Effect of clomiphene citrate treatment on endometrial estrogen and progesterone receptor induction in women

Effect of clomiphene citrate treatment on endometrial estrogen and progesterone receptor induction in women Marc A. Fritz, MD,. Ronald T. Holmes; and EdwardJ. Keenan, PhDb.c Wright-Patterson Air Force Base, Ohio, and Portland, Oregon A direct adverse effect of clomiphene citrate on the endometrium has been presumed, and interference with estrogen receptor-mediated endometrial estrogen receptor and progesterone receptor induction has been implicated as the mechanism responsible for an increased incidence of luteal phase deficiency in association with clomiphene citrate treatment. To clarify the net influence of clomiphene administration on endometrial steroid receptor induction, we studied five normal ovulatory women, in both a spontaneous and clomiphene-induced (150 mg/day, cycle days 5 to 9) ovulatory cycle. From cycle day 11 blood samples were obtained daily and urinary luteinizing hormone determinations were performed twice daily. Endometrial biopsy was performed on the day of the urinary luteinizing hormone surge and again 13 days after the surge. Serum levels of follicle-stimulating hormone and luteinizing hormone were determined by immunoradiometric assay, estradiol and progesterone by radioimmunoassay, and clomiphene citrate isomer concentrations in treatment cycles by reversed-phase high-performance liquid chromatography and fluorescence detection. Total, cytosolic, and salt-extracted nuclear endometrial estrogen receptor and progesterone receptor concentrations were determined by enzyme-linked immunoassay. Serum estradiol was threefold to fivefold higher (p < 0.05) in clomiphene-induced than in spontaneous cycles 8 and 10 days before the luteinizing hormone surge, and progesterone was increased (p < 0.05) from the day of the surge to end of the cycle. Serum enclomiphene rose to plateau between 12 and 6 days before the luteinizing hormone surge (4.1 ± 0.8 ng/ml, mean ± SE, n = 19) and fell thereafter to <1.0 ng/ml. Zuclomiphene levels increased rapidly between 14 and 8 days before the surge (53.9 ± 2.8 ng/ml, mean ± SE, n = 5) and then decreased gradually but remained elevated throughout the luteal phase (29.0 ± 1.2 ng/ml, mean ± SE, n = 33). Late luteal endometrial histology was abnormal in one of four available treatment cycle specimens, but the endocrine characteristics and number and subcellular distribution of estrogen receptor and progesterone receptor in the abnormal cycle were not different from those of normal, in-phase cycles. Total estrogen receptor and progesterone receptor at midcycle were three to four times higher (p < 0.01) than in the late luteal phase in both spontaneous cycles (7.2 ± 2.3 vs 1.9 ± 0.4 pmol/mg deoxyribonucleic acid estrogen receptor and 41.3 ± 13.9 vs 13.4 ± 3.9 pmol/mg deoxyribonucleic acid progesterone receptor) and clomiphene-induced cycles (8.2 ± 2.3 vs 1.8 ± 0.8 pmol/mg deoxyribonucleic acid estrogen receptor and 40.4 ± 9.9 vs 10.1 ± 4.1 pmol/mg deoxyribonucleic acid progesterone receptor), but neither midcycle nor late luteal estrogen receptor or progesterone receptor in any category (total, cytosolic, or salt-extracted nuclear) differed from spontaneous cycle values. These data strongly suggest that clomiphene citrate treatment does not adversely affect endometrial estrogen receptor and progesterone receptor induction. (AM J OSSTET GVNECOL 1991 ;165: 177-85.)

Key words: Clomiphene citrate, estrogen receptors, progesterone receptors, endometrium

From the Department of Obstetrics and Gynecology, United States Air Force Medical Center Wright-Patterson," and the Hormone Receptor Laboratory' and the Department of Obstetrics and Gynecology,' Oregon Health Sciences University. This work was supported by a grant from the Office of the Surgeon General, United States Air Force. Presented in part at the Thirty-seventh Annual Meeting of the Society for Gynecologic Investigation, St. Louis, Missouri, March 21-24, 1990. The opinions expressed in this article are those of the authors and not necessarily those of the United States Air Force or the Department of Defense. Received for publication July 24, 1990; revised December 27, 1990; accepted January 31, 1991. Reprint requests: Marc A. Fritz, Lt. Col. USAF MC, Department of Obstetrics and Gynecology, Uniformed Serivces Univesity of the Health Sciences, 4301 Jones Bridge Road, Bethesda, MD 20814.

