Hormone Profiles in Hormone-Dependent Cancers'

[CANCER RESEARCH35,3365-3373,November1975] Hormone Profiles in Hormone-dependent Cancers'

Barnett Zumoff,2 Jack Fishman, H. Leon Bradlow. and Leon Heilman Department of Oncologyand the Institute for Steroid Research,Montefiore Hospital and Medical Center, Bronx, New York 10467

SUMMARY through ablation of endocrine glands by surgery or radia tion or through administration of antihormones (e.g., Studies on the relationship of urinary excretion of administration of estrogensor androgensin breast cancer; androgen metabolites and estrogens to the natural history of administration of estrogens in prostate cancer; administra breast cancer are reviewed. The importance of distinguish tion of progestagens in endometnial cancer; orchiectomy for ing between â€œwithin-populationâ€•studies(i.e., cancer pa prostate cancer and male breast cancer; oophorectomy, tients versus normal controls) and â€œbetweenpopulationâ€• adrenalectomy, and hypophysectomy for breast cancer; studies (i.e., low-risk versus high-risk populations) is em administration of antiandrogens in prostate cancer; admin phasized, and it is pointed out that â€œqualitativeâ€•agreement istration of antiestrogens in breast cancer). Because of the (i.e., the same direction of differences) between the two therapeutic effect of alterations in the hormonal environ types of studies must be present in order to implicate a ment, considerable interest has developed in the possibility hormonal parameter asa determinant ofthe natural history that the amounts or patterns of the secretion of endogenous of breast cancer. For reasonsdetailed in this paper, it is hormones are related to either the induction or the un concluded that the reported relationship of low urinary treated clinical course of such cancers. This approach has androgen metabolite excretion to increased risk of develop stimulated a large body of work which I should like to ing breast cancer and poor response to adrenalectomy or review for you today. hypophysectomy and the validity of the â€œestniolhypothe The basic methodology for studying the role of hormones sis,â€• namely, that a high urinary ratio of estniol to in hormone-dependent cancer is what might be called estrone-plus-estradiol in early life is protective against â€œbiochemical epidemiology,â€• i.e., the cross-sectional study subsequent development of breast cancer, are both dubious. of hormonal differences between those who do and those A new hypothesisconcerning the relationship of estrogens who do not have a particular type of cancer. There are 2 to breast cancer risk is presented: â€œAperiodoftime, prior to ways in which such biochemical epidemiology can be age 30, during which the amount or biological availability of applied: (a) by comparing hormone findings between popu active estrogens' (i.e., estrone and estradiol) is diminished, lations, one of which has a low incidence of the cancer at protects against subsequent development of cancer.â€• This issue, while the other has a high incidence (e.g., in the case hypothesis is shown to be compatible with the epidemiologi of breast cancer, Chinese or Japanese women in Asia and cal and biochemical data. Caucasian women in North America, respectively), and (b) Reports concerning the influence of nutrition on endo within a population, by comparing hormonal findings of crine parameters are reviewed. Inanition and obesity have individuals who have the cancer at issue with those of beenshownto alter steroid metabolism but it is not known individuals who do not (i.e., the â€œcase-controlâ€•method). whether nutritional â€œmicrodifferencesâ€• (i.e., differences The distinction between these 2 approaches has not always between populations or individuals that are due to cultural, been kept clearly in sight, but it is an important one; in geographic, or socioeconomic factors, but that fall within order to implicate a particular endocrine parameter as a the range of â€œnormalâ€•oradequate nutrition) can also alter determining factor in a particular cancer, it would seem steroid metabolism. essential that the between-population and within-population comparisons of that parameter should yield the same directional or qualitative differences. INTRODUCTION Since the overall topic of the present symposium is â€œNutritionin the Causation of Cancer,â€•it is clearly Certain cancers, notably those ofthe breast, prostate, and incumbent upon us to try to correlate nutritional differences endometnium, are described as hormone dependent on the with hormonal differences. In this connection, it might be basis of 2 findings: (a) that they arise from tissuesthat are pertinent to subdivide nutritional differences into 2 cate normally responsive to endogenous hormones, and (b) that gonies: (a) â€œmicrodifferences,â€•comprisingthose differences their course can often be influenced, favorably or unfavora between populations or individuals that are determined by bly, by administration or removal of hormones, either cultural, geographic, or socioeconomic factors but are still within the overall range of normal or adequate nutrition; (b) 1 Presented at the Conference on Nutrition in the Causation of Cancer, â€œmacrodifferences,â€•consistingof the differences between May 19 to 22, 1975, Key Biscayne, Fla. Supported in part by Grants CA-07304 from the National Cancer Institute and RR-53 from the adequate nutrition on the one hand and gross undernutrition GeneralClinicalResearchCentersBranch,NIH, Bethesda,Md. (inanition) or overnutrition (obesity) on the other hand. I Presenter. Gortner (35), Oiso (65), Hankin (38), and Fernandez (28)

NOVEMBER 1975 3365

have reviewed for us some of the between-population dif 100 ferences in the nutritional patterns of various countries. Combineddata; These fall essentially within the category of microdiffer Guy's Hospital ences. Lipsett's review (52) of the influences of nutritional 80 NCC Tokyo Villejuif factors on endocrine patterns has dealt largely with macro differences. It has been relatively easy to show that gross

