-estrone was confirmed by the 32P-postlabelling technique.
    The authors suggested that the formation of depurinating adducts via
    3,4-quinone followed by misreplication of unrepaired apurinic sites
    are the critical steps in initiation of cancer by estrogens (Cavalieri
    et al., 1997; Stack et al., 1998).

         Studies of the genetic toxicology of estrogens are summarized in
    Table 11.

         The mutagenicity of estradiol and its metabolites was assessed in
     Salmonella typhimurium strain TA100, but the authors concluded that
    none of the compounds was mutagenic in this assay (Liehr et al.,
    1986). Estradiol was evaluated in several short-term tests for
    genotoxic potentiol and at 1, 10, or 100 g/ml was found to cause
    chromosomal aberrations and sister chromatid exchange in cultured
    human lymphocytes with and without metabolic activation. In the
    absence of metabolic activation, the lowest dose caused aberrations
    after 72 h of treatment but not after 24 or 48 h; the intermediate
    dose caused aberrations after 48 and 72 h but not after 24 h of
    treatment. With a 6-h treatment, aberrations were observed at 10 and 
    100 g/ml in the presence of metabolic activation but not in its
    absence. Estradiol caused sister chromatid exchange at most doses with
    or without metabolic activation (Dhillon & Dhillon, 1995).

        Table 5. Carcinogenicity of estradiol and its metabolites in castrated
    male Syrian hamsters
                                                                         

    Compound                      No. of mice with tumours/
                                  total no. of animals (%)
                                                                         

    Estradiol-17                 6/6                      (100)
    Estrone                       8/10                     (80)
    Estriol                       4/7                      (57)
    2-Hydroxy-17-estradiol       0/6                      (0)
    2-Hydroxyestrone              0/6                      (0)
    4-Hydroxy-17-estradiol       5/5                      (100)
    4-Hydroxyestrone              2/6                      (33)
    Ethinylestradiol              3/15                     (20)
    Equilin                       6/8                      (75)
    (+)-Equilenin                 0/9                      (0)
                                                                         

    From Li & Li (1989)
    

        Table 6. Percent incidence of pathological changes in female mice given estradiol for 52 weeks
                                                                                                                    

    Dose           No. of             Effects in the                Effects in the             Effects in the 
    (mg/kg bw      animals            cervix                        uterine horn               mammary gland 
    per day)                                                                                                  
                                 Adenosis in   Adenosis in     Glandular     Hyperplastic    Adeno-      Osseus
                                 upper third   upper and       hyperplasia   alveolar        carcinoma   hyperplasia
                                               lower thirds                  nodules
                                                                                                                    

    0              47            11             0              26            0               4            0
    0.015          35            17             0              24            0               0            0
    0.15           36            15             0              62            3               6           18
    0.75           48            38            36              96            9               8           82
                                                                                                                    

    From Highman et al. (1977)

    Table 7. Incidences (and%) of uterine cervical adenosis and ovarian tubular 
    adenomas in mice fed estradiol
                                                                                                        

    Dose           Week
    (mg/kg bw                                                                                           
    per day)       26            52            78                           104
                                                                                                
                   Adenosis      Adenosis      Adenosis      Tubular        Adenosis        Tubular
                                                             adenoma                        adenoma
                                                                                                        

    0              2/37 (5)      5/46 (11)     1/14 (7)      2/14 (14)      13/19 (16)      12/24 (50)
    0.015          1/37 (3)      5/29 (17)     1/5 (20)      0/5 (0)        3/24 (12)       11/25 (44)
    0.15           5/45 (110)    5/35 (14)     3/14 (21)     2/16 (12)      8/20 (40)       12/20 (60)
    0.75           25/42 (60)    33/45 (73)    4/7 (57)      0/7 (0)        3/6 (50)        3/7 (43)
                                                                                                        
    

        Table 8. Incidences (and%) of pathological changes in the mammary gland of female mice fed estradiol
                                                                                                                                  

    Dose         Hyperplastic alveolar nodules                             Adenocarcinomas
    (mg/kg bw                                                                                                                   
    per day)     Weeks 0-39   Weeks 40-65   Weeks 66-91    Weeks 92-105    Weeks 0-39   Weeks 40-65    Weeks 66-91    Weeks 92-105
                                                                                                                                  

    0            0/91 (0)     0/57 (0)      3/29 (10)      6/50 (12)       4/91 (4)     15/57 (26)     13/29 (45)     19/50 (38)
    0.015        0/89 (0)     0/63 (0)      0/19 (0)       4/31 (13)       0/89 (0)     28/63 (44)     8/19 (42)      13/31 (42)
    0.15         0/94 (0)     2/56 (4)      8/29 (28)      7/21 (33)       4/94 (4)     21/56 (37)     14/24 (58)     6/21 (29)
    0.75         0/93 (0)     5/78 (6)      5/19 (26)      6/17 (35)       5/93 (5)     34/78 (44)     11/19 (58)     8/17 (47)
                                                                                                                                  

    From Highman et al. (1980)

    Table 9. Effects of treatment with estradiol with or without 2-bromo-alpha-ergocryptine on the number of mammary hyperplastic
    nodules and mammary tumours in young nulliparous C3H/HeJ mice
                                                                                                                                

    Treatment                               No. of mice at    No. of mice at    No. of hyperplastic nodules     No. of mice with
                                            start of study    end of study      in inguinal mammary glands      mammary tumours
                                                                                                                                

    Controls                                100               16                3.1                             11
    Estradiol                               100               12                4.8                             27
    Estradiol + 2-bromo-alpha-ergocryptine  100               28                2.8                             9
                                                                                                                                
    
        Table 10. Summary of results of short-term and long-term studies
    of the carcinogenicity of estrogens
                                                                                                                            

    Species               Dose                     Findings                                             Reference
                                                                                                                            

    Short-term study

    Rats                  0.003-4.12 mg/kg bw      Histopathological changes, particularly at           Biegel et al. (1998b)
                          per day for 90 days      intermediate and high doses

    Long-term studies

    Mice                                           Increased incidences of mammary, pituitary,          IARC (1979)
                                                   uterine, vaginal, testicular, lymphoid, and 
                                                   bone tumours

    Rats                                           Mammary and pituitary tumours; statistically         IARC (1979)
                                                   nonsignificant increase in incidence of foci 
                                                   of altered hepatocytes and hepatic nodules 
                                                   after partial hepatotectomy and adminsitration
                                                   of N-nitrosodiethylamine

    Hamsters                                       Malignant renal tumours in intact and                IARC (1979)
                                                   castrated males and in ovariectomized 
                                                   but not intact females

    Castrated male        111 g/day               100% renal tumour incidence                          Li et al. (1995)
    Syrian hamsters, 
    9 months

    Castrated male        96 g/day                100% incidence of renal tumours; modulated by        Li et al. (1998)
    Syrian hamsters,                               ethinylestradiol
    8 months

    B6C3F1 mice           4 daily                  Estrone was protective; estradiol-17 did not        Cavalieri et al. 
                          intraperitoneal          increase incidence over control                      (1997)
                          injections 
                          (30 mol/kg bw)

    Table 10. (continued)
                                                                                                                            

    Species               Dose                     Findings                                             Reference
                                                                                                                            
    Male Syrian           25 mg subcutaneously,    Estradiol-17 and 4-hydroxy-17-estradiol            Liehr et al. (1986)
    hamsters              175 days                 produced tumours, but 2-hydroxy-17-estradiol 
                                                   did not

    Castrated male        Not reported;            Estradiol-17 and 4-hydroxy-17-estradiol            Li & Li (1989)
    Syrian hamsters       9-10 months              produced tumours, but 2-hydroxy-17-estradiol 
                                                   did not

    Castrated male        100-200 g/day           Ethinylestradiol produced a lower incidence of       Li et al. (1983)
    Syrian hamsters                                tumours than estradiol-17

    Virgin C3H/HeJ        0.015-0.75 mg/kg bw      Estradiol-17 caused a dose-dependent increase       Highman et al.
    mice                  per day in feed          in tumour incidence                                  (1977, 1980)

    C3H/HeJ mice          0.5 mg/l in              Estradiol-17 caused tumours                         Welsch (1976)
                          drinking-water
                                                                                                                            
    

         When Swiss albino mice were given a single intraperitoneal
    injection of 100, 1000, or 10 000 g/kg bw, the highest dose increased
    the number of micronucleated polychromatic erythrocytes and the
    frequency of sister chromatid exchange, although the
    polychromatic:normochromatic erythrocyte ratio did not appear to be
    affected. Estradiol did not cause reverse mutation in  Salmonella
    strains TA100, TA1535, TA97a or TA98, at concentrations of 1-10 000
    g/plate in the absence of metabolic activation and 1-1000 g/plate in
    its presence. In a host-mediated assay in which mice were given 100,
    1000, or 10 000 g/kg bw followed 2 h later by injection of
     S. typhimurium, no change in the number of His+ revertants per
    plate was observed (Dhillon & Dhillon, 1995). 

         Estradiol was evaluated in five tests for the induction of
    micronuclei in bone marrow  in vivo in female rats given three daily
    subcutaneous doses of 20 g/kg bw and in mice given a single
    intraperitoneal injection of 10-150 mg/kg bw. The authors concluded
    that estradiol was not genotoxic to the bone marrow of rodents (Ashby
    et al., 1997).

         In male B6C3F1 mice and male Fischer 344 rats that received
    estradiol at 310, 620, or 1250 mg/kg bw in three injections, there was
    no increase in the frequency of polychromatic erythrocytes. In male
    and female B6C3F1 mice treated with various numbers of injections,
    solvents, routes of administration, and killing schedules, no
    significant increase in the frequency of micronuclei was observed
    (Shelby et al., 1997).

         The onset of genomic rearrangements was tested at 10-5 mol/L
    estradiol in two X-ray-transformed cell lines (X-ray-9 and F-17a) and
    two untransformed cell lines (10T1/2b and 10T1/2c). Genomic
    rearrangements (deletions or additions in minisatellites) were
    observed in the transformed but not the untransformed lines. No new
    rearrangements were observed after withdrawal of estradiol (Paquette,
    1996). 

         The ability of estradiol to induce morphological transformation,
    gene mutations, chromosomal aberrations, sister chromotid exchange,
    unscheduled DNA synthesis and other chromosomal changes was assessed
    in Syrian hamster embryo cells. Cell growth was completely inhibited
    at 10-30 g/ml but was not affected at a concentration < 3 g/ml.
    Treatment of cells with 0.3-6 g/ml did not affect their
    colony-forming efficiency, but at 10 g/ml colony formation was 53%
    that of controls. Incubation with estradiol at 0.3-10 g/ml for 48 h
    induced a dose-dependent increase in the frequency of morphological
    changes. Estradiol in this dose range also induced numerical changes
    (chromosome gains and loses). The majority of these cells (94%) were
    diploid. Estradiol had no other effects in this assay (Tsutsui et al.,
    1987). No exogenous metabolic activation was used in these
    experiments.


        Table 11. Genetic toxicity of estrogens
                                                                                                                       

    End-point                   Test system                  Dose                    Result             Reference
                                                                                                                       

    In vitro

    Reverse mutation            S. typhimurium TA100         50-1500 g/plate        Negative           Liehr et al. 
                                                                                                        (1986)

    Reverse mutation            S. typhimurium TA100,        1-10 000 g/plate -S9   Negative           Dhillon & Dhillon
                                TA1535, TA97a, TA98          1-1000 g/plate +S9     Negative           (1995)

    Chromosomal aberration,     Cultured human               1-100 g/ml             Positive           Dhillon & Dhillon
    sister chromatid            lymphocytes                                                             (1995)
    exchange

    Micronucleus formation      Human cells                                          Positive           IARC (1987)

    Aneuploidy, chromosomal     Human cells                                          Negative           IARC (1987)
    aberrations, sister 
    chromatid exchange

    Aneuploidy, unscheduled     Rodent cells                                         Positive           IARC (1987)
    DNA synthesis

    Mutagenicity, DNA damage,   Rodent cells                                         Negative           IARC (1987)
    sister chromatid
    exchange

    Cell transformation,        Syrian hamster               0-10 g/ml              Positive           Tsutsui et al. 
    numerical chromosomal       embryo cells (-S9)                                                      (1987)
    changes

    Gene mutation,              Syrian hamster               0-10 g/ml              Negative           Tsutsui et al.
    chromosomal aberration,     embryo cells (-S9)                                                      (1987)
    sister chromatid 
    exchange, unscheduled 
    DNA synthesis

    Table 11. (continued)
                                                                                                                       

    End-point                   Test system                  Dose                    Result             Reference
                                                                                                                       

    Numerical chromosomal       Cultured human               0.05-75 mol/l          Positive           Schuler et al.(1998)
    changes                     lymphocytes                                                             

    Chromosomal breakage        Cultured human               0.05-75 mol/l          Negative           Schuler et al.(1998)
                                lymphocytes

    DNA damage                  pBR322 (-S9)                 0.01-0.1 mmol/l         Single-strand      Yoshie & Ohshima 
                                                                                     breaks with 2-     (1998)
                                                                                     and 4-hydroxy 
                                                                                     estradiol and
                                                                                     estrone; negative
                                                                                     with estradiol
                                                                                     and estrone

    Microtubule disruption      Chinese hamster              0-100 mol/l            EC50, 10 mol/l    Aizu-Yokota et al.
                                V79 cells                                                               (1995)

    Adduct formation

                                Syrian hamster embryo cells  1 g/ml (-S9)           Increase with      Hayashi et al. 
                                                                                     estradiol and      (1996)
                                                                                     2- and 4-hydroxy   
                                                                                     estradiol

                                Syrian hamsters              2-150 mg/kg bw          Increase in        Han & Liehr 
                                                             intraperitoneally       kidney at 50       (1994a)
                                                                                     mg/kg bw; no
                                                                                     increase in liver  

                                Male Syrian hamsters         50 mg/kg bw             Increase in        Han & Liehr 
                                                             intraperitoneally       kidney; no         (1994a)
                                                                                     time dependence    

    Table 11. (continued)
                                                                                                                       

    End-point                   Test system                  Dose                    Result             Reference
                                                                                                                       

                                Male Syrian hamsters         100 mg/kg bw            Increase in liver  Han & Liehr 
                                                             intraperitoneally       1-2 h after        (1994a)
                                                                                     dosing but not 
                                                                                     later

                                Male Syrian hamsters         25 mg                   Increase in        Han & Liehr 
                                                             subcutaneous implant    kidney on day 3    (1994a)
                                                                                     but not day 6; 
                                                                                     no hepatic 
                                                                                     adducts;
                                                                                     substantial
                                                                                     differences in
                                                                                     adduct levels in
                                                                                     controls between   
                                                                                     days 3 and 6


                                Male Syrian hamsters         100 mg/animal per       Negative for       Han & Liehr 
                                                             day intraperitoneally   kidney with        (1994a)
                                                             for 3 days              estradiol and 
                                                                                     2- and 4-hydroxy
                                                                                     estradiol

                                NBL rats                     Dose not reported;      Unidentified       Han et al. 
                                                             serum level 14 times    adduct after 16    (1995)
                                                             that of control         but not 8 weeks
                                                                                     of treatment

                                Mongrel dogs                 Dose not reported       Decrease in        Winter & Liehr 
                                                                                     prostate adduct    (1996)
                                                                                     level; increase
                                                                                     in carbonyl
                                                                                     content

    Table 11. (continued)
                                                                                                                       

    End-point                   Test system                  Dose                    Result             Reference
                                                                                                                       

                                Human liver                  2 mmol/l                Negative           Seraj et al. 
                                                                                                        (1996)

                                Rat liver                    2 mg/kg bw per day      Adducts in male    Feser et al. (1996)
                                                             by gavage               but not for        female liver
                                                                                     14 days

                                Human breast tissue                                  Positive           Musarrat et al. 
                                                                                     correlation        (1996)

                                Human breast tissue                                  No correlation     Nagashima et al. 
                                                                                                        (1995)

    In vivo

    Micronucleus formation,     Mouse bone marrow            100-10 000 g/kg        Positive at        Dhillon & Dhillon
    sister chromatid exchange                                bw as single            highest dose       (1995)
                                                             intraperitoneal
                                                             injection

    Micronucleus formation      Rat bone marrow              20 g/kg bw as          Negative           Ashby et al. 
                                                             three daily                                (1997)
                                                             subcutaneous
                                                             injections


    Micronucleus formation      Mouse bone marrow            10-10 mg/kg bw          Negative           Ashby et al. 
                                                             as single                                  (1997)
                                                             intraperitoneal
                                                             injection

    Micronucleus formation      Mouse and rat bone           0.1-10 mg/kg bw         Negative           Shelby et al. 
                                marrow                       intraperitoneally                          (1997)

    Table 11. (continued)
                                                                                                                       

    End-point                   Test system                  Dose                    Result             Reference
                                                                                                                       

    Frequency of                Male and female mice         310-1250 mg/kg bw       Negative           Shelby et al. 
    polychromatic                                            as three                                   (1997)
    erythrocytes                                             injections                                 

                                Male Syrian hamster          5-150 mg/kg bw          Negative           Han & Liehr 
                                                             intraperitoneally                          (1994a)

    DNA damage                  Male Syrian hamster kidney   25 mg subcutaneously    Single-strand      Han & Liehr 
                                and liver                    every two weeks         breaks in kidney   (1994a)
                                                             but not liver

    DNA damage                  Male Syrian hamster          250 g/animal per       Single-strand      Han & Liehr 
                                                             day for 7 days by       breaks with        (1994a)
                                                             infusion                estradiol and 
                                                                                     4-hydroxy but 
                                                                                     not 2-hydroxy

    Chromosomal aberration      Male Syrian hamster          20 mg via               Positive in        Banerjee et al.
                                                             subcutaneous capsule    kidney             (1994)

    DNA damage                  NBL rat                      Subcutaneous capsules,  Single-strand      Ho & Roy (1994)
                                                             16 weeks; dose          breaks in 
                                                             not reported            prostate with 
                                                                                     estradiol +
                                                                                     testosterone
    

         The effect of estradiol at 0.05-75 mol/L was examined in
    cultured human lymphocytes by multicolour fluorescence in-situ
    hybridization. DNA probes for the centromere and adjacent
    heterochromatin regions of chromosomes 1, 9, and 16 were used to
    detect hyperdiploidy, polyploidy, and chromosomal breakage. Nonlinear
    increases in hyperdiploidy but no chromosomal breakage was observed.
    The authors concluded that induction of numerical changes in
    chromosomes by estradiol followed a sublinear dose-response
    relationship, probably with a threshold concentration. Binding of
    estradiol to microtubules or saturation of detoxification mechanisms
    are possible explanations for the observation (Schuler et al., 1998).

         Male Syrian hamsters were treated with intraperitoneal injections
    of 5, 15, or 150 mg/kg bw estradiol; subcutaneous implants containing
    25 mg estradiol for two weeks; or continuous infusion of estradiol or
    2-or 4-hydroxy-estradiol at 250 mg/animal per day for seven days. A
    single intraperitoneal injection of estradiol had no effect on DNA
    single-strand breaks in liver or kidney DNA, but the subcutaneous
    implants increased the number of renal single-strand breaks by 10%; no
    effect was seen in liver. Infusion of estradiol or 4-hydroxyestradiol,
    but not 2-hydroxyestradiol, for one week resulted in a 9% increase in
    the number of single-strand breaks relative to untreated controls. The
    authors suggested that estrogen-induced carcinogenesis is mediated by
    free-radical damage (Han & Liehr, 1994a).

         Male Syrian hamsters received capsules containing 20 mg
    diethyl-stilbestrol, estradiol, moxestrol, 17 alpha-estradiol, or
    -dienestrol, and between 94 (dienestrol) and 140 (diethylstilbestrol)
    g/day were obsorbed daily from the pellet. Animals were sacrificed at
    0.5, 1, 2, 3, 4, or 5 months (diethylstil-bestrol) or at 5 months (all
    other treatments). Chromosomal aberrations but not exchanges in
    hamster kidney DNA were cumulative with continued exposure to
    diethylstilbestrol. The kidneys of estradiol-and moxestrol-treated
    animals had chromosomal aberrations at frequencies similar to those
    seen with diethylstilbestrol, whereas the frequency of chromosomal
    aberrations in animals treated with the weaker estrogens were similar
    to those of controls. The authors suggested that estrogen-induced
    chromosomal aberrations are involved in tumorigenesis but that the
    process does not involve metabolic activation, since moxestrol, which
    is poorly metabolized, did induce chromosomal aberrations (Banerjee et
    al., 1994).

         2-Catechol estradiol and 4-catechol estrone at 0.01-0.1 mmol/L
    induced strand breaks in the pBR322 plasmid, and the level was greatly
    enhanced by a nitric oxide-releasing compound. The strand breaks could
    be inhibited by antioxidants such as  N-acetylcysteine and ascorbate
    and by superoxide dismutase. Estradiol, estrone,  O-methylcatechol
    estrogens, and diethylstilbestrol did not induce strand breaks. The
    authors suggest that NO mediates conversion of catechol estrogens to
    quinones, and the oxygen radicals produced by the quinone/hydroquinone
    redox system react with NO to form peroxynitrite, which causes strand
    breaks (Yoshie & Ohshima, 1998).

         Natural estrogens and their derivatives were tested for the
    ability to induce microtubule disruption in Chinese hamster V79 cells
    (which lack cytochrome P450 enzymes) at concentrations of 1-100
    mol/L. The EC50 values were 10 mol/L for estradiol and 9 mol/L for
    17 alpha-estradiol. The most potent disrupting agent tested was
    2-methoxyestradiol (EC50, 2 mol/L). Preincubation of cells with 1
    mol/L taxol for 2 h protected them against microtubule disruption by
    estradiol at doses up to 50 mol/L. The authors concluded that some
    natural estrogens cause microtubule disruption in a nongenomic manner
    (Aizu-Yokota et al., 1995).

         An increase in the frequency of DNA strand breakage and
    accumulation of lipid peroxidation products in the dorsolateral but
    not the ventral prostate were seen in four NBL rats treated
    subcutaneously with testosterone plus estradiol, relative to control
    rats. Treatment of castrated rats with testosterone resulted in a
    slightly lower rate of strand breaks than in untreated controls. The
    authors concluded that estradiol was responsible for the single-strand
    breaks in these animals (Ho & Roy, 1994).

    Studies of adducts

         Male Syrian hamsters were injected intraperitoneally with 2, 10,
    50, or 150 mg/kg bw estradiol and sacrificed 4 h later; with 50 mg/kg
    bw and sacrificed 1-8 h later; or with 100 mg/kg bw and sacrificed 1-8
    h later. Their livers and kidneys were examined for
    8-hydroxy-2'-deoxyguanosine (8-OHdG) as a marker of hydroxy radical
    interaction with DNA. Four hours after dosing with 50 mg/kg bw, the
    renal 8-OHdG levels were double those of controls; adducts were not
    determined in kidneys from animals treated at 150 mg/kg bw. No
    dose-dependence was observed. The levels of hepatic DNA adducts in
    treated animals were similar to those in controls. In hamsters treated
    with 50 mg/kg estradiol and killed 1-8 h later, the level of renal DNA
    adducts was greater than that in controls at 4 h but not at 1, 2, or 8
    h after dosing; hepatic DNA adducts were not determined. In animals
    injected with 100 mg/kg estradiol, the number of hepatic DNA adducts
    was increased 1 and 2 h after dosing. Treatment of hamsters with
    subcutaneous implants containing 25 mg estradiol for three or six days
    increased the renal levels of 8-OHdG by 50% over that in controls by
    day 3, but the levels were no different from those of controls in
    animals implanted with estradiol for six days. No effect was observed
    on the background level of liver DNA adducts at either time. Treatment
    of hamsters for three days by intraperitoneal injection with 100
    g/animal per day estradiol or 2-or 4-hydroxyestradiol also had no
    effect on renal DNA 8-OHdG levels. The authors concluded that the
    mechanism of the carcinogenic action of estrogen occurs through
    generation of free radicals via redox cycling of catechol estrogen
    metabolites (Han & Liehr, 1994b). Substantial differences were
    observed in the levels of adducts in untreated animals after three and
    six days in the implant experiment. While no statistical analysis was
    performed, the differences were sufficient to indicated a substantial

    variation in the background level of adducts. The lack of dose-and
    time-dependence of adduct formation in these experiments is
    inconsistent with the hypothesis that estradiol is a genotoxic
    carcinogen.

         An unidentified adduct was observed in DNA isolated from the
    dorsolateral prostate of NBL rats treated by subcutaneous
    administration of estradiol and testosterone for 16 weeks but not 8
    weeks. The circulating estradiol concentrations were increased
    approximately 14 times over those of controls, while normal plasma
    testosterone concentrations were maintained. The appearance of the
    adduct correlated with the appearance of dysplastic lesions (Han
    et al., 1995).

         Incubation of Syrian hamster embryo cells with 1 g/ml estradiol
    or 2-or 4-hydroxy estradiol for 24 h induced DNA adduct formation in
    parallel with cell transformation. The level of DNA adduct formation
    was greatest with 4-hydroxy estradiol and then with 2-hydroxy
    estradiol and estradiol. No exogenous metabolic activation was used in
    these experiments. In later experiments, diethylstilbestrol increased
    adduct formation in the absence but not in the presence of metabolic
    activation (Hayashi et al., 1996).

         Mongrel dogs were treated with capsules containing 5
    alpha-dihydro-testosterone (DHT) and/or estradiol for 60 days. The
    capsules were of uniform length (7 cm), but the quantity of hormone
    used was not described. Blood sampoles were obtained for the
    measurement of hormone. The 8-OHdG adduct levels in DNA from prostate
    were reduced in all dogs that received DHT, whereas treatment with
    estradiol or DHT plus estradiol had no effect. Free radical-induced
    damage (carbonyl content) of proteins was observed in prostate tissue,
    and the authors concluded that the damage was consistent with injury
    by estrogen metabolites followed by DHT-stimulated growth of altered
    prostatic cells (Winter & Liehr, 1996).

         Sterol-initiated DNA adduct formation was examined in vitro by
    32P-postlabelling. After exposure of human liver DNA to 2 mmol/L
    steroid, several steroids but not estradiol, estrone, or estriol
    formed DNA adducts. The presence of a carbonyl group at C17 (which
    estradiol lacks) was strongly associated with DNA binding (Seraj et
    al., 1996).

         When three male and three female Han:Wistar rats given estradiol
    at 2 mg/kg bw per day intragastrically as an aqueous microcrystalline
    suspension for 14 days, an estrogen-specific DNA adduct was observed
    by 32P-postlabelling in the livers of male but not female rats (Feser
    et al., 1996).

         To assess the hypothesis that estrogen-induced adduct formation
    is related to estrogen-induced tumorigenesis in humans, DNA from
    normal human breast tissue, benign tumours, and malignant tissue with
    invasive ductal carcinoma was examined for the presence of 8-OHdG
    adducts by a novel solid-phase immunoslot-blot assay with
    adduct-specific antibodies. The amounts of 8-OHdG found in the three
    tissues were 0.25, 0.98, and 2.4 pmol/g DNA, respectively; 13 times
    more endogenous formation of 8-OHdG was observed in MCF-7 cells which
    undergo hormone-dependent cell growth and have estrogen receptor s
    than in normal cultured human mammary epithelial cells, but no
    difference in adduct levels was observed between normal cells and
    MDA-MB 231 cells, which undergo receptor-independent growth and lack
    estrogen receptors. 8-OHdG adduct levels also correlated to the
    estrogen receptor status of the tissue, with higher adduct levels in
    malignant tissue with estrogen receptors than in those without. Age
    and smoking status did not correlate to the 8-OHdG content of DNA. The
    authors concluded that accumulation of 8-OhdG adducts in DNA is
    predictive of the risk for breast cancer and may be a major
    contributor to the development of breast neoplasia (Musarrat 
    et al., 1996).

         No difference in 8-OhdG adduct levels was found by
    high-performance liquid chromatography-electrochemical detection in
    breast cancer tissue and adjacent non-cancerous tissue, and no
    correlation was found with expression of estrogen or progesterone
    receptors or with clinical stage or histological grade (Nagashima et
    al., 1995).

    1.3.5  Reproductive toxicity

         The embryotoxicity and teratogenicity of estradiol were reviewed
    by a working group convened by IARC (1979), which concluded that
    "Oestradiol-17 has teratogenic actions on the genital tract and
    possibly on other organs and impairs fertility."

         Estradiol was administered in the feed to female Crl:CD BR rats
    at doses equal to 0, 0.003, 0.17, 0.69, or 4.1 mg/kg bw per day in a
    90-day, one-generation study. As no pups were born to dams at the two
    highest doses, only three dose groups of the F1 generation were
    assessed. The mean daily intakes of the F1 females were 0, 0.005 and
    0.27 mg/kg bw per day, respectively. The F0 rats were 49 days of age
    on test day 0, and serum hormones were evaluated after 7, 28, and 90
    days of feeding; they were evaluated on postnatal day 98 for the F1
    generation. In the F0 generation, estradiol at doses > 0.17 mg/kg bw
    per day produced a dose-dependent increase in serum estradiol
    concentration, and all doses produced a dose-dependent decrease in
    serum progesterone concentration on test day 90, which correlated with
    ovarian atrophy and lack of corpora lutea. The serum concentration of
    luteinizing hormone was decreased at all times at > 0.69 mg/kg bw
    per day and at 0.17 mg/kg bw per day on test day 90. Little change was
    observed in the serum concentrations of follicle-stimulating hormone.