611128540

Clomiphene citrate, a triphenylethylene compound, is the most common agent in clinical use for ovulation induction. Clomiphene citrate binds to the estrogen receptor at numerous sites throughout the reproductive system, including the hypothalamus, pituitary, ovary, endocervix, and endometrium. Its mechanism of action in ovulation induction appears to involve interference with the process of estrogen receptor replenishment at the hypothalamic level, thereby causing a central misinterpretation of circulating estrogen concentrations, the perception of a hypo gonadal state, and, by altering the pattern of pulsatile gonadotropinreleasing hormone secretion, stimulation of increased pituitary gonadotropin release and enhanced follicular

177

178 Fritz, Holmes, and Keenan

development. However, in addition to its desirable central actions, clomiphene citrate exerts the same antiestrogenic influence at sites in the periphery (some of which are undesirable and potentially counterproductive). Clinically, luteal phase deficiency frequently complicates ovulation induction with clomiphene citrate. I. 2 Although the mechanism or mechanisms responsible remain unclear, many consider the problem to be an unavoidable consequence of competition between central and peripheral actions of the drug. Because luteal phase defects may result from inadequate progesterone secretion and/or a failure of endometrial response, levels of luteal phase progesterone production that are typically higher in clomiphene-induced than in spontaneous cycles, suggest that clomiphene may have direct adverse effects on the endometrium that predispose to luteal phase deficiency. The drug's inherently antiestrogenic nature further implies that clomiphene may interfere with estrogen induction of effective endometrial estrogen and progesterone receptor populations. Clearly, inadequate estrogen and/or progesterone receptor or abnormal receptor function could result III delayed, inadequate, or asynchronous endometrial development, but to date the net influence of clomiphene administration on endometrial steroid hormone receptor populations has not been clearly defined. The extent of endometrial estrogen and progesterone receptor induction in cycles induced by clomiphene citrate must reflect the relative balance between two opposing influences. Although receptor induction in clomiphene-induced cycles may well be enhanced by typically elevated preovulatory estradiol levels, it may also be inhibited by a premature rise in circulating progesterone concentrations' and antagonized directly by the drug itself. Previous studies that we 4 and others 57 have conducted to examine this question have been inconclusive. Heretofore, attention has focused on the luteal phase endometrium, in which steroid receptor concentrations have normally already fallen substantially from peak values at midcycle." Ail studies 47 have used conventional steroid-binding assay methods. Unfortunately, antiestrogens or their metabolites have demonstrated the capacity to effectively interfere with in vitro binding of radioligands to estrogen receptor." Thus steroid-binding assay may not be entirely suitable for examinations of endometrial steroid receptor dynamics in clomiphene-treated patients. In this study we have futher investigated the effect of clomiphene citrate on endometrial estrogen receptor and progesterone receptor populations. To avoid the influence of potentially confounding factors present in infertile populations, we studied normal ovulatory women, each of whom was examined in a spontaneous

Julv 1991 Am J Obstet Gynecol

and clomiphene-induced cycle. To better evaluate the net influence of clomiphene treatment on endometrial development and steroid receptor induction, we examined both midcycle and late luteal endometrial tissue specimens in all cycles. To eliminate the possibility that administered antiestrogen might interfere with accurate receptor quantitation, we used an enzyme-linked immunoassay rather than a steroid-binding assay in determining estrogen receptor and progesterone receptor concentrations. Material and methods Subjects. Five normal parous women, aged 26 ± 3 years (mean ± SD), were recruited and gave informed consent for this study, which was approved by our institutional review board. Each woman had regular predictable menstrual cycles and had not used steroid contraception for at least 6 months before the study. The menstrual cycle of each participant was first screened by basal body temperature monitoring to eliminate possible gross ovulatory dysfunction. Study design. Each woman was studied during two separate menstrual cycles. In the first cycle (control), venous blood samples were obtained on alternate days from cycle day 3 to 11. Beginning on day 11, a urine sample was obtained twice daily, the first between 11 AM and 3 PM and the second between 5 and 10 PM. Blood samples were drawn daily until I day after the urinary luteinizing hormone (LH) surge was identified via a rapid semiquantitative commercial enzyme immunoassay kit, Ovustick (Monoclonal Antibodies, Inc., Mountain View, Calif.) that uses the first international standard for menopausal gonadotropins. Thereafter, urine sampling was discontinued and blood sampling returned to alternate days until the onset of menses. An endometrial biopsy speciman was obtained from the uterine fundus with a disposable aspiration cannula (Pipelle, Unimar, Wilton, Conn.) on the day of the urinary LH surge and again 13 days later. In the second cycle blood and urine sampling and endometrial biopsy were carried out in an identical fashion, and 150 mg of clomiphene citrate (Clomid, Merrell, Cincinnati) was self-administered in a single daily dose on cycle days 5 through 9. Venous blood samples were allowed to clot, and the separated serum was frozen at - 20° C until it was assayed for estradiol, progesterone, follicle-stimulating hormone (FSH), LH, and clomiphene isomers. Representative samples of each endometrial tissue specimen were fixed in 10% formalin; the remainder of each specimen was placed in a sealable plastic envelope, immediately snap-frozen by immersion in liquid nitrogen, and stored at - 800 C for determination of cytosolic and salt-extracted nuclear endometrial estrogen and progesterone receptor concentrations. Histologic sec-

