* Department of Biochemical Pharmacology, School of Pharmacy.

Transcription

1 FERTILITY AND STERILITY Copyright ~ 1993 The American Fertility Society Vol. 59, No.1, January 1993 Printed on acid-free paper in U.S.A. Human oviductal fluid proteins. VI. Correlation between alpha-fetoprotein and serum levels of ovarian hormones Jack Lippes, M.D.*t Premanand V. Wagh, Ph.D.:!: State University of New York at Buffalo, Buffalo, New York Objective: To determine the relationship, if any, between the ovarian hormonal status of women and levels of alpha-fetoprotein (AFP) in their oviductal fluid. The threshold questions were the following: (1) does the presence of AFP in human oviductal fluid (hof) represent a generalized phenomenon and (2) if so, is there any relationship between AFP in hof and the ovarian hormones, estradiol (E2) and progesterone (P)? Design: Eleven women who elected minilaparotomy tubal ligations volunteered to donate hof and serum samples for this study. AFP, E2, and P were determined be radioimmunoassay. Setting: Procedures were performed in an academic research environment. Patients: This study encompasses only those patients who elected minilaparotomy tubal ligation and volunteered to donate hof and sera. Clinical features of these patients are described. Interventions: Patients undergoing tubal ligations had Foley catheters inserted into the fimbriated ends of each oviduct, and hof was collected for 24 hours. Concomitantly, blood was drawn for analyses of serum. The Investigational Review Board of the Erie County Medical Center approved this study requiring the collection of oviductal fluid from human subjects. Results: Levels of AFP in hof did not correlate with the serum concentrations of E2 and P per se. However, there was a highly significant correlation between levels of AFP in hof and serum E2:P ratio. The concentration of AFP in hof progressively increased with respect to the ratio of serum E2:P. Although AFP was present in all samples of hof, it was undetectable in corresponding sera. Conclusions: Based on these data, it is concluded that AFP is generally present in hof. Furthermore, it is suggested that AFP in hof may be under the control of ovarian steroids. Fertil Steril1993;59: Key Words: Alpha-fetoprotein, human oviductin-i, human oviductal fluid Since the development in 1972 of a surgical technique for the collection of human oviductal fluid (hof) coincident to an elective sterilization, Received December 23, 1991; revised and accepted September 14,1992. * Reprint requests: Jack Lippes, M.D., Department of Gynecology and Obstetrics, Erie County Medical Center, 462 Grider Street, Buffalo, New York t Department of Gynecology and Obstetrics. * Department of Biochemical Pharmacology, School of Pharmacy. it became feasible to analyze the composition of hof (1). Subsequently, it was demonstrated that hof contains unique proteins (2, 3). Moreover, one of these was isolated in a highly purified state and was designated human oviductin-i (hov-i) (4), which was found to bind selectively to the heads of human spermatozoa (5). In a companion paper in this issue, we presented evidence that hov-i and human alpha-fetoprotein (hafp) are identical molecules (6). This identity was established by comparing amino acid and carbohydrate composition as well as by immunological reaction Vol. 59, No.1, January 1993 Lippes and Wagh AFP and ovarian hormones 157

2 Table 1 Clinical Features of the 11 Patients Donating Oviductal Fluid Patient Age P/G/A* Medical history /3/1 No serious illness /3/1 No serious illness /8/3 Congenital heart /1/0 No serious illness /2/0 Hepatitis B, 3 y PTA t /3/1 No serious illness /1/0 No serious illness /2/0 Cholecystitis 6 y PTA t /3/0 No serious illness /2/1 No serious illness /4/0 No serious illness * Parity/gravidity/abortions. t Before admission. Surgical history Fractured finger, 1 y Herniorraphy 12 y PTA t Tonsillectomy 25 y PTA t Cholecystectomy 6 y PTA t C-section 2 y PTA T and A 24 y PTAt C-section, 1 y PTA t :I: Intrauterine device. Contraceptive history Pills discontinued 6 mo PTA t Pills discontinued 3 mo PTA t Condoms IUD:I: Diaphragm + jelly Periodic abstinence Pills