    The serum concentration of prolactin was increased at 4.1 mg/kg bw per
    day on test day 90. In the F1 generation on postnatal day 28, the
    serum estradiol concentration was increased and that of progesterone
    decreased at 0.27 mg/kg bw per day. No change in serum prolactin,
    follicle-stimulating hormone, or luteinizing hormone concentration was
    noted. Dietary estradiol caused marked effects on the estrus cycle at
    0.17 mg/kg bw per day (F0) and 0.27 mg/kg bw per day (F1) and at
    0.69 and 4.1 mg/kg bw per day (F0 generation) (Biegel et al., 1998b).

         Information on the pups in this study was presented elsewhere.
    The groups at the two highest doses produced no pups, and the weights
    of the pups in the two remaining groups decreased relative to that of
    controls; the weights of pups of the F0 generation at 0.003 mg/kg bw
    per day (0.005 mg/kg bw per day for the F1 generation) recovered
    after birth and remained similar to those of controls throughout the
    study. The mean length of gestation in this dose group was
    statistically nonsignificantly decreased, which the authors suggested
    contributed to the decreased birth weight. The body weights of animals
    at 0.27 mg/kg bw per day remained below control levels throughout the
    study. Parenteral administration of estradiol did not affect the
    anogenital distance in male or female pups. Onset of sexual maturity,
    as measured by prepubertal separation in males and vaginal opening in
    females, was delayed. Some of the histopathological findings were more
    severe in the F1 generation than in the parent generation. The
    authors concluded that additional studies were needed to define the
    dose-response curve more accurately (Biegel et al., 1998a).

         The average litter size of transgenic 'knockout' female mice
    deficient in steroid 5 alpha-reductase type I (SRD5 alpha 1-/-,
    wild-type C57Bl/6J/129Sv) is reduced in comparison with wild-type
    controls (2.7 vs. 8 pups, respectively). In reductase-deficient
    animals, the maternal serum estrogen concentrations were chronically
    increased by two-to threefold relative to control animals. In control
    animals, spikes in DHT and to a much smaller extent testosterone
    concentrations occurred in maternal plasma on day 9 of gestation. In
    the 5a-reductase-deficient animals, the androgen peaks at day 9 were
    reversed. Oogenesis, fertilization, implantation, and placental
    morphology appeared normal in reductase-deficient animals. Fetal loss
    occurred between gestation days 10.75 and 11, commensurate with a
    surge in placental androgen production. Minimal fetal loss was
    observed on gestation day 10.5. To test the hypothesis that steroid
    hormones contribute to fetal loss in reductase-deficient animals,
    pregnant animals were treated with pellets containing various amounts
    of steroid hormone. Table 12 shows the effect on mean embryo survival.
    Bleeding was observed grossly in the uteri of control and experimental
    animals. Administration of estrogen receptor antagonists or inhibitors
    of aromatase prevented the excess fetal loss observed in the
    reductase-deficient mice. Testosterone was mildly protective against
    fetal loss in the knockout mice. The authors concluded that
    5a-reductase guards against the toxic effects of estrogen during
    pregnancy. They also noted that the human placenta, unlike the rodent

    placenta, has high levels of aromatase, resulting in very high
    concentrations of estrogens in the amniotic fluid. They speculated
    that the human fetoplacental unit has developed a mechanism to protect
    itself against estrogens (Mahendroo et al., 1997).

         Pregnant female CF-1 mice were treated subcutaneously on
    gestation day 13 with Silastic capsules containing 0, 25, 100, 200, or
    300 g estradiol. Male fetuses positioned  in utero between a male and
    female (MF males) were examined for treatment-related prostatic
    effects. In some experiments, MF males were obtained from pregnant
    dams killed on gestation day 19 and reared by foster dams. At the age
    of seven months, MF males were castrated, implanted subcutaneously
    with capsules containing 500 g testosterone, and killed three weeks
    later. The total concentration of estradiol in serum was increased in
    a dose-dependent manner in treated MF fetuses collected on day 18,
    with 94, 150, 230, 360, and 530 pg/ml in the animals at the five
    doses, respectively. A 40% increase in the number of prostatic
    glandular epithelial buds was found in MF males from dams treated with
    25 g estradiol, relative to controls. An increase in prostate size
    was also noted, but the size of the individual buds was not changed;
    prostate weight was increased in MF males at the low dose sacrificed
    at eight months but was decreased at the high dose, resulting in an
    inverted-U dose-response curve. The authors concluded that increased
    fetal serum estrogen concentrations affect androgen regulation of
    prostate differentiation, resulting in a permanent increase in the
    number of prostatic androgen receptors and in prostate size (vom Saal
    et al., 1997). The doses used in this study were below the NOEL. 

         Under conditions associated with reduced estrogen synthesis in
    humans (aromatase deficiency, placental sulfatase deficiency, fetal
    anencephaly), estrogen production and concentrations may be reduced by
    80-90%. However, progesterone production and fetal development remain
    normal, indicating that considerably more estrogen is produced during
    normal pregnancy than is necessary (Fisher, 1998).

         The embryotoxicity of steroidal and nonsteroidal estrogens was
    examined in cultured whole embryos obtained from Sprague-Dawley rats.
    Preliminary experiments resulted in steep dose-response curves for all
    estrogens examined at doses ranging from 0.05 to 0.5 mmol/L. For
    example, diethylstilbestrol had no effect at concentrations < 0.15
    mmol/L but was lethal to 100% of cultured embryos at doses > 0.25
    mmol/L. The concentrations tested resulted in low embryolethality
    (2-20%). Estrogens had dysmorpho-genic effects at concentrations of
    0.1-0.2 mmol/L. The commonest effect observed was hypoplasia of the
    prosencephalon. Estradiol and estrone were markedly and statistically
    significantly more toxic in the presence of metabolic activation (from
    livers of pregnant and non-pregant females and Aroclor 1254-treated
    adult male rats), but metabolic activation attenuated the embryotoxic
    effects of ethinylestradiol, tamoxifen, and erythrohexestrol and had
    no effect on other estrogens. In this system, estradiol was more
    efficiently converted to catechol estrogens in male liver, but


        Table 12. Effects of steroid hormones in 21-day release pellets on embryo survival
                                                                                                

    Pellet             Dose      Control females                  Srd5a1-/- females
    treatment                                                                                  

                                 No.        Live       Dead       No.        Live         Dead
                                 litters    embryos    embryos    litters    embryos      embryos
                                                                                                

    None                         4          8.0        0.4        8          3.2          5.1
    Placebo                      3          9.6        0.67       3          4.3          5.0
    Androstenedione    0.5       4          3.3        4.5        5          1.0          7.8
    Testosterone       0.5       2          8.8        1.5        5          6.2          3.2
    Estradiol          0.5       2          0          7.0        5          0            8.4
                       0.08      5          0          7.4        6          0            7.8
                       0.02      2          0          5.0        6          0.3          6.3
                       0.01      2          0          11         6          0            6.8
                       0.005     2          0          8.5        5          0            5.0
                       0.0025    2          4.5        4.0        4          2.8          5.5
                                                                                                
    

    ethinylestradiol was converted to catechol estrogens approximately
    three times more effectively than estradiol when metabolic activation
    systems from pregnant and non-pregnant animals was used. The authors
    concluded that the effects observed are independent of steroid
    structure or estrogen activity and are strongly dependent on the
    pathways and rates of biotransformation of some (but not all) of the
    parent compounds (Beyer et al., 1989).

         Ten mg of estradiol or testosterone were implanted subcutaneously
    into groups of five and seven female Sprague-Dawley rats on day 10 of
    pregnancy, whereas control pregnant rats were given dextran by the
    same method. Implantation with estradiol or testosterone resulted in
    complete resorption of embryos in all treated animals (Sarkar et al.,
    1986).

         A review of the birth certificates and hospital records of 7723
    infants whose mothers had reported using oral contraceptives indicated
    that these compounds present no major teratogenic hazard (Rothman &
    Louik, 1978).

         Studies of reproductive toxicity are summarized in Table 13.

    1.3.6  Special studies of mechanisms of action

         Estrogen-induced tumorigenesis has been the subject of two lines
    of investigation: receptor-based effects and redox cycling and DNA
    adduct formation leading to genetic damage. During the past decade,
    considerable attention has been focused on understanding the molecular
    basis of hormone receptor biology. Recently, transgenic animals that
    overexpress or lack estrogen receptors (Couse et al., 1997b) or
    aromatase (Fisher et al., 1998) have been developed. The role of
    estrogens and other hormones in mammary neoplasia in rodents and their
    relevance to human risk has been reviewed (Russo & Russo, 1996), and
    it was noted that rodent models mimic some but not all of the complex
    external and endogenous factors involved in initiation, promotion, and
    progression of carcinogenesis. Tumour type and incidence are
    influenced by the age, reproductive history, and the endocrine milieu
    of the host at the time of exposure. The spontaneous incidence of
    tumours differs in different strains of rats and mice. In rats, most
    spontaneously developed neoplasias, with the exception of leukaemia,
    are of endocrine organs or organs under endocrine control. Russo &
    Russo (1996) concluded that mechanism-based toxicology is not yet
    sufficient for human risk assessment, and the approach should be
    coupled to and validated by traditional long-term bioassays.

         The estrogen-responsive male Syrian hamster kidney model has been
    widely used to study the carcinogenicity of estrogens in vivo.
    Separation of carcinogenic from hormonal effects in male and
    ovariectomized female Syrian hamster kidney has been reviewed (Yager &
    Liehr, 1996). In hamsters treated chronically with relatively high
    doses by subcutaneous implantation, certain potent synthetic estrogens
    such as ethinylestradiol result in < 10% tumour incidence in kidney,
    whereas treatment with other estrogens result in renal tumour

    development in nearly all animals. Ethinylestradiol also acts at
    different sites from other estrogens in the hamster kidney. Thus, the
    estrogenicity of a compound is essential but not sufficient for renal
    carcinogenesis in hamsters. Other factors that have been suggested to
    contribute to carcinogenicity include cell-type-specific uptake and
    differential estrogen metabolism (such as high 4-hydroxylation rates)
    leading to estrogen-induced damage to cell macro-molecules (DNA
    and protein). Similarly, the progesterone metabolite
    5 alpha-pregnane-3,20-dione successfully competes with progesterone
    for receptor binding and biological effectiveness in some tissues but
    not others (Tsai et al., 1998). Since a second estrogen receptor
    isoform (ERb) has recently been identified, the results described
    below should be interpreted with caution.

         The estrogen metabolite 4-hydroxyestradiol, but not
    2-hydroxyestradiol, was tumorigenic in male hamster kidneys (Yager &
    Liehr, 1996), the proposed mechanism of action being redox cycling
    resulting in oxygen radical formation and subsequent damage to cell
    macromolecules. It is not certain that this pathway is relevant
     in vivo at physiological concentrations of estradiol. For example,
    micromolar concentrations of estradiol are necessary to cause
    microtubular disruption in Chinese hamster V79 cells, and these are
    greatly in excess of the picomolar to nanomolar concentrations
    normally found in serum. At higher concentrations, the lipophilicity
    of estradiol and some metabolites (such as methoxy derivatives) and
    their ability to intercalate into DNA and lipid membranes may be more
    important from a toxicological perspective than the estrogenic
    properties.

         The hormonal (i.e. receptor-mediated) and carcinogenic (i.e.
    genotoxic) properties of synthetic hormones were differentiated by
    measuring the rates of catechol estrogen and methyl ester formation by
    a weak carcinogen, 17a-ethinylestradiol, and by a strong hormonal
    carcinogen, moxestrol. The rates of hydroxylation in comparison with
    that for estradiol were 40-50% for moxestrol and 25-35% for
    ethinylestradiol, the differences being apparent at longer reaction
    times (i.e. 20 min but not 10 min). 2-Hydroxymoxestrol was a poor
    substrate for COMT, proceeding at a rate of about 3% of the
    methylation of 2-hydroxyestradiol. In hamster kidney cytosolic protein
    extracts, 10 nmol/L progesterone decreased binding of 2 nmol/L
    3H-progesterone by 78%, and 10 nmol/L ethinylestradiol inhibited it
    by 35%. Interpretation of this result is complicated, as the assay was
    performed with insufficient excess progesterone. Estradiol and
    moxestrol had no effect. The authors suggested that the decreased
    capacity of ethinylestradiol to form catechol metabolites and its
    progestogen antagonist activity contribute to the low tumour incidence
    seen with this compound (Zhu et al., 1993).

        Table 13. Studies of the reproductive toxicity of estrogens
                                                                                           

    Species             Dose                      Findings                   Reference
                                                                                           

    Mice, rats          0.1-35 mg/day             Teratogenic                IARC (1979)
                        subcutaneously

    Female rats         0.003-4.1 mg/kg bw        No NOEL identified         Biegel et al. 
                        per day orally for                                   (1998b)
                        90 days

    Transgenic mice     Estrogen                  No NOEL; effects           Mahendroo et
                        concentrations            observed between days      al. (1997)
                        reduced two-              10.75 and 11
                        to threefold

    Mice                0-300 g/animal           No NOEL for effects on     vom Saal et
                                                  fetal prostate             al. (1997)

    Cultured whole      0.5-0.5 mmol/l            Dysmorphogenic effects     Beyer et al. 
    embryos of                                    at 0.1-0,2 mmol/L          (1989)
    Sprague-Dawley
    rats

    Sprague-Dawley      10-mg implant             Embryo resorption          Sarkar et al.
    rats                                                                     (1986)

    Humans              Acidental exposure        No effect reported         Rothman & Louk
                                                                             (1978)
                                                                                           
    
         Several steroidal estrogens were tested at doses of 0.1-100
    nmol/L for their ability to increase proliferation of primary renal
    proximal tubular cells in culture. Most of the estrogens tested
    (including 4-hydroxyestradiol and estrone and, to a lesser extent,
    2-hydroxyestradio) increased cell proliferation at a concentration of
    0.1-10 nmol/L but inhibited it at 100 nmol/L. The authors concluded
    that the ability to induce cell proliferation is a more accurate
    predictor of carcinogenicity in this system than estrogen-responsive
    end-points or the amount of catechol metabolites generated (Li et al.,
    1995).

         To determine the role of estrogens in tubular renal damage and
    the subsequent reparative cell proliferation, castrated adult male
    Syrian hamsters were given subcutaneous pellets that released hormones
    at the following rates (g/day): diethylstilbestrol, 145; estradiol,
    134, estrone, 104, ethinylestradiol, 154; tamoxifen, 141;
    progesterone, 147; and DHT, 121. Diethylstilbestrol was administered
    for one to nine months, while the other compounds were administered
    for five months. The severity of tubular damage increased with
    progressive estrogen treatment, with a prominent rise in the number of
    secondary and tertiary lysosomes. The concentration of cathepsin D was
    increased in estrogen-treated kidneys (by approximately 2.7-and
    3.5-fold at four and five months, respectively) and paralleled the
    rise in estrogen receptor content. Progesterone and DHT alone had no
    effect, and concomitant treatment of animals with estrogen and either
    tamoxifen or DHT mitigated the estrogenic effects. The primary form of
    cathepsin D found in the kidneys of control and estrogen-treated
    animals was the 52-kDa isoform, considered to be the inactive form of
    the protein. The 31-and 27-kDa isoforms, believed to be the active
    forms, were found in significant amounts only in the kidneys of
    estrogen-treated animals, primary renal tumours, and their metastases.
    The authors suggested that cathepsin D mediates renal tubular damage
    as a first step in reparative cell proliferation (Li et al., 1997).

         Estradiol- or diethylstilbestrol-induced growth of cultured
    proximal renal tubular cells could be inhibited by ethinylestradiol.
    Expression of estrogen-responsive protooncogene (c- myc, c- fos, and
    c- jun) RNA and protein in kidneys was reduced in animals treated
    concomitantly relative to that found in animals treated with estradiol
    or diethylstilbestrol. The authors concluded that ethinylestradiol
    interferes with estrogen receptor-mediated mitogenic pathways,
    preventing gene dysregulation and tumour development. This effect does
    not appear to be due to differential binding to estrogen receptors by
    estrogenic substances (Li et al., 1998).

         Other hormones, notably progesterone, testosterone, and
    deoxycortico-sterone, and the antiestrogen tamoxifen prevent or
    inhibit the growth of estrogen-induced renal tumours in Syrian
    hamsters (reviewed by Yager & Liehr, 1996). Progesterone and tamoxifen

    exert a protective effect on mammary carcinogenesis (Inoh et al.,
    1985). A review of clinical data indicated that adjuvant progestogen
    therapy for treatment of patients with metastatic renal-cell carcinoma
    is not effective, indicating that carcinogenesis in the Syrian hamster
    model is not representative of human renal carcinogenesis (Linehan et
    al., 1997).

         Rodent tissues that form estrogen-induced tumours have high
    concentrations of the caetcholamine noradrenaline. In a study to test
    the hypothesis that hydrogen peroxide formed by monoamine oxidase
    deamination of catecholamines provides a source of free radicals, in
    addition to that postulated to be provided by metabolic redox cycling
    of catechol estrogen intermediates, Syrian hamsters and Sprague-Dawley
    rats received 25 mg estradiol in a subcutaneous capsule for two weeks.
    Treatment increased monoamine oxidase activity in hamster kidney but
    not liver and had no effect on monoamine oxidase activity in rat liver
    or kidney. The induction of hamster kidney monoamine oxidase activity
    could be prevented by tamoxifen. The authors concluded that
    receptor-mediated induction of monoamine oxidase, which deaminates
    catecholamines, may increase production of hydrogen peroxide and
    hydroxyl radicals, thus contributing to tumour initiation (Sarabia &
    Liehr, 1998).

         Male Syrian hamsters were treated with quercetin, an inhibitor of
    COMT, in order to assess the potentiating effects of this compound on
    renal tumorigenesis. All six animals treated with subcutaneous pellets
    that released estradiol at 61 g/day developed kidney tumours, but no
    tumours were seen in hamsters treated with quercetin at 0.3 or 3% in
    the diet for 5.7 or 6.5 months, respectively. Concomitant
    administration of estradiol and quercetin increased the number of
    large tumours and the incidence of metastases over that seen with
    hormone treatment alone. Quercetin inhibited 2 and 4-catechol estrogen
    methylation by 34 and 22%, respectively. The rates of redox cycling in
    liver and kidney were not affected by treatment with quercetin or
    estradiol (Zhu & Liehr, 1994).

         Male Syrian hamsters received subcutaneously implanted pellets
    containing 25 mg estradiol (which were replaced every three months)
    for seven months. Dietary supplementation with 1% vitamin C decreased
    estrogen-induced renal carcinogenesis by 50% in a small number of male
    Syrian hamsters. In related experiments, the effect of estradiol
    and/or vitamin C was examined on the renal activity of the detoxifying
    enzymes quinone reductase, catalase, superoxide dismutase, glutathione
    peroxidase, glutathione reductase, glucose-6-phosphate dehydrogenase,
    and gamma-glutamyl transpeptidase. Glutathione peroxidase activity was
    increased in the kidneys of hamsters treated with estradiol for 1
    month (141% of control). Quinone reductase activity was reduced in the
    kidneys of estradiol-treated animals (18% of control), but the
    activity was partially restored by dietary supplementation with
    vitamin C for one month (32% of control); in liver, concomitant
    treatment with vitamin C and estradiol reduced the activity of this
    enzyme (6% of control), and estradiol treatment alone caused a smaller

    decrease in activity (68% of control). Differences in catalase
    activities were observed after one month but not by seven months of
    treatment. Vitamin C had no effect on the intensity or specificity of
    estrogen-related kidney DNA adducts. The authors concluded that the
    enzymatic changes observed in estradiol-and/or vitamin C-treated
    animals were insufficient to account for the differences in renal
    tumour incidence. The authors concluded that vitamin C inhibits
    estrogen-induced carcinogenesis by reducing the concentration of
    estrogen quinone metabolites (Liehr et al., 1989).

         COMT is present in the epithelial cells of the proximal
    convoluted tubules of the kidney, predominantly in the juxtamedullary
    region, where estrogen-induced tumours arise. Treatment of male Syrian
    hamsters with estradiol or ethinylestradiol for two or four weeks
    altered the intensity, distribution, and subcellular location of
    immunoreactivity to COMT. Staining for this enzyme in control animals
    was largely of the soluble cytoplasm and nuclear membrane-bound forms,
    whereas staining for the soluble form of nuclear COMT was observed in
    estrogen-treated animals. No differences were observed between the two
    estrogens or between animals treated with estrogen for two or four
    weeks; no difference in nuclear location was observed between treated
    and control animals. Estradiol-induced renal tumours did not stain for
    COMT, and the nuclear signal present in human cells was lacking in
    hamster kidney. The authors suggested that a change in the subcellular
    distribution of COMT is a protective response to catechol estrogen
    metabolic damage to the genome (Weiss et al., 1998).

         Male Noble rats were treated with subcutaneous Silastic implants
    containing testosterone and/or estradiol or diethylstilbestrol for 16
    weeks. In estradiol-treated animals, the plasma testosterone
    concentration, determined after three weeks of treatment, was
    decreased more than 10-fold (from 4.8 ng/ml to < 0.3 ng/ml), whereas
    the estradiol concentration was increased 4.5-fold (from 16 pg/ml to
    75 pg/ml). Diethylstilbestrol but not estradiol caused a statistically
    significant decrease in body weight, and both hormones decreased the
    relative weight of the dorsolateral and ventral prostate and seminal
    vesicles plus coagulating glands. The body weights of animals treated
    for 16 weeks with testosterone and estradiol were significantly lower
    than those of controls, the relative weights of the dorsolateral and
    ventral prostate and seminal vesicles plus coagulating glands were
    increased, and the testicular weight was decreased approximately
    twofold. Multifocal epithelial dysplasia and marked inflammatory
    changes were observed in the lateral prostate. No changes were seen in
    the morphology of the ventral prostate seminal vesicle, coagulating
    gland (anterior prostate), or ampullary gland. Implants of
    testosterone plus diethylstilbestrol induced widespread dysplasia in
    the ventral prostate and lesser or no ventral prostatic dysplasia. In
    explant cultures, animals treated with testosterone plus
    diethylstilbestrol or testosterone plus estradiol showed a reduced
    ability to convert the 5 alpha,3-hydroxysteroid derivative of 3H-DHT
    to the more polar 6 alpha-and 7 alpha-hydroxylated derivatives,
    resulting in accumulation of 3-androstanediol. These metabolic

    changes resulted in a threefold (testosterone plus estradiol) or
    eightfold (testosterone plus diethylstilbestrol) increase in
    accumulation of 3-androstanediol in the dysplastic ventral prostate;
    no accumulation was observed in the explanted dorsolateral prostate.
    In animals treated with testosterone plus diethylstilbestrol, the
    ratio of estrone:estradiol was reversed in the ventral prostate,
    whereas in animals treated with testosterone plus estradiol, estradiol
    metabolism was decreased in the dysplastic dorsolateral prostate but
    not in the ventral prostate. The authors concluded that differences
    between target tissues in the bioavailability of the estrogen
    component determines in which lobe prostate dysplasia develops (Ofner
    et al., 1992).

         Male Noble rats were treated with Silastic implants containing
    testosterone and estradiol for 16 weeks since it had been reported
    previously that such implants increase the plasma estradiol
    concentration threefold while maintaining testosterone at
    physiological concentrations. This treatment regimen produced
    dysplasia in the dorsolateral prostate, without liver dysplasia.
    Microsomes were prepared from the liver, ventral prostate, and
    dorsolateral prostate of control and treated animals to determine
    metabolic conversion of estradiol to catechol estrogens. Catechol
    estrogen formation was observed at high levels in liver microsomal
    incubates and low levels in prostate incubates. Treatment failed to
    alter the extent or profile of hepatic estradiol metabolism, except
    for a significant reduction in estriol production relative to
    controls. A nonsignificant reduction in 2-hydroxyestradiol formation
    was also observed. The authors concluded that catechol estrogen
    formation is not a mediating step in estrogen-induced tumorigenesis
    (Lane et al., 1997).

         Mongrel dogs were treated for 60 days subcutaneously with DHT
    and/or estradiol; however, the quantity of hormone in the implant and
    the resulting plasma concentrations were not measured. Previous
    studies had indicated that such implants maintain plasma hormone
    concentrations at physiological levels. The activities of aryl
    hydrocarbon hydroxylase, 7-ethoxycoumarin  O-deethylase, and
    estradiol 2-and 4-hydroxylase were elevated in the prostate glands of
    animals treated with either hormone or their combination and were
    either decreased or unchanged in liver and kidney. The increase
    observed in estradiol-treated animals was substantially modified by
    concomitant treatment with DHT. The activity of estrogen 2-hydroxylase
    was increased tenfold and fourfold in animals given estradiol and
    estradiol plus DHT, respectively. The activities of
    7-ethoxycoumarin- O-deethylase, aryl hydrocarbon hydroxylase,
    glutathione peroxidase I, and catalase were also increased in the
    prostate. Hormones had variable effects on these enzymes in liver and
    kidney. The free radical-generated carbonyl content of the prostate
    increased 2.5-fold after treatment with estradiol and twofold with
    treatment with estradiol plus DHT. No hormone-related effects on
    carbonyl content were seen in kidney proteins, whereas DHT and the
    combination with estradiol increased the carbonyl content by 60 and

    150% over the control level. Treatment with estradiol alone resulted
    in a substantial but nonsignificant decrease in the hepatic protein
    carbonyl content relative to controls. In DNA hydrolysates, a 45%
    decrease was observed in the 8-OHdG content in DNA from the prostate
    of DHT-treated animals relative to control DNA; no changes were
    observed in DNA from other tissues or in that from animals treated
    with estradiol or estradiol plus DHT. The authors concluded that the
    hormone-induced changes in the activities of catechol estrogen
    synthetase and other enzymes were consistent with the postulated
    generation of free radicals by redox cycling of catechol estrogen
    specifically in the prostate. The protein damage observed is
    consistent with induction of benign prostatic hyperplasia injury by
    estrogen metabolites followed by growth stimulation of altered foci by
    DHT (Winter & Liehr, 1996).

         Adult male Noble rats received subpannicularimplantations of
    pellets that released a daily mean average of 120 g of testosterone
    propionate and 110 g of estradiol. Mammary gland ductal
    adenocarcinomas were induced in 100% of the rats after eight to nine
    months of treatment, whereas neither hormone alone was effective in
    inducing adenocarcinomas. Testosterone propionate disrupted mammary
    gland ducts and caused proliferation of stromal tissue, while
    estradiol induced ductal epithelial growth and nodular atypical
    hyperplasia. The authors concluded that androgens interact with
    androgen receptors or progesterone receptors, or perhaps both, in
    mammary glands induced by estradiol and testosterone propionate to
    cause tumours, reduce the latency, enhance tumour size, and increase
    the incidence to 100% (Liao et al., 1998).

    1.4  Observations in humans

    1.4.1  Therapeutic use

         Estrogens, sometimes in conjunction with progestogens, are used
    for contraception and for hormone replacement. They are among the most
    commonly prescribed drugs and include the parent molecule, its esters,
    and other synthetic steroidal and non-steroidal derivatives. Estrogens
    can be delivered orally, by intramuscular injection, or
    percutaneously. The main concern associated with use of estrogens for
    these purposes is a possibly increased risk for cancer; other
    side-effects may include hypertension, thromboembolic and other
    vascular diseases, breakthrough bleeding, gall-bladder disorders,
    nausea, migraine, and mood changes. The progestational component of
    combined hormonal therapy may be responsible for some of these effects
    (Williams & Stancel, 1996; Carr & Griffin, 1998).

         The effects of various dosages of orally administered conjugated
    estrogens, mestranol, and ethinylestradiol on the binding capacity of
    serum corticosteroid-binding globulin (CBG)-binding capacity was
    assessed in three healthy postmenopausal women, each of whom received
    Premarin(R) orally at 0.3, 0.62, 1.2, or 2.5 mg/day for 14 days.

    Each dose increased serum CBG-binding capacity, but the increase was
    not significant at 0.3 mg/day. The serum concentration was
    approximately doubled with a doubling of the dose. The authors
    concluded that there may be a threshold concentration of exogenous
    estrogens below which there is no associated increase in serum
    concentrations of CBG and above which there is a dose-response
    relationship between the exogenous estrogen and serum CBG-binding
    capacity (Moore et al., 1978).

         Twenty-three healthy postmenopausal women received one or more
    two-week courses of estrogens orally, including fine-particle
    estradiol (1, 2, or 10 mg/day) and conjugated estrogens (Premarin(R) 
    at 0.3, 0.62, 1.2, or 2.5 mg/day). The other estrogens that were
    assessed included piperizine estrone sulfate, ethinylestradiol, and
    diethylstilbestrol. Each dose was ingested by three women, and the
    pre-treatment and post-treatment concentrations of
    follicle-stimulating hormone, luteinizing hormone, CBG-binding
    capacity, SHBG-binding capacity, angiotensinogen, estrone, and
    estradiol were determined. The relative potency of the estrogen
    preparations were assessed. Conjugated estrogens were 3.2 times more
    potent than fine-particle estradiol in increasing SHBG-binding
    capacity, and fine-particle estradiol was 1.9 times more potent than
    piperizine estrone sulfate in inducing CBG-binding capacity, whereas
    conjugated estrogens were 2.5 times more potent than piperizine
    estrone sulfate. Doses < 1.7 mg equivalents of piperizine estrone
    sulfate failed to increase CBG-binding capacity significaqntly.
    Fine-particle estrogens and piperizine estrone sulfate were equipotent
    in more potent than piperizine estrone sulfate. The angiotensinogen
    response exceeded the baseline range only after administration of 2.2
    mg equivalents piperizine estrone sulfate. The natural estrogen
    preparations, piperizine estrone sulfate, fine-particle estradiol, and
    conjugated estrogens were nearly equipotent on affecting serum
    follicle-stimulating hormone concentrations. No dose-response
    relationship for serum luteinizing hormone could be demonstrated with
    any estrogen preparation (Mashchak et al., 1982).

