Assessment of follicular function in women by measurement of urinary estrogen conjugates*

Transcription

1 FERTILITY AND STERILITY Copyright The American Fertility Society Vol. 46, No.4, October 1986 Printed in U.s A. Assessment of follicular function in women by measurement of urinary estrogen conjugates* Nancy M. Czekala, B.S. t James W. Overstreet, M.D., Ph.D.t Frederick W. Hanson, M.D.t George H. Stabenfeldt, D.V.M., Ph.D.11 Bill L. Lasley, Ph.D.1I Zoological Society of San Diego, San Diego and University of California, Davis, Davis, California Daily early morning urine samples from 23 women during their conception cycles were evaluated for urinary estrone (E I ) conjugates and pregnanediol-3-glucuronide (PdG) levels and indexed by creatinine (Cr) concentrations. These 23 cycles were classified into two groups according to the rate of E I conjugates increase following implantation. Urine from 16 nonconception cycles was available for 11 of the women. The two groups differed significantly both before and after, but not during, the midcycle EI conjugates peak. The length of the follicular phase was significantly shorter in the group of women with the lower E I conjugate levels. These data suggest that differences in estrogen excretion at times other than the preovulatory estrogen surge represent the endocrine activity of subdominant ovarian follicles. Further, such urinary estrogen profiles may represent useful parameters for defining ovarian follicular populations and/or indicating the effects of hazards to female reproduction. Fertil Steril 46:604, 1986 Surveillance of reproductive hazards in the workplace has become an important goal in the areas of occupational and environmental medicine. l, 2 There has been significant progress recently in the development of laboratory methods for human male reproductive risk assessment. 3 Application of these methods in the workplace Received February 4, 1986; revised and accepted June 20, *Supported by National Institutes of Health grants HD (JWO) and RRlHD (BLL). tresearch Department, Zoological Society of San Diego. tdepartment of Obstetrics and Gynecology, School ofmedicine, University of California, Davis. Reprint requests: James W. Overstreet, M.D., Ph.D., Division of Reproductive Biology and Medicine, University of California, Davis, California IIDepartment of Reproduction, School of Veterinary Medicine, University of California, Davis, California. 604 Czekala et a1. Assessment of follicular function in women has been demonstrated in longitudinal surveillance programs 4 and in cross-sectional field studies carried out by the National Institute for Occupational Safety and Health. 5 In contrast, there are few methods available to assess potential hazards to the gonads of women other than those hazards that are assessed indirectly in surveillance of pregnancy outcomes. Because potential ovotoxic agents, including industrial pollutants and cigarette smoke, are found widely in the environment, I it is important to identify potential strategies for noninvasive longitudinal surveillance of follicular depletion in women. In a previous study, we measured urinary steroid conjugates in sequential urine samples from women attempting to become pregnant by artificial insemination. 6 Prospective data were obtained on daily urinary concentrations of estrone (E I ) conjugates, pregnanediol-3-glucuronide Fertility and Sterility

2 (PdG) and luteinizing hormone-human chorionic gonadotropin (LH-hCG) in 23 conception cycles as well as in a number of cycles that ended in spontaneous abortions or ectopic pregnancies.6 The excretion of El conjugates was significantly greater in conception cycles than in nonconception cycles by the eighth day after ovulation, suggesting that changes in ovarian steroidogenesis could serve as sensitive indicators of chorionic gonadotropin secretion by the implanting embryo. In some abnormal pregnancies, alterations in urinary steroids suggested the occurrence of conception several days before a rise in urinary chorionic gonadotropin was noted, indicating the possible utility of these assays in surveillance of early pregnancy wastage.6 Although midcycle urinary estrogen levels were similar during ovulatory cycles among the women studied, the levels of estrogen appeared to differ between subjects when the early follicular and luteal phases were examined.6 We hypothesize that the less variable midcycle estrogen values may reflect the activity of the dominant follicle, and estrogen excretion in the early follicular phase may be indicative of steroidogenesis by a larger follicular reserve. Observations of low baseline urinary estrogens in perimenopausal women 7 and in prepubertal girlsb lend further support to the idea that longitudinal measurements of urinary estrogens could be used to monitor changes in follicular function of women over time. In this article we examine further the variability in steroid excretion patterns between women and compare these patterns in conception cycles