1 Division of Reproductive Biology, Department of Obstetrics and Gynecology, Institute of Toxicology and Environmental Health and 2 California Regional Primate Research Center, University of California at Davis, Davis, CA 95616, USA
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Key words: granulosa cell culture/infertility/in-vitro fertilization/pregnancy/relaxin
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
While factors which contribute to the success of IVFembryo transfer have been extensively studied (Andersen et al., 1992; Schalkoff et al., 1993
; Wilcox et al., 1993
), fewer studies have examined the relationship between granulosa lutein cell culture and the characteristics of the cycle from which cells were obtained. One group found that decreased granulosa cell 11ß-hydroxysteroid dehydrogenase activity was associated with higher conception rates (Michael et al., 1993
, 1995
). It was reasoned that exposure of the oocyte to cortisol was necessary for proper functional maturation and high amounts of enzyme in the cumulus cells could prevent this exposure. Another study was based upon observations that the magnitude of rise in progesterone concentrations in response to human chorionic gonadotrophin (HCG) stimulation were correlated with implantation success (Prien et al., 1994
). They found that patients with a >3-fold rise in serum progesterone in response to HCG were more likely to become pregnant (46% conception rate) than patients with a rise of <3-fold (14% conception rate). Granulosa lutein cell culture from these patients showed differences in hormone production which were related to the serum progesterone rise (Prien et al., 1995
). Patients with a large rise in serum progesterone had higher progesterone concentrations in culture. Patients with a low rise in progesterone had more variable oestrogen concentrations in culture but the oestrogen was more responsive to gonadotrophin stimulation.
We have developed a system for culture of human luteinizing granulosa cells which supports the timely and dynamic secretion of oestrogen, progesterone and relaxin (Stewart and VandeVoort, 1997). While other groups have measured relaxin in granulosa cell cultures (Gagliardi et al., 1992
; Mayerhofer et al., 1995
), we were the first to report a similar pattern of secretion of relaxin in culture to that seen in circulation during the luteal phase (Stewart et al., 1995
). It was found that the amount of relaxin produced by cells from different patients was highly variable, with relaxin production on day 10 of culture ranging from undetectable to >1500 pg/ml.
This study compares the magnitude of steroid and relaxin production during granulosa lutein cell culture with the outcome of the cycles from which cells were obtained. We report that the magnitude of relaxin secretion during granulosa lutein cell culture is significantly correlated with pregnancy success, while steroid production is not.
![]() |
Materials and methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Approximately 1.0 ml modified human tubal fluid medium (Irvine Scientific, Santa Ana, CA, USA) containing HEPES buffer (4.0 mM), and antibiotics (100 units/ml penicillin G and 50 µg/ml streptomycin sulphate), was added to the follicular fluid aspirate during the oocyte retrieval procedure. Oocytes and cumulus masses were removed and all granulosa cells from an individual were pooled. Cells from different subjects were not pooled. The cells were prepared by initial centrifugation followed by layering onto 40% Percoll (Sigma, St Louis, MO, USA). The granulosa cell layer was washed twice with 510 ml fresh minimal essential medium (MEM; Gibco, Grand Island, NY, USA) and centrifuged for 10 min at 300 g. The supernatant was discarded, and the pellet was resuspended in 24 ml MEM. Granulosa cells were filtered through an 89 µm polyester filter (Spectra/Mesh; Spectrum Medical, Laguna Hills, CA, USA) just before being counted and plated.
Culture conditions
The culture conditions were identical to those previously described (Stewart and VandeVoort, 1997). Briefly, extracellular matrix (Matrigel®; Becton-Dickinson, Franklin Lakes, NJ, USA) was applied to culture dishes according to the manufacturer's directions on the same day that the cells were collected. MEM was modified with the following additions: sodium bicarbonate, 4.4 mg/100 ml MEM (Sigma); fungizone, 1 ml/100 ml (Gibco); penicillin G, 6 mg/100 ml (Sigma); streptomycin sulphate, 6 mg/100 ml (Sigma) and 10% fetal calf serum (Hyclone; Logan, UT, USA). HCG (Pregnyl®; Organon, W. Orange, NJ, USA) was added to the culture media at 0.02 mIU/ml. The cells were plated at 5x104 cells/well and the medium was changed daily.