         See also section 4, 'Epidemiological studies of women exposed to
    postmenopausal estrogen therapy and hormonal contraceptives'.

    1.4.2  Estradiol-related genetic markers of carcinogenicity

         The A2 allele of CYP17, the cytochrome P450 that catalyses the
    rate-limiting step in androgen biosynthesis, has been suggested to be
    a genetic marker for hormone-dependent disease such as polycystic
    ovarian syndrome, anovulatory infertility, hirsutism, and breast
    cancer. In a nested case-control study, no association was found
    between higher basal transcription of the A2 allele and subsequent
    development of breast cancer, stage of disease at diagnosis, or age at
    menarche. No racial differences in genotype were observed among
    Caucasian, African-American, Asian, and Hispanic women (Helzlsouer et
    al., 1998).

         The data on the effect of COMT polymorphism on human risk for
    hormone-induced cancer are conflicting and are discussed elsewhere in
    this report.

    2.  PROGESTERONE

    2.1  Explanation

         Progesterone (100 or 200 mg) in combination with estradiol
    benzoate (10 mg or 20 mg) is administered to cattle as an ear implant
    to increase the rate of weight gain (i.e. growth promotion) and
    improve feed efficiency. Progesterone is also administered on an
    intravaginal sponge to lactating and non-lactating dairy cows and
    goats to synchronize estrus. Progesterone is considered to be an
    endogenous substance, as exogenously administered progesterone is
    structurally identical to the progesterone produced in animals and
    humans.

    2.2  Biological data

    2.2.1  Absorption, distribution, and excretion

         Progesterone is largely inactive when administered by the oral
    route because of its low systemic bioavailability. High doses of
    certain fine-particle forms have been studied for possible use in
    contraception or in hormone replacement therapy, and the results are
    discussed below.

         The pharmacokinetics of orally administered fine-particle
    progesterone have been reviewed (Sitruk-Ware et al., 1987). Two 100-mg
    capsules of fine-particle progesterone given to women in the early
    follicular phase resulted in a peak plasma concentration at the second
    hour of 13 ng/ml. When compared with crystallized compound, the same
    dose of fine-particle progesterone resulted in twice as the peak serum
    concentration. The plasma profiles of the three major metabolites,
    pregnanediol 3a-glucuronide, 17-hydroprogesterone, and
    20a-dihydroprogesterone, mimicked that of progesterone.

         Six fasting women received 100 mg of fine-particle progesterone
    vaginally or orally in the early follicular phase of the menstrual
    cycle. The maximal plasma concentration was reached 2 h after oral
    administration and had returned to baseline by 6 h. Progesterone was
    metabolized predominantly to 5a derivatives, regardless of the route
    of administration; 5a-pregnanolone was found after oral but not
    vaginal administration. The serum concentrations of
    deoxycorticosterone and deoxycorticosterone sulfate were also
    increased after oral administration of progesterone. The authors
    concluded that ingested progesterone is metabolized mainly in the
    intestine (Nahoul et al., 1993).

         Seven subjects consumed several oral formulations containing 200
    mg progesterone, and blood samples were collected after 0.5, 1, 2, 3,
    4, and 6 h. A mean peak serum concentration of 30 ng/ml progesterone
    was achieved with a fine-particle formulation in oil 2 h after
    administration; the mean peak concentrations achieved with the other
    formulations were 9.6 ng/ml at 4 h with plain-milled, 13 ng/ml at 3.2
    h with fine-particle, 11 ng/ml at 4 h with plain-milled in oil, and 11
    ng/ml at 4.1 h with fine-particle enteric coated capsules (Hargrove et
    al., 1989).

         Five postmenopausal women received 100 mg progesterone per day
    orally for five days. The peak plasma progesterone concentrations,     
    7-11 ng/ml, were reached within 4 h after the last dose, and the
    concentrations remained raised for 96 h. The concentration of the
    metabolite 20a-dihydro-progesterone mimicked that of progesterone
    closely, at 3-5.1 ng/ml. Other metabolites present in plasma included
    pregnanediol-3alpha-glucuronide (550-1000 ng/ml) and
    17-hydroxyprogesterone (1.4-3.2 ng/ml) (Whitehead et al., 1980).

         The effect of food on progesterone absorption was tested in 15
    normal postmenopausal women who received 200 mg fine-particle
    progesterone for five days or placebo under fasting and non-fasting
    conditions; 100, 200, or 300 mg fine-particle progesterone for five
    days under fasting conditions; or 200 mg fine-particle progesterone or
    50 mg progesterone in oil intramuscularly for two days. Ingestion of
    food increased the integrated area under the curve of plasma
    concentration-time and maximal plasma concentration but not the time
    to the maximal concentration, and increased the bioavailability of
    progesterone twofold. Absorption and elimination were independent of
    dose over the range tested. The bioavailability of orally administered
    progesterone was 8.6% that of intramuscularly administered compound
    (Simon et al., 1993).

         The bioavailability of 50-200 mg fine-particle progesterone was
    studied in normally menstruating women treated by sublingual, oral
    (capsule and tablet), vaginal, or rectal administration during the
    follicular phase. Increased serum concentrations persisted for over 8
    h in women given > 100 mg progesterone vaginally or rectally and over
    24 h in those given 200-mg tablets orally. The authors concluded that
    clinically relevant concentrations of progesterone can be reached with
    fine-particle oral formulations (Maxson & Hargrove, 1985).

         Progesterone is bound in serum to CBG and to albumin. Albumin has
    a high capacity but low affinity for progesterone, while CBG has a
    higher affinity but a lower capacity. Approximately 17% of serum
    progesterone is bound to CBG and 80% to albumin, and 2.5% is in the
    free state. CBG is synthesized and secreted by the liver and other
    tissues, and fetal hepatocytes can synthesize it. A receptor for
    unliganded CBG exists on the surface of target cell membranes, which,
    when bound to steroid, can activate adenyl cyclase (Orth & Kovacs,
    1998).

         Progesterone is eliminated via the urine after conversion in the
    liver to pregnane-3 alpha,20 alpha-diol (preganediol) and conjugation
    to glucuronic acid at C3. Pregnanediol is the major urinary metabolite
    of progesterone and serves as an index of progesterone secretion. The
    metabolic clearance rate for progesterone is approximately 2200 l/day
    (Goldfien & Monroe, 1994).

         Elimination of progesterone after intravenous administration of 
    500 mg/kg bw to ovariectomized rats was described by a two-compartment
    model, with distribution and elimination half-lives of 0.13  0.024
    (mean  SD) and 1.2  0.21 h, respectively, and a volume of
    distribution of 2.4 l/kg bw (Gangrade et al., 1992). The plasma
    half-life of progesterone has also been reported as 5 min (Williams &
    Stancel, 1996).

         In five ovariectomized gilts given 47.4 nmol progesterone,
    bi-exponential clearance was observed, with mean half-lives of 2.5 and
    34 min. The authors concluded that the half-life of the first
    compartment indicated that the liver is the principal organ of
    clearance. In two gilts given 44.7 or 64.6 nmol progesterone, a
    secondary increase in the concentration of progesterone in portal and
    mesenteric plasma was observed, indicating that progesterone may
    undergo enterohepatic circulation (Symonds et al., 1994).

         Six adult anestrous bitches were given progesterone
    intramuscularly at 2 mg/kg bw, and the progesterone concentrations in
    serum were monitored for 72 h. A peak serum concentration of 34 ng/ml
    was reached 1.8 h after dosing, but by 72 h the serum concentrations
    had reached the preinjection concentration of 0.9 ng/ml. The mean
    half-life of absorption was 0.5 h and the mean half-life of
    elimination 12 h. The authors used computer modelling to establish
    that a dose of 3 mg/kg bw intramuscularly once a day would maintain
    the serum concentration at > 10 ng/ml (Scott-Moncrieff et al., 1990).

    2.2.2  Biotransformation

         Progesterone is irreversibly converted for subsequent elimination
    by hydroxylation to more polar and less active compounds, particularly
    at C21, C6, and C16. Progesterone hydroxylation is catalysed by
    members of various families of cytochromes P450 and other enzymes
    found in hepatic and extrahepatic tissues. Human 6-hydroxylase
    activity is catalysed most efficiently by CYP3A4 and to a lesser
    extent by CYP2D6, whereas 16 alpha-hydroxylation is catalysed by
    CYP3A4, 1A1, and 2D6 (Niwa et al., 1998).

         Tissue and species differ in the metabolism of progesterone. For
    example, human but not rat kidneys possess 6-hydroxylase activity.
    Steroid 21-hydroxylase in liver is provided primarily by CYP2C5 in
    rabbits (Kedderis & Mugford, 1998) and by CYP2C6 in rats (Endoh et
    al., 1995), which are P450 isozymes that have not been found in human
    liver. Rabbit liver CYP2C3 catalyses the 6-and 16 alpha-hydroxylation
    of progesterone (Hsu et al., 1993), but this enzyme has also not been
    found in human liver.

         It is well recognized that progesterone can serve as a precursor
    for other bioactive steroids, including hydroxylated progestogens,
    corticosteroids, androgens, and estrogens. Progesterone in plasma may
    be 21-hydroxylated to deoxycorticosterone enzymatically, independently
    of the adrenal steroid 21-hydroxylase (Mellon & Miller, 1989).
    Alternatively, progesterone may be reduced to 5
    alpha-dihydroprogesterone (5 alpha-pregnane-3,20-dione) and further to
    5 alpha-pregnan-3 alpha-ol-20-one, which may bind to the GABAA
    receptor on cell surface membranes (reviewed by Baulieu, 1997; see
    also below). Both of these compounds are secreted by the ovary and
    both are also produced by progesterone metabolism in the hypothalamus
    and pituitary (Mahesh et al., 1996).

         Progesterone can be metabolized to 17 alpha-hydroxyprogesterone
    by P450 17 alpha-hydroxylase and further to androstenedione. Cells
    that do not express this enzyme cannot metabolize progesterone to
    androgenic or estrogenic hormones. In human ovarian thecal cells,
    17 alpha-hydroxypro-gesterone was not further metabolized to
    androstenedione or to any other product, whereas pregnenolone and
    17 alpha-hydroxypregnenolone were converted to dihydroepiandrosterone
    and androsterone (McAllister et al., 1989). Thus, the conversion of
    progesterone to androgen is a minor pathway in some tissues or during
    particular stages of follicular development.

         A substantial increase was noted in the ability of the rat uterus
    to metabolize progesterone to 5 alpha-pregnane-3,20-dione and
    3 alpha-hydroxy-5 alpha-pregnan-20-one (allopregnanolone) between days
    11 and 21 of gestation when serum progesterone concentrations are
    high. While the former compound actively binds progesterone receptors
    in some tissues, this ability is apparently lacking in rat uterus
    (Tsai et al., 1998). Some indirect evidence exists that uterine
    contractility is mediated, at least in part, through GABAA receptors
    (Mahesh et al., 1996).

    2.2.3  Biochemical parameters

    2.2.3.1  Synthesis

         Progesterone is synthesized from steroid precursors in the
    gonads, adrenal cortex, and placenta of mammals. Cholesterol obtained
    primarily from circulating low-density lipoprotein serves as the
    precursor for steroid biosynthesis, although steroidogenic cells are
    capable of local  de novo cholesterol synthesis. In non-pregnant women,
    progesterone in the bloodstream is derived principally from secretion
    by the granulosa and thecal cells of the ovary. Gonadal synthesis of
    progesterone is regulated by luteinizing and follicle-stimulating
    hormones secreted by the anterior pituitary. The secretion of these
    hormones, in turn, is regulated by gonadotropin-releasing hormone,
    steroid hormones, and other regulating factors in a complex feedback

    loop which is not completely understood. During the follicular phase,
    about half of the progesterone is of ovarian origin and the other half
    of adrenal origin. During the luteal phase, production shifts
    predominantly to progesterone of ovarian origin.

         The concentrations of circulating progestogens, daily production,
    and metabolic clearance rates have been reported in previous reviews
    (IARC, 1979, specifically 'General remarks on sex hormones') and in
    most textbooks of endocrinology (e.g. Braunstein, 1994; Goldfien &
    Monroe, 1994; Carr, 1998; Griffin & Wilson, 1998). 

    2.2.3.2  Mechanism of action

         In the conventional view of the action of steroid sex hormones,
    they diffuse passively into the cell, interact with specific
    intranuclear receptors, and bind to specific DNA sequence on the
    genome (O'Malley et al., 1991). Ligand-receptor binding in conjunction
    with interactions with other nuclear transcription factors initiates
    transcription of a specific set of genes which are ultimately
    responsible for the translation of hormonal action into physiological
    events. Intracellular receptors for progesterone, a superfamily of
    ligand-activated transcriptional factors, mediate most of its effects.
    In humans, three specific progesterone-binding receptors have been
    defined: the 116-kDa B-receptor, an N-terminal truncated 94-kDa
    A-receptor, and the 60-kDa C-receptor; the last lacks the first zinc
    finger but contains, as do isoforms A and B, the hormone-binding
    domain (Leighton & Wei, 1998). Progesterone receptors, like other
    steroid hormone receptors, form dimers when a hormone binds. The
    various progesterone receptor isoforms have similar binding affinities
    for progesterone but may modulate regulation of a defined set of
    hormone-dependent genes. In addition, receptors for progesterone,
    glucocorticoids, mineralocorticoids, and androgens bind to overlapping
    DNA response elements and can thus regulate the same genes, adding
    complexity to the biological function of the steroid (Tsai et al.,
    1998). The specificity depends on the type or number of receptors
    present in the cell and other DNA sequences outside the
    steroid-binding domain.

         Progesterone and estrogen interact to control the growth and
    development of organ systems, primarily the female reproductive tract.
    Progesterone inhibits estrogen-stimulated endometrial proliferation
    and prepares the endometrium for implantation of the blastocytst. In
    this process, the endometrium is converted from a proliferative to a
    secretative phase, with accompanying morphological and biochemical
    changes. Although progesterone stimulates formation of secretory
    alveoli in the mammary gland, particularly during pregnancy,
    progesterone itself has minimal effect in the absence of estrogen
    priming, as in the uterus. Of the sex steroid hormones, progesterone
    has the smallest identified distribution of receptor tissue (Williams
    & Stancel, 1996) and perhaps the smallest cohort of transcriptionally
    responsive genes. Receptor levels and hormone responsiveness vary

    throughout the estrus cycle in all species: in the follicular phase,
    the levels are low but increase in response to serum estrogen; the
    levels decrease during the luteal phase in response to output of
    progesterone by the corpus luteum. In rats, the progesterone receptor
    levels increase throughout gestation despite high circulating
    concentrations of progesterone, whereas in other species little change
    or a decrease (guinea-pig) is noted (Tsai et al., 1998). An
    estrogen-stimulated increase in progesterone receptors is essential
    for progestogen activity in the uterus. Estrogen is required for the
    maintenance of intracellular progesterone receptor levels, and
    progesterone, in turn, decreases the estrogen receptor concentration
    and the levels of progesterone receptor.

         The metabolic effects of progesterone in humans include increased
    plasma concentrations of cholesterol and low-density lipoprotein and
    decreased high-density lipoprotein, and transient decreases in sodium
    excretion due to its capacity to antagonize aldosterone. Progesterone
    has well-known effects on basal body temperature. Hydroxylated
    metabolites of progesterone, namely the 3 alpha-and 3 alpha,5 alpha
    (allopregnanolone) derivatives, are suspected to act as
    neuropharmacological agents through the GABAA receptor (Mahesh et
    al., 1996; Wiebe et al., 1997; Concas et al., 1998; Smith et al.,
    1998). Synthesis and accumulation of the progestogen-type
    neuropharmacological agents are not related to plasma progesterone
    concentrations (Concas et al., 1998).

    2.3  Toxicological studies

         The toxicological studies reviewed below were reported in the
    scientific literature.

    2.3.1  Acute toxicity

         The therapeutic dose of fine-partcile progesterone is 400 mg/day
    for 10 days 5Physicians' Desk Reference, 1999). A dose of 300 mg/day
    taken by women at bedtime for 10 days per month was well tolerated,
    the only specific side-effect being mild, transient drowsiness (de
    Lignieres).

    2.3.2  Short-term studies of toxicity

         In order to clarify the nature of the vaginal response to
    neonatal treatment with progesterone and to record the responses of
    other target organs, including the mammary gland, to such treatment,
    291 female BALB/cfC3H/Crgl mice bearing the mouse mammary tumour virus
    received daily subcutaneous injections of 5 or 20 g estradiol or 100
    g progesterone; 5 g estradiol and 100 g progesterone; or 20 g
    estradiol and 100 g progesterone for five days beginning 36 h after
    birth. One half of the mice were ovariectomized on day 40. Vaginal
    smears were taken for 25 days when the mice were 40-50 days old. The
    mice were killed and autopsied at the time of onset of mammary tumours

    or at one year of age. The results are shown in Table 14. Mice
    receiving progesterone alone showed ovary-dependent, persistent
    vaginal cornification. When progesterone and estradiol were given
    concurrently, the occurrence of persistent vaginal cornification was
    significantly lower than in mice receiving estradiol alone.
    Progesterone alone produced hyperplastic downgrowths and lesions of
    both vaginal and cervical epithelia but to a lesser degree than in
    mice treated neonatally with estrogen. When progesterone was given
    concurrently with estradiol, the incidence of lesions was lower but
    their severity was greater. The low doses of estradiol and
    progesterone each resulted in an earlier age of onset and a higher
    incidence of mammary tumours; this also occurred after both combined
    estrogen plus progesterone treatments. In treated mice ovariectomized
    on day 40, normal mammary development did not occur and mammary
    tumours failed to appear, regardless of neonatal treatment. The
    authors concluded that progesterone causes ovary-dependent, persistent
    vaginal cornification and hyperplastic changes in the vaginal and
    cervical epithelium, affects the incidence of these lesions in mice
    treated simultaneously with estradiol, and significantly increases the
    incidence of mammary tumours in mammary tumour-virus bearing mice
    (Jones & Bern, 1977).

    2.3.3  Long-term studies of toxicity and carcinogenicity

         The toxicology of progesterone administered by subcutaneous or
    intramuscular injection to mice, rabbits, and dogs or by subcutaneous
    implantation in mice has been extensively reviewed (IARC, 1979).
    Progesterone increased the incidences of ovarian, uterine, and mammary
    tumours in mice and mammary gland tumours in dogs. Neonatal treatment
    increased the occurrence of precancerous and cancerous lesions of the
    genital tract and mammary tumorigenesis in female mice. The IARC
    Working Group concluded that the carcinogenic effects of estrogens and
    progestogens were inextricably linked to the hormonal milieu and to
    dose-effect relationships (IARC, 1979, 1987)

         The role of hormones, including progesterone, in mammary
    neoplasia in rodents and their relevance to human risk assessment have
    been reviewed (Russo & Russo, 1996), and it was noted that rodent
    models mimic some but not all of the complex external and endogenous
    factors involved in initiation, promotion, and progression of
    carcinogenesis. Tumour type and incidence are influenced by the age,
    reproductive history, and endocrine milieu of the host at the time of
    exposure. The spontaneous incidence of tumours varies from strain to
    strain of rats and mice. In rats, most spontaneously developed
    neoplasias, with the exception of leukemia, are of endocrine organs or
    organs under endocrine control. Russo & Russo (1996) concluded that
    mechanism-based toxicology is not yet sufficient for human risk
    assessment, and the approach should be coupled to and validated by
    traditional long-term bioassays.



        Table 14. Incidences of mammary tumours and nodules in mice treated neonatally with progesterone
                                                                                                   

    Neonatal treatment             No. bearing tumours/      Age at           No. bearing nodules/
                               no. examined (%)          appearance       no. examined (%)
    Estradiol  Progesterone                                  (months)
                                                                                                   

    0          100                 23/32 (72)                 9.3             27/32 (84)
    5            0                 17/19 (89)                 9.2             19/19 (100)
    5          100                 20/32 (63)                 9.3             27/32 (84)
    20           0                 4/11 (30)                  9.8             6/11 (55)
    20         100                 33/44 (75)                 9.6             40/44 (91)
    0            0                 5/17 (29)                 11.3             1/17 (6)
                                                                                                   
    

         Female rabbits were treated with large doses of either
    progesterone or testosterone for up to 763 days, and
    hydroxyprogesterone caproate was given intramuscularly every other
    week at an average dose of 13 mg/rabbit to 19 animals; testosterone
    ethanate was given intramuscularly every other week at an average dose
    of 15 mg to 21 animals. Rabbits treated with progesterone developed
    numerous cysts of the endometrium, sometimes associated with atypical
    hyperplasia. Active mammary secretion also occurred. After large doses
    of testosterone, one animal developed two adenomatous polyps of the
    endometrium, but there were no other noteworthy endometrial changes.
    One control animal developed similar polyps. Neither hormone
    significantly altered other tissues, including the ovary, adrenal,
    thyroid, and pituitary glands. No precancerous endometrial changes or
    cancers were found (Meissner & Sommers, 1966).

    2.3.4  Genotoxicity

         Genetic toxicity was not studied to support registration of
    progesterone for use in cattle. A working group convened by IARC
    examined the available information and concluded that: 'Progesterone
    did not induce dominant lethal mutations in mice or chromosomal
    aberrations in rats treated  in vivo. It did not induce chromosomal
    aberrations or sister chromatid exchanges in cultured human cells, nor
    chromosomal aberrations or DNA strand breaks in rodent cells. Studies
    on transformation of rodent cells  in vitro were inconclusive: a
    clearly positive result was obtained for rat embryo cells, a weakly
    positive result for mouse cells and a negative result for Syrian
    hamster embryo cells. Progesterone was not mutagenic to bacteria.'
    (IARC, 1987).

         Several studies on the genetic toxicity of progesterone have been
    reported in the scientific literature since the IARC evaluation. After
    exposure of DNA to 2 mmol/L progesterone, no DNA adducts were formed,
    as determined by 32P-postlabelling. The presence of a carbonyl group
    at C17, which progesterone lacks, was strongly associated with DNA
    binding by other sterols (Seraj et al., 1996).

         No significant increase in the frequency of structural or
    numerical chromosomal aberrations was seen in Syrian hamster embryo
    cells after treatment for 24 h with 3-30 g/ml progesterone.
    Progesterone was able to transform cells only at the highest dose
    tested (30 g/ml), but even at this dose, the frequency of
    transformation was one half that found with 1 g/ml benzo[ a]pyrene
    (Tsutsui et al., 1995). 

         Progesterone was administered by gavage to female rats as a
    single dose of 100 mg/kg bw three days before a two-thirds
    hepatectomy, and the rats were killed for cell sampling two days after
    the hepatectomy. The frequencies of micronucleated and binucleated
    hepatocytes were determined in a suspension of liver cells isolated by
    collagenase perfusion. The frequency of micronucleated hepatocytes was

    3.5-fold higher in progesterone-treated animals than in controls, but
    no increase in the frequency of binucleated hepatocytes was observed.
    The increase in micronucleus frequency may be related to chromosomal
    breakage or to loss of whole chromosomes. A dose of 100 mg/kg bw
    progesterone administered by gavage weekly for six weeks, followed by
    partial hepatocetomy, did not increase the number of
    gamma-glutamyltranspeptidase-positive foci in liver. The authors noted
    that similar findings were made under similar experimental conditions
    with the non-genotoxic liver mitogen 4-acetylaminofluorene,
    progesterone did not form DNA adducts in rat liver, and equivocal
    responses were observed for DNA repair in primary rat heptocytes.
     In vitro, progesterone was not mutagenic in bacteria, did not induce
    chromosomal aberrations or sister chromosome exchange in cultured
    human cells or chromosomal aberrations or DNA strand breaks in rodent
    cells, inconclusive results were obtained with regard to
    transformation of rodent cells and in the L5178Y TK+/- forward
    mutation assay, and a slight but statistically significant increase in
    the frqueny of sister chromatid exchange was observed in lymphocytes
    from children.  In vivo, progesterone did not induce dominant lethal
    mutations in mice or chromosomal aberrations in rat bone marrow
    (Martelli et al., 1998).

         In three male and three female Han:Wistar rats given progesterone
    at 100 mg/kg bw per day intragastrically as an aqueous
    microcrystalline suspension for 14 days, no progesterone-specific
    adducts were observed in the liver, as determined by 32P-postlabelling
    (Feser et al., 1996).

    2.3.5  Reproductive toxicity

         Progesterone at doses of 5-200 mg was administered daily by
    subcutaneous injection to 22 pregnant Long-Evans rats on days 15-20 of
    gestation. The pups were surgically removed from the dams on day 22
    and killed after 20 days. No difference in the anogenital distance was
    observed between treated and control rats, and sex was easily
    determined, whereas animals treated with concentrations of 1.5-2.5 mg
    synthetic progestogens were all masculinized (Revesz et al., 1960).

         Six pregnant rhesus monkeys were treated with progesterone in oil
    intramuscularly at a dose of 50 mg daily for five days each week.
    Treatment was begun on days 24-28 of gestation and was maintained
    until delivery. The newborns were weighed, killed, and completely
    autopsied, with appropriate histological sections. The offspring of
    animals treated with progesterone (three males and thee females) were
    delivered at about term, and no abnormalities were noted. In contrast,
    most of the offspring of monkeys treated with synthetic progestogens
    were born prematurely and were not viable; furthermore, all of the
    females were masculinized. The authors concluded that progesterone in
    comparatively large doses produces no abnormalities and does not
    interfere with the normal conclusion of a pregnancy in rhesus monkeys
    (Wharton & Scott, 1964).

    2.4  Observations in humans

         The commonest uses of progestogens are for contraception and for
    hormone replacement therapy. For these purposes, synthetic
    progestational agents are usually administered, which differ in
    structure and function from natural progesterone (Williams & Stancel,
    1996).

         Changes in endometrial morphology was investigated in 43 patients
    without ovaries who were given progesterone intravaginally,
    intramuscularly, or orally. The response to intramuscular
    administration in oil was heterogeneous, whereas the endometrial
    morphology seen after intravaginal administration of fine-particle
    progesterone closely matched that observed in the natural cycle, and
    pregnancies were supported in two women. Orally administered
    progesterone did not affect endometrial morphology (Bourgain et al.,
    1990).

         Sixty women with oligomenorrhea or amenorrhoea received a 10-day
    course of 200 or 300 mg fine-particle progesterone orally. Withdrawal
    bleeding occurred in 90% of the women receiving 300 mg progesterone,
    58% receiving 200 mg, and 29% of those given placebo. Their lipid
    concentrations were unchanged (Shangold et al., 1991).

         Seven women with luteal phase defects, treatment with 200 mg
    fine-particle progesterone orally three times a day was effective
    (Frishman et al., 1995).

         The side-effects occurring in women receiving fine-particle
    progesterone orally as part of hormone replacement therapy have been
    reviewed. They include minor changes in plasma lipoprotein profile in
    some but not all studies but no effect on haemostatic parameters (such
    as platelet aggregation and fibronolytic capacity) (Sitruk-Ware et
    al., 1987).

         Two groups of six healthy postmenopausal women were given 300 mg
    of fine-particle progesterone once or twice a day on days 1-14 after
    estrogen priming for 30 days. Endometrial biopsy samples were then
    studied for effects on histological appearance, glycogen content,
    ribosomal RNA, and nuclear estrogen receptors in glands, surface
    epithelial, and stroma, and the samples from treated women were
    compared with biopsy samples obtained before therapy. The group given
    progesterone once a day showed incomplete secretory conversion,
    whereas the other group showed full secretory conversion with
    suppression of mitotic activity and the presence of a predecidual
    reaction. Staining for glycogen was significantly increased in both
    groups, by 124% at 300 mg/day and 291% at 600 mg/day. Ribosomal RNA
    was decreased by 50 and 73% in the two groups, respectively. The
    nuclear receptor content decreased in every region of the endometrium
    in both groups; although the decrease in nuclear estrogen receptor
    content in the stroma of the group given 300 mg/day did not reach
    significance, this finding was attributed to insufficient samples (Kim
    et al., 1996).

         Postmenopausal women were treated for 21 of 28 days for five
    years with percutaneous estradiol at a dose of 1.5 mg/day. Within the
    first six months, some women received a dose of 3 mg/day estradiol for
    25 of 28 days. The initial dose of progesterone was 200 mg/day orally
    for the last 14 days of estradiol treatment. This dose was increased
    to 300 mg/day for women in whom the desired clinical effect was not
    achieved at the lower dose (i.e. withdrawal bleeding). The baseline
    plasma concentration was 18 pg/ml, which increased to an average of
    68-89 pg/ml at various times during the treatment period. During the
    last 12 months of treatment, 2% of the women had episodes of irregular
    bleeding, 21% had regular cyclic bleeding, and 77% had no bleeding.
    The highest incidence of cyclic bleeding was observed with the
    high-estradiol plus high-progesterone regimen (3 mg estradiol and 300
    mg progesterone), and the highest incidence of amenorrhoea was seen in
    women given the low-estradiol plus low-progesterone regimen. There was
    no evidence of endometrial hyperplasia after five years of treatment
    with estradiol and progesterone (Moyer et al., 1993).