for 10 y; IUD past 3 y Pills discontinued 4 mo PTA t Pills discontinued 4 mo PTA t Condoms None used employing a purified rabbit polyclonal anti-hov-i immunoglobulin (lg)g and murine monoclonal anti -hafp -IgG. Various aspects of AFP, including ontological, biochemical, physiological, and clinical, have been extensively reviewed (7-10). Under known circumstances, the organs that predominantly produce AFP are the fetal liver, yolk sac, and to a lesser extent, the fetal gastrointestinal tract. Before the evidence provided in the previous paper (6), there has been no report indicating the presence of AFP in any oviductal fluid. The initial hov-i preparation was derived from pooled samples of hof from four patients in a third world country, where medical histories, endometrial biopsies, and steroid evaluations were not available (4). On the other hand, AFP was identified in a sampie of hof obtained from one patient in a 24-hour period (6). In that study the patient was operated on day 13 of the menstrual cycle as noted from the patient's history. The concentration of AFP in this patient's hof was 34.4lLg/mL as measured by noncompetitive enzyme-linked immunosorbent assay (ELISA). Alpha-fetoprotein was undetectable in serum. Estradiol (E 2) and progesterone (P) in this patient's serum were not determined. The undetectable level of AFP in serum and unusually high concentration of AFP in hof were perplexing observations. The lack of identification of AFP in serum was expected in a nonpregnant, healthy patient. However, mild pathological changes were observed in the genital tract as follows: [1] a few dysplastic cells were seen on a Pap smear taken at admission and [2] endocervical metaplasia and dysplasia were evident in a colposcopically directed bi- opsy of the cervix, which was obtained 1 week after tubal ligation. The finding that AFP was present in a single specimen of hof prompted us to investigate whether this protein is generally found in oviductal fluid of normal patients. If so, it was possible that the concentrations of AFP in hof might be related to the woman's ovarian hormonal status. To answer these questions, 11 oviductal fluid samples from nonpregnant patients devoid of cancer and liver disease were analyzed. This allowed us to investigate the relationship between AFP levels in hofs and the concentrations ofe 2 and P in corresponding patients' sera. MATERIALS AND METHODS Collection of hof and Serum Samples The technique for the collection of hof has been described (1). Samples of hof were collected from 11 multiparous subjects ranging in age from 24 to 38 years with a mean of 29.9 ± 4.7 years, an average parity of 2.3, and an average gravidity of 2.9. These patients did not have any complicating illness at the time of admission. All patients, except one, had regular menstrual cycles. A summary of the patients' medical, surgical, obstetric, and contraceptive history is presented in Table 1. The day of the cycle during which hof was collected was recorded (Table 2). There was no factor that might alter the protein constituents of the fallopian tube with the possible exception of patient 5 who had a history of hepatitis. However, this woman had a normal liver profile before surgery. Patients using oral contraceptives were instructed to discontinue their pills at least 3 months 158 Lippes and Wagh AFP and ovarian hormones Fertility and Sterility

3 Table 2 Correlation of AFP in hof Versus E2:P and Versus Day of Cycle Patient Cycle day hofin AFP Serum AFP E2 P E2:P Mean± SD 26t ± 4.8 ng/ml loa ± 4.8 ng/ml pg/ml ng/ml pg/ng UD* UD UD UD UD UD UD UD UD UD UD ± ± ± 46 Statistical comparison Correlation coefficient Probability AFP (hof) versus E. (serum) AFP (hof) versus P (serum) AFP (hof) versus E.:P (serum) AFP (hof) versus cycle day E2:P versus cycle day * Undetectable «5 ng/ml). t Irregular menses. + NS, not significant NS NS < < <0.05 Data from patient 2 are excluded because of irregular menses (n = 10). before surgery. Subjects underwent a minilaparotomy tubal ligation according to the previously described procedure (1). No patient had any serious postoperative complications, and all were discharged on the day