undernutrition alters the metabolism of corticoids (10, 21, 60 48, 63), androgens (10, 21, 25, 48, 61, 63), and estrogens (70), but there is as yet no clear-cut evidence that micro C differences in nutrition can do so as well. After it has been 8. demonstrated that consistent ipsidirectional between-pop 40 ulation and within-population hormonal differences are correlated with cancer incidence, it should then be produc tive to focus research efforts on a study of the influence of 20 nutritional microdifferences between the populations con cerned upon those endocrine parameters that will have been implicated. 0.3 @.O 2.0 3.0 HORMONAL PARAMETERS THAT HAVE BEEN Andmgen + eiiocholanolone in urine (mg /24 hr) STUDIED IN CANCER Chart 2. Relationshipofurinaryandrogenmetaboliteexcretion(i.e., androsterone + etiocholanolone) to the response to endocrine ablation, This discussion will be limited to breast cancer, since consolidating the data of Bullbrook ci al. (2, 5, 6, 13, 14, 16), Juret et al. hormonal data concerning the other types of hormone (47), and Kumaoka et al. (49). The lower the androgen metabolite excretion dependent cancer are as yet fragmentary. There have been the lower is the proportion of patients who respond favorably. Reproduced interesting and extensive findings with respect to 2 sets of from Ref. 13with the permissionof the author and the publishers. homonal parameters, urinary androgen metabolites and urinary estrogens. a!. with respect to response to adrenalectomy, and Beck et Urinary Androgen Metabolites. Extensive within-popula a!. (8) in Canada have failed to confirm a correlation tion studies of urinary androgen metabolite excretion have between androgen metabolite excretion and response to been reported by Atkins, Bulbrook and Hayward and their hypophysectomy. More recently, unpublished studies mon coworkers in Great Britain, in a long series of papers from itored by the Breast Cancer Task Force of the National 1957 to date (2, 5, 6, 13, 14, 16). Basically, there were 2 Cancer Institute have also failed to confirm the predictive major findings: value of urinary androgen excretion for response to endo 1. The daily urinary excretion of androgen metabolites cnine ablation. (etiocholanolone and androsterone (Chart I) was correlated 2. In a group of apparently normal women in Guernsey, with the responsiveness of metastatic breast cancer to the urinary excretion of androgen metabolites predicted, adrenalectomy or hypophysectomy; more specifically, years in advance, the later appearance of breast cancer in women with subnormal excretion of these metabolites certain individuals; more specifically, women who later usually showed a poor response to such surgery while developed breast cancer were found to have been excreting women with normal metabolite excretion usually showed a subnormal amounts of urinary androgen metabolites years good response. These findings have been confirmed by Juret earlier (15) (Chart 3). No other studies have been reported et a!. (47) in France, with respect to response to hypophysec that either confirm or refute these findings. tomy, and by Kumaoka et a!. (49) in Japan, with respect to Bulbrook et a!. (17) also studied between-population response to adrenalectomy (Chart 2). On the other hand, differences in the excretion of urinary androgen metabolites. Wilson et a!. (76) in the United States have reported data Healthy women from the British population (high risk for that are in partial disagreement with those of Bulbrook et breast cancer) were compared with healthy Chinese and Japanese women in Asia (low risk for breast cancer); it was found that the Asian women had much lower urinary androgen metabolite excretion than did the British women,

-O+O@ 3 even after â€œcorrectingâ€•fordifferences in body size (Chart 4). In other words, the directional difference between populations was opposite to the within-population direc tional difference. Lower excretion was associated with higher risk within a population but with lower risk between

4 A E populations. This latter finding is not reassuring concerning the Chart 1. Precursors of urinary I 7-ketosteroids. Compounds I , 2, 3, and 4 are the major androgenic secretory products, respectively, dehy possibility of a biological link, not to mention a cause-and droisoandrosterone, dehydroisoandrosterone sulfate, i@4-androstenedione, effect relationship, between urinary androgen metabolite and testosterone. All of these hormones are metabolized to the same 2 excretion and breast cancer. Furthermore, it has been shown principal urinary metabolites, androsterone (A) and etiocholanolone (E). that subnormal urinary excretion of androgen metabolites

3366 CANCER RESEARCH VOL. 35

Median Urinary Estrogens. Interest in the measurement of un 20 nary estrogens in breast cancer is based on the widespread Matched 16 controls presumption that estrogens are carcinogenic with respect to @12 the breast. This presumption, in turn, is based on the results â€œI of animal experiments and on human epidemiological data, as follows. 1@t9{ffMIrIffh}nirrrr,nn 1. Of the 3 â€œclassicalâ€•estrogens3of human urine, estrone 0, (E1) and estradiol (E2) cause breast cancer in rodents (54, @ 4 . 11 cases 72) and estniol (E3) does not (39). @@ 2 @J @lII 2. Human females castrated prior to their late 30's have A rll@1 1â€¢0 20 30 been shown to have about an 80% decrease in the incidence I4.TIOCHOLAWOLONE(mgper24hr) of breast cancer; no such decrease occurs in women Chart 3. Comparison of the urinary excretion of etiocholanolone in castrated at a later age (27, 43). This has been generally womenwho later developedbreastcancerwith theexcretionin womenwho interpreted to support the conclusion that estrogens are did not develop cancer. The former show significantly lower values. carcinogenic to the human breast also (20). Strictly speak Reproduced from Ref. 15 with the permission of the authors and the ing, it is more accurate to say that the studies show that the publishers. presence of something secreted by the ovary is necessary for the development of breast cancer; this â€œsomethingâ€•(pre 02 - sumably estrogen4) could be merely permissive and not causative. A more direct test of estrogen carcinogenicity is provided by the results of long-term administration of 0 estrogens to intact women; all such studies to date (4, 11, 74) have shown no increase in the incidence of breast cancer, 0 .0 strongly suggesting that estrogens are not carcinogenic to .M ,@ the human breast.5 5. .0,.0@1 - Despite these reservations, it has been considered to be of C. at least heuristic value to measure the urinary estrogens of C. [email protected] various individuals and groups in order to determine whether the degreee of â€œestrogenicityâ€•is related to the development of breast cancer. Lemon et a!. (5 1) and, more recently, Cole and MacMahon (20) have perceived an important problem in this approach. Not only do the 3 @ -I-â€” classical urinary estrogens differ in estrogenic potency, but 20 40 60 80 estniol can be clearly shown, in rodents, to have antiestro Ag.(yr) genic effects' (44, 45, 75). For this reason, these authors Chart 4. Urinary steroids in British and Japanese women. British and have used the relative amounts of the 3 estrogens as a Japanese women excrete identical quantities of corticoid metabolites measure of â€œnetestrogenicity.â€• Lemon et a!. (5 1) have (17-OHCS) per kg of body weight but British womenexcretemuchlarger expressed this relative amount as the â€œestriolquotientâ€• quantities of androgen metabolites (Il-Deoxy-17-OS) than do Japanese [E3/(E1 + E2)]. MacMahon et a!. (56) have used the terms women. Reproduced from Ref. 17 with the permission of the authors and â€œestniolratioâ€• [E3/(E1 + E2)] and â€œestriolproportionâ€• the publishers. [E3/(E1 + E2 + E3)] and have elaborated what has come to