         See also section 4, 'Epidemiological studies of women exposed to
    postmenopausal estrogen therapy and hormonal contraceptives'.

    3.  TESTOSTERONE

    3.1  Explanation

         Testosterone propionate (200 mg) in combination with estradiol
    benzoate (20 mg) is administered to cattle as an ear implant to
    increase the rate of weight gain (growth promotion) and to improve
    feed efficiency. Testosterone propionate is rapidly cleaved to
    testosterone  in vivo and is thus considered an 'endogenous
    substance', as the residues are structurally identical to testosterone
    produced in animals and humans.

    3.2  Biological data

    3.2.1  Absorption, distribution, and excretion

         Testosterone is generally considered to be inactive when given by
    the oral route owing to gastrointestinal and/or hepatic inactivation.
    Maintenance of physiological concentrations after injection of
    testosterone is also difficult because of its rapid clearance (Griffin
    & Wilson, 1998). There is little clinical interest in oral forms of
    testosterone (i.e. fine-particle formulations), as the high doses
    required to overcome first-pass metabolism in the liver result in a
    high hepatic load (WHO, 1992).

         Absorption of the ester testosterone undecanoate has been studied
    for its possible use in testosterone replacement therapy. The plasma
    concentrations of testosterone were measured by radioimmunoassay after
    oral administration of 25 mg testosterone or 40 mg testosterone
    undecanoate (approximately 25 mg testosterone) to young women, and

    were compared with that in subjects receiving testosterone at 1.5
    g/kg bw by intravenous administration. Bioavailable testosterone
    represented 3.6% of administered testosterone and 6.8% of administered
    testosterone undecanoate. The low bioavailability of orally
    aministered testosterone has been attributed to its high metabolic
    clearance rate, 25 ml/min per kg. The estimated bioavailability
    provides justification for the high dose (120-140 mg/day) considered
    to be necessary to replace daily production (5-7 mg) (Tauber et al.,
    1986). The long aliphatic side-chain of testosterone undecanoate is
    thought to facilitate its lymphatic absorption and thus to avoid
    first-pass metabolism (Butler et al., 1992; Griffin & Wilson, 1998).

         In normal men, circulating androgens are bound to SHBG and to a
    lesser extent to serum albumin. Only 1-2% of the testosterone in
    circulation is unbound: 44% is bound to SHBG and 54% is bound to
    albumin and other proteins (Griffin & Wilson, 1998). Testosterone
    binds to albumin with lower affinity than to SHBG. Plasma SHBG is
    secreted from the liver, but adult rodent livers do not produce the
    secretory form of SHBG (Reventos et al., 1993). The plasma
    concentration of SHBG is increased 5-10-fold by estrogens and
    decreased twofold by testosterone (Griffin & Wilson, 1998). A
    non-secretory form of SHBG is present in many tissues, including
    reproductive tissues and the brain. A SHBG receptor on the surface of
    some cell membranes binds unliganded SHBG; steroid hormone then binds
    SHBG, which can then not bind receptor (Hryb et al., 1990). The
    SHBG-receptor complex present on membranes of target tissues may be
    responsible for the interaction between the steroid hormone and cAMP
    pathways (Rosner, 1991). The intracellular form of SHBG may also
    sequester or direct hormones to their target tissue. 

         When radiolabelled testosterone is administered intragastrically
    to rats, about one fourth appears in bile within 12 h as glucuronide
    and sulfate conjugates, indicating enterohepatic circulation (IARC,
    1979). When testosterone is used as a substrate, hepatic
    glycuronsyltransferase and sulfotransferase activities are higher in
    men than in women and show extensive individual variation (Pacifici et
    al., 1997). Individual but not sex differences were observed in the
    activities of the human jejunal estrogen and dehydroepiandrosterone
    sulfotransferases (Her et al., 1996).

         About 90% of an administered dose of radiolabelled testosterone
    is found in the urine, and about 6% undergoes enterohepatic
    circulation and appears in the feces (Wilson, 1996). About 1% of
    testosterone is secreted in urine as glucuronide, while the remainder
    is secreted primarily as 17-ketosteroids (androsterone and
    etiocholanolone). Smaller amounts of glucuronide and sulfate
    conjugates of androstanediol and estrogens are found.

         The plasma half-life of testosterone after intravenous
    administration was 10 min (Tauber et al., 1986).

    3.2.2  Biotransformation

         The metabolism of testosterone has been reviewed (IARC, 1979;
    Goldfien & Monroe, 1994; Wilson, 1996; Griffin & Wilson, 1998).
    17-Hydroxysteroid dehydrogenase catalyses the reversible
    oxido-reduction of androstenedione to testosterone, and of estrone to
    estradiol and dehydroepi-androsterone to D5-androstenediol. Further
    reductions in this pathway (5 alpha and 5 beta followed by 3 keto)
    lead to the formation of androsterone and etiocholanolone, the
    principal urinary products of testosterone.

         Significant differences are found between the metabolic pathways
    of testosterone in rodents and humans. Sex-specific regulation of
    cytochrome P450s has not been found in human liver, although sex
    differences in the metabolism of xenobiotics exist (Kedderis &
    Mugford, 1998). Hepatic rat testosterone 7 alpha-hydroxylase (P450
    2A1) and mouse testosterone 15 alpha-hydroxylase (P450 2A4) are not
    found in human liver. Testosterone is converted to its 6 form by
    human liver CYP3A4 (Lacroix et al., 1997). The major pathways for
    inactivation of testosterone in female mouse liver are hydroxylation
    at the 16-, 6 alpha-, and 16 alpha-positions followed by conjugation
    (Wilson & LeBlanc, 1998). Metabolism of testosterone in rat liver,
    used as a diagnostic tool for P450 isozyme activity, produced
    7 alpha-, 6-, 16 alpha-, and 2 alpha-hydroxytestosterones (Swales et
    al., 1996). Testosterone 6-and 16 alpha-hydroxylation is carried out
    in rat liver by CYP3A18 and by other members of the 3A family (Nagata
    et al., 1996). No 16-hydroxylase (CYP2B1) activity was detected in
    rat liver (Swales et al., 1996).

         Extensive reductive metabolism of testosterone occurs in the
    liver and extrahepatic tissues. The rat small intestine is capable of
    oxidizing testosterone, and the enzyme is present in homogenates of
    rat gastric and duodenal mucosa at significantly higher levels than in
    jejunum; it was also found in the ileum and colon. The major portal
    vein metabolite was androstenedione, indicaing that the gut rather
    than the liver is primarily responsible for the metabolism of orally
    administered testosterone (Farthing et al., 1982)

         Testosterone is a precursor of other physiologically relevant
    steroids. The active metabolite DHT, formed through the action of
    5 alpha-reductase, is an important physiological substrate for the
    androgen receptor. DHT binds with higher affinity to the androgen
    receptor than testosterone and is considered to be the main cellular
    mediator of androgen activity in some tissues. DHT is metabolized to
    androsterone, androstanedione, and 3 alpha- and 3-androstanediol.
    3-Hydroxysteroid dehydrogenase also catalyses the conversion of
    dehydroepiandrosterone to androstenedione in the testosterone
    biosynthetic pathway. Estradiol is formed by aromatization of the A
    ring of testosterone. DHT cannot be converted to estrogen.

    3.2.3  Biochemical parameters

    3.2.3.1  Synthesis

         Testosterone is synthesized in testicular Leydig cells, ovarian
    thecal cells, and the adrenal cortex. The principal regulator of
    gonadal steroidogenic synthesis is luteinizing hormone, secreted by
    the anterior pituitary. DHT, derived from reduction of testosterone in
    peripheral tissues, and D5-androstenediol are also physiologically
    relevant androgens (Griffin & Wilson, 1998; Miyamoto et al., 1998).
    Testosterone synthesis begins in the fetus during the first trimester
    of pregnancy after stimulation by maternal chorionic gonadotropin. In
    males, the plasma concentrations of testosterone are high during the
    prenatal period, peak at around three months of age, and then remain
    very low until about 11 years of age. The testosterone concentrations
    reach a plateau at around 17 years of age, remain elevated for
    decades, and then gradually decline (see Table 15). The blood
    concentrations of testosterone oscillate in response to the pulsatile
    secretion of luteinizing hormone throughout the day. Testosterone
    decreases the pulsatile release of gonadotropin-releasing hormone from
    the hypothalamus and luteinizing hormone from the pituitary by
    negative feedback (Griffin & Wilson, 1998). A diurnal trend in the
    blood concentrations of testosterone is found as well, the
    concentrations in plasma being approximately 25% higher in the early
    morning than in the afternoon (Orth & Kovacs, 1998). 

         Plasma testosterone concentrations also increase in females
    during puberty, although to a lesser extent than in males. In adult
    premenopausal women, 70-80% of circulating testosterone is derived
    from conversion of dehydroxyepiandrosterone and androstenedione from
    the adrenals and ovaries. In postmenopausal women, ovarian secretion
    may account for up to 50% of testosterone production, as bilateral
    oophorectomy results in a significant reduction in circulating
    androgens. Women in their 40s have about 50% less circulating
    androgens than women in their 20s, probably because of declining
    adrenal dehydroxyepiandrosterone secretion (reviewed by Gelfand &
    Wiita, 1997).

         The concentrations of circulating androgens, their daily
    production and metabolic clearance rates are given in reviews (e.g.
    IARC, 1979, specifically 'General remarks on sex hormones') and in
    most textbooks of endocrinology (e.g. Braunstein, 1994; Goldfien &
    Monroe, 1994; Carr, 1998; Griffin & Wilson, 1998).

    3.2.3.2  Mechanism of action

         Testosterone and DHT readily diffuse into target cells and bind
    with high affinity to the androgen receptor. Only one intracellular
    receptor type has so far been identified, and molecular and genetic
    analyses indicate that multiple androgen receptor subtypes are

    unlikely to exist. The receptor-hormone complex binds to specific
    target DNA sequences and initiates transcription of androgen-regulated
    genes, translating exposure to hormones into physiological events
    (Wilson, 1996).

         Testosterone is necessary during pubertal development for the
    initiation and maintenance of spermatogenesis and for the libido.
    Androgens interact with other growth factors to regulate the growth
    and differentiation of the prostate. Androgens also stimulate
    erythropoietin production in the kidney, stimulate stem cells in the
    haematopoietic system, increase nitrogen retention, and regulate male
    hair growth, including balding in genetically susceptible individuals.
    Androgens in conjunction with growth hormone may accelerate linear
    growth during puberty. Androgens decrease serum SHBG concentrations
    and shift the high-or low-density lipoprotein:cholesterol ratio
    (Griffin & Wilson, 1998; WHO, 1992)

         Testosterone and DHT produce different biological effects in
    reproductive and non-reproductive organs and during development. In
    tissues such as bone, muscle, and testis, testosterone is thought to
    have a direct androgenic action, while conversion to DHT is essential
    for androgenic action in other target tissues. In the fetus,
    testosterone is responsible for development of the seminal vesicles
    and epididymis and for virilization of the mesonephric ducts, whereas
    DHT is required for external virilization. 

    3.3  Toxicological studies

         The toxicological studies reviewed below were reported in the
    scientific literature.

    3.3.1  Acute toxicity

         Fine-particle orally administered testosterone has not been
    approved for clinical use. In studies with 400 mg of fine-particle
    testosterone administered orally for 21 days to six healthy male
    volunteers, testosterone was well tolerated, with few effects on
    lmiver enzyme systems, but induction of the heaptic drug-metabolizing
    system, including testosterone metabolism, was reported in all six
    subjects (Johnsen et al., 1976).

         In one report of an adverse event, a hypogonadal 21-year-old man
    began treatment with 200-mg intramuscular injections of testosterone
    enanthate every two weeks. He had a cerebral accident which was
    attributed to treatment. His plasma testosterone concentration was 11
    400 ng/dL, whereas the range in normal men is 200-1000 ng/dL. The
    authors concluded that cerebral infarct is a potential complication of
    over-zealous andrigen replacement (Nagelberg et al., 1986).


        Table 15. Concentrations, metabolic clearance rates, and daily production of circulating androgens
                                                                                                     

    Person              Serum concentration   Metabolic clearance   Total daily production
                        (ng/ml)               (l/day)               (mg/day)
                                                                                                     

    Male                                      690

      Prepubertal       0.08-0.14                                   0.06-0.1
      Pubertal          0.84-1.8                                    0.58-1.2
      Adult             3-10                                        2.1-6.9

    Female                                    690

      Prepubertal       0.05-0.13                                   0.03-0.09
      Pubertal          0.09-0.24                                   0.06-0.17
      Adult             0.3-0.7                                     0.21-0.48
                                                                                                     
    


    3.3.2  Short-term studies of toxicity

         Six intact adult male baboons received weekly intramuscular
    injections of 200 mg testosterone enanthate (equivalent to 8 mg/kg bw)
    for up to 28 weeks, while two control animals received weekly
    injections of the vehicle only. Quantitative increases in the weight
    and volume of both prostatic lobes were seen after 15 weeks of
    treatment, and by week 28 there was an increase in stromal tissue with
    papillary ingrowth or invagination of glandular epithelium in the
    caudal lobe of the prostate. The serum concentrations of testosterone
    and DHT were significantly elevated, from 10 and 2-3 ng/ml to 30-40
    and 5-6 ng/ml, respectively. The androstenedione concentrations were
    increased by three to four times and that of estradiol from 20 to
    80-90 pg/ml. The authors concluded that these steroids play a direct
    role in inducing early benign prostate hypertrophy in baboons and that
    their observations were similar to those in human benign prostate
    hypertrophy (Karr et al., 1984).

    3.3.3  Long-term studies of toxicity and carcinogenicity

         The effects of subcutaneously injected or implanted testosterone
    and its esters have been reviewed extensively (IARC, 1979). The
    working group convened by IARC concluded that: "There is sufficient
    evidence for the carcinogenicity of testosterone in experimental
    animals. In the absence of adequate data in humans, it is reasonable,
    for practical purposes, to regard testosterone as if it presented a
    carcinogenic risk to humans."

         The relevance of animal models to human prostate disorders has
    been reviewed (WHO, 1992). Besides humans, dogs are the only animals
    that develop prostatic cancer and benign prostatic hyperplasia at a
    high frequency. In this model, long-term treatment with androgens and
    estrogens is required to produce hyperplasia, although such synergism
    is not observed in other species. ACI rats spontaneously develop
    histologically evident prostatic cancer which does not progress to
    clinically relevant disease when pharmaco-logically relevant amounts
    of exogenous androgen are administered. Prostate cancer has been
    induced only in the Noble and Lobund-Wistar strains of rat. 

         The role of hormones, including androgens, in the development of
    mammary neoplasia in rodents and their relevance to human risk
    assessment have been reviewed (Russo & Russo, 1996). Endogenous
    androgens are necessary for mammary development in rodents, and it was
    noted that rodent models mimic some but not all of the complex
    external and endogenous factors involved in initiation, promotion, and

    progression of carcinogenesis. Tumour type and incidence are
    influenced by the age, reproductive history, and the endocrine milieu
    of the host at the time of exposure. The spontaneous incidence of
    tumours differs in different strains of rats and mice. In rats, most
    spontaneous neoplasias, with the exception of leukaemia, are of
    endocrine organs or organs under endocrine control. Russo & Russo
    (1996) concluded that mechanism-based toxicology is not yet sufficient
    for human risk assessment, and the approach should be coupled to and
    validated by traditional long-term bioassays.

         Fischer 344 rats were given 3,2'-dimethyl-4-aminobiphenyl (a
    prostate carcinogen) at 50 mg/kg bw 10 times at two-week intervals,
    and then, from week 20, testosterone propionate and/or
    diethylstilbestrol by subcutaneous Silastic implant for 40 weeks, as
    seven cycles of 30 days' treatment and 10 days' withdrawal.
    Intermittent administration of testosterone resulted in suppression of
    the development of ventral prostate adenocarcinomas and slight
    (non-significant) increases in the incidences of invasive carcinomas
    of the lateral prostate and seminal vesicles. Diethylstilbestrol
    completely suppressed tumorigenesis, and the combination with
    testosterone propionate inhibited prostate tumour development (Cui
    et al., 1998).

         Hydroxyprogesterone caproate was given intramuscularly every
    other week at an average dose of 13 mg to 19 female rabbits, and
    testosterone ethanate was given intramuscularly every other week at an
    average dose of 15 mg to 21 animals; both treatments were given for up
    to 763 days. Rabbits treated with progesterone developed numerous
    endometrial cysts, sometimes associated with atypical hyperplasia;
    active mammary secretion was also seen. Treatment with testosterone
    induced two adenomatous polyps of the endometrium in one animal, but
    no other noteworthy endometrial changes were found and one control
    animal developed similar polps. Neither significantly altered other
    tissues such as the ovary, adrenal, thyroid, or pituitary gland. No
    precancerous endometrial changes or cancers were found (Meissner &
    Sommers, 1966).

    3.3.4  Genotoxicity

         No studies of genetic toxicity were available to the IARC working
    group (IARC, 1979). Summaries of more recent studies from the
    scientific literature are presented below.

         Testosterone at concentrations of 1-100 g/ml had weak
    transforming activity, but with no dose-response relationship, in
    Syrian hamster embryo cells. It inhibited the transforming potential
    of benzo[ a]pyrene when incubated concomitantly or sequentially. The
    transforming effect of testosterone was amplified by phorbol acetate
    but was completely inhibited by dexamethasone (Lasne et al., 1990).

         The growth of Syrian hamster embryo cells was reduced by
    incubation with testosterone or testosterone propionate at 10-30 g/ml
    in a dose-dependent manner, and the two compounds were equipotent in
    producing morphological transformation. The propionate induced
    dose-dependent cell transformation at concentrations of 1-30 g/ml,
    while testosterone was effective only at 30 g/ml. At that
    concentration, the transformation frequency induced by testosterone
    and the propionate was one-half that induced by 1 g/ml
    benzo[a]pyrene. Neither steroid significantly increased the frequency
    of chromosomal aberrations or aneuploidy. Testosterone did not induce
    gene mutations at the  hprt or Na+/K+ ATPase locus (Tsutsui et
    al., 1995).

         When testosterone and other steroids were added at a final
    concentration of 2 mmol/L to DNA obtained from human surgical
    resections, rat liver, HepG2 cells, and calf thymus, testosterone did
    not form adducts with naked DNA. Furthermore, no adducts were observed
    in DNA isolated from HepG2 cells incubated with 10-100 mol/L
    testosterone for 24 h. The presence of a carbonyl group at C17 (which
    testosterone lacks) was predictive of DNA reactivity (Seraj et al.,
    1996). DNA strand breaks were observed in the dorsolateral prostate of
    Nobel rats treated with a combination of testosterone and estradiol
    but not in those treated with testosterone alone (Ho & Roy, 1994).

         NBL/Cr rats received two Silastic implants subcutaneously, one of
    which contained testosterone and the other estradiol, at the age of
    eight to nine weeks, and were killed 8, 16, or 24 weeks after the
    start of treatment. The quantity of steroid in the implant was not
    described, but the release rate was predicted to increase the
    estradiol concentrations by approximately 14-fold while maintaining
    normal plasma testosterone concentrations. Adduct levels were
    determined by 32P-postlabelling. The prostate glands were found to
    contain two major endogenous adducts and several minor spots,
    irrespective of treatment, but in treated animals the relative levels
    of the two adducts were decreased from 1 and 10  109 to 0.5 and 3 
    109 in the ventral but not in the dorsolateral or anterior prostate.
    A major adduct spot was observed in the dorsolateral prostate of
    animals treated with both hormones for 16 and 24 weeks but not in

    those treated for 8 weeks or in control animals. The presence of this
    adduct coincided with the occurrence of dysplastic lesions in the
    dorsolateral prostate. It appears to be different from the indirect
    estrogen-DNA adducts found in the kidneys of male hamsters treated
    with estrogen implants and from I-compounds. The authors suggested
    that testosterone enhances prostatic carcinogenesis, perhaps by
    generating estradiol through peripheral conversion (Han et al., 1995).

    3.3.5  Reproductive toxicity

         Complete resorption of embryos was seen in female Sprague-Dawley
    rats that received 10 mg estradiol (five rats) or testosterone (seven
    rats) as a subcutaneous implant on day 10 of gestation. No effect was
    seen in control pregnant rats given dextran by the same route (Sarkar
    et al., 1986).

    3.4  Observations in humans

         Androgens are used therapeutically in men with deficient
    testicular function to replace normal testosterone. In women, an
    estrogen-androgen combination is used in postmenopausal hormone
    replacement therapy (Sands & Studd, 1995; Wilson, 1996). The effects
    of excess androgens particularly in young boys and women include
    virilizing effects (such as hair loss, deepened voice, acne, growth of
    facial hair, and menstrual irregularities); estrogenic effects (such
    as gynecomastia in men), and toxic effects (such as oedema and
    hepatotoxic effects, especially with 17 alpha substitutions). The 
    serious side-effects of orally administered androgens are particularly 
    common in the liver in the treatment of aplastic anaemia but are also
    observed during treatment for other conditions, including
    hypogonadism. The complications include blood-filled cysts and
    heaptomas. Increasing the plasma concentration of testosterone to
    above physiological levels results in decreased secreation of
    follicle-stimulating and luteinizing hormones and decreased testicular
    volume and sperm production (Wilson, 1996; Griffin & Wilson, 1998).

         In postmenopausal women treated with androgen replacement therapy
    at doses up to 10 mg/day for more than six months, the virilizing
    effects of orally administered alkylated andrigens such as
    methyltestosterone were found to be deoendent on dose and duration.
    The masculinizing effects abated when exposure to the andrigens was
    decreased or eliminated (reviewed by Gelfand & Wiita, 1997).
    Methylation of testosterone at the 17a position effectively reduces
    its first-pass metabolism in the liver, thus increasing its
    bioavailability. It can thus reach the systemic circulation when given
    at therapeutic concentrations (Friffin & Wilson, 1998).

         Testosterone undecanotate, the only oral testosterone ester
    preparation available, has been used successfully for the management
    of delayed puberty in boys by administration of 40 mg daily for 15-21
    months. No side-effects were observed, but progression of
    virilization, testicular growth, and acceleration of growth associated
    with puberty were observed (Butler et al., 1992).

         Subcutaneously implanted pellets containing 100 mg crystalline
    testosterone release the compound at a rate of 1 mg/day, which falls
    to 0.5 mg by three months and 0.35 mg at six months. In postmenopausal
    women, such implants raised the serum testosterone concentration
    threefold over that of controls. Development of downy facial hair,
    hirsutism, acne, and, rarely, voice changes and clitoromegaly have
    been reported as adverse effects. Estrogen and estrogen plus
    testosterone implants lowered low-density lipoprotein cholesterol and
    raised high-density lipoprotein cholesterol in the same proportion
    (Sands & Studd, 1995).

         The role of endogenous androgens in the development of benign
    prostatic hyperplasia and prostate cancer and the effectiveness of
    androgen ablation have been reviewed (WHO, 1992; Osterling et al.,
    1997). These disorders arise from different parts of the prostate but
    are highly associated statistically. The prevalence of histologically
    discernible prostate cancer appears to be similar in various
    geographic locations, but progression to clinical prostate cancer is
    associated with environmental and genetic factors. These disorders
    develop later in life only in men who had normal concentrations of
    androgens throughout puberty and possibly up to 40 years. Differences
    in serum testosterone concentrations have been studied as a possible
    cause for the geographical differences in the occurrence of prostate
    cancer. Dutch men were found to have a higher average plasma
    testosterone concentration than Japanese men, which was statistically
    associated with body weight. Nevertheless, the serum testosterone
    concentrations of Japanese men with prostate cancer were not different
    than those in men without this disease (WHO, 1992).

         Epidemiological studies of the relationship between serum
    concentrations of androgens and estrogens and prostate cancer have
    yielded inconsistent but largely negative results (for references, see
    Dorgan et al., 1998; Ross et al., 1998). Since a possible association
    with markers of androgen metabolism (androstanediol glucuronide and
    DHT) has been observed in some studies, a complex, androgen-dependent,
    multi-gene model involving both androgen biosynthesis and metabolic
    pathways has been suggested to be critical for susceptibility to
    prostatic cancer (Ross et al., 1998).

         The effectiveness of orally administered testosterone was
    assessed in five eunuchs in a trial in which controls received a
    placebo and neither the participants nor the investigators was aware
    of which group received treatment. Initial studies indicated that a
    dose of 25 mg testosterone had no effect on any clinical parameter,
    including sexual desire, erection, ejaculation, and general
    well-being. In a second study, a dose of 100 mg/day did not alter the
    clinical parameters, whereas 400 mg/day (100 mg four times a day) was
    fully effective (Johnsen et al., 1974). Oral administration of
    testosterone at 30 mg/day did not restore sexual function to
    castrates, while 100 mg/day had a slight effect; 20 mg of testosterone
    propionate administered twice a week by injection had moderate
    potency. Thus, testosterone is more than 16 times as active by
    injection as it is orally (Foss & Camb, 1939).

    4.  EPIDEMIOLOGICAL STUDIES OF WOMEN EXPOSED TO POSTMENOPAUSAL
        ESTROGEN THERAPY AND HORMONAL CONTRACEPTIVES

    4.1  Methods

         Although no large epidemiological studies on estradiol,
    progesterone, or testosterone alone are available, millions of women
    have been exposed, chiefly since the 1960s, to a range of exogenous
    estrogens and/or progestogens in the course of postmenopausal estrogen
    therapy or hormonal contraception. A few extensive reviews on these
    compounds have been published recently (WHO, 1998; IARC, 1999) which
    provide many more details of the studies summarized here. The findings
    of a large number of case-control and cohort studies and clinical
    trials in which the relationship between use of exogenous estrogens
    and progestogens and the risk for disease was evaluated are reviewed.
    Each study design has its strengths and weaknesses: cohort studies are
    considered to be less susceptible to bias in recall or information,
    but case-control studies are useful for investigating relatively rare
    conditions. Whenever possible, the results of pooled analyses of
    original data or meta-analyses of published results are given, in
    order to provide the best estimate of the relationship between hormone
    use and cancer risk. The strength of an association between exposure
    to exogenous hormones and the development of disease is given, as is
    customary in epidemiological studies, in terms of the relative risk
    (RR), which is the ratio of the incidence rate of the disease among
    subjects exposed to the factor to that among unexposed subjects. A
    relative risk of 1.0 indicates no association, whereas a relative risk
    significantly greater than 1.0 implies a positive association between
    exposure to the factor and the development of disease; conversely, a
    relative risk of less than 1.0 implies that exposure to the factor may
    have a protective effect. In order to provide some indication of the
    variation of relative risks, 95% confidence intervals (CIs) are
    usually calculated. If the interval does not include 1.0, the finding
    is considered to be statistically significant. In a few instances,
    such as for breast cancer, the estimated excess number of breast
    cancer among 10 000 women who had used hormones is also given.

    4.2  Postmenopausal estrogen therapy

    4.2.1  Exposure

         The numbers of women who have used postmenopausal estrogen
    therapy varies among countries and within regions of countries (IARC,
    1999). The prevalence of use has been greater in the United States
    than in most other countries, and use of estrogen therapy after the
    menopause is rare in developing countries, although it is increasing.
    Conjugated equine estrogens at a standard dose of 0.625 mg daily are
    the most widely prescribed preparation for estrogen therapy in women
    in the United States, but estradiol and its esters such as estradiol
    valerate (1-2 mg daily) are more widely used in most of Europe (Table
    16). Conjugated equine estrogens and estradiol compounds yield similar

    plasma concentrations of estradiol and estrone. Progestogens such as
    medroxyprogesterone acetate, levonorgestrel, and norethindrone acetate
    (Table 17) are added to estrogens, typically for 10 days of each
    21-day cycle, in combined regimens. Oral administration is the most
    popular route, although percutaneous methods are becoming commoner;
    use of injections, the first form of postmenopausal estrogen therapy,
    has been declining.

    4.2.2  Human carcinogenicity

    4.2.2.1  Breast cancer

         The relationship between estrogen use and breast cancer risk is
    of paramount importance because of the frequency of this neoplasm, of
    which there are approximately 800 000 new cases per year worldwide,
    representing 21% of all cancers in women (Parkin et al., 1999).
    Information on the relationship between postmenopausal estrogen
    therapy and the risk for breast cancer is available from many
    epidemiological studies (Colditz, 1998; IARC, 1999). A pooled analysis
    of the original data for 52 705 women with breast cancer and 108 411
    women without breast cancer from 51 studies in 21 countries showed a
    small increase in risk in current and recent users with longer
    duration of use (five years or more) (Collaborative Group on Hormonal
    Factors in Breast Cancer, 1997). 

         The relative risk for having breast cancer was 1.02 (95% CI,
    1.01-1.04) for each year of use among current users or those who had
    ceased use one to four years previously and 1.4 (95% CI, 1.2-1.5) for
    women who had used postmenopausal estrogen therapy for five years or
    longer (average duration, 11 years). This increase is comparable with
    the effect on risk for breast cancer of delaying menopause, since the
    risk for women who have never used such preparations is 1.03 (95% CI,
    1.02-1.03) for each additional year at menopause. Five or more years
    after cessation of estrogen use there was no significant excess risk
    for breast cancer overall or in relation to duration of use.