and in repeated non conception cycles. We also present data that suggest that low baseline estrogen levels may influence the time of follicular maturation. MATERIALS AND METHODS Early morning urine samples were collected daily from women who were artificially inseminated with donor semen because of infertility in their male partner. These women had no apparent abnormalities of reproductive functon and received no medications in the insemination cycles. The urine was frozen immediately and stored until assayed. In a previously reported study,6 23 women conceived and delivered live, full-term infants. Nonconception, but ovulatory menstrual cycles were available for 16 cycles of 11 of these women. In this study, three additional nonconception cycles were assayed and the data were reanalyzed to compare respective conception and nonconception cycles. Urinary El conjugates and PdG, indexed by daily creatinine (Cr) levels, were evaluated for each woman.6 Urinary El conjugates were evaluated as previously described.6, 9 Ten microliters of unprocessed urine were used in the El conjugate assay and combined with antisera (1:5000) and tritiated estrone-3-sulfate (7000 cpm, specific activity 55 Ci/mmol). After 1 hour of incubation at 15 C, 0.3 ml of charcoal dextran was used to separate bound and free steroid. Tubes were centrifuged following 30 minutes of incubation at 15 C, and the supernatant was decanted and counted in Ready-Solv HP (Beckman, Fullerton, CA). PdG was evaluated as previously described.6, 10 Unprocessed urine (0.1 J..Ll) was combined with antisera (1:180,000; Courtauld, London, UK) and tritiated pregnanediol-3-glucuronide (6000 cpm, specific activity 20 Ci/mmol). After overnight incubation at 15 C, bound and free steroids were separated with 0.5 ml of charcoal dextran. The tubes were centrifuged following 45 minutes of incubation at 15 C. The supernatant was decanted and counted in Ready-Solv HP. Standard curves for both the PdG and El sulfate assays ranged from 31 to 5000 pg. Interassay coefficients of variation were 22.2% and 16.4% at 85% and 44% binding for PdG, and 9.9% and 8.1 % at 32% and 16% binding for El conjugates. Conception cycles reported previously as one group6 were divided into two groups according to the day of the cycle in which the urinary El conjugates reached 200 ng/ml after ovulation. In the first group of conception cycles (n = 11), El conjugates reached a level of at least 200 ng/mg Cr by the 11th day after the preovulatory El conjugates peak; conception cycles with lower values formed the second group (n = 12). The nonconception cycles of the women in the high estrogen group and low estrogen group also were examined. Rates of hormone increase were calculated by linear regression from the nadir that followed the preovulatory El conjugates peak (day 0) to day + 22, unless otherwise noted. Standard statistical procedures were used to derive the composite mean profiles of each group; data are expressed as mean ± standard error of the mean (SEM). Daily urines were aligned to the highest preovulatory level of El conjugates (day 0). The length of the follicular phase of the cycle (first day of menses to Vol. 46, No.4, October 1986 Czekala et al. Assessment of follicular function in women 605

3 OJ 280 E ---- OJ c 240 (J) Q) ~ 200 ::J c 0 U 160 W ! Days From E1 Conjugate Peak Figure 1 Daily mean (± SEM) urinary El conjugate levels indexed by Cr at the time of ovulation and implantation in two groups of women: a high estrogen group (n = 11) 0--0 and a low estrogen group (n = 12) ---e. Day 0 is the day of peak El conjugates. The asterisks indicate significant elevation of El conjugates concentration in the high estrogen group, compared with the low estrogen group (P values: * ; ** ; *** ; ****< 0.001). The number of samples represented in each mean is shown at the top for high estrogen group and at the bottom for low estrogen group. the day of the El conjugates peak) was determined in the 28 cycles for which adequate data were available, including conception and nonconception cycles from the two groups of women. Included in this comparison were three additional nonconception cycles from other women for which daily El conjugate values were available. Statistical significance was calculated with group Student's t-test. RESULTS Urinary El conjugate profiles for the conception cycles of the two groups of women are shown in Figure 1. In the high estrogen group, El conju- gate concentrations reached 200 ng/mg Cr by day 8 ± 0.7 after El conjugates peak (range, 5 to 11 days). In the low estrogen group, 200 ng/mg Cr was reached at day 20.6 ± 1.6 (range, 14 to 30 days). The mean level of estrone conjugates differed significantly (P ~ 0.05 to ) between the groups before ovulation on days -10, - 9, and - 8 before the El conjugates peak and from day +1 onward. The rates of increase of El conjugates from the postovulatory El conjugate nadir (days + 2 to + 5) to day + 22 were significantly different (P ~ 0.002) for the high estrogen