Volume 165 Number I

tions were examined by three independent gynecologic pathologists, all using the criteria of Noyes et al. 10 and without knowledge of the experimental conditions. Each late luteal phase biopsy specimen was assigned a histologic date encompassing a 2-day interval and interpreted according to the most advanced day; the mean result was judged with reference to both the serum LH surge and onset of menses. Endometrial specimens that were ~3 days out of phase were considered abnormal and consistent with the diagnosis of luteal phase defect. Hormone, clomiphene, and receptor assays. Serum estradiol and progesterone concentrations were measured by specific radioimmunoassay; concentrations of FSH and LH were measured by immunoradiometric assay with commercial assay kits (Diagnostic Products, Los Angeles). To reduce interassay variability, all serum samples from the control and clomiphene treatment cycles of each individual woman were assayed together for each hormone. The intraassay and interassay coefficients of variation were < 10% in all four assay systems. Detection limits were 0.12 and 0.2 mIU/ml for LH and FSH, respectively, and 8.0 pg/ml and 0.5 ng/ml for estradiol and progesterone, respectively. Gonadotropin values were expressed in international units of the first international reference preparation 68/40 for LH and the second international reference preparation of human menopausal gonadotropin for FSH. Serum concentrations of enclomiphene and zuclomiphene isomers in treatment cycles were determined by methods previously described in detail. II Briefly, samples were extracted with tert-butyl methyl ether, and the drug and its metabolites were separated by reversed-phase high-performance liquid chromatography. The eluent was fed into a braided or woven, cylindric Teflon reaction coil into which a low-energy mercury lamp had been inserted. The result was a photoinduced stilbene-to-phenanthrene oxidation that yielded highly fluorescent analytes that were then quantified with fluorescence spectrophotometry. The assays were performed by Hazelton Laboratories (Madison, Wis.) in affiliation with Serono Laboratories (Norwell, Mass.). Intraassay and interassay coefficients of variation were < 10% for both the enclomiphene and zuclomiphene isomers. Frozen endometrial tissue specimens were thawed on ice and homogenized (l: 10, wt/vol) in TEDG-Mo buffer (0.05 mollL Tris hydrochloride, 1.5 mmollL sodium ethylenediaminetetraacetic acid, 0.5 mmollL dithiothreitol, 10% glycerol, and 10 mmoll L sodium molybdate; pH 7.5) with two bursts (10 seconds) of a precooled Polytron (Brinkman Instruments, Westbury, N.Y.) at a setting of 5 with a 10-second cooling interval. Tissue homogenates and all reagents were maintained at 0° to 4° C during processing and assay. Homogenates

Clomiphene effect on endometrial steroid receptors

179

were centrifuged at 800 g for 10 minutes to provide a crude nuclear pellet. To prepare the cytosol fraction, the supernatant was removed and centrifuged at 40,000 g for 20 minutes; it was then diluted with TEDGMo buffer to yield a protein concentration of 0.5 to 1.0 mgl ml as determined by the method of Lowry et al. 12 Crude nuclear pellets were washed once with TEDGMo buffer (5 ml), and a nuclear extract was obtained by resuspending pellets in hypertonic buffer (0.6 mollL potassium chloride-TEDG-Mo buffer) and incubating them for 60 minutes. Before recentrifugation (40,000 g for 20 minutes), nuclear extracts were diluted 1: 1 with TEDG-Mo buffer to yield a protein concentration in a range from 0.5 to 1.0 mg/ml. Estrogen receptor and progesterone receptor enzyme-linked immunoassay kits (Abbott Laboratories, North Chicago) were obtained, and the assays were conducted according to the manufacturer's specifications. Endometrial estrogen and progesterone receptor concentrations were expressed in picomoles per milligram of deoxyribonucleic acid after quantitation of tissue deoxyribonucleic acid content by the method of Burton. 13 Data analysis. Hormone data for all cycles and drug data from treatment cycles were synchronized to the day of the serum LH peak (LH + 0), and significant differences in hormone concentrations between the control and clomiphene treatment cycles were determined by two-way analysis of variance (day of cycle and experimental groups) for repeated measures. Results of histologic endometrial dating were compared with the McNemar test. Receptor concentrations within and between control and treatment cycles were compared with paired t tests; p < 0.05 was considered significant. Results

Clinical findings. Each of the menstrual cycles of all five women was ovulatory, as judged by typical biphasic basal body temperature patterns, the identification of an LH surge in both urine and serum, a postovulatory rise in serum progesterone concentrations, and the finding of secretory endometrium in the second biopsy specimen obtained in each of the studied cycles (both control and clomiphene treatment cycles. A urinary LH surge was identified in all cycles and correlated well with that in serum. In all women peak serum LH concentrations occurred on or 1 day after the day of urinary LH surge onset. As a result, midcycle endometrial biopsies in both control and clomiphene-induced cycles were performed on LH - lor LH + 0, and late luteal endometrial biopsy specimens were uniformly obtained on LH + 12 or LH + 13 in 9 of 10 study cycles. One patient experienced onset of menses before a scheduled late luteal biopsy after clomiphene treatment. Therefore, data derived from treatment cycle LH + 13 tissue

180

Fritz, Holmes, and Keenan

.July 1991 Am.J Obstet Gynccol

20.-------------------------------------, 800

-:J' :::;;

'-- 600 t.'J

~ ---'

§ « 0::

I U1 LL

til

5

w

400 200

-:J' :::;; '-t.'J

30 ':]

~ w z o

~ 20 ::::)

E

0:: W

til w

'-' 10 I

--.J

t.'J

o

o

0::

CL

-12

-8

-4

0

4

8

12

DAYS FROM LH PEAK

-12

-8

-4

o

4

8

12

DAYS FROM LH PEAK

Fig. 1. Serum FSH and LH concentrations (mean ± SE) in five women during a spontaneous cycle (control) and a cycle in which they received clomiphene citrate (150 mg/d, cycle days 5 to 9). FSH values are expressed in terms of second international reference preparation human menopausal gonadotropin and those of LH in terms of the first international reference preparation 68/40.