         Of the many factors examined that might affect the relationship
    between risk for breast cancer and estrogen use, only a woman's weight
    and body-mass index had an effect: the increase in the risk for breast
    cancer of current and recent users associated with long duration of
    use was greater for women of lower weight than of higher weight or
    higher body-mass index. There was no marked variation in the results
    according to hormonal type or dose, but little information was
    available about long duration of use of any specific preparation.
    Cancers diagnosed in women who had ever used postmenopausal estrogen
    therapy tended to be less advanced clinically than those diagnosed in
    women who had never used them.

        Table 16. Routes of administration and dosages of commonly used estrogens
                                                                                       

    Estrogen                            Route                 Dosage
                                                                                       

    Natural and equine estrogens

      Conjugated equine estrogens       Oral                  0.3-2.5 mg/day
                                        Intramuscular         25 mg/day
                                        Intravenous           25 mg/day
                                        Vaginal               2-4 g/day

      Piperazine estrone sulfate        Oral                  0.625-2.5 mg/day
                                        Vaginal               2-4 g/day

      Estradiol                         Patch                 0.05-0.1 mg every 
                                                              3-4 days
                                        Vaginal               1-4 g/day
      Micronized, valerate              Oral                  1-2 mg/day
      Valerate                          Intramuscular         10 mg/month
      Cypionate                         Intramuscular         1-5 mg/month

    Synthetic estrogens

      Ethinylestradiol                  Oral                  20-50 g/day
      Mestranol                         Oral                  50-100 g/day
      Diethylstilbestrol                Oral                  1-5 mg/day
      Quinestrol                        Oral                  0.1 mg/day
                                                                                       

    Table 17. Routes of administration and dosages of commonly used progestogens
                                                                                    

    Progestogen                      Route                  Dosage
                                                                                    

    Progesterone

      Suppositories                  Vaginal                25-200 mg/day
                                     Rectal                 25-200 mg/day
      In oil                         Intramuscular          50-200 mg/day
      Micronized                     Oral                   100-300 mg/day

    17 alpha-Hydroxy derivatives

      Medroxyprogesterone acetate    Oral                   2.5-10 mg/day
                                     Intramuscular          250-1000 mg/day or week
      Megestrol acetate              Oral                   20-320 mg/day
      17 alpha-Hydroxyprogesterone   Intramuscular          125-250 mg/week
      caproate

    Table 17. (continued)
                                                                                    

    Progestogen                      Route                  Dosage
                                                                                    
    19-Nortestosterone derivatives

      Ethynodiol diacetate           Oral                   1 mg/day
      Norethindrone                  Oral                   0.35-10 mg/day
      Norethindrone acetate          Oral                   1-10 mg/day
      Norethyndrol                   Oral                   2.5-5 mg/day
      Norgestrel                     Oral                   0.3-0.5 mg/day
      Levonorgestrel                 Oral                   0.075-0.5 mg/day

    Halogenated progesterone

      Cyproterone acetate            Oral                   10-50 mg/day

    Retroprogesterone

      Dydrogesterone                 Oral                   5-20 mg/day
                                                                                    
    
         In Europe and North America, the cumulative incidence of breast
    cancer among women aged 50-70 who have never used postmenopausal
    estrogen therapy is about 45 per 1000. The cumulative excess numbers
    of breast cancers diagnosed per 1000 women of these ages who began use
    of hormones at age 50 and used them for 5, 10, or 15 years were
    estimated to be 2 (95% CI, 1-3), 6 (3-9), and 12 (5-20), respectively
    (Collaborative Group on Hormonal Factors in Breast Cancer, 1997).

    4.2.2.2  Endometrial cancer

         The association with use of postmenopausal estrogen therapy is
    most consistent for endometrial cancer (IARC, 1999). In the United
    States, the incidence of endometrial cancer began to rise in the
    1960s, reached a peak in the mid-1970s, and then declined until the
    1990s. The increased incidence occurred primarily among postmenopausal
    women and followed and then paralleled an increase in use of
    postmenopausal estrogen therapy.

         At least eight cohort and 30 case-control studies have been
    conducted to investigate the impact of postmenopausal estrogen therapy
    on the occurrence of endometrial cancer. Virtually all (6/8 cohort
    studies and 29/30 case-control studies) reported elevated risks
    associated with any use of postmenopausal estrogen therapy, and the
    excess was statistically significant in the majority (IARC, 1999). The
    published results from 30 such studies have been summarised by
    meta-analytic methods (Table 18; Grady et al., 1995). The summary
    relative risk for an increased incidence of this cancer was 2.3 (95%
    CI, 2.1-2.5) for estrogen users when compared to non-users, and a much

    higher relative risk was associated with prolonged duration of use
    (RR = 9.5 for 10 or more years' use); the risk remained elevated five
    or more years after discontinuation of use of unopposed (without the
    addition of progestogen) estrogen therapy (RR = 2.3). Interruption of
    estrogen use for five to seven days per month did not result in a
    lower risk than daily use. Users of unopposed conjugated equine
    estrogens had a greater increase in the relative risk of developing
    endometrial cancer than users of synthetic estrogens. The risk for
    dying from endometrial cancer was increased among users of unopposed
    estrogens (RR = 2.7; 95% CI, 0.9-8). Cohort studies showed a decreased
    risk for endometrial cancer among users of estrogen plus
    progestogen(RR = 0.4), whereas case-control studies showed a small
    increase (RR = 1.8). Information on the risk for endometrial cancer
    among users of estrogen plus progestogen is, however, still limited
    (Grady et al., 1995; IARC, 1999).

         In conclusion, the risk for endometrial cancer increases
    substantially with long duration of use of unopposed estrogens, and
    the increased risk persists for several years after discontinuation of
    use. Although not statistically significant, the risk of users of
    unopposed estrogens of dying from endometrial cancer is also
    increased, so that the association with use of estrogens cannot be
    attributed exclusively to greater medical attention for users.

    4.2.2.3  Cervical cancer

         Only one cohort and two case-control studies are available on the
    relationship between use of postmenopausal estrogen therapy and the
    risk for invasive cervical cancer (IARC, 1999); in none of them were
    the possible confounding effects of oncogenic human papillomaviruses
    (the most important cause of cervical cancer, IARC, 1995) considered.
    On balance, the limited evidence available suggests that
    postmenopausal estrogen therapy is not associated with an increased
    risk for invasive cervical carcinoma. The results provide some
    suggestion that it is associated with a reduced risk for cervical
    cancer, but that may be due to more active screening for preinvasive
    disease among women who have received postmenopausal estrogen therapy.

    4.2.2.4  Ovarian cancer

         The four cohort and 12 case-control studies that addressed the
    risk for ovarian cancer (largely epithelial) among women undergoing
    postmenopausal estrogen therapy gave somewhat variable results (IARC,
    1999). One cohort study and one large case-control study showed a
    significant excess risk for ovarian cancer among women who used this
    therapy. A pooled analysis of individual data from 12 studies in the
    United States based on 2197 cases of invasive epithelial ovarian
    cancer among white women and 8893 white controls (Whittemore et al.,
    1992) gave pooled multivariate relative risks for invasive ovarian
    cancer among women who had ever used postmenopausal estrogen therapy
    for more than three months of 0.9 (95% CI, 0.7-1.3) in comparison with
    hospital-based controls and 1.1 (95% CI, 0.9-1.4) in comparison with

    population-based controls; no consistent duration-risk relationship
    was seen. The relative risk for use for > 15 years was 0.5 (95% CI,
    0.2-1.3) in the hospital-based and 1.5 (95% CI, 0.8-3.1) in the
    population-based studies. The overall trend per year of use was 1.0
    for both types of study, and both risk estimates were nonsignificant.
    Allowance was made in the analysis for age, study, parity, and use of
    combined oral contraceptives. The relative risk in a similar pooled
    analysis of individual data for 327 cases of epithelial ovarian
    tumours of borderline malignancy associated with any use of
    postmenopausal estrogen therapy was 1.1 (95% CI, 0.7-2) (Harris et
    al., 1992). 

         The original data from the four available European studies on
    ovarian cancer (from Greece, Italy, and the United Kingdom) were also
    reassessed. For a total of 1470 ovarian cancer cases and 327 controls,
    a pooled relative risk of 1.7 (95% CI, 1.3-2.2) was found (Negri et
    al., in press). Although this relative risk appears to be higher than
    those reported from North America, it is not incompatible.

        Table 18. Pooled relative risks (RRs) and 95% confidence intervals (CIs)
    derived in a meta-analysis of studies of post-menopausal estrogen therapy
    and endometrial cancer
                                                                            

    Analysis                      RR         95% CI        No. of studies
                                                                            

    Any use of estrogens

      All eligible studies        2.3a       2.1-2.5       29
      Cohort studies              1.7a       1.3-2.1       4
      Case-control studies        2.4a       2.2-2.6       25
      Hospital controls           2.2        2.0-2.5       10
      Gynaecologcial controls     3.3b       2,7-4.0       6
      Community controlsc         2.4a       2.0-2.9       10

    Dose of conjugated
    estrogens (mg)

      0.3                         3,9        1.6-9.5       3
      0.625                       3.4        2.0-5.6       4
      > 1.25                      5.8        4.5-7.5       9

    Duration of use (years)

      < 1                         1.4        1.0-1.8       9
      1-5                         2.8        2.3-3.5       12
      5-10                        5.9        4.7-7.5       10
      > 10                        9.5b       7.4-12        10

    Table 18. (continued)

                                                                            

    Analysis                      RR         95% CI        No. of studies
                                                                            
    Regimen

      Intermittent and cyclic     3.0b       2.4-3.8       8
      Continuous                  2.9b       2.2-3.8       8

    Type of estrogen

      Conjugated                  2.5d       2.1-2.9       9
      Synthetice                  1.3b       1.1-1.6       7

    Time since last use
    (years)

      < 1                         4.1b       2.9-5.7       3
      1-4                         3.7        2.5-5.5       3
      > 5                         2.3        1.8-3.1       5

    Stage of tumour

      0-1                         4.2        3.1-5.7       3
      2-4                         1.4        0.8-2.4       3
      Not invasive                6.2        4.5-8.4       4
      Invasive                    3.8b       2.9-5.1       6

    Death from endometrial        2,7        0.9-8.0       3
    cancer
                                                                            

    a p for homogeneity, < 0.0001

    b p for homogeneity, < 0.01

    c Residential-, neighbourhood- and population-based controls

    d p for homogeneity, < 0.005

    e Primarily ethinylestradiol, estradiol valerate, estriol,
      and other, unspecified estrogens; diethylstilbestrol and
      estrogens combined with androgen were excluded, except
      when such use was included with use of all synthetic estrogens
    
    4.2.2.5  Cancers of the liver and biliary tract

         Two cohort and two case-control studies that addressed the
    association between use of postmenopausal estrogen therapy and the
    risk for cancers of the liver or biliary tract showed no alteration in
    risk.

    4.2.2.6  Colorectal cancer

         Seven cohort and 12 case-control studies provide information on
    use of postmenopausal estrogen therapy and the risk for colorectal
    cancer (Franceschi & La Vecchia, 1998a; IARC, 1999). The risk was not
    increased and appeared to be reduced in half of the studies. In a
    meta-analysis of the 20 independent estimates of the association
    between any use of postmenopausal hormones and colorectal cancer
    published up to December 1996 (Herbert-Croteau, 1998), the summary
    relative risk was 0.85 (95% CI, 0.73-0.99), although there was
    substantial heterogeneity among the studies. The protective effect of
    hormones was found to be stronger in more recent publications. The
    relative risks were lower for current or recent users (RR = 0.69; 95%
    CI,0.52-0.91) and among users of > 5 years (RR = 0.73; 95% CI,
    0.53-1) than for short-term users (RR = 0.88; 95% CI, 0.64-1.2). These
    data suggest that use of postmenopausal estrogen therapy has a
    protective effect against colorectal carcinogenesis. Inadequate
    assessment of exposure, poor control of confounding factors such as
    lifestyle, and changing patterns of use over time possibly contributed
    to the results of studies conducted so far.

    4.2.2.7  Cutaneous malignant melanoma

         One cohort and nine case-control studies addressed the risk for
    cutaneous malignant melanoma in relation to use of postmenopausal
    estrogen therapy. Most suggested no alteration (IARC, 1999).

    4.2.2.8  Thyroid cancer

         Seven case-control studies on thyroid cancer and use of
    postmeno-pausal estrogen therapy showed no clear effect on risk (IARC,
    1999). La Vecchia et al. (1999) carried out a pooled analysis of the
    original data from eight studies on estrogens and thyroid cancer,
    comprising a total of 1305 cases and 2300 controls: 110 (8%) cases and
    205 (9%) controls reported ever having used postmenopausal estrogen
    therapy (RR = 0.78; 95% CI, 0.58-1.0).

    4.2.2.9  Summary and conclusions

         The working group convened by IARC (1999) concluded that there is
    sufficient evidence in humans for the carcinogenicity of
    postmenopausal estrogen therapy, largely based on the clear evidence
    and substantial strength of the association with risk for endometrial
    cancer. The association with breast cancer is weak but consistent with
    biological mechanisms such as the adverse effects of delay in age at

    menopause and obesity in postmenopausal women. Users of postmenopausal
    estrogen therapy had no excess risk for cancers at other sites. The
    evidence for the carcinogenicity of postmenopausal combined
    estrogen-progestogen therapy was deemed to be limited.

    4.2.3  Cardiovascular disease

         Many cohort (Grodstein et al., 1996) and case-control (Rosenberg
    et al., 1993) studies and three meta-analyses (Stampfer & Colditz,
    1991; Grady et al., 1992, US Congress, 1995) have concluded that
    postmenopausal estrogen therapy decreases the risk for coronary heart
    disease by 35-50%. The predicted increase in the life expectancy of
    hormone users, based on estimates from observational studies, is two
    to three years (Grady et al., 1992). In a few studies of the outcomes
    of estrogen therapy, with or without the addition of a progestogen, in
    women with established coronary disease, lower risks were found among
    users for a second infarct, death related to coronary heart disease,
    and a second coronary artery stenosis, due to a reduction in lipids
    and increased fibrinolysis (Petitti, 1998).

         A pooled analysis (Hemminki & McPherson, 1997) of 22 randomized
    trials involving 4124 women who received oestrogen therapy, placebo,
    no therapy, or vitamins and minerals did not support the favourable
    findings of the observational studies. The relative risk for women
    taking hormones compared with those not taking them was 1.4 (95% CI,
    0.48-4) for cardiovascular events without pulmonary embolus and deep
    vein thrombosis and 1.6 (95% CI, 0.65-4.2) for cardiovascular events
    with these aspects. The largest trial so far on the effect of
    postmenopausal therapy with estrogen plus progestoge on the risk for
    coronary heart disease involved 2763 women under the age of 80 (mean
    age, 67) with established coronary disease (Hulley et al., 1998).
    During an average follow-up period of 4.1 years, hormonal treatment
    did not reduce the overall rate of coronary heart disease (RR = 0.99;
    95% CI, 0.80-1.2) but did increase the frequency of thromboembolic
    events and gall-bladder disease. The difference between the findings
    of the observational studies and the trials may reflect some selection
    bias in the former (Petitti, 1998). Trials, however, reflect
    exclusively the short-term effects of postmenopausal estrogen therapy
    on cardiovascular disease, while the long-term effects, reflected in
    observational studies may be substantially different. An early
    thrombogenic effect of estrogen may be counterbalanced by an
    anti-atherogenic effect that occurs with longer duration of use
    (Rosenberg et al., 1993).

    4.2.4  Osteoporosis

         Osteoporosis, a systemic skeletal disease characterized by low
    bone mass and microarchitectural deterioration with a consequent
    increase in bone fragility and susceptibility to fractures, affects an
    estimated 75 million people in developed countries. Evidence that
    postmenopausal estrogen therapy prevents fractures comes from cohort

    and case-control studies, the typical relative risk for non-spinal
    fracture being about 0.7 for women currently taking hormones. The
    beneficial effect of postmenopausal estrogen therapy was more marked
    in women who began therapy within five years after the menopause, and
    it appears to be unaffected by age and concomitant progestogen
    therapy. Evidence from randomized, controlled clinical trials of a
    long-term favourable effect of postmenopausal estrogen therapy is
    scanty (Eastell, 1998).

    4.2.5  Overall mortality

         The estimates of the excess numbers of cases of endometrial
    cancer and, possibly, breast cancer diagnosed in women who used
    postmenopausal estrogen therapy should be considered in the context of
    other effects of such therapy on health (IARC, 1999). Use of
    postmenopausal estrogens affects organs other than the breast and may
    well decrease the incidences of coronary heart disease and
    osteoporotic fractures, although it increases the incidence of venous
    thromboembolism (Petitti, 1998).

         A few studies of mortality in women showed an inverse association
    with postmenopausal estrogen therapy use, most of the relative risks
    ranging from 0.4 to 0.8 (Folsom et al., 1995; Sturgeon et al., 1995;
    Grodstein et al., 1997; Schairer et al., 1997). In a prospective study
    of nurses' health in the United States, current users of hormones had
    a lower risk for death (RR = 0.63; 95% CI, 0.56-0.7) than subjects who
    had never taken them, after adjustment for confounding variables;
    however, the apparent benefit decreased with long-term use (RR = 0.8;
    95% CI, 0.67-0.96 after > 10 years) because of an increase in the
    rate of mortality from breast cancer. Current hormone users with
    coronary risk factors (69% of the women) had the greatest reduction in
    mortality (RR = 0.51; 95% CI, 0.45-0.57), and those at low risk had
    substantially less benefit (RR = 0.89; 95% CI, 0.62-1.3). The rate of
    mortality among women who use postmenopausal hormones is thus lower
    than that of non-users, but the survival benefit diminishes with
    longer duration of use and is lower for women at low risk of coronary
    disease (Grodstein et al., 1997). Since women who use hormones may be
    basically healthier than women who do not (Sturgeon et al., 1995), the
    benefits attributed to use of estrogens in non-randomized studies of
    postmenopausal estrogen users may be overestimated.

    4.3  Hormonal contraceptives

    4.3.1  Exposure

         Oral contraceptives have been used since the early 1960s and are
    now used by about 90 million women worldwide (IARC, 1979, 1999 and
    related references). 'The pill' is given as a combination of an
    estrogen and progestogen or, formerly, as sequential therapy.
    Progestogen-only contraceptives have been available for over 20 years
    (IARC, 1999) and are administered primarily by intramuscular depot
    injections and subcutaneous implants in developing countries, where
    there is widest use. Oral progestogen-only 'mini pills' are used

    primarily in Europe and North America, but fewer women use these
    preparations than parenterally administered progestogens and combined
    oral contraceptives. The hormone content of combined preparations has
    decreased over time, multiphase formulations have been marketed, and
    various progestogens have been introduced (WHO, 1998). Formulations
    containing 30-50 g of estrogen, called 'low-dose' pills, are
    available. Oral contraceptives may be used for emergency post-coital
    contraception, and the components of oral contraceptives are used to
    treat peri-and postmenopausal symptoms and number of other conditions.
    Oral contraceptive preparations containing 50 g of ethinylestradiol
    contain 1, 2.5, 3, or 4 mg of norethisterone, while preparations
    containing 30 g of ethinylestradiol contain 150 or 250 g of
    levonorgestrel. The scientific validity of classifying different oral
    contraceptive formulations by 'potency' is a controversial issue.
    Combined oral contraceptives tend to have progestogenic, estrogenic,
    and anti-oetrogenic effects which vary according to the target organ
    and the hormonal content of the formulation (WHO, 1998). The estrogen
    component of combined oral contraceptives is either ethinylestradiol
    or mestranol and the progestogens used are chlormadinone acetate,
    cyproterone acetate, desogestrel, ethynodiol diacetate, gestodene,
    levonorgestrel, lynestrenol, megestrol, norethisterone, norethisterone
    acetate, norethynodrel, and norgestrel. The estrogen used most
    commonly is ethinylestradiol, and the most commonly used progestogens
    are desogestrel, levonorgestrel, and norethisterone.

         Use of oral contraceptives varies widely (IARC, 1999). They were
    already being used extensively in the Netherlands, Sweden, the United
    Kingdom, and the United States in the 1960s; extensive use of oral
    contraceptives by adolescents was documented in Sweden and the United
    Kingdom as early as 1964. In contrast, very little use of oral
    contraceptives is reported in Japan, the countries of the former
    Soviet Union, and most developing countries. The type of oral
    contraceptives prescribed also differs among countries, and both the
    type of oral contraceptive and the doses of estrogens and progestogens
    have changed between and within countries over time.

         It is important to stress that use of oral contraceptives is a
    relatively recent human activity, and the health benefits and adverse
    effects in women have not yet been followed over a complete
    generation, even though these preparations are some of the most widely
    used and best studied drugs in the world. Women who began using oral
    contraceptives before the age of 20 in the 1960s are only now reaching
    the ages (50-60 years) at which the incidences of most malignancies
    begin to increase.

         Estrogens and progestogens belonging to the same chemical groups
    may have different estrogenic, androgenic, and progestogenic effects.
    Little is known about the long-term health risks and potential
    protective effects of the individual components. The effects become
    increasingly complex as women grow older and may be exposed to
    different types and doses of hormones, starting with oral
    contraceptives and progressing to postmenopausal hormone therapy.

    4.3.2  Human carcinogenicity

    4.3.2.1  Breast cancer

         More than 10 cohort and 50 case-control studies have assessed the
    relationship between use of combined (estrogen plus progestogen) oral
    contraceptives and the risk for breast cancer (IARC, 1999). Individual
    data on 53 297 women with breast cancer and 100 239 women without
    breast cancer from 54 studies conducted in 25 countries were
    collected, checked, and analysed centrally by the Collaborative Group
    on Hormonal Factors in Breast Cancer (1996). The studies included in
    the collaboration represented about 90% of the epidemiological
    information on the topic. The results provided strong evidence of a
    small increase in the relative risk for a diagnosis of breast cancer
    among women who are taking combined oral contraceptives and during the
    10 years after stopping: RR = 1.2; 95% CI, 1.12-1.3 for current users;
    1.2; 95% CI, 1.1-1.2 one to four years after stopping; 1.1; 95% CI,
    1-1.1 five to nine years after stopping. The results also showed no
    significance excess risk of having breast cancer diagnosed 10 or more
    years after stopping use (RR = 1.01; 95% CI, 0.96-1.05; Figure 2).

         There was no pronounced variation in the results for recent use
    among women with different background risks for breast cancer,
    including women from different countries and ethnic groups, women with
    different reproductive histories, and those with or without a family
    history of breast cancer.

         Other features of hormonal contraceptive use such as duration of
    use, age at first use, and the dose and type of hormone in the
    contraceptives had little additional effect on risk once recent use
    had been taken into account. Women who began use before the age of 20
    had higher relative risks for breast cancer being diagnosed while they
    were using combined oral contraceptives and in the five years after
    stopping than women who began use when older, but the higher relative
    risks apply at ages when breast cancer is rare and for a given
    duration of use earlier use does not result in more cancers being
    diagnosed than use begun later.

         Because the incidence of breast cancer rises steeply with age,
    the estimated excess number of cancers diagnosed between starting use
    and 10 years after stopping increases with age at last use: for
    example, among 10 000 women in Europe and North America who used oral
    contraceptives between the ages of 16 and 19, 20 and 24, and 25 and
    29, the estimated excess numbers of cancers diagnosed up to 10 years
    after stopping use were 0.5 (95% CI, 0.3-0.7), 1.5 (0.7-2.3), and 4.7
    (2.7-6.7), respectively. Up to 20 years after cessation of use, the
    difference between women who have ever used oral contraceptives and
    those who have never used them is not so much in the total number of
    cancers diagnosed but in their clinical presentation: breast cancers
    diagnosed in women with any use are less advanced clinically than
    those diagnosed in women with no use, the relative risk for tumours
    that had spread beyond the breast compared with localized tumours

    being 0.88 (95% CI, 0.81-0.95). The relationship observed between the
    risk for breast cancer and use of hormones is unusual, and it is not
    possible to infer from these data whether it is due to earlier
    diagnosis of breast cancer in women who have used oral contraceptives,
    the biological effects of hormonal contraceptives, or a combination
    (Collaborative Group on Hormonal Factors in Breast Cancer, 1996).

         Data on injectable progestogen-only contraceptives were available
    from two case-control studies and a pooled analysis of original data,
    overall involving about 350 women with breast cancer who had used
    these drugs (IARC, 1999). Data on oral progestogen-only contraceptives
    were available from a pooled analysis of original data on 725 women
    with breast cancer. Overall, there is no evidence of an increased risk
    for breast cancer in users of progestogen-only contraceptives
    (Collaborative Group on Hormonal Factors in Breast Cancer, 1996). 

    4.3.2.2  Endometrial cancer

         Three cohort and 16 case-control studies addressed the
    relationship between use of combined oral contraceptives and risk for
    endometrial cancer (IARC, 1999). The results of these studies
    consistently show that the risk for endometrial cancer of women who
    have taken these pills is approximately halved. The reduction in risk
    is generally stronger the longer the oral contraceptives have been
    used and persists for at least 10 years after cessation of use. Few
    data are available for the more recent, low-dose formulations. Use of
    sequential oral contraceptives which were removed from the consumer
    market in the 1970s was associated with an increased risk for
    endometrial cancer. One case-control study addressed the relationship
    between use of oral progestogen-only contraceptives and risk for
    endometrial cancer. Less than 2% of the control women had used these
    preparations, and women with endometrial cancer were less likely to
    have used oral progestogen-only contraceptives than control women but
    not significantly so.

         The effects of use of depot medroxyprogesterone acetate on the
    risk for endometrial cancer have been evaluated in one cohort and one
    case-control study. No reduction in risk was seen in the cohort study,
    whereas a strong reduction was observed in the case-control study.
    Although the evidence is based on small numbers of women, the results
    of these studies suggest that women who use progestogen-only
    contraceptives have a reduced risk for endometrial cancer.

         In biological terms, the favourable effect of oral contraceptive
    use with regard to endometrial cancer is attributed to the daily
    presence of progestogens, whereas in the normal cycle they are present
    for only 14 days.

    FIGURE 4

    Risk (and 95% confidence interval) relative to that of women who have 
    never used combined oral contraceptives, stratified by study, age at
    diagnosis, parity, age at birth of first child, and age at which risk
    for conception ceased


    4.3.2.3  Cervical cancer

         Two cohort and 16 case-control studies of use of combined oral
    contraceptives and invasive cervical cancer have been published (IARC,
    1999); these consistently show relative risks of approximately 1.5-2
    associated with long duration of use. Similar associations were seen
    in four studies in which some analyses were restricted to cases and
    controls who had human papillomavirus infection, the most important
    cause of cervical cancer (IARC, 1995). Biases related to differences
    in sexual behaviour, screening practices, and other factors between
    users and non-users of oral contraceptives cannot be ruled out as
    possible explanations for the observed associations. There is little
    evidence that use of depot medroxyprogesterone acetate or other
    progestational injectable contraceptives alters the risk for either
    squamous-cell carcinoma or adenocarcinoma of the uterine cervix.

    4.3.2.4  Ovarian cancer

         Four cohort and 21 case-control studies addressed the
    relationship between ovarian cancer and use of combined oral
    contraceptives (IARC, 1999). Most of the information available on the
    topic has been summarized in two pooled analyses (Table 19) of studies
    involving 971 cases and 2258 controls in three European countries
    (Franceschi et al., 1991) and 2197 cases and 8893 controls in 12
    studies in the United States (Whittemore et al., 1992), giving a total
    of over 3100 cases and 11 000 controls. In the pooled analysis of
    original data from three hospital-based European studies, the relative
    risk was 0.6 (95% CI, 0.4-0.8) for any use and 0.4 (0.2-0.7) for
    longest (> 5 years) use. Allowance was made in the analysis for age
    and other sociodemographic factors, menopausal status, and parity. The
    protection persisted for at least 15 years after use ceased. In the
    pooled analysis of original data from studies in the United States,
    adjustment was made for age, study, and parity, and the corresponding
    values were 0.7 (95% CI, 0.6-0.8) for any use and 0.3 (0.2-0.4) for
    use for more than six years. Similar results were obtained in
    population-based and hospital-based studies considered separately,
    with a relative risk of 0.7 in both types of study for any use of
    combined oral contraceptives, 0.6 in hospital-based studies and 0.3 in
    population-based ones for use for more than six years, and 0.95
    (nonsignificant) and 0.90 (p < 0.001), respectively, per additional
    year of use.

         An inverse association was also observed in a further analysis of
    seven studies of 110 black cases and 251 black controls from the
    United States pooled analysis, with a relative risk of 0.7 for any use
    and 0.6 for longest use (John et al., 1993). In an analysis of data on
    327 epithelial ovarian neoplasms of borderline malignancy in white
    women, the relative risks were 0.8 (95% CI, 0.6-1) for any use of
    combined oral contraceptives and 0.6 (0.4-0.9) for more than five
    years of use (Harris et al., 1992).