after surgery. Two patients each produced >50 ml of fluid from two oviducts in a 24-hour period. These represent the largest quantities of hof from single patients that we have recorded in 20 years of collecting hof. Samples of hof were centrifuged at 18,000 X g for 15 minutes at 4 C, and the supernatants were stored at -70 C. Sera were obtained from the same patients on the day of surgery and were similarly stored until used. Determination of AFP, E 2, and P Each sample of hof and serum was thawed at 25 C and analyzed for AFP using a commercial radioimmunoassay (RIA) kit (Amersham, Chicago, IL) according to the manufacturer's specifications. The lowest detection limit was 5 ng AFP /ml by this procedure. Hormonal status of each patient was evaluated by measuring Ez and P concentrations in the subjects' sera. These two parameters were quantitated by RIA using commercially available kits from Pan- tex (Santa Monica, CA) and Diagnostic Products (Los Angeles, CA), respectively. Statistics Regression analyses were carried out by the method of least squares to determine the correlation between AFP in hof and levels of Ez, P, and Ez:P ratios in serum samples. Correlation coefficients were also determined between the day of the menstrual cycle and hafp values in hof and Ez:P ratios. There was considerable variability in the levels of Ez, P, and Ez:P ratios in serum samples. However, based on median values, each of these parameters was separated into two groups for statistical comparison of AFP concentration in hof. The groups with values larger than the median were included in the high level category. All of the remaining values were grouped into the low level category. The mean values of AFP in hof in the two groups were compared by Student's t-test. Probability < 0.05 was considered significant. RESULTS The concentrations of AFP in hof and the levels of AFP, Ez, P, and Ez:P ratios in sera of 11 patients Vol. 59, No.1, January 1993 Lippes and Wagh AFP and ovarian hormones 159

4 u. o.c E C) C D.. u. c:c O~----=25~--~5~O-----7=5----~10LO-----1~25----~15~O Serum E2 / P ( pg / ng ) Figure 1 Correlation between concentration of AFP in oviductal fluids of 11 patients and E 2 :P ratio in corresponding sera. Each data point is indicated by patient's number (see Table 2). The straight line represents regression analysis plot of AFP versus E 2:P. are given in Table 2. Values of the mean with SD for each parameter are presented including correlation coefficients for the following: [1] AFP in hof versus serum E2, P, and E2:P ratio and [2] cycle day versus E2:P and hafp. Because patient 2 had irregular menses (Table 2), the data on this patient were excluded in the calculation of correlation coefficients for cycle day versus E2:P and hafp in hof. Both E2:P ratio and hafp in hof declined as the days of the menstrual cycle increased. Higher values for hafp in hof were found in menstrual days before ovulation. Elevated levels of hafp in hof were found associated with lower serum levels of P as seen in patients 2, 3, and 4 (Table 2). AFP was undetectable in sera of all patients by RIA, which had a detection limit of 5 ng AFP /ml. On the other hand, oviductal fluid of each patient contained AFP that ranged from as low as 5.8 ng/ml (patient 7) to as high as 20.0 ng/ml (patient 4), with an average of 10.2 ± 4.8 ng/ml for all patients. Accordingly, the data were evaluated after regression analysis. There was no correlation between P or E2 in serum andafp in hof. However, the ratio ofe2:p showed a highly significant correlation with AFP in hof (r = 0.842; P < 0.001). The linear regression plot of AFP in hof versus E2:P in serum is shown in Figure 1. A general survey of the diagram revealed the following features: [1] 7 of the 11 patients' sera having E2:P ratios between 14 and 36 had 5.8 to 7.9 ng AFP /ml in their oviductal fluids, and these E2:P ratios were consistent with the postovulatory phase ofthe menstrual cycle; [2] the remaining 4 patients with E2:P ratios ranging from 42 to 150 had 10.4 to 20.0 ng AFP /ml in their hof. These patients were in their preovulatory phase; and [3], AFP levels in hof increased progressively in a linear