group (11.7 ± 1.8 ng/mg Cr/day) and the low estrogen group (4.6 ± 0.7 ng/mg Cr/day). Mean PdG and Cr levels on the same samples from the two groups were not statistically different (P ~ 0.05) and the levels of LH-hCG did not differ between the groups (data not shown). A comparison of the height, weight, age, diet, and ethnic origin revealed no difference between the two groups of women. However, the height/weight ratio (38.04 ± 1.07 kg/m) of the high estrogen group was statistically different (P ~ 0.02) from the low estrogen group (33.75 ± 1.25 kg/m). Urinary El conjugate levels in nonconception cycles are illustrated in Figure 2. Mean El conjugate levels from the low estrogen group (n = 5) were significantly lower than those of the high estrogen group (n = 11) during the midfollicular phase, day - 7 through day - 4 (P ~ 0.05 to 0.003) and during luteal phase days + 2, + 3, and + 5 through + 12 in the non conception cycles (P ~ 0.05 to ). In contrast to conception cycles, a difference was found between the PdG levels in the nonconception cycles (Fig. 3; P ~ 0.05 to 0.002). The absolute values of El conjugate levels (without indexing by Cr) from the low estrogen group were lower and the levels from the high estrogen group were higher than with the respective values normalized by Cr concentration. Four mean daily Cr levels (days - 9, - 5, - 4, and + 7) were also significantly lower (P ~ 0.05 to 0.002) in the nonconception cycles in the low estrogen group versus the high estrogen group. The mean concentrations of urinary El conjugates in the follicular phase (days - 12 to - 1) of nonconception cycles differed by 62 ng/mg Cr between the two groups. Approximately the same differences were observed between the groups during the luteal phase (days + 1 to + 11) of the nonconception cycles. This difference was accentuated in the conception cycles following implan- 606 Czekala et al. Assessment of follicular function in women Fertility and Sterility

4 o 200 E c:: (/) Q.) -co ::J 'c 120 o uj (N ) : -:: ** ** :: :::::!* f (N _) 0~ ~2~3~3~4~4~4~5~5~5~4_4r-~ ~ Days From E1 Conjugate Peak Figure 2 Daily mean (± SEM) urinary El conjugate levels indexed by Cr during 16 menstrual cycles from 11 women who later conceived (and are represented in Figure 1). Each woman has been previously grouped according to her El conjugates profile of pregnancy and is similarly grouped in this figure: high estrogen group (n = 11) and low estrogen group (n = 5) --. Day 0 is the day of peak El conjugates. The asterisks indicate a statistically significant elevation of El conjugates concentration in the high estrogen group, compared with the low estrogen group (P values: * ; ** ; *** ; ****< 0.001). The number of samples represented in each mean is shown at the top for the high estrogen group and at the bottom for the low estrogen group. tation, when mean levels were 166 ng/mg Cr higher in the high estrogen group, compared with the low estrogen group, 13 to 26 days after the El conjugates peak. Profiles of consecutive nonconception cycles from individual women were consistent in terms of high estrogen excretion patterns or low estrogen excretion patterns (n = 7; two to three cycles each), as indicated by the mean and standard errors shown in Figure 2. All individuals in the high estrogen group had El conjugate levels > 160 ng/mg Cr in the early to midluteal phase, and all individuals in the low estrogen group had El conjugate levels < 140 ng/mg Cr during the entire luteal phase. In the conception cycles of the high estrogen group, mean El conjugates were significantly elevated (indicating pregnancy) at day + 8 (P ~ 0.05) in comparison with nonconception cycles and at days + 10, + 11, and + 12 (P ~ 0.02, 0.01, 0.01, respectively). Similarly, in conception cycles of the low estrogen group, El conjugates were significantly elevated over the nonconception levels on day + 8 and after day + 11 (P ~ 0.05). The length of the follicular phase was compared in 17 cycles from individuals in the high estrogen group and 11 cycles from individuals in the low estrogen group by combining both conception and non conception cycles. There was a significant difference between the groups in the length of the follicular phase, as measured from the first day of menstrual bleeding to the day of the El conjugates peak (high estrogen group = 17.1 ± 0.7 days; low estrogen group = o 10 E 8... ::t 6 CJ II II II II II II II 10 II II 10 2 O~~'-'P'PT~:r..::r-'~~~T--ir-r--'-' -10 o Days From E1 Conjugate Peak Figure 3 Daily mean (± SEM) urinary PdG levels indexed by Cr during 15 menstrual cycles from 11 women who later conceived (E 1 conjugates for the same cycles are shown in Figure 2). Each woman has been previously grouped according to her pregnancy profile of El conjugates and is similarly grouped in this figure: high estrogen group (n = 11) and low estrogen group (n = 5) --. Day 0 is the day of peak El conjugates. The asterisks indicate a statistically significant elevation of PdG concentration in the high estrogen group, compared with the low estrogen group (P values: * ; ** ; *** ; ****< 0.001). The number of samples represented in each mean is shown at the top for the high estrogen group and at the bottom for the low estrogen group. Vol. 46, No.4, October 1986 Czekala et al. Assessment of follicular function in women 607