        Table 19. Results of pooled analysis of the association between use of oral contraceptives
    and ovarian cancer
                                                                                                                                            

    Reference,            Type of study          No. of cases      Relative risk (95% CI)                        Comments
    country                                      (age)                                                     
                                                                   Any use          Longest use     Duration 
                                                                                                    (years)
                                                                                                                                            

    Franceschi et al.     Three                  971 (< 65)        0.6 (0.4-0.8)    0.4 (0.2-0.7)   > 5          Protection still present
    (1991), Greece,       hospital-based                                                                         > 15 years after use
    Italy, United                                                                                                ceased (odds ratio, 0.5)
    Kingdom 

    Whittemore et al.     Pooled analysis        2197 (all ages)   0.7 (0.6-0.87)   0.3 (0.2-0.4    > 6          Invasive epithelial 
    (1992),               of 12                                                                                  neoplasms in white women;
    United States         population- and                                                                        protection seen in both
                          hospital-based                                                                         types of study
                          case-control 
                          studies

    Harris et al.         Same as above          327               0.8 (0.6-1.1)    0.6 (0.4-0.9)   > 5          Epithelial tumours of low
    (1992),                                                                                                      malignant potential in
    United States                                                                                                white women
                                                                                                                                            
    

         The most convincing aspect of the inverse relationship found
    between use of combined oral contraceptives and risk for ovarian
    cancer is the consistency of the results, independently of the type of
    study, geographical area (Australia, Europe, North America, and
    developing countries), or type of analysis, although the covariates
    differed from study to study. Likewise, the inverse relationship was
    observed for most formulations considered, including low-dose ones
    (Rosenberg et al., 1994), even though relatively few data are
    available on the more recent formulations.

         One case-control study addressed use of progestogen-only oral
    contraceptives and one case-control study specifically addressed any
    use of depot medroxyprogesterone acetate. Neither showed any
    alteration in risk, either overall or in relation to duration of use.

         In biological terms, the favourable effect of oral contraceptives
    with regard to the risk for ovarian cancer can be interpreted in the
    context of the 'incessant ovulation theory'. Thus, any event such as
    pregnancy or eaqrly menopause that diminishes the number of ovulations
    over a lifetime reduces the risk for ovarian cancer.

    4.3.2.5  Cancers of the liver and biliary tract

         Three cohort studies showed no significant association between
    use of combined oral contraceptives and the incidence of or mortality
    from liver cancer, but the expected numbers of cases were very small,
    resulting in low statistical power. The association between use of
    combined oral contraceptives and liver cancer was evaluated in 11
    case-control studies (IARC, 1999). Three European and two North
    American studies showed relative risks that were significantly
    increased.

         The two largest investigations on the topic were carried out in
    countries at intermediate to high risk for liver cancer, i.e.
    developing countries and southern Europe, where the relative risks for
    use of oral contraceptives for two years or more were 0.2 and 0.8,
    respectively. Use of oral contraceptives does not therefore enhance
    the risk for liver cancer in populations where the prevalence of
    hepatitis B and C is high, but they may be a rare cause of cancer in
    developed countries where the prevalence of hepatitis is low (IARC,
    1999). Positive associations have been reported chiefly with use of
    high-dose formulations.

         Two case-control studies addressed the association between risk
    for liver cancer and use of injectable progestogen-only
    contraceptives. In neither study did the risk for liver cancer differ
    significantly between women who had ever or never used these
    contraceptives. Both studies were conducted in areas endemic for
    hepatitis viruses.

    4.3.2.6  Colorectal cancer

         Four cohort and 10 case-control studies provided information on
    use of combined oral contraceptives and the risk for colorectal cancer
    (IARC, 1999; Franceschi & La Vecchia, 1998b). None showed
    significantly elevated risks in women who used these preparations for
    any duration. Relative risks below 1.0 were found in nine studies, and
    the risk was significantly reduced in two.

    4.3.2.7  Cutaneous malignant melanoma

         Four cohort and 18 case-control studies provided information on
    use of combined oral contraceptives and the risk for cutaneous
    malignant melanoma (IARC, 1999). The relative risks were generally
    close to 1.0 and not related to duration of use. A meta-analysis of
    published data from 18 case-control studies on melanoma, involving
    3796 cases and 9442 controls, yielded a relative risk for any use of
    oral contraceptives of 0.95 (95% CI, 0.87-1). Further analyses in
    different subgroups defined by personal characteristics and study
    design did not materially alter this finding (Gefeller et al., 1997).
    One case-control study of cutaneous malignant melanoma showed no
    increase in risk among users of progestogen-only contraceptives.

    4.3.2.8  Thyroid cancer

         Ten case-control studies published information on use of combined
    oral contraceptives and risk for cancer of the thyroid gland. In
    general there was no elevation in risk associated with oral
    contraceptive use Original data from 13 studies from North America,
    Asia and Europe were reanalysed (La Vecchia et al., 1999). Based on
    2,132 cases of thyroid cancer and 3,301 controls, relative risk for
    ever use of oral contraceptives was 1.15 (95% CI, 0.97-1.37). There
    was no relationship with duration of use, or age at first use, or use
    before first birth. The relative risk was significantly increased for
    current oral contraceptive users (RR = 1.45; 95% CI, 1.02-2.07), but
    declined with increasing time since stopping.

    4.3.2.9  Summary and conclusions

         The working group convened by IARC in 1998 (IARC 1999) concluded
    that there was sufficient evidence in humans for the carcinogenicity
    of combined oral contraceptives. Conversely, the evidence for the
    carcino-genicity of progestogen-only contraceptives was deemed
    inadequate. It is impossible to infer whether the association between
    breast cancer and use of combined oral contraceptives is due to
    earlier diagnosis of breast cancer in users, to the biological effects
    of contraceptives, or to a combination.

    4.3.3  Cardiovascular disease

    4.3.3.1  Acute myocardial infarct

         Myocardial infarct is uncommon in women of reproductive age (WHO,
    1998 and related references). Age, cigarette smoking, diabetes,
    hypertension, and raised cholesterol concentration in blood are
    important risk factors for myocardial infarct in young women.

         After the first reports of a roughly threefold increase in risk
    for myocardial infarct in current users of combined oral
    contraceptives, three cohort and six case-control studies were
    reported which included data collected after 1980. The relative risks
    varied widely, from 0.9 to 5 (WHO, 1998), but six of the studies
    reported elevated relative risks, three of which were statistically
    significant. None of the studies found a relationship between
    myocardial infarct and duration of current use. in nine studies that
    included data on past use of combined oral contraceptives, there was
    no evidence of either an increase or a decrease in the risk for
    myocardial infarct among previous users when compared with women who
    had never used oral contraceptives.

         The available data do not allow an evaluation of the effect of
    dose of estrogen on the relative risk for myocardial infarct
    independently of the type and dose of progestogen. There are
    insufficient data to assess whether the risk of users of low-dose
    combined oral contraceptives is modified by the type of progestogen,
    and the suggestion that users of low-dose combined oral contraceptives
    containing gestodene or desogestrel have a lower risk for myocardial
    infarct than users of low-dose formulations containing levonorgestrel
    remains to be substantiated.

    4.3.3.2  Stroke

         Strokes can be divided into two broad categories, infarcts and
    haemorrhages, according to the nature of the cerebral lesion. The
    relationship between use of combined oral contraceptives and ischaemic
    stroke has been demonstrated in a number of epidemiological studies,
    from as early as 1968 (WHO, 1998). Two cohort and 11 case-control
    studies show a fairly consistent pattern of association, especially if
    greater weight is given to larger studies conducted in areas where the
    prevalence of use of oral contraceptives among women in the control
    group was more than 10%. Most of the studies found that current use of
    combined oral contraceptives was associated with an overall increase
    in the risk for ischaemic stroke that was approximately 3.4 times that
    of non-users. An association with transient ischaemic attacks was also
    reported.

         None of the studies of the relationship between risk for stroke
    and duration of use of contraceptives provided strong evidence for an
    association. Four studies found no significantly elevated risk among
    past users, and two others found that the relative risk for ischaemic
    stroke among former users of combined oral contraceptives was lower

    than that among women who had never used them (WHO, 1998). The risk
    for ischaemic stroke among women who do not smoke, who have their
    blood pressure checked regularly, and who do not have hypertension is
    increased by about 1.5-fold with current (albeit not past) use of
    low-dose combined oral contraceptives when compared with non-users.

         Two recent cohort and four case-control studies examined the risk
    for haemorrhagic stroke in current users and found relative risks of
    1-2. The incidence of fatal and non-fatal haemorrhagic stroke was not
    increased in current users of combined and oral contraceptives who did
    not smoke and did not have hypertension, but the incidence of
    haemorrhagic stroke increases with age, smoking, and hypertension, and
    current use of oral contraceptives appears to magnify these effects.

         Use of progestogen-only oral and injectable contraceptives does
    not appear to increase the relative risk for all types of stroke.

    4.3.3.3  Venous thromboembolism

         Thrombosis arises from the interaction of disturbances of the
    wall of the blood vessels, of the components of blood, and of the
    dynamics of blood flow (stasis). The risk factors for venous
    thromboembolism include pregnancy and puerperium, recent surgery,
    immobilization, obesity, and various chronic diseases. Venous
    thromboembolic disease is the most common cardiovascular event among
    users of oral contraceptives, but it is associated with low rates of
    mortality and long-term disability. Three cohort and six case-control
    studies which included data collected after 1980 found that current
    users of combined oral contraceptives had a risk for thromboembolic
    disease that was three to four times higher than that of women who
    were not using oral contraceptives (WHO, 1998). The risk of current
    users is probably highest in the first year of use but falls to that
    of non-users within three months of stopping use of oral
    contraceptives.

    4.3.4  Overall mortality

         Since use of oral contraceptives probably increases the risks for
    cancers of the breast, cervix uteri, and liver but decreases those for
    cancers of the ovary and endometrium, it is important to evaluate the
    net effect of use of oral contraceptives. This obviously depends on a
    woman's age and the background frequencies of cancers affected by oral
    contraceptives at various sites.

         The risks for developing cancers of the breast, cervix,
    endometrium, ovary, and liver among women in the United States aged
    20-54 and using oral contraceptives was evaluated in a meta-analysis
    of 79 relevant studies (Schlesselman, 1995). Women who had used oral
    contraceptives for eight years were estimated to have a small net
    benefit, with 73 fewer cancer cases (approximately 2%) than women who
    had never used these preparations.

         In a 25-year follow-up of 46 000 British women, half of whom had
    used oral contraceptives in 1968-69 (Beral et al., 1999), the risks
    for dying from all cancer were 1.0 (95% CI, 0.8-1.1) for any use and
    1.3 (1.0-1.6) for use for > 10 years. The relative risk for death
    from all causes combined, adjusted for age, parity, social class, and
    smoking, was 1.0 (95% CI, 0.9-1.1) for any use and 1.1 (95% CI,
    0.9-1.3) for long-term users. Women who had stopped using oral
    contraceptives > 10 years previously had no significant excess or
    deficits of deaths from any specific cause or from all causes combined
    (Beral et al., 1999).

    5.  COMMENTS AND EVALUATION

    5.1  Estradiol-17

         The Committee considered data in the published literature from
    studies on the bioavailability, metabolism, short-term and long-term
    toxicity, reproductive toxicity, geneotoxicity, and carcinogenicity of
    exogenous estrogens administered orally. Numerous reports on studies
    of the use of exogenous estrogens in women were considered, as were
    studies in experimental animals on the mechanisms of action of
    estradiol. The extensive database derived from the results of
    epidemiological studies of women taking oral contraceptive
    preparations containing estrogens or postmenopausal estrogen
    replacement therapy was also used to evaluate the safety of estradiol.

         Estradiol is an 18-carbon steroid and the most potent of the
    natural estrogens. It exerts its biological effects largely by
    receptor-mediated mechanisms, as it binds with high affinity and high
    specificity to intracellular receptors. Binding of estradiol directly
    affects the growth and development of the reproductive tract and
    breast and the appearance of secondary sex characteristics. In
    non-pregnant females, estradiol acts synergistically with progesterone
    during the luteal phase of the menstrual cycle to initiate events
    leading to a new cycle. Continued estradiol production is essential
    for the normal growth and development of the fetus. In addition to its
    effects on reproductive tissues, estradiol is an important metabolic
    hormone, particularly because of its effects on the cardiovascular,
    skeletal, and gastrointestinal systems.

         In general, estradiol is inactive when given orally because it is
    inactivated in the gastrointestinal tract and liver, although
    fine-particle formulations of estradiol are effective when given
    orally and are used therapeutically. The bioavailability of a single
    4-mg dose of fine-particle estradiol administered orally to 14 young
    women was 5% of that of a dose administered intravenously. At least
    60% of the absorbed dose appeared in the serum as estrone and estrone
    sulfate and was available as part of the endogenous pool.

         Estradiol shows little toxicity when given as a single oral dose.
    Few conventional short-and long-term studies of the systemic toxicity
    of estradiol in animals treated orally were available, but there is
    sufficient information to demonstrate that the adverse effects of
    estradiol seen in animals are associated with estrogenic activity.
    Because of the specificity and affinity with which estradiol binds to
    its receptors, the hormonal effects occur at much lower doses than
    other toxicological responses and hence are the most appropriate for
    use in evaluating the safety of the compound.

         In studies of developmental toxicity in rats, all embryos were
    resorbed when estradiol was administered subcutaneously at a dose
    equivalent to 25 mg/kg bw on day 10 of gestation. In an interim report
    of a multigeneration study of reproductive toxicity in female rats,
    estradiol was administered in the feed a a dose of 0.003, 0.17, 0.69,
    or 4.1 mg/kg bw per day. No viable pups were observed at the two
    higher doses. As the concen tration of progesterone was altered at
    various times in the treated groups, a NOEL was not identified in this
    study.

         The Committee reviewed studies of the genotoxic potential of
    estradiol. The compound did not cause gene mutations  in vitro. In
    some other assays, sporadic but unconfirmed positive results were
    obtained. There was more consistent evidence for the induction of
    micronuclei  in vitro, aneuploidy  in vitro, cell transformation
     in vitro, oxidative damage to DNA  in vivo, and DNA single-strand
    breakage  in vivo by estradiol. The Committee concluded that
    estradiol has genotoxic potential.

         A normal biological function of estrogens is to increase the
    number of proliferating cells in the endometrium and breast. This
    effect is exerted by binding with high affinity to estrogen receptors.
    These receptors, of which there are several forms, are found in many
    tissues. In cultured human breast cancer cells containing estrogen
    receptors, estradiol stimulated growth when added at concentrations of
    10-12 mol/L and above, with a maximal response at about 10-10 mol/L.
    Estradiol does not stimulate the growth of cultured human breast
    cancer cells that do not contain estrogen receptors. At higher
    concentrations, estrogens also stimulate cell proliferation in rat
    liver in vivo and in vitro. Any factor that increases mitotic
    activity reduces the time available for repair of DNA damage before
    the next cell division. An agent that causes cell proliferation may
    not, however, induce the mutagenic events that are required for
    neoplasia. If receptor-mediated stimulation of cell growth is an
    important mechanism in the induction of neoplasia by estradiol,
    late-stage carcinogenic activity would be expected to predominate.
    Experimental studies in rodents in which estradiol was administered in
    conjunction with known carcinogens support this mechanism, as do
    observations of an increased incidence of cancer among women taking
    postmenopausal hormone replacement therapy. In long-term studies of

    carcinogenicity in animals, revieiwed at the thirty-second meeting,
    oral and parenteral administration of estradiol increased the
    incidences of tumours only in hormone-dependent tissues, including the
    kidneys of male Syrian hamsters. The Committee concluded that the
    carcinogenicity of estradiol is most probably a result of its
    interaction with hormonal receptors.

         The commonest uses of estrogens in humans are for oral
    contraception and postmenopausal hormone replacement therapy. For oral
    contraception, the xenobiotic estrogen ethinylestradiol is usually
    used. Fine-particle estradiol and conjugated equine estrogen
    preparations are commonly used in postmenopausal estrogen replacement
    therapy. When healthy postmeno-pausal women were given four courses of
    0.3, 0.62, 1.2, or 2.5 mg/day of conjugated equine estrogens for two
    weeks followed by no medication for three weeks, the NOEL was 0.3
    mg/day on the basis of changes in the serum concentrations of
    corticosteroid-binding globulin (CBG). These results indicate that
    there is a threshold concentration of estrogen administered orally,
    below which there is no increase in serum concentrations of CBG. In a
    study involving 23 healthy postmenopausal women receiving various
    estrogen preparations orally, 0.3 mg/day of conjugated equine
    estrogens or fine-particle estradiol had no effect on the serum
    concentrations of follicle-stimulating hormone, angiotensinogen, sex
    hormone-binding globulin, or CBG. Thus, 0.3 mg/day, equivalent to 5
    g/kg bw per day, was the NOEL for these hormonal effects of
    estradiol. A further indication that this is the NOEL is that 0.3
    mg/day of estradiol administered orally to women did not relieve
    symptoms of the menopause.

         A study of approximately 7700 infants whose mothers had reported
    taking oral contraceptives while pregnant with them showed no evidence
    that estrogens present a teratogenic hazard.

         Because of differences in the pharmacokinetics and
    pharmaco-dynamics of natural and xenobiotic estrogen preparations, the
    Committee concluded that data on the use of estrogens for
    postmenopausal hormone replacement therapy are more appropriate for
    evaluating the safety of estradiol than data on their use for oral
    contraception. Epidemiological studies on women who took estrogens,
    either alone or in combination with progestogens and androgens, showed
    that the risks for cancers at most sites were unaffected; the risks
    for cancers of the endometrium and breast were increased. In a
    meta-analysis of 30 epidemiological studies of women who had ever used
    postmenopausal estrogen-only therapy, the relative risk for
    endometrial cancer was 2.3 (95% confidence interval, 2.1-2.5). The
    relative risk of women who had taken postmenopausal estrogen-only
    therapy for more than 10 years was 9.5 (95% confidence interval,
    7.4-12). The addition of progestogens to postmenopausal estrogen
    therapy reduces the excess risk substantially, althoug it may not be
    completely eliminated. In a review of 51 epidemiological studies of
    women taking hormonal replacement therapy, the relative risk for

    breast cancer was increased by a factor of 1.023 (95% confidence
    interval, 1.011-1.036) for each year of use. These estimates of
    relative risk are based on the results of studies of women who used
    postmenopausal estrogen replacement therapy preparations containing
    either conjugated equine estrogens (average dose, 0.625 mg/day) or
    estradiol   (1-2 mg/day). Overall, the available data suggest that the
    increased incidences of cancers of the breast and endometrium observed
    among women receiving postmenopausal estrogen replacement therapy is
    due to the hormonal effects of estrogens.

         See also section 4, 'Epidemiological studies of women exposed to
    postmenopausal estrogen therapy and hormonal contraceptives'.

    5.2  Progesterone

         The Committee considered published data from studies on the
    bioavailability, metabolism, short-term toxicity, reproductive
    toxicity, genotoxicity, and the long-term toxicity and carcinogenicity
    of orally administered progesterone. Numerous reports of studies on
    progesterone in humans were considered. In addition, the extensive
    database derived from studies of women taking progestogens as a
    component of oral contraception, as injectable progestogen-only
    contraception, and in postmenopausal hormone replacement therapy was
    used to support the safety evaluation.

         Progesterone is a 21-carbon steroid that is the only important
    natural progestogen. Its normal role is to prepare the uterus for
    implantation and to maintain pregnancy. Continued production of
    progesterone is necessary to maintain pregnancy. The production of
    progesterone by the corpus luteum is controlled by release of
    luteinizing hormone from the pituitary gland. In non-pregnant females,
    an elevated concentration of progesterone inhibits the cyclic release
    of luteinizing hormone, and higher levels inhibit production of
    follicle-stimulating hormone. Progesterone opposes some of the effects
    of estrogens, but prior stimulation with estrogens is essential for
    progesterone to elicit biological responses. Progesterone binds with
    high afinity and high specificity to an intracellular receptor
    protein; binding of progesterone activates the receptor, resulting in
    activation of specific genes.

         Less than 10% of progesterone is bioavailable after oral
    administration, as it is inactivated in the gastroinetestinal tract
    and/or the liver. Progesterone has little toxicity when given as a
    single oral dose.

         No conventional studies of toxicity in animals treated with
    progesterone orally were available, and few other studies were found.
    Because progesterone binds specifically and with high affinity to its
    receptor, the hormonal effects are the most sensitive toxicological
    end-points in these studies. The results of studies of toxicity after
    administration by other routes suggest that the effects seen in
    animals are associated with hormonal activity. Although equivocal

    results have been reported for the induction of single-strand DNA
    breaks and DNA adducts have been seen  in vitro and  in vivo in some
    studies, progesterone was not mutagenic. The Committee concluded that,
    on balance, progesterone has no genotoxic potential.

         Mouse pups given subcutaneous injections of 100 g of
    progesterone on five consecutive days (equivalent to 200 mg/kg bw per
    day), beginning 36 h after birth and observed for up to one year had
    an increased incidence of mammary gland tumours. Female rabbits given
    an average dose of 8 mg/kg bw intramuscularly every second week for
    two years developed endometrial cysts, which were sometimes associated
    with atypical hyperplasia, but no significant changes were observed in
    other tissues.

         No multigeneration study of reproductive toxicity with
    progesterone was available. Developmental toxicity was not seen in
    studies in rats and rhesus monkeys. Rats dosed at 5-25 mg/kg bw per
    day on days 14-19 of gestation delivered pups that showed no evidence
    of masculinization. Rhesus monkeys given progesterone at 5 mg/kg bw
    per day intramuscularly on five days per week beginning after one
    month of gestation and continuing to delivery had healthy offspring
    with no evident abnormalities. The Committee noted that exogenous
    progesterone has been used to maintain pregnancy, with no evidence of
    toxicity and with no effect on the normal outcome of pregnancy.

         The commonest uses of progesterone by humans are in contraception
    and in postmenopausal hormone replacement therapy in which synthetic
    progestogens are used either alone or in combination with estrogens.
    In a study designed to explore anti-proliferatory and secretory
    end-points in the endometrium, women received fine-particle
    progesterone at 300 or 600 mg/day (300 mg twice a day) orally for two
    weeks after pretreatment with estrogens for 30 days. The group
    receiving 300 mg/day showed incomplete conversion of the uterus to
    full secretory activity, while the group receiving 600 mg/day showed
    full secretory conversion of the uterus with suppression of mitotic
    activity. In studies in which women were given 200 or 300 mg of
    progesterone orally for one or five years, there was no evidence of
    endometrial hyperplasia or carcinoma. Oral administration of a single
    dose of 200 mg of fine-particle progesterone (equivalent to 3.3 mg/kg
    bw) to women provided concentrations in blood similar to those found
    during the luteal phase of the ovulatory cycle. This dose was
    considered to be the lowest-observed-effect level (LOEL) in humans.

         There is extensive literature on the use of synthetic
    progestogens in combination with estrogens for oral contraception.
    While synthetic progestogens differ from natural products in their
    pharmacokinetics, pharmacodynamics, tissue specificity, and potency, a
    prominent feature of the synthetic agents is a protective effect
    against the untoward effects of estrogens. No increase in the risk for
    cancer at any site was found in women taking only progesterone or

    other progestogens orally for contraception. When used in combination
    with estrogens in postmenopausal replacement therapy, progesterone
    reduced the excess risk for endometrial cancer found with estrogen
    alone but did not alter the increased risk for breast cancer.

         See also section 4, 'Epidemiological studies of women exposed to
    postmenopausal estrogen therapy and hormonal contraceptives'.

    5.3  Testosterone

         The Committee considered published data from experimental studies
    on the bioavailability, metabolism, short-term toxicity, reproductive
    toxicity, genotoxicity, and the long-term toxicity and carcinogenicity
    of testosterone administered orally. Reports of studies of human use
    were also considered.

         Testosterone is a 19-carbon steroid which has potent androgenic
    properties, including maintenance of testicular function and growth
    and differentiation of secondary sex characteristics. It exerts its
    biological effects through receptor-mediated mechanisms, as it binds
    with high affinity and high specificity to an intracellular receptor
    protein, the androgen receptor, which is consequently activated,
    resulting in activation of specific genes. In certain target tissues,
    testosterone is metabolized to 5 alpha-dihydrotestosterone, which has
    greater binding affinity for the androgen receptor.

         Androgens have marked anabolic effects which include increased
    protein synthesis in muscle and bone. This results in an increased
    rate of body growth. In females, androgens have actions in the breast,
    uterus, and vagina that are similar to those of progestogens.
    Luteinizing hormone and follicle-stimulating hormone from the
    pituitary gland control the production of testosterone by the testes.
    Testosterone in turn modulates the concentration of
    follicle-stimulating hormone and luteinizing hormone, thus controlling
    the circulating levels of testosterone through a feedback mechanism.

         Little testosterone is bioavailable after oral administration, as
    it is inactivated in the gastrointestinal tract and the liver. After
    oral administration of 25 mg of testosterone to young women,
    approximately 4% of the dose was found to be bioavailable.

         Testosterone has very little toxicity when given as a single oral
    dose. Few studies have been conducted of the toxicity of testosterone
    in animals treated by the oral route. Short-and long-term studies in
    animals demonstrate that its adverse effects are due to its hormonal
    activity. Therefore, the most sensitive toxicological targets are
    hormone-sensitive tissues such as the prostate. Six adult male baboons
    received intramuscular injections of testosterone enanthate to provide
    a dose equivalent to 8 mg/kg bw each week for up to 28 weeks. At the
    end of the study, histological evidence of non-neoplastic alterations
    was found in the prostate. Female rabbits given an average dose of

    testosterone equivalent to 6 mg/kg bw intramuscularly every second
    week for two years developed endometrial cysts and secretions from the
    mammary gland. No significant changes were found in tissues from
    organs other than those of the reproductive system.

         In studies of developmental toxicity, testosterone was
    embryotoxic; in rats, a dose of testosterone equivalent to 25 mg/kg bw
    administered as subcutaneous implants on day 10 of gestation resulted
    in resorption of all embryos. No multigeneration studies of
    reproductive toxicity have been conducted with testosterone, since
    while normal circulating concentrations are required for normal
    reproductive function in males eleveated concentrations of
    testosterone interfere with normal reproductive function in both males
    and females.

         In mammalian cells, no chromosomal aberrations, mutations, or DNA
    adducts were found with testosterone alone, and the Committee
    concluded that testosterone has no genotoxic potential. No studies to
    investigate the carcinogenicity of testosterone in experimental
    animals had become available since the previous evaluation. The
    Committee re-affirmed that the increased rate of prostatic cancer
    detected in rats is consistent with the hormonally mediated effects of
    testosterone and its metabolites.

         In men, the physiological concentrations of circulating
    testosterone are 3-10 ng/ml. In women, the value is less than 1 ng/ml,
    and most of the circulating testosterone is derived from the
    conversion of dehydroepiandrosterone and androsterone from adrenal and
    ovarian sources. Androgens are used therapeutically in men with
    deficient testicular function to restore normal testosterone levels.
    The effects of excess androgens, particularly in young boys, include
    deepened voice, acne, and growth of facial hair, while in women hair
    loss and menstrual irregularities may be seen. In a clinical trial
    involving five eunuchs, an oral dose of 100 mg/day of a fine-particle
    formulation of testosterone had no effect on sexual function, while an
    oral dose of 400 mg/day was effective in restoring full sexual
    function. Thus, oral administration of 100 mg/day (equivalent to 1.7
    mg/kg bw per day) was the NOEL in this study. In studies in which
    postmenopausal women received the testosterone analogue
    methyltestosterone, alone or in combination with estrogens, a dose of
    10 mg/day induced virilizing signs such as acne and hirsutism in a
    sizeable proportion of the women. The effects were dose-and
    time-dependent. The Committee noted that methyltestosterone is more
    potent than testosterone when given orally.

         No epidemiological studies of long-term treatment of humans were
    available. Therapeutic doses of testosterone given for the treatment
    of aplastic anaemia or hypogonadism have resulted in the induction of
    liver cysts and hepatomas.

         See also section 4, 'Epidemiological studies of women exposed to
    postmenopausal estrogen therapy and hormonal contraceptives'.

    6  EVALUATION

    6.1  Estradiol-17

         The Committee established an ADI of 0-50 ng/kg bw on the basis of
    the NOEL of 0.3 mg/day (equivelent to 5 g/kg bw per day) in studies
    of changes in sev eral hormone-dependent parameters in postmenopausal
    women. A safety factor of 10 was used to account for normal variation
    among individuals, and an additional factor of 10 was added to protect
    sensitive populations.

    6.2  Progesterone

         The Committee established an ADI of 0-30 g/kg bw for
    progesterone on the basis of the LOEL of 200 mg/day (equivalent to 3.3
    mg/kg bw) for changes in the uterus. A safety factor of 100 was used
    to allow for extrapolation from a LOEL to a NOEL and to account for
    normal variation among individuals.

    6.3  Testosterone

         The Committee established an ADI of 0-2 g/kg bw for testosterone
    on the basis of the NOEL of 100 mg/day (equivalent to 1.7 mg/kg bw per
    day) in the study of eunuchs and a safety factor of 1000. The large
    safety factor was used in order to protect sensitive populations and
    because of the small number of subjects in the study from which the
    NOEL was identified.

    7.  REFERENCES

    Aizu-Yokota, E., Susaki, A. & Sato, Y. (1995) Natural estrogens induce
    modulation of microtubules in Chinese hamster V79 cells in culture.
     Cancer Res., 55, 1863-1868.

    Aldercreutz, H., Gorbach, S.L., Goldin, B.R., Woods, M.N., Dwyer, J.T.
    & Hamalainen, E. (1994) Estrogen metabolism and excretion in oriental
    Caucasian women.  J. Natl Cancer Inst., 86, 1076-1082.

    Angsusingha, K., Kenny, F.M., Nankin,H.R. & Taylor, F.H. (1974)
    Unconjugated estrone, estradiol and FSH and LH in prepubertal and
    pubertal males and females.  J. Clin. Endocrinol. Metab., 39,
    63-68.

    Ashby, J., Fletcher, K., Williams, C., Odum, J. & Tinwell, H. (1997)
    Lack of activity of estradiol in rodent bone marrow micronucleus
    assays.  Mutat. Res., 395, 83-88.