Fig. 2. Serum estradiol and progesterone concentrations (mean ± SE) in five women during a spontaneous cycle (control) and during a cycle in which they received clomiphene citrate (150 mg/ day, cycle days 5 to 9). Asterisk, p < 0.05.

specimens reflect results obtained in four of the five women studied. Overall cycle length, day of LH surge, and luteal phase duration are summarized in Table I. None of these paramters in clomiphene treatment cycles was significantly different from corresponding control cycle values. Hormone data. Serum LH and FSH concentrations in the control and clomiphene treatment cycles of the five women are shown in Fig. I. Although midfollicular-phase FSH levels tended to be higher and luteal phase concentrations lower in clomiphene cycles, the differences were not significant. Similarly, LH levels throughout the cycle of clomiphene administration were not different from those observed in control cycles. Serum concentrations of estradiol and progesterone in control cycles and after clomiphene treatment are illustrated in Fig. 2. As expected, serum estradiol levels in clomiphene cycles were significantly higher (P < 0.05) than corresponding control cycle values. Estradiol concentrations remained significantly elevated throughout the interval from day LH - 8 to day LH + 10 in clomiphene cycles; levels in the follicular phase were threefold to fivefold greater than they were without clomiphene treatment. Serum progesterone concentrations were also significantly higher (p < 0.05)

after clomiphene administration. This difference first became apparent in the periovulatory phase; significantly increased progesterone levels were observed on days LH + 0 and LH + 1. Mean serum progesterone level remained elevated throughout the luteal phase but to a widely varying extent among patients; concentrations in one individual rose as high as 166 ng/ml. Clomiphene isomers. Serum concentrations of enclomiphene and zuclomiphene in treatment cycles are shown in Fig. 3. Enclomiphene levels rose abruptly to a plateau (4.1 ± 0.8 ng/ml, mean ± SE, n = 19) between day LH - 12 and day LH - 6. Thereafter, concentrations fell equally as abruptly to <1.0 ng/ml, where they remained for the rest of the cycle. In contrast, levels of zuclomiphene increased rapidly from day LH - 14 to peak on day LH - 8 (53.9 ± 2.8 ng/ml, mean ± SE, n = 5) and then very gradually fell until onset of the subsequent menses. The overall mean luteal phase concentration of zuclomiphene was 29.0 ± 1.2 ng/ml (mean ± SE, n = 33). Receptor concentrations. Endometrial estrogen receptor concentrations are shown in Fig. 4. As we had anticipated, midcycle (day LH + 0) levels of total, cytosolic, and salt-extracted nuclear estrogen receptor were significantly higher (p < 0.01) than late luteal

Clomiphene effect on endometrial steroid receptors

Volume 165 :-lumber I

8

::J

7

"<..?

6

:::;;; Z

~

w w I a..

z

~ 0

~

u zw

5 4 3 2

0

10.0

I

0: 0

1

1

f-~

a.. <

w Z U O wCl

rl\(~

0::::;;

Z"-. w....J ClO

0:::;; 0: a..

"<..? z ~

60 50

w z 40 w I a.. 30 ~ 0

'-i-.-._._._._._._._.

~

u

::J N

20 10 0

6.0

i

w

j'l-!'!,T---Ll-: 1

J.

/1

•

-12

-8

a a

2.0

T T • .-e_.-".1. J. •

J."-.T J.

4

8

12

DAYS FROM LH PEAK

b b

0.0 t---'---L..l.-R.....J-

a

a

a

a

b

0.0 0

CONTROL CLOMIPHENE

a

If)

1/+,1

-4

a

4.0

50.0

1

D -

f-~

::J 70 :::;;;

8.0

a a

181

b

b b TOTAL CYrO NUCL LH + 13

Fig. 3. Serum enclomiphene and zuclomiphene concentrations (mean ± SE) in five women during a cycle in which they received clomiphene citrate (150 mg/day, cycle days 5 to 9),

Fig. 4. Endometrial concentrations (mean ± SE) of total, cytosolic ((yta), and salt-extracted nuclear (nuci) estrogen and progesterone receptor in five women during a spontaneous cycle and during a cycle in which they received clomiphene citrate (150 mg/day, cycle days 5 to 9). In each class (total, nto, and nucl) bars with different superscrijJ/s are different (p < 0.0 I).

phase (LH + 13) values in both control and clomiphene treatment cycles, However, despite the significant sustained increase in circulating estradiol levels associated with clomiphene administration, neither midcycle nor late luteal estrogen receptor concentrations in any of the three categories differed significantly from control cycle values. In both cycles the proportion of total estrogen receptor recovered from nuclear extracts was significantly greater (p < 0.01) in the LH + 0 tissue specimens than in the LH + 13 biopsy specimens, but neither the mid cycle luteal values nor the late luteal values were altered by the clomiphene treatment. Immunoassayable endometrial progesterone receptor concentrations are also illustrated in Fig. 4. As was true of estrogen receptor, total, cytosolic, and saltextracted nuclear progesterone receptor on day LH + 0 were all significantly elevated (p < O.(ll) when compared with day LH + 13 values in both control and clomiphene treatment cycles. Again, however, both midcycle and late luteal progesterone receptor concentrations in all three categories remained unchanged after clomiphene administration. The proportion of total endometrial progesterone receptor residing within the nuclear fraction was relatively low in all tissue specimens (6.5% ± 2.6%, mean ± SE, n = 19) and did not vary with phase of cycle or clomiphene treatment.