Assays of culture fluid
Oestradiol and progesterone were measured by commercial kits (Diagnostic Products Corp., Los Angeles, CA, USA) and relaxin was measured by an enzyme immunoassay as previously reported (Stewart and VandeVoort, 1997). Controls were run in each assay and inter-assay coefficients of variation were calculated (Rodbard, 1974
).
Patient and cell culture data
We report only cycles for which we had both granulosa cell culture relaxin data and conception success as reported to us by the clinic. We also requested further information from the clinic about the characteristics of these cycles. We received data on some cycles but, due to closing of the clinic, we did not receive data on all cycles. Thus, the n is reduced for cycle parameters. The data included patient age, serum follicle stimulating hormone (FSH) concentrations on day 3 of the cycle, serum oestradiol 2 days prior to follicular aspiration (serum oestrogen), number of oocytes collected, number of oocytes fertilized, number of embryos graded I, II or III, number of embryos transferred, whether or not conception occurred in the cycle, and pregnancy outcome if conception did occur. Embryo grade I indicated no fragmentation, evenly divided blastomeres and five or more cells by day 3. Grade II indicated some fragmentation, evenly divided blastomeres and three or four cells by day 3 and grade III showed substantial fragmentation and uneven blastomeres.
Evidence of a conception was determined by measurement of HCG in serum from samples collected on days 8, 10 and 20 following embryo transfer. Cycles without detectable HCG or falling HCG concentrations were defined as non-conceptive. Conceptive cycles were defined as those with rising serum HCG concentrations. Ultrasound analysis was done on days 28 and 42 following embryo transfer to determine the presence of a gestational sac. A biochemical pregnancy was defined as a cycle with rising HCG in serum (conceptive) but no intrauterine gestational sac in the uterus at any time and no gestation outside the uterus (ectopic pregnancy). A spontaneous abortion (SAB) was defined as a cycle with rising HCG and the presence of a gestational sac but failing to deliver a live birth. Term cycles were defined as cycles ending in a live birth, whereas non-term cycles were defined as either a non-conceptive cycle or a conceptive cycle ending prior to a live birth (biochemical or SAB).
The cell culture parameters collected were: the progesterone concentration on the first day of culture, progesterone and oestradiol concentrations on day 11 of culture and the relaxin concentration on day 10 of culture. These parameters were collected from cultures run using various experimental designs for different studies lasting 12 or 20 days. The granulosa cell culture system was originally developed for toxicology experiments and the correlation with conception success arose in retrospective examination of patients' cycles to help explain the large variation in relaxin production between cycles. All experiments used an identical protocol for control wells, which ranged from one to three, while additional wells were used for different treatments. Multiple control wells were averaged for this study, so the observational unit is the mean control value from the cells from one patient at the designated time point.
Data analysis
Patient and culture data were grouped by presence or absence of conception and analysed using commercial statistical software (Statistica 98 for Windows, StatSoft, Inc., Tulsa, OK, USA). The cycle and culture data were compared using a correlation matrix (P < 0.01) to look for significant interactions of these parameters. The data from patient's cycles and granulosa cell culture hormone production were grouped by either conception success or term pregnancy success and analysed by the MannWhitney U-test for significance. The parameters that showed significance in the MannWhitney U-test were further analysed by logistic regression, either individually or combined and the level of significance was determined for each. Contingency tables were constructed based upon the results of the logistic regression and the odds ratio determined. A small number of donor/recipient cycles (n = 7) were analysed for day 10 relaxin concentrations and a contingency table constructed. A 2-test was utilized to determine significance.
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Patient and culture data were examined from 69 IVFembryo transfer cycles. A correlation matrix of these parameters showed significant interactions of primarily the quality of oocytes (Table I). None of the culture parameters showed significant correlation with any of the other parameters.
|
|
|
|
|
Mean hormone concentrations in culture were assessed according to cycle outcome (Figure 3). Mean oestradiol and progesterone concentrations were not significantly different between non-conceptive and term cycles while relaxin concentrations were significantly higher in term cycles than in all other cycles.
|
|
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Both the number of grade I embryos and the day 10 relaxin concentrations were significantly related to conception success. It is possible that the quality of the follicular phase sets up both the quality of the embryo and the capacity of granulosa cells to produce relaxin. As both embryo grade and relaxin production in culture were correlated to conception success, it is possible that the same characteristics of the follicular phase which produce the best oocytes also develop granulosa cells which produce the most relaxin in culture. These may be independent factors in determining conception success because the two parameters were not significantly correlated (Table I) and the use of both of these parameters produced the best P value in the logistic regression analysis of conception success (Table III
), indicating that each parameter may be contributing to a successful implantation.