    Banerjee, S.K., Banerjee, S., Li, S.A. & Li, J.J. (1994) Induction of
    chromosome aberrations in Syrian hamster renal cortical cells by
    various estrogens.  Mutat. Res., 311, 191-197.

    Baulieu, E.E. (1997) Neurosteroids: Of the nervous system, by the
    nervous system, for the nervous system.  Recent Prog. Horm. Res.,
    52, 1-32.

    Behl, C., Widmann, M., Trapp, T. & Holsboer, F. (1995) 17-Estradiol
    protects neurons from oxidative stress-induced cell death.  Biochem.
     Biophys. Res. Commun., 216, 473-482.

    Beral, V., Hernon, C., Kay, C., Hannaford, P., Darby, S. & Reeves, G.
    (1999) Mortality associated with oral contraceptive use: 25 year
    follow up of cohort or 46000 women from Royal College of General
    Practitioners' oral contraception study.  Br. Med. J., 318, 96-100.

    Bernstein, L. & Ross, R.K. (1993) Endogenous hormones and breast
    cancer risk.  Epidemiol. Rev., 15, 48-65.

    Beyer, B.K., Stark, K.L., Fantel, A.G. & Juchau, M.R. (1989)
    Biotransformation, estrogenicity, and steroid structure as
    determinants of dysmorphogenic and generalized embryotoxic effects of
    steroidal and nonsteroidal estrogens.  Toxicol. Appl. Pharmacol.,
    98, 113-127.

    Bhat, H.K., Hacker, H.J., Bannasch, P., Thompson, E.A. & Liehr, J.G.
    (1993) Localization of estrogen receptors in interstitial cells of
    hamster kidney and in estradiol-induced renal tumors as evidence of
    the mesenchymal origin of this neoplasm.  Cancer Res., 53,
    5447-5451.

    Biegel, L.B., Flaws, J.A., Hirshfield, A.N., O'Connor, J.C., Elliot,
    G.S., Ladics, G.S., Silbergold, E.K., Van Pelt, C.S., Hurtt, M.E.,
    Cook, J.C. & Frame, S.R. (1998a) 90-day feeding study and
    one-generation reproduction study in Crl:CD BR rats with 17
    alpha-estradiol.  Toxicol. Sci., 44, 116-142.

    Biegel, L.B., Cook, J.C., Hurtt, M.E. & O'Connor, J.C. (1998b) Effects
    of 17 beta-estradiol on serum hormone concentrations and estrous cycle
    in female Crl:CD BR rats: Effects on parental and first generation
    rats.  Toxicol. Sci., 44, 143-154.

    Blum-Degen, D., Haas, M., Pohli, S., Harth, R., Romer, W., Oettel, M.,
    Riederer, P. & Gotz, M.E. (1998) Scavestrogens protect IMR 32 cells
    form oxidative stress-induced cell death. Toxicol.  Appl. Pharmacol.,
    152, 49-55.

    Bourgain, C., Devroey, P., Van Waesverghe, L., Smitz, J. & van
    Steirteghem, A.C. (1990) Effects of natural progesterone on the
    morphology of the endometrium in patients with primary ovarian
    failure.  Hum. Reprod., 5, 537-543.

    Braunstein, G.D. (1994) Testes. In: Greenspan, F.S. & Baxter, J.D.,
    eds,  Basic and Clinical Endocrinology, 4th Ed., Norwalk,
    Connecticut, Appleton & Lange, pp. 391-418.

    Butler, G.E., Sellar, R.E., Walker, R.F., Kelnar, C.J.H. & Wu, F.C.W.
    (1992) Oral testosterone undecanoate in the management of delayed
    puberty in boys: Pharmacokinetics and effect on sexual maturation.
     J. Clin. Endocrinol. Metab., 75, 37-44.

    Butterworth, M., Lau, S.S. & Monks, T.J. (1996) 17-Estradiol
    metabolism by hamster hepatic microsomes: Implications for the
    catechol-O-methyltransferase-mediated detoxication of catechol
    estrogens.  Drug. Metab. Disposition, 24, 588-594.

    Carr, B.R. (1992) Fertilization, implantation, and endocrinology of
    pregnancy. In: Griffin, J.E. & Ojeda, S.R., eds,  Textbook of
     Endocrine  Physiology, 2nd Ed., New York, Oxford University Press,
    pp.189-209.

    Carr, B.R. (1998) Disorders of the ovary and female reproductive
    tract. In: Wilson, J.D., Foster, D.W., Kronenberg, H.M. & Larsen,
    P.R., eds,  Williams Textbook of Endocrinology, 9th Ed.,
    Philadelphia, W.B. Saunders Co., pp. 751-817.

    Carr, B.R. & Griffin, J.E. (1998) Fertility control and its
    implications. In: Wilson, J.D., Foster, D.W., Kronenberg, H.M. &
    Larsen, P.R., eds,  Williams Textbook of Endocrinology, 9th Ed.,
    Philadelphia, W.B. Saunders Co., pp. 901-925.

    Cavalieri, E.L., Stack, D.E., Devanesan, P.D., Todorovic, R., Dwivedy,
    I., Higginbotham, S., Johansson, S.L., Patil, K.D., Gross, M.L.,
    Gooden, J.K., Ramanathan, R., Cerny, R.L. & Rogan, E.G. (1997)
    Molecular origin of cancer: Catechol estrogen-3,4-quinones as
    endogenous tumor initiators.  Proc. Natl Acad. Sci. USA, 94,
    10937-10942.

    Colditz, G.A. (1998) Relationship between estrogen levels, use of
    hormone replacement therapy, and breast cancer.  J. Natl Cancer
     Inst., 90, 814-823.

    Collaborative Group on Hormonal Factors in Breast Cancer (1996) Breast
    cancer and hormonal contraceptive reanalysis of individual data on 53
    297 women with breast cancer and 100 239 women without breast cancer
    from 54 epidemiological studies.  Lancet, 347, 1713-1727.

    Collaborative Group on Hormonal Factors in Breast Cancer (1997) Breast
    cancer and hormone replacement therapy: Collaborative reanalysis of
    data from 51 epidemiological studies of 52 705 women with breast
    cancer and 108 411 women without breast cancer.  Lancet, 350,
    1047-1059.

    Concas, A., Mostallino, M.C., Porcu, P., Follesa, P., Barbaccia, M.L.,
    Trabucchi, M., Purdy, R.H., Grisenti, P. & Biggio, G. (1998) Role of
    brain allopregnanolone in the plasticity of gamma-aminobutyric acid
    type A receptor in rat brain during pregnancy and after delivery.
     Proc. Natl Acad. Sci. USA, 95, 13248-13289.

    Coppoc, G.L., Bottoms, G.D., Monk, E., Moore, A.B. & Roesel, O.F.
    (1982) Metabolism of estrogens in the gastrointestinal tract of swine.
    II. Orally administered gestadiol-17-D-glucuronide.  J. Anim. Sci.,
    55, 135-144.

    Couse, J.F., Lindzey, J., Grandien, K., Gustafsson, J.A. & Korach,
    K.S. (1997a) Tissue distribution and quantitative analysis of estrogen
    receptor-alpha (ERalpha) and estrogen receptor-beta (ERbeta) messenger
    ribonucleic acid in the wild-type ERalpha-knockout mouse.
     Endocrinology, 138, 4613-4621.

    Couse, J.F., Davis, V.L. & Korach, K.S. (1997b) Physiological findings
    from transgenic mouse models with altered levels of estrogen receptor
    expression. In: Pavlik, E.J., ed.,  Estrogens, Progestogens and Their
     Antagonists, Boston, Birkhauser, Vol. 2, pp. 69-98.

    Cui, L., Mori, T., Takahashi, S., Imaida, K., Akagi, K., Yada, H.,
    Yaono, M. & Shirai, T. (1998) Slight promotion effects of intermittent
    administration of testosterone propionate and/or diethylstilbestrol on
    3,3'-dimethyl-4-aminobiphenyl-initiated rat prostate carcinogenesis.
     Cancer Lett., 9, 195-199.

    Das, S.K., Talor, J.A., Korach, K.S., Paria, B.C., Dey, S.K. & Lubahn,
    D.B. (1997) Estrogenic responses in estrogen receptor-a deficient mice
    reveal a distinct estrogen signaling pathway.  Proc. Natl Acad. Sci.
     USA, 94, 12786-12791.

    Dhillon, V.S. & Dhillon, I.K. (1995) Genotoxicity evaluation of
    estradiol.  Mutat. Res., 345, 87-95.

    Dorgan, J.F., Albanes, D., Virtamo, J., Heinonen, O.P., Chandler,
    D.W., Galmarini, M., McShane, L.M., Barrett, M.J., Tangrea, J. &
    Taylor, P.R. (1998) Relationship of serum androgens and estrogens to
    prostate cancer risk: Results from a prospective study in Finland.
     Cancer Epidemiol. Biomarkers Prev., 7, 1069-1074.

    Eastell, R. (1998) Treatment of postmenopausal osteoporosis. 
     N. Engl. J. Med., 338, 736-746

    Endoh, A., Natsume, H. & Igarashi, Y. (1995) Dual regulation of
    21-hydroxylase activity by sex steroid hormones in rat hepatocytes.
     J. Steroid Biochem. Mol. Biol., 54, 163-165.

    Farthing, M.J.G., Vinson, G.P., Edwards, C.R.W. & Dawson, A.M. (1982)
    Testosterone metabolism by the rat gastrointestinal tract,  in vitro
    and  in vivo.  Gut, 23, 226-234.

    Feser, W., Kerdar, R.S., Blode, H. & Reimann, R. (1996) Formation of
    DNA-adducts by selected sex steroids in rat liver.  Hum. Exp.
     Toxicol., 15, 556-562.

    Fisher, D.A. (1998) Endocrinology of fetal development. In: Wilson,
    J.D., Foster, D.W., Kronenberg, H.M. & Larsen, P.R., eds,  Williams
     Textbook of Endocrinology, 9th Ed., Philadelphia, W.B. Saunders Co.,
    pp. 1273-1301.

    Fisher, C.R., Graves, K.H., Parlow, A.F. & Simpsor, E.R. (1998)
    Characterization of mice deficient in aromatase (ArKO) because of
    targeted disruption in cyp19 gene.  Proc. Natl Acad. Sci. USA, 95,
    6965-6970.

    Folsom, A.R., Mink, P.J., Sellers, T.A., Hong, C.P., Zheng, W. &
    Potter, J.D. (1995) Hormonal replacement therapy and morbidity and
    mortality in a prospective study of postmenopausal women.  Am. J.
     Public Health, 85, 1128-1132.

    Foss, G.L. & Camb, M.B. (1939) Clinical administration of androgens.
     Lancet, i,502-504.

    Fotis, T., Zhang, Y., Pepper, M.S., Aldercreutz, H., Montesano, R.,
    Nawroth, P.P. & Schwelgerer, L. (1994) The endogenous estrogen
    metabolite 2-methoxyestradiol inhibits angiogenesis and suppresses
    tumor growth.  Nature, 368, 237-239.

    Franceschi, S. & La Vecchia, C. (1998a) Colorectal cancer and hormone
    replacement therapy: An unexpected finding.  Eur. J. Cancer Prev.,
    7, 427-438.

    Franceschi, S. & La Vecchia, C. (1998b) Oral contraceptives and
    colorectal tumors.  Contraception, 58, 335-343.

    Franceschi, S., Parazzini, F., Negri, E., Booth, M., La Vecchia, C.,
    Beral, V., Tzonou, A. & Trichopoulos, D. (1991) Pooled analysis of 3
    Euopean case-control studies of epithelial ovarian cancer. III. Oral
    contraceptives use.  Int. J. Cancer, 49, 61-65.

    Frishman, G.N., Klock, S.C., Luciano, A.A. & Nulsen, J.C. (1995)
    Efficacy of oral micronized progesterone in the treatment of luteal
    phase defects.  J. Reprod. Med., 40, 521-524.

    Gangrade, N.K., Boudinot, F.D. & Price, J.C. (1992) Pharmacokinetics
    of progesterone in ovariectomized rats after single dose intravenous
    administration.  Biopharm. Drug Disposition, 13, 703-709.

    Gefeller, O., Hassan, K. & Wille, L. (1997) A meta-analysis on the
    relationship between oral contraceptives and melanoma: Results and
    methodological aspects.  J. Epidemiol. Biostat., 2, 225-235.

    Gelfand, M.M. & Wiita, B. (1997) Androgen and estrogen-androgen
    hormone replacement therapy: A review of the safety literature,
    1941-1996.  Clin. Ther., 19, 383-404.

    Goldfien, A. & Monroe, S.E. (1994) Ovaries. In: Greenspan, F.S. &
    Baxter, J.D., eds,  Basic and Clinical Endocrinology, 4th Ed.,
    Norwalk, Connecticut, Appleton & Lange, pp. 419-470.

    Grady, D., Rubin, S.M., Petitti, D.B., Fox, C.S., Black, D., Ettinger,
    B., Ernster, V.L. & Cummings, S.R. (1992) Hormone therapy to prevent
    disease and prolong life in postmenopausal women.  Ann. Intern. Med.,
    117, 1016-1037.

    Grady, D., Gebretasadik, T., Kerlikowske, K., Ernster, V. & Petitti,
    D. (1995) Hormone replacement therapy and endometrial cancer risk: A
    meta-analysis.  Obstet. Gynecol., 85, 304-313.

    Griffin, J.E. & Wilson, J.D. (1998) Disorders of the testes and the
    male reproductive tract. In: Wilson, J.D., Foster, D.W., Kronenberg,
    H.M. & Larsen, P.R., eds,  Williams Textbook of Endocrinology, 9th
    Ed., Philadelphia, W.B. Saunders Co., pp. 819-875.

    Grodstein, F., Stampfer, M.J., Manson, J.-A.E., Colditz, G.A.,
    Willett, W.C., Rosner, B., Speizer, F.E. & Hennekens, C.H. (1996)
    Postmenopausal estrogen and progestin use and the risk of
    cardiovascular disease.  New Engl. J. Med., 335, 453-461.

    Grodstein, F., Stampfer, M.J., Colditz, G.A., Willett, W.C., Manson,
    J.-A.E., Joffe, M., Rosner, B., Fuchs, C., Hankinson, S.E., Hunter,
    D.J., Hennekens, C.H. & Speizer, F.E. (1997) Postmenopausal hormone
    therapy and mortality.  New Engl. J. Med., 336, 1769-1775.

    Grumbach, M.M. & Styne, D.M. (1998) Puberty: Ontogeny,
    neuroendocrinology, physiology and disorders. In: Wilson, J.D.,
    Foster, D.W., Kronenberg, H.M. & Larsen, P.R., eds,  Williams
     Textbook of Endocrinology, 9th Ed., Philadelphia, W.B. Saunders
    Co., pp. 1509-1625.

    Hammond, D.K., Zhu, B.T., Wang, M.Y., Ricci, M.J. & Liehr, J.G. (1997)
    Cytochrome P540 metabolism of estradiol in hamster liver and kidney.
     Toxicol. Appl. Pharmacol., 145, 54-60.

    Han, X. & Liehr, J.G. (1994a) DNA single strand breaks in the kidneys
    of Syrian hamsters treated with steroidal estrogens: Hormone-induced
    free radical damage preceeding renal malignancy.  Carcinogenesis,
    15, 997-1000.

    Han, X. & Liehr, J.G. (1994b) 8-Hydroxylation of guanine bases in
    kidney and liver DNA of hamsters trested with estradiol: Role of free
    radicals in estrogen-induced carcinogenesis.  Cancer Res., 54,
    5515-5517.

    Han, X., Liehr, J.G. & Bosland, M.C. (1995) Induction of a DNA adduct
    detectable by 32P-postlabeling in the dorsolateral prostate of NBL/Cr
    rats treated with estradiol-17b and testosterone.  Carcinogenesis,
    16, 951-954.

    Hargrove, J.T., Maxson, W.S. & Wentz, A.C. (1989) Absorption of oral
    progesterone is influenced by vehicle and particle size.  Am. J.
     Obstet. Gynecol., 161, 948-951.

    Harris, R., Whittemore, A.S., Intyre, J. & the Collaborative Ovarian
    Cancer Group (1992) Characteristic relating to ovarian cancer risk:
    Collaborative analysis of 12 US case-control studies.  Am. J.
     Epidemiol., 136, 1204-12011.

    Hayashi, N., Hasegawa, K., Komine, A., Tanaka, Y., McLachian, J.A.,
    Barrett, J.C. & Tsutsui, T. (1996) Estrogen-induced cell
    transformation and DNA adduct formation in cultured Syrian hamster
    embryo cells.  Mol. Carcinog., 16, 149-156.

    Helzlsouer, K.J., Huang, H.Y., Strickland, P.T., Hoffman, S., Alberg,
    A.J., Comstock, G.W. & Bell, D.A. (1998) Association between CYP17
    polymorphisms and the development of breast cancer.  Cancer
    Epidemiol.  Biomarkers Prev., 7, 945-949.

    Hemminki, E. & McPherson, K. (1997) Impact of postmenopausal hormone
    therapy on cardiovascular events and cancer: Pooled data from clinical
    trials.  Br. Med. J., 315, 149-153.

    Her, C., Szumlanski, C., Aksoy, I.A. & Weinshilboum, R.M. (1996) Human
    jejunal estrogen sulfotransferase and dehydroepiandrosterone
    sulfotrans-ferase.  Drug Metab Disposition, 24, 1328-1335.

    Herbert-Croteau, N. (1998) A meta-analysis of hormone replacement
    therapy and colon cancer among women.  Am. J. Epidemiol., 147, 87.

    Highman, B., Norvell, M.J. & Shellenberger, T.E. (1978) Pathological
    changes in female C3H mice continuously fed diets containing
    diethylstilbestrol or 17beta-estradiol.  J. Environ. Pathol.
     Toxicol., 1, 1-30.

    Highman, B., Greenman, D.L., Norvell, M.J., Farmer, J. &
    Shellenberger, T.E. (1980) Neoplastic and preneoplastic lesions
    induced in female C3H mice by diets containing diethylstilbestrol or
    17beta-estradiol.  J. Environ. Pathol. Toxicol., 4, 81-95.

    Ho, S.M. & Roy, D. (1994) Sex hormone-induced nuclear DNA damage and
    lipid peroxidation in the dorsolateral prostates of Noble rats.
     Cancer  Lett., 84, 155-162.

    Hoogerbrugge, N., Zillikens, M.C., Jansen, H., Meeter, K., Deckers,
    J.W. & Birkenhager, J.C. (1998) Estrogen replacement decreases the
    level of antibodies against oxidized low-density lipoprotein in
    postmenopausal women with coronary heart disease.  Metabolism, 47,
    675-680.

    Hryb, D.J., Khan, M.S., Romas, N.A. & Rosner, W. (1990) The control of
    the interaction of sex hormone-binding globulin with its receptor by
    steroid hormones.  J. Biol. Chem., 265, 6048-6054.

    Hsu, M.H., Griffin, K.J., Wang, Y., Kemper, B. & Johnson, E.F. (1993)
    A single amino acid substitution confers progesterone 6
    beta-hydroxylase activity to rabbit cytochrome P4502C2.  J. Biol.
     Chem., 268, 6939-6944.

    Huang, Z., Guengerich, F.P. & Kaminsky, L.S. (1998)
    16-alpha-Hydroxylation of estrone by human cytochrome P4503A4/5.
     Carcinogenesis, 19, 867-872.

    Hulley, S., Grady, D., Bush, T., Furberg, C., Herrington, D., Riggs,
    B. & Vittinghoff, E. for the Heart and Estrogen/Progestin Replacement
    Study (HERS) Research Group. (1998) Randomized trial of estrogen plus
    progestin for secondary prevention of coronary heart disease in
    postmenopausal women.  J. Am. Med. Assoc., 280, 605-613.

    IARC (1979)  IARC Monographs on the Evaluation of the Carcinogenic
    Risk  of Chemicals to Humans, Vol. 21,  Sex Hormones (II), Lyon,
    IARCPress, pp. 35-82, 139-547.

    IARC (1987)  IARC Monographs on the Evaluation of the Carcinogenic
    Risk  of Chemicals to Humans, Suppl. 7,  Overall Evaluations of
     Carcinogenicity: An Updating of IARC Monographs  Volumes 1 to 42,
    Lyon, IARCPress, pp. 272-310, 395-400. 

    IARC (1995)  IARC Monographs on the Evaluation of Carcinogenic Risks
    to  Humans, Vol. 64,  Human Papillomaviruses, Lyon, IARCPress.

    IARC (1999)  IARC Monographs on the Evaluation of Carcinogenic risks
    to  Humans, Vol. 72,  Hormonal Contraception and Postmenopausal
    Hormone  Therapy Lyon, IARCPress.

    Inoh, A., Kamiya, K., Fujii, Y. & Yokoro, K. (1985) Protective effects
    of progesterone and tamoxifen in estrogen-induced mammary
    carcinogenesis in ovariectomized W/Fu rats.  Jpn. J. Cancer Res.,
    76, 699-704.

    John, E.M., Whittermore, A.S., Harris, R., Intyre, J. & the
    Collaborative Ovarian Cancer Group (1993) Characteristics relating to
    ovarian cancer risk: Collaborative analysis of seven US case-control
    studies. Epithelial ovarian cancer in black women.  J. Natl Cancer
     Inst., 85, 142-147.

    Johnsen, S.G., Bennett, E.P. & Jensen, V.G. (1974) Therapeutic
    effectiveness of oral testosterone.  Lancet, ii, 1473-1475.

    Johnsen, S.V., Kampmann, J.P., Bennett, E.P. & Jorgensen, F.S. (1976)
    Enzyme induction by oral testosterone.  Clin. Ther., 20, 233-237.

    Jones, L.A. & Bern, H.A. (1977) Long-term effects of neonatal
    treatment with progesterone alone and in combination with estrogen, on
    the mammary gland and reproductive tract of female BALB/cfc3H mice.
     Cancer Res., 37, 67-75.

    Karr, J.P., Kim, U., Resko, J.A., Schneider, S., Chai, L.S., Murphy,
    G.P. & Sandburg, A.A. (1984) Induction of benign prostatic hypertrophy
    in baboons.  Urology, 3, 276-289.

    Kedderis, G.L. & Mugford, C.A. (1998) Sex-dependent metabolism of
    xenobiotics.  CIIT Activities, 18, 1-7.

    Kim, S., Korhonen, M., Wilborn, W., Foldesy, R., Snipes, W., Hodgen,
    G.D. & Anderson, F.D. (1996) Antiproliferative effects of low-dose
    micronized progesterone.  Fertil. Steril., 65, 323-331.

    Krege, J.H., Hodgin, J.B., Couse, J.F., Enmark, E., Warner, M.,
    Mahler, J.F., Sar, M., Korach, K.S., Gustafsson, J.A. & Smithies, O.
    (1998) Generation and reproductive phenotypes of mice lacking estrogen
    receptor .  Proc. Natl Acad. Sci. USA, 95, 15677-15682.

    Kuhnz, W., Gansau, C. & Mahler, M. (1993) Pharmacokinetics of
    estradiol, free and total estrone, in young women following single
    intravenous and oral administration of 17beta-estradiol.
     Arzeimittelforschung, 43, 966-973.

    Kuiper, G.G.J.M., Lemmen, J.G., Carlsson, B., Corton, J.C., Safe,
    S.H., van der Saag, P.T., van der Burg, B. & Gustafsson, J.-A. (1998)
    Interaction of estrogenic chemicals and phytoestrogens with estrogen
    receptor .  Endocrinology, 139, 4252-4263.

    Lacroix, D., Sonnier, M., Moncion, A., Cheron, G. & Cresteil, T.
    (1997) Expression of CYP3A in the human liver--Evidence that the shift
    between CYP3A7 and CYP3A4 occurs immediately after birth.  Eur. J.
     Biochem., 247, 625-634.

    Lane, K.E., Ricci, M.J. & Ho, S.M. (1997) Effect of combined
    testosterone and estradiol-17 treatment on the metabolism of E2 in
    the prostate and liver of Noble rats.  Prostate, 30, 256-262.

    Larsen, M.C., Angus, W.G.R., Brake, P.B., Eltom, S.E., Sukow, K.A. &
    Jefcoate, C.R. (1998) Characterization of CYP1B1 and CYP1A1 expression
    in human mammary epithelial cells: Role of the aryl hydrocarbon
    receptor in polycyclic aromatic hydrocarbon metabolism.  Cancer Res.,
    58, 2366-2374.

    Lasne, C., Lu, Y.P., Orfila, L., Ventura, L. & Chouroulinkov, I.
    (1990) Study of various transforming effects of the anabolic agents
    trenbolone and testosterone on Syrian hamster embryo cells.
     Carcinogenesis, 11, 541-547.

    La Vecchia, C., Ron, E., Franceschi, S., Dal Maso, L., Mark, S.D.,
    Chatenoud, L., Braga, C., Preston-Martin, S., McTiernan, A., Kolonel,
    L., Mabuchi, K., Yin, F., Wingren, G., Galanti, M.R., Hallquist, H.,
    Lund., E., Levi, F., Linos, D. & Negri, E. (1999) A pooled analysis of
    case-control studies of thyroid cancer. III. Oral contraceptives,
    hormonal replacement therapy and other female hormones.  Cancer
     Causes Control, 10, 157-166.

    Lavigne, J.A., Helzlsouer, K.J., Huang, H.Y., Strickland, P.T., Bell,
    D.A., Selmin, O., Watson, M.A., Hoffman, S., Comstock, G.W. & Yager,
    J.D. (1997) An association between the allele coding for a low
    activity variant of catechol-O-methyltransferase and the risk of
    breast cancer.  Cancer Res., 57, 5493-5497.

    Leighton, J.K. & Wei, L.L. (1998) Progesterone and development. In:
    Dickson, R.B. & Salomon, D.S., eds,  Hormones and Growth Factors in
     Development and Neoplasia, New York, Wiley-Liss, pp. 177-190.

    Lewis, S.J., Heaton, K.W., Oakey, R.E. & McGarrigle, H.H. (1997) Lower
    serum oestrogen concentrations associated with faster intestinal
    transit.  Br. J. Cancer, 76, 395-400.

    Lewis, S.J., Oakey, R.E. & Heaton, K.W. (1998) Intestinal absorption
    of oestrogen: The effect of altering transit-time.  Eur. J.
     Gastroenterol. Hepatol., 10, 33-39.

    Li, J.J. & Li, S.A. (1989) Estrogen carcinogenesis in Syrian hamster
    tissues: Role of metabolism.  Fed. Proc., 46, 1858-1863.

    Li, J.J., Li, S.A., Klicka, J.K., Parsons, J.A. & Lam, L.K.T. (1983)
    Relative carcinogenic activity of various synthetic and natural
    estrogens in the Syrian hamster kidney.  Cancer Res., 43,
    5200-5204.

    Li, J.J., Li, S.A., Oberly, T.D. & Parsons, J.A. (1995) Carcinogenic
    activities of various steroidal and nonsteroidal estrogens in the
    hamster kidney: Relation to hormonal activity and cell proliferation.
     Cancer Res., 55, 4347-4351.

    Li, S.A., Liao, D.Z.J., Yazlovitskaya, E.M., Pantazis, C.G. & Li, J.J.
    (1997) Induction of cathepsin D protein during estrogen
    carcinogenesis: Possible role in estrogen-mediated kidney tubular cell
    damage.  Carcinogenesis, 18, 1375-1380.

    Li, J.J., Hou, X., Bentel, J., Yaslovitskaya, E.M. & Li, S.A. (1998)
    Prevention of estrogen carcinogenesis in the hamster kidney by ethinyl
    estradiol: Some unique properties of a synthetic estrogen.
     Carcinogenesis, 19, 471-477.

    Liao, D.Z., Pantazis, C.G., Hou, X. & Li, S.A. (1998) Promotion of
    estrogen-induced mammary gland carcinogenesis by androgen in the male
    Noble rat: Probable mediation by steroid receptors.  Carcinogenesis,
    19, 2173-2180.

    Liehr, J.G. & Ricci, M.J. (1996) 4-Hydroxylation of estrogens as
    marker of human mammary tumors.  Proc. Natl Acad. Sci. USA, 93,
    3294-3296.

    Liehr, J.G., Fang, W.F., Sirbasku, D.A. & Ari-Ulubelen, A. (1986)
    Carcinogenicity of catecholestrogens in Syrian hamsters.  J. Steroid
     Biochem., 24, 353-356.

    Liehr, J.G., Roy, D. & Gladek, A. (1989) Mechanism of inhibition of
    estrogen-induced renal carcinogenesis in male Syrian hamsters by
    vitamin C.  Carcinogenesis, 10, 1983-1988.

    Liehr, J.G., Ricci, M.J., Jefcoate, C.R., Hannigan, E.V., Hokanson,
    J.A. & Zhu, B.T. (1995) 4-Hydroxylation of estradiol by human uterine
    myometrium and myoma microsomes: Implications for the mechanism of
    uterine tumorigenesis.  Proc. Natl Acad. Sci. USA, 92, 9220-9224.

    de Lignieres, B. (1999) Oral micronized progesterone.  Clin. Ther.,
    21, 41-60.

    Linehan, W.M., Shipley, W.U. & Parkinson, D.R. (1997) Cancer of the
    kidney and ureter. In: DeVita, V.T., Jr, Hellman, S. & Rosenberg,
    S.A., eds,  Cancer: Principles and Practice of Oncology, 5th Ed.,
    Philadelphia, Lippincott-Raven, pp. 1271-1300.