(19, 32, 33, 58, 60, 67) and subnormal conversion of 3 More recent studies have shown the presence of many other estrogens in human urine, among them 2-hydroxyestrone (3 1), 2-methoxyestrone androgenic hormones to these metabolites (26, 78) occur in (30), l6a-hydroxyestrone (80) and 16-epiestriol (81) (Chart 5). Few several other types of cancer and a wide variety of other studies of these compounds in breast cancer have been done, although 2- chronic illnesses. A possible common factor may be the hydroxyestrone is quantitatively the major estrogen in both plasma (77) more or less marked undernutrition that is a feature of and urine (79). 4 Grattarola et a!. (36) have suggested that the ovarian secretion that many of these diseases, since it has been shown that favors the development of breast cancer is androgen, not estrogen.

undernutrition, varying from simple â€œdietingâ€•(25)to actual 5 It is still theoretically possible that estrogens might be carcinogenic if starvation (48, 63), can cause a decrease in the urinary administered to very young females; against this possibility is the fact that excretion of androgen metabolites. administration of stilbestrol to fetal females in utero causes vaginal cancer To summarize, therefore, the opposite directional find (41) but has not been shown to cause breast cancer. 6 Estriol anatgonizes the uterotrophic and transhydrogenase-stimulating ings of between-population and within-population studies in effects of estrone and estradiol (44, 45, 75) but has not been shown to the hands of Bulbrook et a!., the contradictary reports of antagonize the carcinogenic effect (20). other workers concerning within-population studies and the 7 The estriol ratio is distributed log-normally whereas the estriol nonspecificity of subnormal androgen metabolite excretion proportion is distributed normally (24). The latter, therefore, lends itself in many unrelated diseases leave room for considerable more readily to graphing and statistical manipulation. A practical advantage of using these relative expressions is that they are presumably doubt that androgen metabolite excretion is an important independent of the duration of a urinary collection and of the absolute determinant of the natural history of breast cancer. excretion of each individual estrogen, so that studies can be carried out on

NOVEMBER 1975 3367

@ ,%@cÃ @vco@â€” I Ii I Ho#A@*@,t10%,.,) HO'@'@â€•@' ;o,cÃª:s@oH ESTRADIOL.0 ESTRONEOH 16-EPIESTRIOL I â€˜0 Chart 5. Major pathways and major metabolites of II estradiol metabolism in man. ,Â©5@y.@0H

HO 2-HYDROXYESTRONE 160C- HYDROXYESTRONE

c:x:3@5:@ HO_X@'@@ 2- METHOXYESTRONE ESTRIOL

be known as the estniol hypothesis, namely, that â€œtheCaucasian women (Table 2). Once again, the excretion of relative levels of the estrogen fractions produced between estriol was identical in all 3 groups; it was the differences in puberty and about the 30th year of age are crucial excretion of estrone and estradiol that determined the determinants of a women's life-time breast cancer riskâ€• differences in estniol ratio. Chart 7 shows the estriol (20), more specifically that a higher estriol ratio is associ proportion of these 3 populations. Each group showed ated with a lower risk of developing breast cancer. The essentially a normal distribution of this parameter, with no reference to â€œbetweenpuberty and about the 30th year of suggestion of bimodality. ageâ€•derives from another important piece of epidemiologi There have been a number of within-population studies cal data. Women who become pregnant prior to age 30 show comparing the urinary estrogen excretion of Caucasian a decreased risk of breast cancer (relative to nulliparous women with and without breast cancer. A serious potential women) whereas women whose 1st pregnancy occurs after problem exists in the interpretation of such studies, since age 30 fail to show such a decreased risk and may indeed Cole and MacMahon (20) have concluded that only the have an increased risk (55) (Chart 6). Cole and MacMahon

(20) have concluded from this that some hormonal configu ,.,T ration characteristic of pregnancy must develop prior to age 30 in order to protect against breast cancer. Since these 1.4 5 workers feel that it is the characteristic elevation of the estniol ratio in pregnancy that mediates its protective effect, they have generalized this to the conclusion that an elevated 1.2 5 estriol ratio not due to pregnancy is also protective only if present prior to age 30. NtLLC*ROUS I.e â€” @ A careful between-population study of the estriol ratio x a a was carried out by MacMahon et a!. (56), using Chinese and In a @@@ Japanese women in Asia as the low-risk population and at . Caucasian women in North American as the high-risk population. Table 1 shows that the Asian women had much higher estriol ratios than did the Caucasian women, as called for by the estriol hypothesis. Careful note should be made of the fact that the higher ratio was due to lower excretion of estrone and estradiol; the excretion of estriol was identical in the 2 populations. This important point will be discussed in more detail later in this paper. Another between-population study by Dickinson et a!. (24) provides further support for the estriol hypothesis. $5 20 25 30 35 AGE AT FIRST @RTH These workers found that Asian women living in Hawaii (intermediate-risk population) showed estriol ratios inter Chart 6. Risk of developing breast cancer (relative to nulliparous mediate between those of Asian women in Asia and women) in women who have their 1st full-term pregnancy at various ages. There is a linear increase in relative risk with age at 1st birth, from about an overnight urine sample instead of a 24-hr collection; this is the 0.3 at age 15 to 1.0 in the early 30's to about 1.4 at age 40. Reproduced procedure that MacMahon's group has used for its studies. from Ref. 55 with the permission of the authors and the publishers.