5 14.2 ± 0.6 days; P :0;;; 0.02). The baseline El conjugates levels in these cycles were 51.1 ± 2.9 ng/mg Cr (n = 17) for the high estrogen group (average of two lowest values before day 10 for each subject ± SEM) versus 31.4 ± 3.3 ng/mg Cr (n = 11) for the low estrogen group (P = ). DISCUSSION These data demonstrate that urinary El conjugate levels vary significantly between women during the follicular and luteal phases of the menstrual cycle. This difference was continued and amplified under the influence of chorionic gonadotropin from the implanting embryo. Although the group definitions in this study were arbitrary, influences other than estrogen production seem unlikely, because both the absolute concentrations of El conjugates (not indexed by Cr) and their rate of increase after implantation were consistently and statistically different between the groups. Further, if the observed difference in El conjugate excretion were a result of any factor other than production, then the two groups would exhibit a consistent difference throughout the complete ovarian cycle. In fact, differences in urinary El conjugate levels were greatest during the early follicular phase (days -10 to - 5) before ovulation. Taken together, these results indicate that observed differences in urinary El levels reflect differences in ovarian estrogen production in the two groups of women. The lack of difference in urinary El. conjugate levels betweeen the two groups when the dominant follicle was the major influence on estrogen production suggests that the basis for the different estrogen profiles is related to another ovarian component. Because the difference in the El conjugate profiles was seen in both the follicular and the luteal phase, it seems unlikely that either the corpus luteum or the midcycle gonadotropin surge was a major contributing factor. The relative large difference (51.1 versus 31.4 ng/mg Cr) observed early in the menstrual cycle when few if any well-developed, tertiary follicles were present suggests that differences in estrogen excretion may be attributable to immature follicles that are destined for atresia. This is consistent with the observation that there were no significant differences between the groups in urinary estrogen excretion during the 4 days before and 1 day after the preovulatory El conjugates peak, when the dominant follicle appeared to contribute a similar amount of estrogen in both groups of women. There are few data available on circulating levels of estrogens during conception cycles with which to compare our observations of urinary estrone concentrations. Mishell et al. ll monitored daily serum estradiol during the conception cycles of three women from before ovulation to day 50 after LH peak. One woman appeared to have a relatively high ~erum estrogen response following conception, whereas the other two were distinctly lower; these patterns are similar to the urinary patterns in the present study. At the least, the data of Mishell andcolleagues ll indicate that there is a range of variation in estrogen production and/or clearance rates in conception cycles. Because differences in clearance rates of estrogen have not been observed, we interpret the present data to reflect differences in estrogen production. If the levels of El conjugates in urine can be shown to reflect differences in ovarian follicular activity, then it is possible that such measurement could be used in assessing the impact of pharmacologic agents or environmental toxicants on female reproductive potential. Urinary estrogen levels may prove to be useful in monitoring changes in ovarian follicular populations over time and/or to indicate the relative number of endocrinologic ally active follicles remaining in the ovary. Support for this concept is provided by our observations of El conjugate levels in perimenopausal women7 andperipubertal girls. 8 In both of these populations, the number of developing follicles is known to be relatively small, and in both groups the baseline urinary El conjugate levels are below the lowest levels demonstrated in the present study. In concordance with the present observations, urinary El conjugate levels in perimenopausal women and in peripubertal girls reach or exceed the normal range only when preovulatory follicles are contributing to estrogen production.7, 8 The present data relating to follicular phase length and E1conjugate levels are also in agreement with observations we have made of endocrine profiles in sequential ovarian cycles of peri menopausal women. In these women, low baseline estrogen levels are associated with gonadotropin escape, ovarian hyperstimulation, and dysfunctional uterine bleeding.7 In contrast to the perimenopausal women who exhibit reduced fertility, the women in the present study 608 Czekala et al. Assessment of follicular function in women Fertility and Sterility