manner with increasing E2:P ratios in sera. Statistical comparisons between low and high levels of E2, P, and E2:P ratios (see Materials and Methods) in serum samples and AFP concentrations in hof are presented in Table 3. The differences between the mean values of low and high level groups of each parameter were significant. The means of the high levels of E2, P, and E2:P were 2.5-, 3.7-, and 4.0-fold greater then the corresponding mean of the low level of E2, P, and E2:P, respectively. In spite of the small sample size, the division of the values for E2, P, and E2:P ratios into two groups is justified in view of the statistically significant difference between the high and low averages. There was no significant difference in AFP concentrations in hof with respect to the serum E2 levels. In fact, Table 3 Statistical Comparisons Between Low and High Levels of E 2, P, and E 2:P Ratio in Serum and AFP Concentration in Oviductal Fluid' Serum E2 Serum P E,:P Parameter Low High Probability Low High Probability Low High Probability pg/ml 17 to to to 4.0 (n ~ 6) (n ~ 5) (n ~ 6) E, (pg/ml) 72 ± 49t 179 ± 36 t ~ 4.04 (P < 0.005) P (ng/ml) 1.88 ± 1.42 E,:P (pg/ng) AFP (ng/ml) ± ± 4.39 t ~ ± 5.89 (NS)t ng/ml 4.1 to 10.1 (n ~ 50) 6.88 ± 2.43 t ~ 4.27 (P < 0.005) 7.82 ± 1.51 t ~ 1.58 (NS) 14 to 33 (n ~ 6) ± ± to 150 (n ~ 5) ± t ~ 2.90 (P < 0.025) ± 4.96 t ~ 3.43 (P < 0.01) * For comparison of means by Student's t-test, the data in Table 2 were collated. See Materials and Methods for definition of low and high levels of E 2 P, and E 2:P ratio in serum. t Values are means ± SD. :I: NS, not significant. 0 Lippes and Wagh AFP and ovarian hormones Fertility and Sterility

5 AFP concentrations were similar at both high and low levels of serum E2. AFP associated with the low level of P was 1.5 times greater than AFP found with high P levels. This increase was not statistically significant. However, there was a significant difference (P < 0.01) in the concentration of AFP with regard to the E2:P ratio: at high E2:P ratios, AFP concentrations were twice that found at low ratios. These observations suggest that the concentration of AFP is elevated at low levels of serum P. AFP levels in hof ~ 14 ng/ml were observed in hof samples collected during cycle days < 13. DISCUSSION Data presented herein conclusively demonstrate that hof generally contains AFP. Radioimmunoassay was used to quantitate AFP concentrations in hof and sera of 11 patients. The detection limit for AFP by RIA was 5 ng/ml, a 60-fold higher sensitivity than the ELISA procedure used in our previous study (6). The undetectable levels of AFP in sera of healthy, nonpregnant women as found here is a well-established fact. However, the demonstration that AFP is generally found in hof constitutes a new finding. Concentrations of AFP in hof are related to the before and after ovulatory phases of the menstrual cycle (Fig. 1 and Table 3). It seems that a slight increase in P may trigger the release or the secretion of hafp into the luminal contents of the oviduct. We postulate that oviductal AFP is controlled by ovarian steroids. This hypothesis is in any case based on a limited sample size. At this time, the relationship between AFP and ovarian hormones cannot be precisely defined. Studies in animals, wherein one can control the presence or absence of specific ovarian hormones, may help to clarify questions about the possible endocrine control of oviductal AFP. There are at least two functions ascribed to AFP in mammalian systems. First, AFP binds estrone and E2 and acts as a carrier for these compounds (11, 12). Nonestrogenic steroids are not bound to AFP (13, 14). This phenomenon of AFP binding to estrogen has been found in rats and mice but not in humans. Second, AFP has immunosuppressive characteristics demonstrated by in vitro and in vivo assays (15-20). The earliest postulate we know of concerning immunosuppression by AFP was made by Ogra et al. (15) in They found that mice, administered amniotic fluid, exhibited a depressed immune response to injections of sheep red blood cells. Subsequently, the work of Murgita and Tomasi detailed the immunosuppressive properties of AFP (,17), and