3368 CANCER RESEARCH VOL.35

TableI Comparison of urinary estrogens in Asian and Caucasian women (56) â€”Estrone (E,) E5

Follicular Luteal Follicular Luteal Follicula r Luteal Follicular Luteal Follicular Luteal phaseAsian phase(E1) phaseEstradiophaseI phaseEstriolphase(E5) phaseEstriolphaseratiophaseE5 phase+

women4.56.92.43.78.315.41.331.536.910.6(15-39)Caucasian

women6.912.83.56.96.115.90.650.7110.419.7(15â€”19)Caucasian

women10.612.45.55.99.614.00.560.7516.118.3(30â€”39)

Table2 Table 3 Comparison ofâ€•estriol ratiosâ€•andestriol excretion between 3 populations â€œEstriolproportionâ€•in Caucasian women of various ages (24) of normalwomen(24) proportion proportion Estrio ratio cx phase)15-190.400.4420-240.390.4330-390.380.15AgeEstnol (follicular phase)Follicular(luteal

Follicular Luteal Follicular Luteal phaseAsian phase1 phaseEstriolphasecretion

Asia1.331.538.315.4Asianwomenin Hawaii0.831.038.616.4Caucasianwomen in women0.060.737.715.9

FREQUENCY1%)

0ASIANS IN ASIA A ASIANS IN HAWAII . CAUCASIANS

0.2 04 06 0.8 ESTRIOL PROPORTION @A5TC @ (EVEIE2+E5) (6) (10) Chart 7. Distributionof theestriolproportionin 3 groupsof normal Chart 8. Women with breast cancer show about twice the normal women (drawn from data in Ref. 24). The 3 populations studied all show an excretion of endogenousestriol. Excretion of estrone and estradiol (not essentially normal distribution of the estriol proportion. The mode shifts shown) is approximately normal, so that the estriol ratio is approximately from about 0.6 for Asians in Asia to about 0.45 for Asians in Hawaii to twice normal. Reproduced from Ref. 40. about 0.35 for Caucasians.None of the 3 population distributions show evidence of more than I mode. expected from the estniol hypothesis, but several other groups (12, 40, 57, 64, 66), including Brown et a!. (12) and estriol ratio that is present before age 30 influences the risk our group (40), have found the opposite, namely, higher of developing breast cancer. Thus in order to interpret the estriol ratios in women with breast cancer than in controls estriol ratio in women of breast cancer age, it would have to (Charts 8 and 9). Studies in Japanese women, cited by be shown that this parameter does not change with age. A Lemon (50), showed no difference in estniol ratios between definitive study of this problem has not been reported, but controls and patients with breast cancer. Without going into Dickinson et a!. (24) have shown that the estniol proportion a detailed analysis of the relative merits of the various based on endogenous estrogens does not change over the age studies, it is certainly fair to say that the results of range 15 to 39 (Table 3) and our group has shown (79) that within-population studies ofthe estriol ratio in breast cancer the estriol proportion basedon metabolites of radioactive are in considerable dispute and do not clearly support the estradiol tracers does not change over the age range 21 to estriol hypothesis. 80. It appears not unreasonable, therefore, to examine An interesting sidepoint concernsthe findings of Lemon, whether the results of within-population studies are in who has reported lower estniol quotients in women with agreement with those of between-population studies. breast cancer than in controls (51). He has more recently Unfortunately, the data are highly contradictory. Several reported (50) that normal women showed a polymodal groups (7, 46, 51, 71) have found lower estriol ratios in distribution of the estriol quotient (Chart 10) and has women with breast cancer than in controls, as would be interpreted this in terms of 3 genotypes, a homozygous

NOVEMBER 1975 3369

(e.g., 2-hydroxyestrone) may possessantiestrogenic proper 45, . ties should be kept in mind. This would seem to be a very productive area for further research. . The within-population studies that show an increased PERCENT . --37 OF 35@ I . estriol ratio in cancer patients are difficult to fit into the NEUTRAL estriol hypothesis but can be readily reconciled with the STEROID I . revised hypothesis, since the increased ratio in the cancer EXTRACT I 25 I patients is on a different basis from that of the Asian women . despite the comparable magnitude of the ratios in the 2 . . . â€”20 groups (Chart I 1). In the Asian women, excretion of estriol S S I. is normal and excretion of estrone and estradiol is de I5 creased, so that total estrogen excretion is decreased; in the â€”12 â€¢1â€¢ I. cancer patients, excretion of estrone and estradiol is normal S . and excretion of estniol is increased, so that total estrogen 5 S excretion is increased.9 Thus the directional change in the excretion of estrogens in the cancer patients (higher) is NORMAL FEMALE MALE WOMEN 8REAST CANcER BREASTCANCER opposite to the change in patients with low cancer incidence (12) (23) (6) (lower). Chart 9. Women with breast cancer show about double the normal Three further caveats concerning estniol ratios must be conversion of [â€˜4Cjestradioltracers to estriol. Conversion to estrone and expressed: (a) it is possible that the elevated estriol ratio in estradiol (not shown) is approximately normal, so that the estriol ratio women with breast cancer is a nonspecific effect of illness, based on metabolites of radioactive estriol tracers is approximately twice since an elevated ratio has also been reported in men who normal. Reproducedfor Ref. 40. have had myocardial infarctions (9) or prostate cancer (59) as well as in patients with a variety of other illnesses low-quotient group, a homozygous high-quotient group, and (unpublished data from this laboratory); (b) the â€œlow-T3 a heterozygous intermediate-quotient group. Chart 7, which syndromeâ€•reported (68) in many illnesses including cancer presents data from the normal population studied by might, if present, result in an increased relative excretion of Dickinson et a!. (24), fails to show any indication of more estriol, since the latter is characteristic of hypothyroidism than 1 mode in the distribution of the estniol proportion. (30); (c) it is possible that the cancer itself may produce It appears possible to reconcile the findings of the some of the excess estniol, since breast cancer tissue has between-population studies with the epidemiological data been shown to have 16a-hydroxylase activity (1). concerning the protective effects of early castration and Finally, it is pertinent to consider the possible involve early pregnancy by a revised hypothesis that we would like ment of nutritional factors in estrogen metabolism. Marked to suggest: â€œAperiodof time, prior to age 30, during which undernutrition (as exemplified by women with anorexia the amount or biological availability of â€˜activeestrogens' nervosa) causes a decrease in the urinary excretion of all the (i.e., estrone and estradiol) is diminished protects against classical estrogens and a decrease in the estriol ratio (70). subsequent development of breast cancer.â€•The diminution The effects of the conversestate, obesity, appearevenmore may be due to surgical removal of the source of estrogens pertinent, since obesity has been shown to be associated with (castration), environmentally determined decreases in estro an increased risk of breast cancer1Â°(23). Obesity alters gen (Asian women in Asia) or the presence of an â€œantiestro estrogen metabolism in a manner that is compatible with genâ€•(pregnancy). From this point of view, emphasis upon the revised hypothesis. It causes increased conversion of the central role of estriol itself in all these situations appears androstenedione (an ovarian and adrenocortical secretory somewhat misplaced. In the 1st 2 cases, simple decrease in product) to estrone (37) and thereby increases the availabil the amounts of estrone and estradiol constitutes @nade ity of active estrogen. Thus another piece of epidemiological quate explanation; furthermore, the plasma levels of un data, the increased breast cancer risk in obesity, can be conjugated estriol are probably too low in nonpregnant fitted into a single overall hypothesis centering on the individuals [ < 10 pg/mI, i.e., 3 to 5% ofestradiol levels (53)] availability of active estrogen. to exert any significant antiestrogenic effect.8 Only in pregnancy do we need to invoke the existence of an ADDENDUM antiestrogen; estriol may well fulfill this function, since the plasma estriol/estradiol ratios in pregnancy [about 1/2 Several very recent publications are highly pertinent to the discussion in (73)] are comparable to those required for the demonstra this paper concerning the estriol hypothesis. The biological basis for this tion of antiestrogenic effects of estriol in rodents [about 1/1 hypothesis consists largely of 2 observations in rodents: (a) that estriol has (75)], but the possibility that other estrogen metabolites a Dao ci a!. (22) have reported increased total urinary estrogens in men I MacMahon ci al. (55) have cited unpublished data from their with breast cancer also; the estriol ratio was normal. laboratory indicating that in nonpregnant women the plasma concentration 10Increased dietary fat intake, a microdifference in nutrition, has also of unconjugated,non-protein-bound estriol is approximately two-thirds been shown to be associated with an increased risk of breast cancer ( I 8); that of unconjugated, non-protein-bound estradiol. Publication of these since increased fat intake is often associated with obesity, it is not clear data is awaitedwith interest. whether these are independent risk factors.