6 conceived and maintained pregnancies, demonstrating the integrity of the homeostatic mechanisms of the hypothalamic-pituitary-gonadal axes. Nevertheless, it is possible that the shortened follicular phases demonstrated in this study in association with lower urinary estrogen levels could be a result of elevated gonadotropin levels during the early follicular growth phase. Differences were observed in mean PdG levels between the groups during nonconception menstrual cycles, and the higher levels were found in the group of women with higher E1 conjugates levels. This may indicate a relationship between follicular-phase estrogen production and subsequent function of the corpus luteum, although no other difference (luteal phase length or fertility) was noted between the two groups of women. The possibility remains, however, that diminished luteal activity subsequent to lower follicular phase estrogen production 12 could contribute to reduced fertility, which was not revealed in the present study because women were studied only near the time of a conception. The lack of difference in PdG levels in conception cycles is difficult to reconcile with this concept, because follicular-phase estrogen conjugate levels were consistently low in the women assigned to the low estrogen group. It may be that progesterone production and luteal function are related to follicular phase estrogen production but not on an absolute basis. Lower estrogen production may lead to reduced luteal function in some, but not all, cycles. Further study involving repetitive menstrual cycles in both groups of women will be necessary to provide an answer to this question. These data demonstrate different excretion profiles of urinary estrogen in subgroups of women, and this identifiable variation should be taken into account when urinary steroid assays are used clinically. For example, it is doubtful whether the absolute level of El conjugates, particularly as a single analysis, can be used to detect early pregnancy.6 Further, it appears that women with lower estrogen levels are approaching a point of nonequilibrium in terms of gonadotropin entrainment. These women appear to have less margin for error in estrogen negative feedback in the early follicular phase of the menstrual cycle; as a consequence, many of their cycles may be infertile. The lower levels of PdG in the nonconception cycles of the low estrogen group are consistent with a lower level of corpus luteum function, although these values are still within the clinically normal range. This observation is consistent with the observed reduction of fertile cycles in premenopausal women older than 35 years of age and may explain, at least in part, the mechanism involved. Acknowledgment. We thank Dr. Delwood C. Collins for the gift of antiserum to estrone conjugates. REFERENCES 1. Mattison DR, ed: Reproductive toxicology. Am J Indus Med 4:1, Lockey JE, Lemasters GK, Keye WR Jr, eds: Reproduction: The New Frontier in Occupational and Environmental Health Research. New York, Alan Liss, Overstreet JW: Laboratory methods for human male reproductive risk assessment. Teratogenesis Carcinog Mutagen 4:67, Perkins C, Schenker M, Sanuels S, Overstreet JW: Prospective monitoring of semen quality in chemical workers. Am J Epidemiol 120:492, Schrader SM, Turner JW, Clapp DE, Ratcliff JM: Sperm velocity as a measure of the toxic effects of acute EDB exposure. J Androl (Suppl) 6:38P, Lasley BL, Stabenfeldt GH, Overstreet JW, Hanson FW, Czekala N, Munro C: Urinary hormone levels at the time of ovulation and implantation. Fertil Steril 43:861, DeVane GW, Shideler SE, Kalra PS, Lasley BL: Ovarianpituitary interaction during the perimenopause. Clin Endocrinol Submitted 8. DeVane GW, Soto D, Lasley BL: Daily urinary estrogen and progesterone metabolites at the time of menarche. (Abstract) Presented at the Annual Meeting of the Society of Gynecologic Investigation, Phoenix, Arizona, March 20 to 23, Shideler SE, Czekala NM, Kasman LH, Lindburg DG, Lasley BL: Monitoring ovulation and implantation in the lion-tailed macaque (Macaca silenushhrough urinary estrone conjugate evaluations. Bioi Reprod 29:905, Mitchell WR, Presley S, Czekala NM, Lasley BL: Urinary immunoreactive estrogen and pregnanediol-3-glucuronide during the normal menstrual cycle of the female lowland gorilla (Gorilla gorilla). Am J Primatol 2:167, Mishell DR, Thorneycroft IH, Nagata Y, Murata T, Nakamura RM: Serum gonadotropin and steroid patterns in early human gestation. Am J Obstet Gynecol 117:631, Sherman BM, Korenman SG: Measurement of plasma LH, FSH, estradiol and progesterone in disorders of the human menstrual cycle, the short luteal phase. J Clin Endocrinol Metab 38:89, 1974 Vol. 46, No.4, October 1986 Czekala et.a1. Assessment of follicular function in women 609