If relaxin concentrations of donor granulosa lutein cell cultures were correlated with the recipients' pregnancy success, it would argue that the quality of the embryo is the most significant factor in pregnancy success and that elevated granulosa relaxin production is just a by-product of better follicular development. Conversely, if the correlation of donor granulosa relaxin production and recipient pregnancy success is lost in donor/recipients (as preliminary data suggest), it would argue that some factor produced by the donor ovary is correlated with implantation success and that this relationship is not transduced through the embryo to the recipient. While the number of donor/recipient cycles examined is too small to draw valid conclusions as to whether embryo quality or some ovarian factor is the critical factor in implantation success, the cycles currently available support the second hypothesis.
Relaxin is found in serum during the mid and late luteal phase, beginning around day 8 after the luteinizing hormone surge (Stewart et al., 1990, 1995
). In conceptive cycles relaxin increases above non-conceptive concentrations in close accord with rising HCG (Stewart et al., 1990
, 1993
, 1995
). Serum relaxin is believed to be of luteal origin (Sherwood, 1994
). We have previously demonstrated wide variations in serum relaxin concentrations during non-conceptive luteal phases and that relaxin varied more than progesterone concentrations between patients (Stewart et al., 1992
). These variations in serum could be due to differences in granulosa lutein production of relaxin, similar to that observed in culture from IVF patients. We also showed that cycles with lowered relaxin concentrations were those with out-of-phase endometrial biopsies (Stewart et al., 1992
) indicating that relaxin may play a role in normal endometrial development. It is possible that patients whose granulosa cells do not produce relaxin in culture will have abnormally lowered circulating relaxin and that this in some way accounts for lowered implantation and pregnancy success. We are currently determining if patients with cells that produce less relaxin in culture have less relaxin in serum. While it is possible that in-vivo feedback controls would regulate relaxin production from granulosa lutein cells in the corpus luteum to compensate for low relaxin production, we believe that relaxin production during the luteal phase is not in a direct endocrine feedback loop (manuscript in preparation). Thus, lowered relaxin production by granulosa cells would not be compensated for in vivo and would lead to inadequate serum relaxin during the luteal phase.
There are several means by which lowered circulating relaxin concentrations might influence implantation success. Relaxin induces dilation of superficial endometrial blood vessels and proliferation of the endothelial cells in monkeys (Dallenbach-Hellweg et al., 1966). `The effects produced seem to be a direct response of the endothelium to relaxin, as they occur only when this hormone is administered' (Hisaw et al., 1967
). It has recently been demonstrated that recombinant human relaxin stimulates vascular endothelial growth factor in human endometrial stromal cells (Unemori et al., 1997
).
Another possible mechanism by which relaxin could be involved in aiding a successful implantation is through the stimulation of glycodelin release by the endometrium. We have shown that relaxin is a direct stimulus for glycodelin secretion (Stewart et al., 1997), a potent immunosuppressive which may be needed to prevent maternal rejection of the embryo (Dell et al., 1995
). Relaxin is highly effective in reducing the amplitude of spontaneous and induced uterine contractions in several species (Sherwood, 1994
), but the role of relaxin in uterine quiescence in the human is less clear. Synthetic human relaxin has little or only a slight effect on spontaneous contractility of human myometrial tissue in vitro (Petersen et al., 1991
; MacLennan et al., 1995
).
Steroid concentrations from cultured cells did not vary as much as relaxin concentrations and were not predictive of conception. While oestrogen and progesterone administration are sufficient to allow implantation in cases of ovarian failure (Emmi et al., 1991; Johnson et al., 1991
), that situation may not be comparable to cycles of IVFembryo transfer due to the hormone stimulation protocols used. The stimulation in IVF cycles may cause endocrine excesses such as over-production of oestrogen. It is known that elevated oestradiol at the time of ovulation induction relates to a lowered success of conception (Forman et al., 1988
) and this would not occur in hormone replacement therapies. It is possible that relaxin could have corrective actions to some of these endocrine disturbances. For example, excess oestrogen has been associated with reduced uterine blood flow (Greiss et al., 1986
), so relaxin may aid implantation by increasing uterine blood flow.