Despite the fact that mean periovulatory progesterone levels were significantly higher in clomiphene cycles (Fig. 2), there was no associated significant shift in the overall mean percent nuclear progesterone receptor observed between the two cycles in day LH + 0 specimens. There was, however, a tendency toward increased percent nuclear progesterone receptor in clomiphene cycle day LH + 0 tissues in those patients who had the greatest relative increases in periovulatory progesterone over the respective control cycle values. Endometrial histology. Late luteal phase endometrial histology revealed no significant difference between control and clomiphene treatment cycles overall. All control cycle biopsy specimens and three of the four treatment cycle specimens were judged normal, whether they were evaluated with reference to the serum LH surge or to the onset of the next menses. The treatment cycle biopsy specimen in one patient was abnormal because it exhibited a maturation delay of 4 and 5 days relative to onset of menses and LH surge, respectively, but neither the endocrine characteristics nor the number and subcellular distribution of midcycle or late luteal phase endometrial estrogen or progesterone receptor in this abnormal cycle were significantly different from those values in the remaining three treatment cycles in which the histology was normal.

182

Fritz, Holmes, and Keenan

July 1991 Am J Obstet Gynecol

Table I. Cycle characteristics before and during clomiphene citrate treatment (mean ± SEM) Cycle Characteristic

Control

Clomiphene

Cycle length (days) LH surge (cycle day) Luteal phase duration (days)

30.8 ± 1.3

29.8 ± 0.9 16.4 ± 0.8

17.2 ± 1.2 13.6 ± 0.2

13.4 ± 0.4

Comment

Clomiphene citrate has been implicated as a cause of luteal phase deficiency and, paradoxically, also is advocated for treatment of the same disorder. Its therapeutic potential derives from luteal phase steroid hormone production that is consistently greater in clomiphene-induced than in spontaneous ovulatory cycles. The higher circulating steroid levels almost certainly result from a drug-induced increase in gonadotropin secretion, improved folliculogenesis in the single dominant follicle and! or the larger cohort of follicles recruited to participate in the treatment cycle, II and a subsequent increase in overall luteal cell mass. Although the drug's extended half-life may alternatively afford it the opportunity to influence corpus luteum steroidogenesis more directly, available evidence indicates that such effects, if they are exerted at all, are inhibitory in nature." However, the ability of clomiphene to antagonize estrogen actions in the periphery may at the same time prevent an appropriate endometrial response to normal or even increased steroid concentrations. Because normal secretory endometrial maturation requires the combined actions of estrogen and progesterone and because both steroids exert their effects through specific estrogen-induced endometrial receptors, direct interference with estrogen induction of effective endometrial estrogen and progesterone receptor populations is one mechanism that might explain the increased incidence of luteal phase deficiency associated with clomiphene use. 1.2 Clomiphene treatment typically increases late follicular phase serum estradiol and progesterone concentrations; both may exert significant, if opposite, influence on endometrial steroid receptor induction and thus may modify any direct antiestrogenic actions of clomiphene itself. The effect of these many competing direct and indirect clomiphene-induced changes has not been clear. Clomiphene has been shown to reduce concentrations of endometrial estrogen and progesterone receptor induced by estrogen replacement in postmenopausal women. Moreover, decreased concentra-

tions of both receptors have been identified in late luteal phase endometrium after clomiphene induction of ovulation," after continuous luteal phase clomiphene administration in normal women, I and in tissue specimens obtained at the time of oocyte retrieval after combined treatment with clomiphene, human menopausal gonadotropins, and human chorionic gonadotropin for in vitro fertilization." Together, these data suggest that clomiphene does interfere with the induction and! or maintenance of effective endometrial steroid receptor populations. Other studies, however, deny that clomiphene has an adverse effect on the endometrium. A recent report that periimplantation phase nuclear estrogen and progesterone receptor levels and binding affinity in normal ovulatory women were no different after both low-dose (50 mg) and high-dose (150 mg) clomiphene treatment than in spontaneous cycles is of particular interest. 7 Interpretation of these conflicting data has been confounded by diversity in study populations and methodologic differences and is now further complicated by recent evidence that conventional steroid-binding assay methods used in these studies may yield inaccurate estrogen receptor determinations in tissues exposed to antiestrogens in vivo. To clarify the effect of clomiphene on endometrial estrogen and progesterone receptor induction, we elected to examine tissue specimens that would reflect both the peak (midcycle) and nadir (late luteal phase) in receptor populations. We used a recently developed enzyme-linked immunoassay to measure total, cytosolie, and salt-extracted nuclear estrogen and progesterone receptor concentrations in endometria derived from both spontaneous and clomiphene-induced cycles in the same patients. The enzyme-linked immunoassay is based on the recognition of two distinct antigenic determinants in the receptor molecule that are distinct from the hormone-binding region and that require the presence of an intact receptor molecule for detection. The method is therefore independent of the state of receptor occupancy and applies equally well to receptor quantitation in both cytosol and nuclear fractions of tissue homogenates. Enzyme-linked immunoassay determinations of cytosolic and nuclear estrogen and progesterone receptor generally correlate well with data generated by steroid-binding assay in both normal endometrium ' " and in breast cancer tissue" and have been demonstrated to accurately predict the endocrine dependence of human breast cancers. However, only immunoassay detected significant levels of cytosolic estrogen receptor in breast tumor tissue from patients under treatment with tamoxifen, a potent antiestrogen 9 ; steroid-binding assay failed to detect estrogen receptor in these specimens. The mechanism