3370 CANCER RESEARCH VOL.35

IS NORMAL CAUCASIANWOMEN

w.5

Chart 10. Distribution of estriol quotients (equiva U lent to estriol ratios) found by Lemon in normal Cauca 0 sian women. There is a suggestion of a polymodal distri bution. Reproduced from Ref. 50 with the permission p A â€˜-â€˜7

@ of the authors and the publishers. NINO,@,,fi@@P*UUL) I

0.5 1.0 1.5 2.0 2.5 3.0 3.6 QUOTIENTCAUCASIAN ESTRIOL EXCRETION

causes increased luteinizing hormone secretion; this is gener NORMAL ASIAN WOMEN WOMEN ally considered to represent an antiestrogen effect, i.e., blockage of tissue IN ASIA WITH BREAST CANCERmetabolites, uptake of estradiol in the hypothalamus. Studies from this laboratory (C. [ESTRIOL RATiOâ€˜2XNORMALI RATIO@2XN0RMAL1this 200% I Martucci and J. Fishman, to be published) have shown that 2-hydroxyes 5' I endows@STRIOL trone binds well to the cytosol estrogen receptor, which necessarily steroid with antiestrogen properties since it has no uterotrophic activ V)@ I bindingsites.ity of its own (42) and it would compete with active estrogens for @ wI5O% I Gelbke et al. (34) have shown that urinary excretion of 2-hydroxyes L I @ ct@ I beingdisproportionatelytrone rises early and markedly during pregnancy, with the increase tJV) z andâ€”midpregnancy; greater than that of other estrogens during early @cI4J 5 â€˜lJ @,) I if 2-hydroxyestrone is indeed an antiestrogen, this observa V@ 4 100% @ )@.:@3 I tionsubsequentbreast may be related to the protective effect of pregnancy against @ .4 â€˜@ U 2-hydroxyestronecancer. Finally, Fishman et al. (29) have shown that the ratio of @I@I&I_J I theincreased to estradiol is greatly decreased in obese women; thus @ ____ @ 50% I theâ€”â€”decreased risk of breast cancer in these women (23) might be related to @zz I relative amount of the protective antiestrogen, 2-hydroxyestrone. 4@ @E1@E2 I El+E2 E3 !2 I REFERENCESâ€”Chart estriolratio.I 1. Two different mechanisms for achieving an elevated HumantheThe values for normal Asian women are drawn from data in Ref. 24;1 Adams, J. B., and Wong, M. S. F. Paraendocrine Behavior of values for Caucasianwomen with breast cancer are drawn from aBreast Carcinoma: In Vitro Transformation of Steroidsto Physiologi composite1968.ratio ofthe data from Ref. 40, 57, and 66. Both groups show an estriolcally Active Hormones. J. Endocrinol., 41: 41 -52, Excretionthisthat is approximately twice normal, but in the normal Asian women2. Allen, B. J., Hayward, J. L., and Merrivale, W. H. H. The derives from normal estriol excretion and decreased estrone +of 17-Ketosteroids in the Urine of Patients with Generalized Car I:derivesestradiol excretion, whereas in the Caucasian women with breast cancer itcinomatosis Secondary to Carcinoma of the Breast. Lancet, 1957.excretion.3.from normal estrone + estradiol excretion and increased estriol496499, Estrogen-inducedUterineAnderson, J. N., Peck, E. J., and Clark, J. H. Estrogenminimal Responses and Growth. Relationship to Receptor 1975.antagonistestrogenicity (i.e., uterotrophic activity) of its own and acts as anBinding by Uterine Nuclei. Endocrinology, 96: 160-165, Estrogensadministrationto the uterotrophic effect of estradiol (44, 45, 75); (b) that4. Arthes, F. G., Sartwell, P. E., and Lewison, E. F. The Pill, I.estrone of estriol in contrast to administration of estradiol orand the Breast. Epidemiologic Aspects. Cancer, 28: 1931- 1394, 197 L.,tions does not produce mammary cancers (39). Both of these observa 5. Atkins, H., Bulbrook, R. D., Falconer, M. A., Hayward, J. thechronichave now been challenged. Anderson et al. (3) have reported thatMacLean, K. C., and Schuer, P. H. Urinary Steroid Estimations in administration of estriol, a more â€œphysiologicalâ€•approachthanPrediction of Response to Adrenalectomy or Hypophysectomy. Lan 1964.hasthe acuteadministration usedin previousstudies,demonstratesthatestriolcet, 2: 1133â€”1136, 1.,(69)the sameuterotrophic activity as estradiol, mg for mg; Rudali et al.6. Atkins, H., Bulbrook, R. D., Falconer, M. A., Hayward, J. Predictionestriolhave reported just as high an incidence of mammary cancer afterMacLean, K. C., and Schuer, P. H. Urinary Steroids in the Secondreinforceadministration as after estradiol administration. These new findingsof Response to Adrenalectomy or Hypophysectomy: A 1968.us the doubts that we expressed earlier in this paper and suggest toClinical Trial. Lancet, 2: 1261-1263, inConversly,that the estriol hypothesis should probably be abandoned at this time.7. Bacigalupo, G., and Schubert, K. Studies on the Estrogen Excretion 804-805,the there is new evidence that 2-hydroxyestrone might well playUrine in Mastopathy (in German). KIm. Wochschr., 38: estriol.1960.Morishitaantiestrogen role currently ascribed (apparently erroneously) to andto et al. (62) have shown that administration of 2-hydroxyestrone8. Beck, J. C., Blair, A. J., Griffiths, M. M., Rosenfield, M. W., male rats, like clomiphene but unlike other natural estrogensor estrogenMcGarry, E. E. In Search of Hormonal Factors as an Aid in