The role of endometrial relaxin must also be considered. Relaxin is produced in the endometrial glands during the luteal phase of the natural menstrual cycle (Boyers et al., 1992; Bryant-Greenwood et al., 1993
) and in the stroma of the late luteal phase and early pregnancy (Bryant-Greenwood et al., 1993
). It is possible that this is normally sufficient for relaxin's role(s) in early pregnancy and that luteal relaxin is may be unnecessary or redundant. This is supported by the establishment of successful pregnancies in cases of premature ovarian failure through embryo donation (Emmi et al., 1991
; Johnson et al., 1991
; and our own unpublished observations). In this situation, a corpus luteum is not formed and serum relaxin is undetectable. The presence or absence of endometrial relaxin in IVFembryo transfer cycles has not been reported but it is possible that with IVFembryo transfer treatment protocols, endometrial relaxin is suppressed and/or requires supplementation, making additional relaxin from the corpus luteum beneficial.
This study supports the hypothesis that relaxin is involved in the normal implantation process and that lowered relaxin concentrations may be a partial cause of poor pregnancy rates after IVF treatment. In patients with granulosa lutein cell cultures which produce low amounts of relaxin, one would expect to find lowered circulating relaxin and this low relaxin may prevent optimal implantation.
![]() |
Acknowledgments |
---|
![]() |
Notes |
---|
3 To whom correspondence should be addressed
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Boyers, S., Stewart, D., Douglas, G. et al. (1992) Immunocytochemical localization of relaxin and prorelaxin in endometrial biopsies from normal and out-of-phase cycles. Paper presented at the Society for Gynecologic Investigations, March 20, 1992 San Antonio, TX, USA.
Bryant-Greenwood, G., Rutanen, E., Partanen, S. et al. (1993) Sequential appearance of relaxin, prolactin and IGFBP-1 during growth and differentiation of the human endometrium. Mol. Cell. Endocrinol., 95, 2329.[ISI][Medline]
Dallenbach-Hellweg, G., Dawson, A.B. and Hisaw, F.L. (1966) The effect of relaxin on the endometrium of monkeys. Histological and histochemical studies. Am. J. Anat., 119, 6178.[ISI]
Dell, A., Morris, H.R., Easton, R.L. et al. (1995) Structural analysis of the oligosaccarides derived from glycodelin, a human glycoprotein with potent immunosuppressive and contraceptive activities. J. Biol. Chem., 270, 2411624126.
Emmi, A.M., Skurnick, J., Goldsmith, L.T. et al. (1991) Ovarian control of pituitary hormone secretion in early human pregnancy. J. Clin. Endocrinol. Metab., 72, 13591363.[Abstract]
Ezra, Y. and Schenker, J.G. (1995) Abortion rate in assisted reproductiontrue increase? Early Pregnancy, 1, 171175.[Medline]
Forman, R., Fries, N., Testart, J. et al. (1988) Evidence for an adverse effect of elevated serum estradiol concentrations on embryo implantation. Fertil. Steril., 49, 118122.[ISI][Medline]
Gagliardi, C.L., Goldsmith, L.T., Saketos, M. et al. (1992) Human chorionic gonadotropin stimulation of relaxin secretion by luteinized human granulosa cells. Fertil. Steril., 58, 31420.[ISI][Medline]
Greiss, F., Rose, J., Kute, T. et al. (1986) Temporal and receptor correlates of the estrogen response in sheep. Am. J. Obstet. Gynecol., 154, 831838.[ISI][Medline]
Hisaw, F.L., Hisaw, F.L. and Dawson, A.B. (1967) Effects of relaxin on the endothelium of endometrial blood vessels in monkeys (Macaca mulatta). Endocrinology, 81, 375385.[ISI][Medline]
Johnson, M.R., Abdalla, H., Allman, A.C.J. et al. (1991) Relaxin levels in ovum donation pregnancies. Fertil. Steril., 56, 5961.[ISI][Medline]