Volume 165 Number 1

responsible for this discrepancy is not yet clear, but these data demonstrate the potential for steroid-binding assay to underestimate estrogen receptor in tissues of patients treated with antiestrogens. The increase in follicular phase estradiol levels observed in treatment cycles was striking and remarkable in the absence of a distinct rise in gonadotropin secretion. Mean FSH and LH levels were indeed higher than corresponding control values during the interval of clomiphene administration, but the differences were not statistically significant, probably because the rise in each was subtle and the sample size small. The possibility remains that clomiphene administration may impair the metabolic clearance of estradiol or may result in the release of an altered form of FSH that has enhanced biologic activity, but there is no evidence to indicate that clomiphene might stimulate ovarian steroidogenesis directly. I; The increase in estradiol production after clomiphene treatment probably reflects the recruitment of additional follicles more than increased estrogen synthesis in the dominant follicle already selected. 14 Regrettably, however, we did not monitor follicular development with serial ultrasonography in this study, and we acknowledge that the mechanism responsible for the marked rise in estradiol levels requires additional investigation. As anticipated, the concentrations of total, cytosolic, and salt-extracted nuclear estrogen and progesterone receptor were markedly higher at mid cycle than during the late luteal phase. These findings are consistent with earlier studies 8 and reflect the largely unopposed influence of estrogen on estrogen and progesterone receptor induction during the follicular phase of the menstrual cycle. Despite an approximately threefold increase in circulating estradiol levels after clomiphene administration, no significant net change in the concentration of estrogen or progesterone receptor in any category was observed in treatment cycles. The failure of elevated estradiol to induce increased concentrations of endometrial estrogen and progesterone receptor at midcycle may result from the inhibition of circulating clomiphene and/or the premature periovulatory rise in serum progesterone, because progesterone inhibits estrogen receptor production and, indirectly, the induction of its own receptor. Alternatively, already maximal concentrations of endometrial estrogen and progesterone receptor may be induced by normal ovulatory threshold levels of estradiol, and elevated estrogens may, conversely, buffer an otherwise inevitable decline in receptor levels caused by clomiphene and/or elevated progesterone. Our data are in general agreement with those of Hecht et a\.7 in a recent study of midluteal phase endometrium that uses a similar

Clomiphene effect on endometrial steroid receptors

183

design. Their somewhat lower absolute levels of receptor may be attributed to the use of crude nuclear suspensions and elevated incubation temperatures (which are characteristic of steroid-binding assay) that may lead to underestimation of nuclear steroid receptor content, in contrast to our use of high-salt nuclear extracts and enzyme-linked immunoassay. Although we observed no change in the number and subcellular distribution of estrogen and progesterone receptor in whole endometrium after clomiphene treatment, a shift in the distribution of estrogen and progesterone receptor between the stromal and glandular compartments may have occurred and gone unrecognized. Complex and dyssynchronous fluctuations in the secretory phase progesterone receptor content of endometrial glands and stroma have recently been demonstrated. I7 Such changes would not be detected by enzyme-linked immunoassay and could be demonstrated only by an immunocytochemical approach to steroid receptor detection. The late luteal phase endometrial histology was normal in three of four available treatment cycle specimens, exhibited a clinically significant maturation delay consistent with luteal phase deficiency in one individual, and was not evaluated in another patient because of the early onset of menses. The endocrine characteristics and number and subcellular distribution of endometrial estrogen and progesterone receptor in the abnormal cycle were indistinguishable from those of in-phase cycles. Although the number of individuals studied is small, these data clearly argue against druginduced disruption of endometrial steroid receptor induction as the mechanism of luteal phase deficiency associated with the administration of clomiphene citrate. Nevertheless, the clincial relevance of our findings in normal ovulatory women may be questioned because numerous studies of endometrial steroid receptor levels have failed to establish inadequate or dysfunctional receptors as a cause of luteal phase deficiency and because the link drawn between clomiphene treatment and the increased risk of luteal phase inadequacy derives from observations made in anovulatory infertile patients undergoing long-term therapy. The widely varying estrogen and progesterone receptor content of endometrial tissue derived from infertile women with spontaneous luteal phase defects probably reflects only the many and varied mechanisms that may cause luteal phase inadequacy, as well as the unavoidable study of etiologically diverse populations within and between such studies. In any case the lack of a characteristic endometrial steroid receptor profile in luteal phase deficiency does not exclude altered receptor dynamics as a cause of the disorder. However,