NOVEMBER 1975 3371

Predicting the Outcome of Breasi Carcinoma. Can. Cancer Conf., 6: 31. Fishman,J., Hellman, L., Zumoff, B., and Gallagher,T. F. Effect of 3â€”35,1966. Thyroid on Hydroxylation of Estrogen in Man. J. Clin. Endocrinol., 9. Bersohn, I., and Oelofse, P. J. Urinary Oestrogen Levels in Myocar 25:365-368,1965. dial Infarction. S. African Med. J., 32: 979-983, 1958. 32. Forbes,A. P., Donaldson,E. C., Reifenstein,E. C., Jr., and Albright, 10. Bliss, E. L., and Migeon, C. J. Endocrinology of Anorexia Nervosa. J. E. The Effect of Trauma and Disease on the U rinary 17-Ketosteroid Clin. Endocrinol., 17: 766-776, 1957. Excretion in Man. J. Clin. Endocrinol., 7: 264-288, 1947. I I . Boston Collaborative Drug Surveillance Program. Surgically Con 33. Fraser, R. W., Forbes, A. P., Albright, F., Sulkowitch, F., and firmed Gallbladder Disease, Venous Thromboembolism and Breast Reifenstein, E. C., Jr. Colorimetric Assay of 17-Ketosteroids in Urine. Tumors in Relation to Postmenopausal Estrogen Therapy. New Engl. A Survey of the Use ofThis Test in Endocrine Investigation, Diagnosis J. Med., 290: 15-19, 1974. and Therapy. J. Clin. Endocrinol., 1: 234-250, 1941. 12. Brown, J. B. In: A. R. Currie (ed.), Endocrine Aspects of Breast Can 34. Gelbke, H. P., Bottger, M., and Knuppen,R. Excretion of 2-Hydrox cer, p. 197. Edinburgh: E & S Livingstone, Ltd., 1958. yestrone in Urine throughout Human Pregnancies.J. Clin. Endo 13. Bulbrook, R. D. Prediction of Responseof BreastCancer to Treat crinol., in press. ment. In: J. F. Holland and Emil Frei, Ill (eds.), Cancer Medicine pp. 35. Gortner, W. A. Nutrition in the United States, 1900â€”1974.Cancer 907-91 1. Philadelphia: Lea & Febiger, 1973. Res., 35: 3246â€”3253,1975. 14. Bulbrook, R. D., Greenwood,F. C., and Hayward, J. L. Selectionof 36. Grattarola, R., Secreto,G., and Recchione,C. Androgensin Breast Breast-CancerPatients for Adrenalectomy or Hypophysectomy by Cancer. III. Breast Cancer Recurrences Years after Mastectomy and Determination of Urinary l7-Hydroxycorticosteroids and Aetio Increased Androgenic Activity. Am. J. Obstet. Gynecol., 121: cholanolone. Lancet, 1: ll54-ll57, 1960. 169â€”172,1975. 15. Bulbrook, R. D., Hayward, J. L., and Spicer,C. C. Relation between 37. Grodin, J. M., Siiteri, P. K., and MacDonald, P. C. Source of Urinary Androgen and Corticoid Excretion and Subsequent Breast Estrogen Production in Postmenopausal Women. J. Clin. Endocrinol., Cancer. Lancet, 3: 395-398, 1971. 36: 207-214, 1973. 16. Bulbrook, R. D., Hayward, J. L., Spicer, C. C., and Thomas, B. S. A 38. Hankin, J. H., Nomura. A., and Rhoads, G. C. Dietary Patterns Comparison between the Urinary Steroid Excretion of Normal among Men of Japanese Ancestry in Hawaii. Cancer Res., 35: Women and Women with Advanced Breast Cancer. Lancet, 2: 1235â€” 3259-3264, 1975. 1237,1962. 39. Hartwell, J. Survey of Compounds Which Have Been Tested for 17. Bulbrook, R. D., Thomas, B. S., Utsonomiya, J., and Hamaguchi, E. Carcinogenic Activity. United States Public Health Service Publica The Urinary excretion of I l-Deoxy-17-oxosteroids and 17-Hydrox tion 149. Washington, D. C.: United States Government Printing ycorticosteroids by Normal Japaneseand British Women. J. Endo Office, 1951. crinol., 38: 401 -406, 1967. 40. Hellman, L., Zumoff, B., Fishman,J., andGallagher,T. F. Peripheral 18. Carroll, K. K., Gammel, E. B., and Plunkett, E. R. Dietary Fat and Metabolism of â€˜H-Estradiolandthe Excretion of EndogenousEstrone Mammary Cancer. Can. Med. Assoc. J., 98: 590-594, 1968. and Estriol Glucosiduronate in Women with Breast Cancer. J. Clin. 19. Chou, C. Y., and Wang, C. W. Excretion of Male Sex Hormone in Endocrinol., 33: 138-144, 1971. Health and Disease. Chinese J. Physiol., 14: 151-159, 1939. 