    Mahendroo, M.S., Cala, K.M., Landrum, D.P. & Russell, D.W. (1997)
    Fetal death in mice lacking 5alpha-reductase type I caused by estrogen
    excess.  Mol. Endocrinol., 11, 917-927.

    Mahesh, V.B., Brann, D.W. & Hendry, L.B. (1996) Diverse modes of
    action of progesterone and its metabolites.  J. Steroid Biochem. Mol.
     Biol., 56, 209-219.

    Malayer, J.R. & Gorski, J. (1993) An integrated model of estrogen
    receptor action.  Domest. Anim. Endocrinol., 10, 159-177.

    Markides, C.S.A., Roy, D. & Liehr, J.G. (1998) Concentration
    dependence of prooxidant and antioxidant properties of
    catecholestrogens.  Arch. Biochem. Biophys., 360, 105-112.

    Martelli, A., Mereto, E., Ghia, M., Orsi, P., Allavena, A., De
    Pascalis, C.R. & Brambilla, G. (1998) Induction of micronucli and
    enzyme-altered foci in liver of female rats exposed to progesterone
    and three synthetic progestins.  Mutat. Res., 9, 33-41.

    Mashchak, C.A., Lobo, R.A., Dozono-Takano, R., Eggena, P., Nakamura,
    R.M., Brenner, P.F. & Mishell, D.R., Jr (1982) Comparison of
    pharmaco-dynamic properties of various estrogens.  Am. J. Obstet.
     Gynecol., 144, 511-518.

    Maxson, W.S. & Hargrove, J.T. (1985) Bioavailability of oral
    micronized progesterone.  Fertil. Steril., 44, 622-626.

    McAllister, J.M., Kerin, J.F.P., Trant, J.M., Estabrook, R.W., Mason,
    J.I., Waterman, M.R. & Simpson, E.R. (1989) Regulation of cholesterol
    side-chain cleavage and 17 alpha-hydroxylase/Lyase activities in
    proliferating human theca interna cells in long term monolayer
    culture.  Endocrinology, 125, 1959-1966.

    Meissner, W.A. & Sommers, S.C. (1966) Endometrial changes after
    prolonged progesterone and testosterone administration to rabbits.
     Cancer Res., 26, 474-478.

    Mellon, S.H. & Miller, W.L. (1989) Extraadrenal steroid
    21-hydroxylation is not mediated by P450c21.  J. Clin. Invest., 84,
    1497-1502.

    Michnovicz, J.J. & Bradlow, H.L. (1990) Induction of estradiol
    metabolism by dietary indole-3-carbinol in humans.  J. Natl Cancer
     Inst., 82, 947-949.

    Michnovicz, J.J., Hershcopf, R.J., Haley, N.J., Bradlow, H.L. &
    Fishman, J. (1989) Cigarette smoking alters hepatic estrogen
    metabolism in men: Implications for athersclerosis.  Metabolism,
    38, 537-541.

    Millikan, R.C., Pittman, G.S., Tse, C.K., Duell, E., Newman, B.,
    Savitz, D., Moorman, P.G., Boissy, R.J. & Bell, D.A. (1998)
    Catechol-O-methyltransfe-rase and breast cancer risk.
     Carcinogenesis, 19, 1943-1947.

    Miyamoto, H., Yeh, S., Lardy, H., Messing, E. & Chang, C. (1998)
    Delta5-androstenediol is a natural hormone with androgenic activity in
    human prostate cells.  Proc. Natl Acad. Sci. USA, 15, 11083-11088.

    Moore, D.E., Kawagoe, S., Davajan, V., Nakamura, R.M. & Mishell, D.R.
    (1978) An  in vivo system in man for quantitation of estrogenicity.
    II. Pharmacologic changes in binding capacity of serum
    corticosteroid-binding globulin induced by conjugated estrogens,
    mestranol, and ethinyl estradiol.  Am. J. Obstet. Gynecol., 130,
    482-486.

    Moore, A.B., Bottoms, G.D., Coppoc, R.C., Pohland, R.C. & Roesel, O.F.
    (1982) Metabolism of estrogens in the gastrointestinal tract of swine.
    1. Instilled estradiol.  J. Anim. Sci., 55, 124-134.

    Moss, R.L., Gu, Q. & Wong, M. (1997) Estrogen: Nontranscriptional
    signaling pathway.  Recent Prog. Horm. Res., 52, 33-69.

    Moyer, D.L., de Lingnieres, B., Driguez, P. & Pez, J.P. (1993)
    Prevention of endometrial hyperplasia by progesterone during long-term
    estradiol replacement: Influence of bleeding pattern and secretory
    changes.  Fertil. Steril., 59, 992-997.

    Musarrat, J., Arezina-Wilson, J. & Wani, J.J. (1996) Prognostic and
    aetiolo-gical relevance of 8-hydroxyguanosine in human breast
    carcinogenesis.  Eur. J. Cancer, 32A, 1209-1214.

    Nagashima, M., Tsuda, H., Takenoshita, S., Nagamachi, Y., Hirohashi,
    S., Yokota, J. & Kasai, H. (1995) 8-Hydroxydeoxyguanosine levels in
    DNA of human breast cancers are not significantly different from those
    of non-cancerous breast tissues by the HPLC-ECD method.  Cancer
     Lett., 90, 157-162.

    Nagata, K., Murayama, N., Miyata, M., Shimada, M., Urahashi, A.,
    Yamozoe, Y. & Kato, R. (1996) Isolation and characterization of a new
    rat P450 (CYP3A18) cDNA encoding P450(6)beta-2 catalyzing testosterone
    6 beta-and 16 alpha-hydroxylations.  Pharmacogenetics, 6, 103-111.

    Nagelberg, S.B., Laue, L., Loriaux, D.L., Liu, L. & Sherins, R.J.
    (1986) Cerebrovascular accident associated with testosterone therapy
    in a 21-year-old hypogonadal man.  New Engl. J. Med., 314, 649-650.

    Nahoul, K., Dehennin, L., Jondet, M. & Roger, M. (1993) Profiles of
    plasma estrogens, progesterone and their metabolites after oral or
    vaginal administration of estradiol or progesterone.  Maturitas,
    16, 185-202.

    Negri, E., Tzonou, A., Beral, V., Lagiou, P., Trichopoulos, D.,
    Parazzini, F., Franceschi S., Booth, M. & La Vecchia C. (in press)
    Hormonal therapy for menopause and ovarian cancer in a collaborative
    re-analysis of European studies.  Int. J. Cancer

    Niwa, T., Yabusaki, Y., Honma, K., Matsuo, N., Tatsuta, K., Ishibashi,
    F. & Katagiri, M. (1998) Contribution of human hepatic cytochrome P450
    isoforms to regioselective hydroxylation of steroid hormones.
     Xenobiotica, 28, 539-547.

    Oberley, T.D., Gonzalez, A., Lauchner, L.J., Oberley, L.W. & Li, J.J.
    (1991) Characterization of early kidney lesions in estrogen-induced
    tumors of the Syrian hamster.  Cancer Res., 51, 1922-1929.

    Ofner, P., Bosland, M.C. & Vena, R.L. (1992) Differential effects of
    diethylstilbestrol and estradiol-17 in combination with testosterone
    on rat prostate lobes.  Toxicol. Appl. Pharmacol., 112, 300-309.

    O'Malley, B.W., Tsai, S.Y., Bagchi, M., Weigel, N.L., Schrader, W.T. &
    Tsai, M.J. (1991) Molecular mechanisms of action of a steroid hormone
    receptor.  Recent Prog. Horm. Res., 47, 1-26.

    Orth, D.N. & Kovacs, W.J. (1998) The adrenal cortex. In: Wilson, J.D.,
    Foster, D.W., Kronenberg, H.M. & Larsen, P.R., eds,  Williams
     Textbook of Endocrinology, 9th Ed., Philadelphia, W.B. Saunders
    Co., pp. 517-664.

    Osterling, J., Fuks, Z., Lee, C.T. & Scher, H.I. (1997) Cancer of the
    prostate. In: DeVita, V.T., Jr, Hellman, S. & Rosenberg, S.A., eds,
     Cancer: Principles and Practice of Oncology, 5th Ed., Philadelphia,
    Lippincott-Raven, pp.1322-1386.

    Pacifici, G.M., Gucci, A. & Giuliani, L. (1997) Testosterone
    sulphation and glucuronidation in the human liver: Interindividual
    variability.  Eur. J. Drug Metab. Pharmacokinet., 22, 253-258.

    Paquette, B. (1996) Enhancement of genomic instability by
    17beta-estradiol in minisatellite sequences of X-ray-transformed mouse
    10T1/2 cells.  Carcinogenesis, 17, 1221-1225.

    Paria, B.C., Chakraborty, C. & Dey, S.K. (1990) Catechol estrogen
    formation in the mouse uterus and its role in implantation.  Mol.
     Cell Endocrinol., 69, 25-32.

    Parkin, D.M., Pisani, P. & Ferlay, J. (1999) Estimates of the
    worldwide incidence of twenty-five major cancers in 1990.  Int. J.
     Cancer

    Petitti, D.B. (1998) Hormone replacement therapy and heart disease
    prevention. Experimentation trumps observation.  J. Am. Med. Assoc.,
    280, 650-651.

    Philip, A. & Murphy, B.E. (1986) Relative binding of certain steroids
    of low polarity to human sex-hormone binding globulin: Strong binding
    of 2-methoxyestrone, a steroid lacking the 17beta-OH group.
     Steroids, 47, 373-379.

     Physicians' Desk Reference (1999), 53rd Ed., pp. 830-833, 3124-3126.

    Pohland, R.C., Coppoc, G.L., Bottoms, G.D. & Moore, A.B. (1982)
    Metabolism of estrogens in the gastrointestinal tract of swine. III.
    Estradiol-17-D-glucuronide instilled into sections of intestine.
     J. Anim. Sci., 55, 145-152.

    Punnenon, R. & Salmi, T. (1983) Effects of a massive single oral dose
    of oestradiol valerate in a young woman.  Ann. Clin. Res., 15,
    134-136.

    Renoir, J.-M., Mercier-Bodard, C. & Baulieu, E.-E. (1980) Hormonal and
    immunological aspects of the phylogeny of sex steroid binding plasma
    protein.  Proc. Natl Acad. Sci. USA, 77, 4578-4582.

    Reventos, J., Sullivan, P.M. Josheh, D.R. & Gordon, J.W. (1993)
    Tissue-specific expression of the rat androgen-binding protein/sex
    hormone-binding globulin gene in transgenic mice.  Mol. Cell
     Endocrinol., 96, 69-73.

    Revesz, C., Chappel, C.I. & Gaudry, R. (1959) Masculinization of
    female fetuses in the rat by progestational compounds.
     Endocrinology, 66, 140-144.

    Rifici, V.A. & Kachadurian, A.K. (1992) The inhibition of low-density
    lipoprotein oxidation by 17- estradiol.  Metabolism, 41,
    1110-1114.

    Rosenberg, L., Palmer, J.R. & Shapiro, S. (1993) A case-control study
    of myocardial infacton in relation to use of estrogen supplements.
     Am. J. Epidemiol., 137, 54-63.

    Rosenberg, L., Palmer, J.R., Zauber, A.G., Washauer, M.E., Lewis,
    J.L., Jr, Strom, B.L., Harlap, S. & Shapiro, S. (1994) A case-control
    study of oral contraceptive use and invasive epithelial ovarian
    cancer.  Am. J. Epidemiol., 139, 654-661.

    Rosner, W. (1991) Plasma steroid-binding proteins.  Endocrinol.
     Metab. Clin. North Am., 20, 697-720.

    Ross, R.K., Pike, M.C., Coetzee, G.A., Reichardt, J.K.V., Yu, M.C.,
    Feigelson, H., Stanczyk, F.Z., Kolonel, L.N. & Henderson, B.E. (1998)
    Androgen metabolism and prostate cancer: Establishing a model of
    genetic susceptibility.  Cancer Res., 58, 4497-4505.

    Rothman, K.J. & Louik, C. (1978) Oral contraceptives and birth
    defects.  New Engl. J. Med., 210, 522-524.

    Roy, D., Weisz, J. & Liehr, J.G. (1990) The O-methylation of
    4-hydroxyestradiol is inhibited by 2-hydroxyestradiol: Implications
    for estrogen-induced carcinogenesis.  Carcinogenesis, 11, 459-462.

    Ruoff, W.L. & Dziuk, P.J. (1994) Absorption and metabolism of
    estrogens from the stomach and duodenum of pigs.  Domest. Anim.
     Endocrinol., 11, 197-208.

    Russo, I.H. & Russo, J. (1996) Mammary gland neoplasia in long-term
    rodent studies.  Environ. Health Perspectives, 104, 938-967.

    vom Saal, F.S., Timms, B.G., Montano, M.M., Palanza, P., Thayer, K.A.,
    Nagel, S.C., Dhar, M.D., Ganjam, V.K., Parmigiani, S. & Welshons, W.V.
    (1997) Prostate enlargement in mice due to fetal exposure to low doses
    of estradiol or diethylstilbestrol and opposite effects at high doses.
     Proc. Natl Acad. Sci. USA, 94, 2056-2061.

    Sands, R. & Studd, J. (1995) Exogenous androgens in postmenopausal
    women.  Am. J. Med., 98 (Suppl. 1A), 76S-79S.

    Sarabia, S.F. & Liehr, J.G. (1998) Induction of monoamine oxidase B by
    17beta-estradiol in the hamster kidney preceding carcinogenesis.
     Arch.  Biochem. Biophys., 355, 249-253.

    Sarabia, S.F., Zhu, B.T., Kurosawa, T., Tohma, M. & Liehr, J.G. (1997)
    Mechanism of P450-catalyzed aromatic hydroxylation of estrogens.
     Chem.  Res. Toxicol., 10, 767-771.

    Sarkar, K., Kinson, G.A. & Rowsell, H.C. (1986) Embryo resorption
    following administration of steroidal compounds to rats in
    mid-pregnancy.  Can. J. Vet. Res., 50, 433-437.

    Savas, U., Carstens, C.P. & Jefcoate, C.R. (1997) Biological
    oxidations and P450 reactions. Recombinant mouse CYP1B1 expressed in
     Escherichia coli exhibits selective binding by polycyclic hydrocarbons
    and metabolism which parallels C3H10T1/2 cell microsomes, but differs
    from human recombinant CYP1B1.  Arch. Biochem. Biophys., 347,
    181-192.

    Schairer, C., Adami, H.-O., Hoover, R. & Persson, I. (1997)
    Cause-specific mortality in women receiving hormone replacement
    therapy.  Epidemiology, 8, 59-65.

    Schleicher, F., Tauber, U., Louton, T. & Schunack, W. (1998) Tissue
    distribution of sex steroids: Concentration of 17-oestradiol and
    cyproterone acetate in selected organs of female Wistar rats.
     Pharmacol. Toxicol., 82, 34-39.

    Schlesselman, J.J. (1995) Net effect of oral contraceptive use on the
    risk of cancer in women in the United States.  Obstet. Gynecol.,
    85, 793-801.

    Schneider, J., Huh, M.M., Bradlow, H.L. & Fishman, J. (1984)
    Antiestrogen action of 2-hydroxy estrone on MCF-7 breast cancer cells.
     J. Biol. Chem., 259, 4840-4845.

    Schuler, M., Hasegawa, L., Parks, R., Metzler, M. & Eastmond, D.A.
    (1998) Dose-response studies of the induction of hyperdiploidy and
    polyploidy by diethylstilbestrol and 17-estradiol in cultured human
    lymphocytes using multicolor fluorescense  in situ hybridization.
     Environ. Mol. Mutag., 31, 263-273.

    Schultze, N., Vollmer, G., Wunsche, W., Grote, A., Feit, B. & Knuppen,
    R. (1994) Binding of 2-hydroxyestradiol and 4-hydroxyestradiol to the
    estrogen receptor of MCF-7 cells in cytosolic extracts and in nuclei
    of intact cells.  Exp. Clin. Endocrinol., 102, 399-405.

    Scott-Moncrieff, J.C., Nelson, R.W., Bill, R.L., Natlock, C.L. &
    Bottoms, G.D. (1990) Serum disposition of exogenous progesterone after
    intramuscular administration in bitches.  Am. J. Vet. Res., 51,
    893-895.

    Seraj, M.J., Umemoto, A., Tanaka, M., Kajikawa, A., Hamada, K. &
    Monden, Y. (1996) DNA adduct formation by hormonal steroids  in
     vitro. Mutat. Res., 370, 49-59.

    Shangold, M.M., Tomai, T.P., Cook, J.D., Jacobs, S.L., Zinaman, M.J.,
    Chin, S.Y. & Simon, J.A. (1991) Factors associated with withdrawal
    bleeding after administration of oral micronized progesterone in women
    with secondary amenorrhea.  Fertil. Steril., 56, 1040-1047.

    Shelby, M.D., Tice, R.R. & Witt, K.L. (1997) 17b-Estradiol fails to
    induce micronucli in the bone marrow cells of rodents.  Mutat. Res.,
    395, 89-90.

    Shou, M., Korzekwa, K., Brooks, E.N., Krausz, K.W., Gonzalez, F.J. &
    Gelboin, H.V. (1997) Role of human hepatic cytochrome P450 1A2 and 3A4
    in the metabolic activation of estrone.  Carcinogenesis, 18,
    207-214.

    Simon, J.A., Robinson, D.E., Andrews, M.C., Hildebrand, J.R., III,
    Rocci, M.L., Jr, Blake, R.E. & Hodgen, G.D. (1993) The absorption of
    oral micronized progesterone: The effect of food, dose
    proportionality, and comparison with intramuscular progesterone.
     Fertil. Steril., 60, 26-33.

    Sitruk-Ware, R., Bricaire, C., de Lignieres, B., Yaneva, H. &
    Mauvais-Jarvis, P. (1987) Oral micronized progesterone.
     Contraception, 36, 373-402.

    Smith, S.S., Gong, Q.H., Hsu, F.C., Markowitz, R.S., ffrench-Mullen,
    J.M.H. & Li, X. (1998) GABAA receptor alpha 4subunit suppression
    prevents withdrawal properties of an endogenous steroid.  Nature,
    302, 926-930.

    Spicer, L.J., Kao, L.-C., Strauss, J.F., III & Hammond, J.M. (1990)
    2-Hydroxyestradiol enhanced progesterone production by porcine
    granulosa cells: Dependence on  de novo cholesterol synthesis and
    stimulation of cholesterol side-chain cleavage activity and cytochrome
    P450scc messenger ribonucleic acid levels.  Endocrinology, 127,
    2736-2770.

    Spink, D.C., Spink, B.C., Cao, J.Q., Gierthy, J.F., Hays, C.L., Li, Y.
    & Sutter, T.R. (1997) Induction of cytochrome P450 1B1 and catechol
    estrogen metabolism in ACHN human renal adenocarcinoma cells.
     J. Steroid Biochem. Mol. Biol., 62, 223-232.

    Stack, D.E., Cavalieri, E.L. & Rogan, E.G. (1998) Catecholestrogens as
    procarcinogens: Depurinating adducts and tumor initiation.
     Adv. Pharmacol., 42, 833-836.

    Stampfer, M.J. & Colditz, G.A. (1991) Estrogen replacement therapy and
    coronary heart disease: A quantitative assessment of the epidemiologic
    evidence.  Prev. Med., 20, 47-63.

    Sturgeon, S.R., Schairer, C., Brinton, L.A., Pearson, T. & Hoover,
    R.N. (1995) Evidence of a healthy estrogen user survivor effect.
     Epidemiology, 6, 227-231.

    Swales, N.J., Johnson, T. & Caldwell, J. (1996) Cryopreservation of
    rat and mouse hepatocytes.  Drug Metab. Disposition, 24, 1224-1230.

    Symonds, H.W., Prime, G.R. & Pullar, R.A. (1994) Preliminary evidence
    for the enterohepatic circulation of progesterone in the pig.  Br.
     Vet. J., 150, 585-593.

    Tauber, U., Schroder, K., Dusterberg, B. & Matthes, H. (1986) Absolute
    bioavailability of testosterone after oral administration of
    testosterone-undecanoate and testosterone.  Eur. J. Drug Metab.
     Pharmacokinet., 11, 145-149.

    Thompson, P.A., Shields, P.G., Freudenheim, J.L., Stone, A., Vena,
    J.E., Marshall, J.R., Grahm, S., Laughlin, R., Nemoto, T., Kadlubar,
    F.F. & Ambrosone, C.B. (1998) Genetic polymorphisms in
    catechol-O-methyltransferase, menopausal status and breast cancer
    risk.  Cancer Res., 58, 2107-2110.

    Tsai, M.J., Clark, J.H., Schrader, W.T. & O'Malley, B.W. (1998)
    Hormones that act as transcription-regulatory factors. In: Wilson,
    J.D., Foster, D.W., Kronenberg, H.M. & Larsen, P.R., eds,  Williams
     Textbook of Endocrinology, 9th Ed., Philadelphia, W.B. Saunders Co.,
    pp. 55-94.

    Tsutsui, T., Suzuki, N., Fukuda, S., Sato, M., Maizumi, H., McLachlan,
    J.A. & Barrett, J. (1987) 17-Estradiol-induced cell transformation
    and aneuploidy of Syrian hamster embryo cells in culture.
     Carcinogenesis, 11, 1715-1719.

    Tsutsui, T., Konine, A., Huff, J. & Barrett, J.C. (1995) Effects of
    testosterone, testosterone propionate, 17beta-trenbolone and
    progesterone on cell transformation and mutagenesis in Syrian hamster
    embryo cells.  Carcinogenesis, 16, 1329-1333.

    Ursin, G., London, S., Stanczyk, F.Z., Gentzschein, E., Paganini-Hill,
    A., Ross, R.K. & Pike, M.C. (1997) A pilot study of urinary estrogen
    metabolites (16alpha-OHE1 and 2-OHE1) in postmenopausal women with and
    without breast cancer.  Environ. Health Perspectives, 105 (Suppl 
    3), 601-605.

    US Congress (1995)  Effectiveness and Costs of Osteoporosis Screening
     and Hormone Replacement Therapy, Vol. II,  Evidence of Benefits,
     Risks, and Costs (Publication OTA-BP-H-144), Washington DC, Office
    of Technology Assessment.

    Wang, M.Y. & Liehr, J.G. (1994) Identification of fatty acid
    hydroperoxide cofactors in the cytochrome P450-mediated oxidation of
    estrogens to quinone metabolites.  J. Biol. Chem., 269, 284-291.

    Wehling, M. (1997) Specific, nongenomic actions of steroid hormones.
     Annu. Rev. Physiol., 59, 365-393.

    Weiss, J., Fritz-Wolz, G., Clawson, G.A., Benedict, C.M., Abendroth,
    C. & Creveling, C.R. (1998) Induction of nuclear
    catechol-O-methyltransferase by estrogens in hamster kidney:
    Implications for estrogen-induced renal cancer.  Carcinogenesis,
    19, 1307-1312.

    Weisz, J., Bui, Q.D., Roy, D. & Leihr, J.G. (1992) Elevated
    4-hydroxylation of estradiol by hamster kidney microsomes: A potential
    pathway of metabolic activation of estrogens.  Endocrinology, 131,
    655-661.

    Welsch, C.W. (1976) Interaction of estrogen and prolactin in
    spontaneous mammary tumorigenesis of the mouse.  J. Toxicol. Environ.
     Health, Suppl. 1, 161-175.

    Wharton, L.R. & Scott, R.B. (1964) Experimental production of genital
    lesions with norethindrone.  Am. J. Obstet. Gynecol., 89, 701-715.

    White, C.M., Ferraro-Borgida, M.J., Fossati, A.T., McGill, C.C.,
    Ahlberg, A.W., Feng, Y.J., Heller, G.V. & Chow, M.S. (1998) The
    pharmacokinetics of intravenous estradiol-A preliminary study.
     Pharmacotherapy, 18, 1343-1346.

    Whitehead, M.I., Townsend, P.T., Gill, D.K., Collins, W.P. & Campbell,
    S. (1980) Absorption and metabolism of oral progesterone.  Br. Med.
     J., 280, 825-827.

    Whittemore, A.S, Harris, R., Intyre, J. & the Collaborative Ovarian
    Cancer Group (1992) Characteristics relating to ovarian cancer risk:
    Collaborative analysis of 12 US case-control studies. II. Invasive
    epithelial ovarian cancers in white women.  Am. J. Epidemiol., 136,
    1184-1203.

    WHO (1992)  Guidelines for the Use of Androgens in Men
    (WHO/HPR/MALE/92), Geneva, Special Programme of Research, Development
    and Research Training in Human Reproduction.

    WHO (1998)  Cardiovascular Disease and Steroid Hormone Contraception.
     Report of a WHO Scientific Group (WHO Technical Report Series 877),
    Geneva.

    Wiebe, J.P., Boushy, D. & Wolfe, M. (1997) Synthesis, metabolism and
    levels of the neuroactive steroid, 3alpha-hydroxy-4-pregnen-20-one
    (3alpha HP), in rat pituitaries.  Brain Res., 764, 158-166.

    Williams, C.L. & Stancel, G.M. (1996) Estrogens and progestins. In:
    Hardman, J.G., Limbird, L.E., Molinoff, P.B., Ruddon, R.W. & Gilman,
    A.G., eds,  Goodman and Gilman's The Pharmacological Basis of
     Therapeutics, 9th Ed., pp. 1411-1440.

    Wilson, J.D. (1996) Androgens. In: Hardman, J.G., Limbird, L.E.,
    Molinoff, P.B., Ruddon, R.W. & Gilman, A.G., eds,  Goodman and
     Gilman's The Pharmacological Basis of Therapeutics, 9th Ed.,
    pp.1441-1457.

    Wilson, V.S. & LeBlanc, G.A. (1998) Endosulfan elevated testosterone
    biotransformation and clearance in CD-1 mice.  Toxicol. Appl.
     Pharmacol., 148, 158-168.

    Wilson, J.D., Foster, D.W., Kronenberg, H. & Larsen, P.R. (1998)
    Principles of endocrinology. In: Wilson, J.D., Foster, D.W.,
    Kronenberg, H.M. & Larsen, P.R., eds,  Williams Textbook of
     Endocrinology, 9th Ed., Philadelphia, W.B. Saunders Co., pp. 1-10..

    Wingard, L.M., Brody, T.M., Larner, J. & Schwartz, A. (1991)
    Estrogens, progestins, and oral contraceptives. In:  Human
     Pharmacology: Molecular to Clinical, St Louis, Mosby-YearBook, pp.
    494-514.

    Winter, M.L. & Liehr, J.G. (1996) Possible mechanism of induction of
    benign prostatic hyperplasia by estradiol and dihydrotestosterone in
    dogs.  Toxicol. Appl. Pharmacol., 136, 211-219.

    World Cancer Research Fund in association with the American Institute
    for Cancer Research (1997)  Food, Nutrition and the Prevention of
     Cancer: A Global Perspective, Washington DC.

    Yager, J.D. & Liehr, J.G. (1996) Molecular mechanisms of estrogen
    carcinogenesis.  Annu. Rev. Pharmacol. Toxicol., 36, 203-232.

    Yamazaki, H., Shaw, P.M., Guengerich, F.P. & Shimada, T. (1998) Roles
    of cytochromes P450 1A2 and 3A4 in the oxidation of estradiol and
    estrone in human liver microsomes.  Chem. Res. Toxicol., 11,
    659-665.

    Yoshie, Y. & Ohshima, H. (1998). Synergistic induction of DNA strand
    breakage by catechol-estrogens and nitrous oxide: Implications for
    hormonal carcinogenesis.  Free Radicals Biol. Med., 15, 341-348.

    Yue, T.L., Wang, X., Louden, C.S., Gupta, S., Pillarisetti, K., Gu,
    J.L., Hart, T.K., Lysko, P.G. & Feuerstein, G.Z. (1997)
    2-Methoxyestradiol, an endogenous estrogen metabolite, induces
    apoptosis in endothelian cells and inhibits angiogenesis: Possible
    role for stress-activated protein kinase signaling pathway and Fas
    expression.  Mol. Pharmacol., 51,951-962.

    Zhu, B.T. & Conney, A.H. (1998a) Functional role of estrogen
    metabolism in target cells: Review and perspectives.  Carcinogenesis,
    19, 1-27.

    Zhu, B.T. & Conney, A.H. (1998b) Is 2-methoxyestradiol an endogenous
    estrogen metabolite that inhibits mammary carcinogenesis?  Cancer
     Res., 58, 2269-2297.

    Zhu, B.T. & Liehr, J.G. (1994) Quercitin increases the severity of
    estradiol-induced tumorigenesis in hamster kidney.  Toxicol. Appl.
     Pharmacol., 125, 149-158.

    Zhu, B.T., Roy, D. & Liehr, J.G. (1993) The carcinogenic activity of
    ethinyl estrogens is determined by both their hormonal characteristics
    and their conversion to catechol metabolites.  Endocrinology, 132,
    577-583.

    Zhu, B.T., Evaristus, E.N., Antoniak, S.K., Sarabia, S.F., Ricci, M.J.
    & Liehr, J.G. (1996) Metabolic deglucuronidation and demethylation of
    estrogen conjugates as a source of parent estrogens and
    catecholestrogen metabolites in Syrian hamster kidney, a target organ
    of estrogen-induced tumorignesis.  Toxicol. Appl. Pharmacol., 136,
    186-193.
    


    See Also:
       Toxicological Abbreviations