184

Fritz, Holmes, and Keenan

the relative endocrine impact of experimental acute clomiphene administration in normal women may indeed significantly differ from that in anovulatory individuals who receive continuous treatment. The extent to which the estrogen agonistic and antagonistic actions of clomiphene isomers are expressed may vary with dose, duration of therapy, and the differing endocrine milieus of normal and anovulatory women. The extended half-life of clomiphene and its corresponding potential for significant systemic accumulation have been demonstrated. Even in our isolated treatment cycle, we observed a striking, sustained elevation of serum ~ zuclomiphene throughout the luteal phase. Circulating concentrations that are higher still almost certainly exist in patients who fail to conceive promptly and who therefore continue treatment for extended periods of time. Therefore dose- and time-dependent adverse effects of clomiphene on endometrial steroid receptor populations cannot be excluded, in spite of the use of high-dose clomiphene in this study and in other studies that have failed to demonstrate such an effect. If clomiphene treatment does, in fact, predispose to an increased risk of luteal phase deficiency, mechanisms that are unrelated to endometrial steroid receptor concentrations and/or function are likely to be involved. These mechanisms may be unique to clomiphene cycles or they may be similar to those that operate in women with spontaneous luteal phase defects. Direct adverse endometrial effects of clomiphene that operate through post-receptor mechanisms remain a possibility. The occasional marked elevation of luteal phase estradiol that follows clomiphene-induced ovulation may directly interfere with endometrial decidualization,18 and there is evidence that clomiphene may directly inhibit corpus luteum steroidogenesis. 15 Second, inadequate luteal function is often only a natural consequence of similarly inadequate folliculogenesis; luteal phase deficiency consistently accompanies follicular phase FSH deficiency that is spontaneous or experimentally induced. Because clomiphene induction of ovulation is an empiric incremental titration in which success depends on the extent of induced pituitary FSH release, clomiphene-associated luteal phase deficiency may at times reflect only inadequate treatment rather than a specific deleterious action of the drug itself. An abnormally increased follicular phase LH pulse frequency is a subtle aberration of gonadotropin secretion that is also implicated in the pathogenesis of luteal phase defects. 19 Interestingly, clomiphene appears to induce a similar increase in LH pulse frequency which presumably reflects an increase in pulsatile hypothalamic gonadotropin-releasing hormone secretion."O Excess clomiphene, whether it results from frank overdosage or accumulation after long-term administration, may

July 1991

Am J Obstet Gynecol

provoke an abnormally rapid pattern of gonadotropin secretion that predisposes to inadequate corpus luteum function. Regardless of the mechanism involved, the potential for luteal phase deficiency to accompany either too much or too little drug suggests that the effective therapeutic range for clomiphene may be narrower than that previously appreciated, at least in some individuals. In summary, clomiphene citrate treatment (150 mg) did not affect concentrations of immunoassayable midcycle or late luteal phase endometrial estrogen receptor and progesterone receptor in normal ovulatory women. Although an adverse effect of low-dose or longterm clomiphene citrate administration cannot be excluded, these data refute a drug-induced disruption of endometrial steroid receptor induction as the mechanism of luteal phase deficiency associated with clomiphene citrate. We thank Dr. Vinod Chungi (Serono Laboratories, Inc., Norwell, Mass.) and Mr. Jerry Brisson (Hazleton Laboratories, Madison, Wis.) for their generous assistance and support in providing serum assays of clomiphene isomers, and Unimar, Inc. (Wilton, Conn.) for providing the Pipe lie aspiration cannulas used in this study. REFERENCES 1. Cook CL, Schroeder ]A, Yussman MA, Sanfilippo ]S. Induction of luteal phase defect with clomiphene citrate. AMJ OBSTET GYNECOL 1984;149:613-6. 2. Garcia], Jones GS, Wentz AC. The use of clomiphene citrate. Ferti! Steril 1977;28:707-17. 3. Fedele L, Brioschi D, Marchini M, Dorta M. Parazzini F. Enhanced preovulatory progesterone, levels in clomiphene citrate-induced cycles . .I Clin Endocrinol Metab 1989;69:681-3. 4. Fritz MA, Westfahl PK, Graham RL. Effect ofluteal phase estrogen antagonism on endometrial development and luteal function in women. J Clin Endocrinol Metab 1987;65: 1006-13. 5. Aksel S, Saracoglu OF, Yeoman RR, Wiebe RH. Effects of clomiphene citrate on cytosolic estradiol and progesterone receptor concentrations in secretory endometrium. AMJ OBSTET GYNECOI. 1986;155:1219-23. 6. "Ronnberg L, Isotalo H, Kauppila A, Martikainen H, Vihko R. Clomiphene-induced changes in endometrial receptor kinetics on the day of ovum collection after ovarian stimulation: a study on cytosol and nuclear estrogen and progestin receptors and 17f\-hydroxysteroid dehydrogenase. Ann NY Acad Sci 1985;442:408-15. 7. Hecht BR, Khan-Dawood FS, Dawood MY. Peri-implantation phase endometrial estrogen and progesterone receptors: effect of ovulation induction with clomiphene citrate. AM] OBSTET GYNECOL 1989;161:1688-93. 8. Edman CD. The effects of steroids on the endometrium. Semin Reprod Endocrinol 1983; I : 179-88. 9. Keenan E], Corbin D. Significance of steroid receptor immunoassay in breast cancer. In: Ceriana RL, ed. Breast cancer immunodiagnosis and immunotherapy. New York: Plenum Press, 1988:149. 10. Noyes RW, Hertig AT, Rock J. Dating the endometrial biopsy. Ferti! Steril 1950; 1:3-23.