41. Herbst, A. L., Ulfelder, H., and Poskonzer, D. C. Adenocarcinoma of 20. Cole, P., and MacMahon, B. Oestrogen Fractions during Early the Vagina: Association of Material Stilbestrol Therapy with Tumor Reproductive Life in the Etiology of Breast Cancer. Lancet, I: Appearance in Young Women. New Engl. J. Med., 284: 878-881, 604-606, 1969. 1971. 21. Cooke, J. N. C., James, V. H. T., Landon, J., and Wynn, V. 42. Hilgar, A. G. in: A. G. Hilgar and L. C. French (eds.), Uterotropic Adrenocortical Function in Chronic Malnutrition. Brit. Med. J., 1: Bioassay Data. Bethesda, Md.: National Cancer Institute, 1968. 662-666, 1964. 43. Hiroyama, T., and Wynder, E. L. A Study of the Epidemiology of 22. Dao, T. L., Morreal, C., and Nemoto, T. Urinary Estrogen Excretion Cancer of the Breast. II. The Influence of Hysterectomy. Cancer, 15: in Men with Breast Cancer. New EngI. J. Med., 289: 138-140, 1973. 28â€”38,1962. 23. DeWaard, F., Baanders-van-Halewign, E. A., and Huizinga, J. The 44. Hisaw, F. L., Velardo, J. T., and Goolsby, S. M. Interaction of Bimodal Age Distribution of Patients with Mammary Carcinoma Estrogens on Uterine Growth. J. Clin. Endocrinol., 14: 1134-1143, Evidence of the Existence of 2 Types of Human Breast Cancer. Can 1954. cer, 17: 141â€”151,1964. 45. Huggins, C., and Jensen,E. V. The Depressionof Estrone-induced 24. Dickinson, L. E., MacMahon, B., Cole, P., and Brown,J. B. Estrogen Uterine Growth by PhenolicEstrogenswith OxygenatedFunctionsat Profiles of Oriental and Caucasian Women in Hawaii. New EngI. J. Position 6 or 16; the Impeded Estrogens.J. Exptl. Med., 102: 335-346, Med.,291:1211-1213,1974. 1955. 25. Durant, A., and Miller, H. Non-specific FactorsThat May Influence 46. Irvine, W. T., Aitken, E. H., Rendelman, D. H., and Folea, P. J. Significance of Urinary Steroid Excretion in Breast Cancer. Brit. Med. Urinary Oestrogen Measurements after Oophorectomy and Adrenal J., 2: 767-770,1973. ectomy for Advanced Breast Cancer. Lancet, 2: 791-796, 1961. 26. Engstrom,W. W., Munson, P. L., Cook, L., and Costa, P. J. Studies 47. Juret, P., Hayem, M., and Flaisler, A. A Proposde 150Implantations on the Conversion of Testosterone to Urinary 17-Ketosteroids in d'Yttrium Radio-actif Intra-hypophysaires dans Ic Traitement du Normal Human Malesand Maleswith Hepatic andOther Diseases.J. Cancer du Scm a un Stade Avanc6. Technique; Bilan Clinique; Clin. Endocrinol., ii: 416â€”426,1951. Indications. J. Chir., 87: 409â€”433,1964. 27. Feinleib, M. Breast Cancer and Artificial Menopause: A Cohort 48. Keys, A., Brozek, J., Henschel, A., Mickelson, 0., and Taylor, H. L. Study. J. Natl. Cancer Inst., 4!: 315â€”329,1958. TheBiologyof HumanStarvation,Vol.1,p.755.Minneapolis,Minn.: 28. Fernandez,N. A. Nutrition in Puerto Rico. Cancer Res.,35: 3272- University of Minnesota, 1950. 3291,1975. 49. Kumaoka,S.,Sakauchi,N.,Abe,N., Kusama,M.,andTakatani,0. 29. Fishman, J., Boyar, R. M., and Hellman, L. Influence of Body Weight Urinary 17-Ketosteroid Excretion of Women with Advanced Breast on Estradiol Metabolism in Young Women. J. Clin. Endocrinol., in Cancer. J. Clin. Endocrinol., 28: 667-672, 1968. press. 50. Lemon, H. M. Genetic Predispositionto Carcinoma of the Breast: 30. Fishman,J., Hellman, L., Zumoff, B., andGallagher,T. F. Influence Multiple Human Genotypesfor Estrogen16-AlphaHydroxylase of Thyroid Hormone on Estrogen Metabolism in Man. J. Clin. Activity in Caucasians. J. Surg. Oncol., 4: 255-273, 1972. Endocrinol., 22: 389-392, 1962. 51. Lemon, H. M., Wotiz, H. H., Parsons, L., and Mozden, P. J. Reduced