MacLennan, A.H., Grant, P. and Bryant-Greenwood, G. (1995) hRlx-1. In vitro response of human and pig myometrium. J. Reprod. Med., 40, 7036.[ISI][Medline]
Mayerhofer, A., Engling, R., Stecher, B. et al. (1995) Relaxin triggers calcium transients in human granulosalutein cells. Eur. J. Endocrinol., 132, 507513.[ISI][Medline]
Michael, A.E., Gregory, L., Walker, S.M. et al. (1993) Ovarian 11 beta-hydroxysteroid dehydrogenase: potential predictor of conception by in-vitro fertilisation and embryo transfer. Lancet, 342, 711712.[ISI][Medline]
Michael, A.E., Gregory, L., Piercy, E.C. et al. (1995) Ovarian 11 beta-hydroxysteroid dehydrogenase activity is inversely related to the outcome of in vitro fertilizationembryo transfer treatment cycles. Fertil. Steril., 64, 590598.[ISI][Medline]
Petersen, L., Svane, D., Uldbjerg, N. et al. (1991) Effects of human relaxin on isolated rat and human myometrium and uteroplacental arteries. Obstet. Gynecol., 78, 757762.[Abstract]
Prien, S., Canez, M. and Messer, R. (1994) Increases in progesterone after human chorionic gonadotropin administration may predict cycle outcome in patients undergoing in vitro fertilizationembryo transfer. Fertil. Steril., 62, 10661068.[ISI][Medline]
Prien, S.D., Canez, M.S. and Messer, R.H. (1995) Hormone release from cultured luteinized-granulosa cells mimics differences seen in vivo in patients undergoing IVF-ET. J. Assist. Reprod. Genet., 12, 1806.[ISI][Medline]
Rodbard, D. (1974) Statistical quality control and routine data processing for radioimmunoassays and immunoradiometric assays. Clin. Chem., 20, 12551270.
SART/ASRM (1996) Assisted reproductive technology in the United States and Canada: 1994 results generated from the American Society for Reproductive Medicine/Society for Assisted Reproductive Technology Registry. Fertil. Steril., 66, 697705.[ISI][Medline]
Schalkoff, M.E., Oskowitz, S.P. and Powers, R.D. (1993) A multifactorial analysis of the pregnancy outcome in a successful embryo cryopreservation program. Fertil. Steril., 59, 10701074.[ISI][Medline]
Sherwood, O. (1994) Relaxin. In Knobil, E. and Neill, J. (eds), The Physiology of Reproduction. Raven Press, New York, pp. 8611009.
Stewart, D. and VandeVoort, C. (1997) Simulation of human luteal endocrine function with granulosa cell culture. J. Clin. Endocrinol. Metab., 82, 30783083.
Stewart, D.R., Celniker, A.C., Taylor, C.A. et al. (1990) Relaxin in the peri-implantation period. J. Clin. Endocrinol. Metab., 70, 17711773.[Abstract]
Stewart, D.R., Cragun, J.R., Boyers, S.P. et al. (1992) Serum relaxin concentrations in patients with out-of-phase endometrial biopsies. Fertil. Steril., 57, 453455.[ISI][Medline]
Stewart, D.R., Overstreet, J.W., Celniker, A.C. et al. (1993) The relationship between hCG and relaxin secretion in normal pregnancies vs periimplantation spontaneous abortions. Clin. Endocrinol., 38, 379385.[ISI][Medline]
Stewart, D., Nakajima, S., Overstreet, J. et al. (1995) Relaxin as a biomarker for human pregnancy detection. In MacLennan, A. Treager, G. and Bryant-Greenwood, G. (eds), Progress in Relaxin Research. Singapore, pp. 214224
Stewart, D., Erikson, M., Erikson, M. et al. (1997) Role of relaxin in glycodelin secretion. J. Clin. Endocrinol. Metab., 82, 839846.
Unemori, E., Erikson, M. and Grove, B. (1997) Recombinant relaxin stimulates expression of vascular endothelial growth factor (VEGF) in normal human endometrial cells. Paper presented at the Annual Meeting of the Society for the Study of Reproduction, Portland, Oregon, USA, Aug 25, 1997.
Wilcox, L.S., Peterson, H.B., Haseltine, F.P. et al. (1993) Defining and interpreting pregnancy success rates for in vitro fertilization. Fertil. Steril., 60, 1825.[ISI][Medline]
Submitted on March 10, 1998; accepted on October 13, 1998.