Volume 165 Number 1

11. Baustian CL, Mikkelson TJ. Analysis of clomiphene isomers in human plasma and detection of metabolites using reversed-phase chromatography and fluorescence detection. J Ph arm Biomed Anal 1986;4;237-46. 12. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. Protein measurement with the Folin phenol reagent. J Bioi Chern 1951; 193:265-75. 13. Burton K. A study of the conditions and mechanism of the diphenylamine reaction for the colorimetric estimation of deoxyribonucleic acid. Biochem J 1956;62:315. 14. Guzick DS, Zeleznik A. Efficacy of clomiphene citrate in the treatment of luteal phase deficiency; quantity versus quality of preovulatory follicles. Fertil Steril 1990;54:20610. 15. Yuen BH, Mari N, Duleba AJ, Moon YS. Direct effects of clomiphene on the steroidogenic capability of human granulosa cells. Ferti! Steril 1988;49:626-31. 16. Senekjian EK, Press MF, Blough RR, Herbst AL, DeSombre ER. Comparison of the quantity of estrogen receptors in human endometrium and myometrium by

Clomiphene effect on endometrial steroid receptors

steroid-binding assay and enzyme immunoassay based on monoclonal antibodies to human estrophilin. AM J OllSTET GYNECOL 1989;160:592-7. 17. Lessey BA, Killam AP, Metzger DA, Haney AF, Greene GL, McCarty KS. Immunohistochemical analysis of human uterine estrogen and progesterone receptors throughout the menstrual cycle. J Clin Endocrinol Metab 1988;67:334-40. 18. Daly DC, Maslar lA, Riddick DH. Prolactin production during in vitro decidualization of proliferative endometrium. AM J OBSTET GYNECOL 1983; 145:672-8. 19. Soules MR, Clifton DK, Cohen NL, Bremner WJ, Steiner RA. Luteal phase deficiency: abnormal gonadotropin and progesterone secretion patterns. J Clin Endocrinol Metab 1989;69:813-20. 20. Kerin JF, Liu JH, Phillipou G, Yen SSC. Evidence for a hypothalamic site of action of clomiphene citrate in women. J Clin Endocrinol Metab 1985;61 :265-8. A complete list of references is available from the authors on request.

The effect of the inhibition of prostaglandin synthesis on renal blood flow in fetal sheep Susan A. Arnold-Aldea, MD, Ron A. Auslender, MD, and Julian T. Parer, MD, PhD San Francisco, California Treatment with prostaglandin synthesis inhibitors has been associated with oligohydramnios in the fetus. The presumed mechanism is a reduction in fetal renal blood flow. We examined the effect of meclofenamate administration on renal blood flow in chronically catheterized fetal sheep during normoxia and during moderate and severe hypoxia. Ten fetal sheep were made hypoxic twice at least 4 days after surgery: once in the presence and once in the absence of meclofenamate infusion. Renal blood flow and combined ventricular output were measured with radioactive microspheres. Prostaglandin synthesis blockade with meclofenamate caused no significant change in blood pressure, combined ventricular output, renal blood flow, or renal vascular resistance in either the normoxic or hypoxic animals. These data challenge the contention that prostaglandin activity protects the renal vascular bed of the fetus from vasoconstriction during hypoxia and they also do not support the hypothesis that prostaglandin synthesis inhibition causes oligohydramnios through reduction of fetal renal blood flow. (AM J OBSTET GVNECOL 1991 ;165:185-90.)

Key words: Fetal renal blood flow, prostaglandin synthetase inhibitors, meclofenamate, fetal hypoxia, fetal sheep

From the Cardiovascular Research Institute and the Department of Obstetrics, Gynecology and Reproductive Sciences, University of California, San Francisco. Supported by National Institutes of Health Grant No. HD 18016. Presented at the Thirty-sixth Annual Meeting of the Society for Gynecologic Investigation, San Diego, California, March 15-19, 1989. Received for publication June 17, 1990; revised January 16, 1991; accepted February 1, 1991. Reprint requests: Julian T. Parer, MD, HSE 1462, Department of Obstretrics, Gynecology and Reproductive Sciences, University ofCalifornia, San Francisco, CA 94143-0132. 611128488

Prostaglandin synthetase inhibitors have emerged as useful agents in the therapy of preterm labor. '·3 In spite of their tocolytic efficacy and ease of administration, widespread clinical use has been limited by fetal side effects. Oligohydramnios has developed in as many as 50% of patients treated with prostaglandin synthetase inhibitors. 4 . 5 Prostaglandin synthetase inhibitors are known to decrease fetal urine output" and thereby lead to oligo-

185