3372 CANCER RESEARCH VOL.35

Estriol Excretion in Patients with Breast Cancer Prior to Endocrine Patients with Breast Cancer. Acta Endocrinol., 47: 15-26, 1964. Therapy. J. Am. Med. Assoc., 196: 112-120, 1966. 67. Rao, L. G. S. Urinary Steroid Excretion Patterns after Acute 52. Lipsett, M. B. Hormones, Nutrition, and Cancer. Cancer Res., 35: Myocardial Infarction. Lancet, 2: 390-391, 1970. 3359â€”3361,1975. 68. Reichlin, S., Bollinger, J., Nejad, I., and Sullivan, P. Tissue Thyroid 53. Loriaux, D. L., Ruder, H. J., Knob, D. R., and Lipsett, M. B. Estrone Concentration of Rat and Man Determined by Radioimmunoassay; Sulfate, Estrone, Estradiol and Estriol Plasma Levels in Human Biologic Significance.Mt. Sinai Med. J., 40: 502-510, 1973. Pregnancy. J. Clin. Endocrinol., 35: 887-891, 1972. 69. Rudali, G., Apiou, F., and Mud, B. Mammary Cancer Produced in 54. MacKenzie, I. The production of Mammary Cancer in Rats Using Mice with Estriol. European J. Cancer, 4: 39-43, 1975. Estrogens. Brit. J. Cancer, 9: 284-299, 1955. 70. Russell,G. F. M., Loraine, J. A., Bell, E. T., and Harkness, R. A. 55. MacMahon, B., Cole, P., and Brown, J. B. Etiology of Human Breast Gonadotropin and Oestrogen Excretion in Patients with Anorexia Cancer: A Review. J. NatI. Cancer Inst., 50: 21-42, 1973. Nervosa. J. Psychosomat. Res., 9: 79-85, 1965. 56. MacMahon, B., Cole, P., Brown,J. B., Aoki, K., Lin, T. M., Morgan, 71. Schweppe,J. S., Jungman, R. A., and Lewin, I. Urine Steroid R. W., andWoo,N. C. UrineOestrogenProfilesofAsianandNorth Excretion in Postmenopausal Cancer of the Breast. Response to American Women. Intern. J. Cancer, /4: 161-167, 1974. Corticotropic Stimulation and DexamethasoneSuppression.Cancer, 57. Marmorston, J., Crowley, L. G., Myers, S. M., Stern, E., and 20: 155-163,1967. Hopkins, C. E. Urinary Excretionof Estrone,Estradiol and Estriol by 72. Silberberg, M., and Silberberg, R. Age and Susceptibility of the Patientswith BreastCancerand BenignBreastDisease.Am.J. Obstet. Mammary Gland of Mice to Estrogen. Cancer Res., 11: 279-280, Gynecol., 92: 460-467, 1965. 1958. 58. Marmorston, J., Lombardo, L. J., Jr., Myers, S. M., Gierson, H., 73. Tulchinsky, D. Placental Secretion of Unconjugated Estrone, Stern, E., and Hopkins, C. E. Urinary Excretion of Neutral 17-Keto Estradiol and Estriol into the Maternal and Fetal Circulation. J. Clin. steroids and Pregnanediol by Patients with Prostatic Cancer and Endocrinol., 36: 1079- 1087, 1973. Benign Prostatic Hypertrophy. J. Urol., 93: 276-286, 1965. 74. Vessey, M. P., Doll, R., and Sutton, P. M. Investigation of the 59. Marmorston, J., Lombardo, L. J., Jr., Myers, S. M., Gierson, H., Possible Relationship betweenOral Contraceptivesand Benign and Stern, E., and Hopkins, C. E. Urinary Excretionof Estrone,Estradiol Malignant Breast Disease. Cancer, 28: 1395- 1399, 1971. and Estriol by Patients with Prostate Cancer and Benign Prostatic 75. Wicks, A. E., and Segal, S. J. Time and Dose Relationship in Hypertrophy. J. Urol., 93: 287-295, 1965. Estriol-Estradiol Interactions. Proc. Soc. Exptl. Biol. Med., 93: 60. Marmorston, J., Weiner, J. M., Hopkins, C. E., and Stern, E. 270â€”273,1956. Abnormalities in Urinary Hormone Patterns in Lung Cancer and 76. Wilson, R. E., Crocker, D. W., Fairgrieve, J., Bartholomay, A. F., Emphysema.Cancer,19: 985-995, 1965. Emerson, K., Jr., and Moore, F. D. Adrenal Structure and Function in 61. Metcalf, M. G. Effect of Age, Obesity, Starvation and Changes in Advanced Carcinoma of the Breast. II. The Relation of Steroid Adrenocortical and Ovarian Activity on the Urinary Excretion of Excretion to Adrenal Morphology and the Outcome of Adrenalectomy Dehydroepiandrosterone Relative to Androsterone + Etio with Description of a New Discriminant Function. J. Am. Med. cholanolone: Relevance to the Measurement of Breast Cancer Dis Assoc., 199: 474-482, 1967. criminants in Women. Clin. Biochem., 7: 119â€”130,1974. 77. Yoshizawa, I., and Fishman, J. Radioimmunoassay of 2-Hydroxyes 62. Morishita, H., Naftolin, F., Davies, I. J., Todd, R., Ryan, K. J., and trone in Human Plasma. J. Clin. Endocrinol., 32: 3-6, 1971. Fishman, J. Elevation of Serum LH in Male Rats by Catechol 78. Zumoff, B., Bradlow, H. L., Gallagher, 1. F., and Hellman, L. Estrogens. Endocrinology, 96 (Suppl.) 164, 1975. Decreased Conversion of Androgens to Normal I 7-Ketosteroid Me 63. Neuwirth, R. S., Philip, B. A., and Bondy, P. K. The Effect of tabolites: A Non-specific Consequence of Illness. J. Clin. Endocrinol., Prolonged Starvation on Adrenal Cortical Function. Yale J. Biol. 32:824â€”832,1971. Med., 36: 445-454, 1964. 79. Zumoff, B., Fishman,J., Cassouto,J., Gallagher,T. F., and Hellman, 64. Nissen-Meyer, P., and Sanner, T. The Excretion of Oestrone, L. InfluenceofAgeandSexonNormalEstradiolMetabolism.J.Clin. Pregnanediol and Pregnanetriol in Breast Cancer Patients. I. Excre Endocrinol., 28:937-941, 1968. tion after Spontaneous Menopause. Acta Endocrinol., 44: 334-345, 80. Zumoff, B., Fishman,J., Cassouto,J., Hellman, L., andGallagher,T. 65. Oiso, T. Incidence of Stomach Cancer and its Relation to Dietary F. EstradiolTransformationin Men with BreastCancer.J. Clin. Habits and Nutrition in Japan between 1900 and 1975. Cancer Res., Endocrinol., 26: 960-966, 1966. 35:3254â€”3258,1963. 81. Zumoff, B., Fishman,J., Gallagher,T. F., and Hellman, L. Estradiol 66. Persson, B. H., and Risholm, L. Oophorectomy and Cortisone Metabolism in Cirrhosis. J. Clin. Invest., 47: 20â€”25,1968. Treatment as a Method of Eliminating Oestrogen Production in

NOVEMBER 1975 3373

Barnett Zumoff, Jack Fishman, H. Leon Bradlow, et al.

Cancer Res 1975;35:3365-3373.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/35/11_Part_2/3365

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

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

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/35/11_Part_2/3365. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.