several reports followed confirming this (18-20). It is possible that AFP, in light of these immunosuppressive properties, may not only prevent the production of antisperm antibodies within the oviduct but also protect the nidating embryo from immunological rejection. A possible third function for hafp was suggested in a previous paper (5) that this protein by selectively binding to the heads of human spermatozoa may act as an acrosome-stabilizing factor preventing a premature acrosome reaction. Whether the binding of AFP to the sperm head as we observed by immunofluorescent technique is because of native AFP or its isoforms arising from microheterogeneity of AFP is also a question worthy of further research. The presence of AFP in hof raises important questions about the reproductive physiology of the oviduct, some of which have been addressed here. Further investigations about the function of AFP in reproduction are warranted. Acknowledgment. We thank Patrick J. Carmody, Ph.D., for his assistance with the RIAs reported in this study and Maria Zielesny, Ph.D., for her help with the statistical analyses. REFERENCES 1. Lippes J, Enders RG, Pragay DA, Bartholomew WR. The collection and analysis of human fallopian tube fluid. Contraception 1972;5: Lippes J, Krasner J, Alfonso LA, Dacalos ED, Lucero R. Human oviductal fluid proteins. Fertil Steril 1981;36: Lippes J, van Oss CJ, Bronson PM, Alfonso LA, Dacalos EA, Lucero R. Human oviductal fluid proteins. II. Preparation of an antiserum to a human oviductal fluid protein: existence of autoantibodies against it in some sera. Fertil Steril1983;39: Wagh PV, Lippes J. Human oviductal fluid proteins. III. Identification and partial purification. Fertil Steril 1989;51: Lippes J, Wagh PV. Human oviductal fluid (hof) proteins. IV. Evidence for hof proteins binding to human sperm. Fertil Steril 1989;51: Wagh PV, Lippes J. Human oviductal fluid proteins: V. Identification of human oviductin I as alpha-fetoprotein. Fertil Steril 1993;59: Ruoslahti E, Hirai H. Alpha-fetoprotein. Scand J Immunol 1978;8 (8 Suppl): Adinolfi M. Human alpha-fetoprotein In: Harris H, Hirschhorn K, editors. Advances in human genetics. New York: Plenum Press, 1979;9: Crandall BF. Alpha-fetoprotein: a review. In Batsakis J, Savory J, editors. CRC critical reviews in clinical laboratory sciences. Boca Raton, FL: CRC Press, 1981;15: Vol. 59, No.1, January 1993 Lippes and Wagh AFP and ovarian hormones 1

6 10. Bergstrand CG. Alpha-fetoprotein in paediatrics. Acta Paediatr Scand 1986;75: Aussel C, Masseyeff R. Binding of estrogens to the molecular variants of rat alpha-fetoprotein. FEBS Lett 1977;81: Aussel C, Uriel J, Mercier-Bodard C. Rat alpha-fetoprotein. Isolation, characterization and estrogen binding properties. Biochimie 1973;55: Nunez E, Vallette G, Benassayag C, Jayle MF. Comparative study on the binding of estrogens by human and rat serum proteins in development. Biochem Biophys Res Commun 1974;57: Raynaud JP, Mercier-Bodard C, Baulieu EE. Rat estradiol binding plasma protein (EBP). Steroids 1971;18: Ogra SS, Murgita RA, Tomasi TB. Immunosuppressive activity of mouse amniotic fluid. Immunol Commun 1974;3: Murgita RA, Tomasi TB. Suppression of the immune response by AFP. I. The effect of mouse AFP on the primary and the secondary antibody response. J Exp Med 1975;141: Murgita RA, Tomasi TB. Suppression of the immune response by AFP. II. The effect of mouse AFP on mixed lymphocyte reactivity and mitogen-induced lymphocyte transformation. J Exp Med 1975;141: Auer 10, Kress MG. Suppression of the primary cell-mediated immune response by human alpha fetoprotein in vitro. Cell Immunol 1977;30: Yachnin S, Lester E. Inhibition of human lymphocyte transformation by human alpha-fetoprotein (hafp): comparison of fetal and hepatoma hafp and kinetic studies of in vitro immunosuppression. Clin Exp Immunol 1976;26: Wajner M, Papiha SS, Wagstaff TI. Response of human peripheral blood lymphocytes in the presence of cord sera: relationship of lymphocyte transformation with number of pregnancies and levels of alpha-fetoprotein. Clin Exp Immunol1983;52: Lippes and Wagh AFP and ovarian hormones Fertility and Sterility